Citation
The synthesis, chemistry, and solvolytic behavior of tetracyclo [4.3.0.02 4.03 8]nonan-9-ol, and related compounds

Material Information

Title:
The synthesis, chemistry, and solvolytic behavior of tetracyclo [4.3.0.02 4.03 8]nonan-9-ol, and related compounds
Creator:
McRitchie-Tichnor, Donna, 1934-
Publication Date:
Language:
English
Physical Description:
xi, 128 leaves : ill. ; 28cm.

Subjects

Subjects / Keywords:
Alcohols ( jstor )
Chlorides ( jstor )
Esters ( jstor )
Ethers ( jstor )
Infrared spectrum ( jstor )
Ions ( jstor )
Ketones ( jstor )
Protons ( jstor )
Sodium ( jstor )
Solvolysis ( jstor )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Tetracyclononanol ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 124-127.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Donna Dembaugh McRitchie.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
This item is presumed in the public domain according to the terms of the Retrospective Dissertation Scanning (RDS) policy, which may be viewed at http://ufdc.ufl.edu/AA00007596/00001. The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
025243555 ( ALEPH )
02707387 ( OCLC )

Downloads

This item has the following downloads:


Full Text












THE SYNTHESIS, CHEMISTRY, AND SOLVOLYTIC BEHAVIOR
OF TETRACYCLO[4-.3.0.02,4 .03,]NONAN-9--OL,
AND RELATED COMPOUNDS











By

DONNA DEMBAUGH McRITCHIE


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY










UNIVERSITY OF FLORIDA


1975









































To Ruth















ACKNOWLEDGMENTS


The author wishes to acknowledge the creative contributions and helpful discussions of Dr. Merle A. Battiste during the course of this research. She is especially grateful for his uncanny ability to be available when needed and yet nonintrusive when not needed.

Sincerest thanks are due Dr. Rocco Fiato, Dr. Warren Nielsen,

John Timberlake, Dick Galley, and Bob Posey for their suggestions, advice, humor, and most importantly for their friendship, freely given, without which the author's circumstances would have been far less enjoyable. I

A large measure of thanks to Judy Romanik and Judy Nielsen for being there when the author needed someone to giggle or cry with.

Finally, and perhaps most importantly, the author acknowledges her husband, John, for his love, unflagging support, and unfailing assistance over all the little rough places.


iii
















TABLE OF CONTENTS


ACKNOWLEDGMENTS..............................

LIST OF TABLES...............................

LIST OF FIGURES..............................

ABSTRACT.....................................

INTRODUCTION.................................

SYNTHESIS AND CHEMISTRY......................

KINETIC AND PRODUCT STUDIES..................

The Stereochemistry of the Solvolysis and of the Carbonium ,Ion......................

Geometrical Considerations................

Effects of Strain.........................

EXPERIMENTAL.................................


Page iii vii viii ix

1

14 35


..............

..............

..............

..............

..............

..............

..............

the Nature ..............

..............

..............


46 48 56


........... 64


A. Synthesis.................................................

Preparation of 1,2,3,4-Tetrachloro-5,5dimethoxycyclopenta-1,3-diene (29)........................

Preparation of 1,2,3,4-Tetrachloro-5-endo-(1,1diethoxymethyl)-7,7-dimethoxybicyclo[2.2.1]hept-2ene (30)...................................................

Preparation of endo-5-(1,1-diethoxymethyl)-7,7dimethoxybicyclo[2.2.1]hept-2-ene (31)....................

Preparation of endo-bicyclo[2.2.1]hept-2-en-7one-5-carboxylic acid (33)................................

Preparation of endo-anti-bicyclo[2.2.1]hept-2-en7-ol-5-carboxylic acid (34)...............................

Preparation of endo-anti-7-Acetoxybicyclo[2.2.1]hept-2-ene-5-carboxylic acid (35).........................

iv


64 65 66 67 67 69 69


......

......

......

......

......

......

......









Page


Preparation of endo-anti-7-Acetoxybicyclo[2.2.1]hept2-ene-5-carboxylic acid chloride (36)....................... 71

Preparation of Diazoketone (37).............................. 72

Preparation of Diazomethane.................................. 72

Preparation of 9-Acetoxytetracyclo[4.3.0.02,4.03,8]nonan-5-semicarbazone (39)................................... 73

Preparation of CuO-Cu Catalyst.............................. 74

Preparation of anti-9-Acetoxytetracyclo[4.3.0.02,4.03,8]nonan-5-one (38)............................ 75

Preparation of anti-Tetracyclo[4.3.0.02,4.03,8Inonan-9-ol (26)............................................ 76

Preparation of anti-Tetracyclo[4.3.0.02,14.03,8]nonan-9-yl p-bromophenylurethan (26)-PBPU................... 77

Preparation of anti-Tetracyclo[4.3.0.02,4.03,8Inonan-9-yl p-nitrobenzoate (26)-OPNB........................ 78

Preparation of anti-Tetracyclo[4.3.0.02,4.03,8] nonan-9-yl 3,5-dinitrobenzoate (26)-ODNB.................... 79

Preparation of Tetracyclo[4.3.0.02,4.03,8 nonan-9-one (45)........................................... 79

Preparation of Chromium Trioxide-dipyridine Complex................................................... 80

Preparation of Bicyclo[3.2.1]octa-2,6-diene (46)............ 81

Preparation of Tetracyclo[4.3.0.02,4.03,8inonan-5-semicarbazone (43)................................... 81

Preparation of Tetracyclo[4.3.0.02,4.03,8 nonane (44)............................................... 82

Preparation of anti-Tetracyclo[3.3.1.02,4.03,7inonan-9-ol (27)............................................ 83

Preparation of anti-Tetracyclo[3.3.1.02,4.03,7]nonan-9-yl p-nitrobenzoate (27)-OPNB........................ 84

Preparation of anti-Tetracyclo[3.3.1.02,4.03,7 nonan-9-yl 3,5-dinitrobenzoate (27)-ODNB.................... 84


V









Page


Preparation of Tetracyclo[3.3.1.02,4.03,7]nonan-9-one (47)........................................... 85

Preparation of Tetracyclo[3.3.1.02,4.03,7
nonane (49)............................................... 86

Preparation of 7-Norbornadienyl p-nitrobenzoate
(12)-OPNB................................................. 86

Preparation of 7-Norbornadienyl 3,5-dinitrobenzoate
(12)-ODNB................................................. 87

Preparation of anti-7-Norbornenyl p-nitrobenzoate
(2)-OPNB.................................................. 87

Preparation of anti-7-Norbornenyl 3,5dinitrobenzoate (2)-ODNB..................................... 88

B. Kinetic Studies.............................................. 88

Preparation of Kinetic Solutions............................ 88

Kinetic Procedures........................................... 89

Analysis of Data............................................. 91

C. Solvolysis Product Studies.................................. 110

anti-Tetracyclo[4.3.0.02,4.03'8]nonan-9-yl pnitrobenzoate (26)-OPNB................................... 110

anti-Tetracyclo[3.3.1.02,4.03,7]nonan-9-yl esters
(27)-OPNB and (27)-ODNB..................................... 112

Auxiliary Experiments........................................ 114

APPENDIX......................................................... 116

BIBLIOGRAPHY..................................................... 124

BIOGRAPHICAL SKETCH.............................................. 128


vi
















LIST OF TABLES


Table Page

I. Comparative Rates in the Bicyclo[2.2.llheptan-7-yl
Series, and the Tricyclo[3.2.1.02,4]-oct-8-yl Series....... 5

II. Comparative Rates in the Basic anti-Bicyclo[2.1.l]
Series..................................................... 11

III. 13CNMR Chemical Shift Values (ppm)........................... 24

IV. Comparative Fragmentation Patterns (70 eV)................... 25

V. Comparative Ketone Fragmentation Patterns.................... 34

VI. Kinetic Results............................................... 38

VII. Kinetic Values for Structurally Related Systems............ 40

VIII. Comparison of Relative Rates of p-Nitrobenzoate Esters
in 70% Acetone:Water at 25*C.................................. 41

IX. Product Analysis.............................................. 44

X. Comparative Rates of Nitrogen Extrusion...................... 53

XI. Rates of Ketone Decarbonylation.............................. 54

XII. Comparative 13Cnmr Chemical Shift Values (ppm)............... 61

Ia. Bond Lengths............................................... 116

Ha. Interatomic Angles............................................ 117

IIIa. Interatomic Distances......................................... 117


vii















LIST OF FIGURES


Figure Page

I. Molecular Structure of (26)-PBPU........................... 28

II. Internal Plane Angles....................................... 51

Ia. Plot of log[C] Versus Time to 10 x t for (26)-OPNB...... 118 IIa. Plot of ln [C] Versus Time to 1 x t for (26)-OPNB ....... 119 IlIa. Plot of in [C] Versus Time to 10 x t for (2)-OPNB...... 120 IVa. Plot of log k Versus l/T for (26)-OPNB and (27)-OPNB..... 121 Va. Plot of log k Versus l/T for (12)-OPNB and (12)-ODNB..... 122 VIa. Plot of log k Versus l/T for (2)-OPNB...................... 123


viii















Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy


THE SYNTHESIS, CHEMISTRY, AND SOLVOLYTIC BEHAVIOR OF
TETRACYCLO[4.3.0.02,4.03',8NONAN-9-OL, AND RELATED COMPOUNDS By

Donna Dembaugh McRitchie

August, 1975

Chairman: Dr. Merle A. Battiste Major Department: Chemistry

Solvolysis studies of endo-anti-tricyclo(3.2.1.02,4]octan-8-yl esters have revealed an enormous rate acceleration compared to 7norbornyl due to the stabilization afforded by the tris-homocyclopropenyl cation. The products of the solvolysis were greater than 99% rearranged, which has led some investigators to rationalize the large acceleration as partly due to strain relief on ionization leading to a classical ion with charge concentrated alternately on C2 and C4.

Studies of pentacyclo[4.3.0.02,4.01,8.05,7]nonan-9-yl esters have also demonstrated accelerated solvolysis rates, but with no rearrangement possible. Here there could be no doubt of the intermediacy of a tris-homocyclopropenyl ion and there was no strain relief upon going from reactant to product.

The alignment of the cyclopropane edge bond orbitals with the

site of developing positive charge could be a more important influence


ix









on the rate of reactivity than relief of ground state strain in the rate determining ionization. To resolve this question the tetracyclo[4.3.0.02,4.03'8]nonan-9-yl esters, which appeared to be more highly strained than those of the tricyclo series, were prepared and studied.

The synthesis of the title compound was accomplished through the key intermediate endo-bicyclo[2.2.1]hept-2-ene-7-one 5-carboxylic acid, obtained by Diels-Alder addition of acrolein diethylacetal to 1,2,3,4tetrachloro-5,5-dimethoxycyclopentadiene followed by hydrolysis and oxidation. Reduction of the key intermediate with sodium borohydride was followed by esterification to give endo-anti-7-acetoxybicyclo[2.2.l]hept-2-ene 5-carboxylic acid. Treatment of the acetoxy-acid with oxalyl chloride followed by diazomethane, and then cupric oxide in boiling tetrahydrofuran gave anti-9-acetoxytetracyclo[4.3.0.02 4.03,8]nonan-5-one. Wolff-Kishner reduction of the semicarbazone derivative of the tetracyclic acetoxy-ketone gave the title compound.

Solvolysis of the p-nitrobenzoate esters of the title compound yielded a less strained rearranged product, but the relative rate acceleration was only a factor of four slower than the endo-anti-tricyclo[3.2.1.02,4]octan-8-yl esters. Relief of strain in the rate determining ionization step of the tetracyclic system does not appear to be responsible for part of the observed acceleration. The methylene bridge which transformed tricyclo[3.2.1.02,4]octan-8-ol into tetracyclo[4.3.0.02' 4.03,8] nonan-9-ol has been shown by x-ray structural determination to have distorted the orbital geometry in two ways. The edge bond orbitals have been pulled out of alignment with the site of developing positive charge and their through space distance to that site has been made unequal.


x








The methylene bridge has made C2 - C9 overlap more favorable than C - C overlap, leading to a cation which resembles product. Solvent
3 9
attack on this intermediate ion, after the rate determining step, gives rise to the less strained product. Additional documentation for the effect of orbital topography on chemical reactivity was sought through a study of the 13Cnmr spectra of the reactants and products, and from a study of the relative rate of decarbonylation of the respective ketones.


xi















CHAPTER I


INTRODUCTION

The assistance to ionization by a neighboring , y-double bond is most dramatically revealed by a comparison of the solvolytic rates of the syn and atnti-7-norbornenyl tosylates (1)-OTs and (2)-OTs with the 7-norbornyl system (3).1 The acceleration factor, which was recorded for posterity as 1011, was attributed to backside homoconjugative stabilization to the developing carbonium ion center. This phenomenon was most succinctly expressed as the delocalized, or nonclassical, bishomocyclopropenyl ion (4);2a,b although others preferred the concept of rapidly equilibrating carbonium ions.2c,d

X X X




5+


12 3 4



More recently, research attention has been focused on the neighboring group reactivity of the cyclopropyl a-bond. 3 The early research of Winstein and his coworkers in the cis and trans-bicyclo[3.1.0]hexyl systems (5) and (6) established the fact of assistance to ionization by the cyclopropyl bond.4 This, now historic, study of the solvolytic rates (k,. /k = 35), the demonstrated existence of a special salt
sus trans
effect in only the cis isomer,4 and a detailed product analysis, led to
21





2


the postulation of the non-classical trishomocyclopropenyl ion (7) as the reactive intermediate.1,5 Further support for (7) was obtained via


H
s2 3

AOTs


5


OTs


H


6


deuterium labeling studies. Cis-tosylate (5) labeled at C3 was found to give products completely scrambled in the C., C3, and C5 positions. There was no scrambling observed when the correspondingly deuterated trans-tosylate (6) was solvolyzed under the same conditions.4a,d,e









7


Cyclopropyl participation was further probed in the 8-tricyclo[3.2.1.02,4]octyl systems (8-11). 6,7 These systems were chosen because,


X
8
2
7
4 5
6
3
8


X







9


X


X


10


11


at this point, evidence for ion (7) was limited to the bicyclohexyl system. The maximum rate enhancement reported to that point was 9228 which fell short of previously published calculations and theoretical predictions.9 In particular, if the C2 and C4 positions of the chair





3


and boat conformations of both the cis and trans isomers were connected by an ethano bridge, the four possible combinations were effectively frozen, allowing an efficient probe of the influence of the cyclopropyl group. The 8-tricyclooctyl series would also demonstrate that the alignment of the cyclopropyl edge bond orbitals with that of the developing cationic center was of critical importance. One further advantage of this series was the direct comparison which could now be made with the

7-norbornyl (3), anti-7-norbornenyl (2), and 7-norbornadienyl (12) systems.


X







12


This makes possible the evaluation, in a systematic manner, the stabilizing effects of the normal a-bond, the cyclopropane a-bond, and the classic Tr or double bond. The normal c-bonded compound (3) can now be designated as a reference, or "parent-model" for a series of structurally related compounds.

Pincock reported that the p-bromobenzenesulfonate (10)-OBs underwent acetolysis at a rate 2.7 times slower than (3)-OBs.7 The observed retardation could be due to steric interference to solvation at the backside of the leaving group but, in any case, the cyclopropane ring clearly provided no rate acceleration, thus ruling out "face" participation. The cyclopropane "edge" orbitals are oriented downward, away from the incipient carbonium ion site precluding edge participation. If orientation of the edge orbitals were of importance, then compound (8) should be expected to show a definite rate enhancement when compared





4


to (3)-OBs. This was found to be the case in several laboratories simultaneously.6 The results which were available at that time are summarized in Table I.

The rate enhancement of the endo-anti-p-nitrobenzoate ester (8)OPNB, due to direct participation in the ionization step, was, at that time, the largest reported in the literature. Product studieslOa,b which showed the overwhelming formation of rearranged product, e.g.,

(14)-OH, and only 0.1% of retained alcohol (8)-OH imply that most of the charge resides on C2 and C4.


OPNB


SOH


HO H 8-OPNB 13 14-OH


OH



+8



-18-OH


Particularly noteworthy here was the observation that both

solvent capture and internal return occur only from the endo direction. This result was in marked contrast to results observed when the ketone

(15) corresponding to (14) was reduced with lithium aluminum to yield a product of exclusively eXo attack.10b Internal return to a rearranged


endo lexo
0O


HO H H OH
62%
14 16


15





5


TABLE I. Comparative Rates in the Bicyclo[2.2.1]heptan-7-yl Series,
and the Tricyclo[3.2.1.02,4]-oct-8-yl Series



k 250 k 103
Compound Reference rel. rel.



x





3



10 109

2
x


1 104 10 1

1



7 0.4 1.7

10
x



111

x



6 15 37

9 X



6 10415 1012

8





6


ester requires the existence of a stable cation whose lifetime permits migration of the anion, and approach from another direction. 10a Equilibration experiments were reported to produce a mixture containing 62% exo-alcohol which would thus appear to be both thermodynamically and kinetically preferred on steric grounds. This is analagous to the ion generated by the anti-norbornenyl esters and is supportive of the nonclassical ion (13).

The rate of acetolysis of (9)-OBs is only 15 times that of (3)OBs and gives a complex mixture of products. 6,10a,b In contrast, (11)OBs is 104 faster than (3)-OBs. Here the methylene group of the exo cyclopropane ring may sterically aid the departure of the brosylate anion. 10a,ll Alternative possibilities involving a concerted CI-C bond shift to C have been discussed.10a
8
While indeed (8)-OPNB was locked in position so as to allow extensive backside participation via cation (13), it had been suggested10 that relief of strain could account in part for the lower activation energy observed in the solvolysis of (8).

Further research in cyclopropyl participation led Coates!2 to synthesize the pentacyclic alcohol (17)-OH. Here, overall relief of




OPNB OH


4+



45


1 7-OPNB


18


1 7-OH





7


strain cannot be a major factor in the observed rate acceleration (xlO11) since the rearrangement is degenerate, i.e. produces a structure identical with the original. The sole product of hydrolysis in the presence of base was (17)-OH, parent alcohol. Deuterium labeling studies yielded products which pointed to a triply symmetric trishomocyclopropenyl cation (18). It was suggested that the greater reactivity of (8)-OPNB relative to (17)-OPNB could be attributed to some relief of strain in the solvolytic transition state coupled with a more favorable orientation of the cyclopropyl edge bond. It was Coates' opinion that anchimeric assistance to ionization, forming the non-classical, symmetrical ion

(18), supplied the main driving force in his compound (17).

The concept that precise orientation would be required for maximum assistance to ionization by cyclopropane was advanced by Battiste13 and further corroborated by the investigations of Sargent14 and Masamune.15a



NsO




19


Sargent's work with 2-(trans-3-bicyclo[3.1.0]) ethyl p-nitrobenzenesulfonate (19) gave no evidence, either product or kinetic, for participation.

Conversely, Masamune's study of the exo and endo-cmti-tricyclo[3.1.1.02,4]heptan-6-ols provided additional insight into the geometrical requirements for assistance. He showed that (20)-OPNB was but 56 times faster than (21)-OPNB. However, (20)-OPNB yielded only the two rearranged products expected from the trishomocyclopropenyl cation while

(21)-OPNB produced an olefinic mixture via ring cleavage.





8


6 OPNB
5
4

2 1
7
H
3
H
20

BNPO


+1'





I'\


OH


OH



+\


+


OPNB







+


H OH


21


Gassmanl5b independently concluded that both (20) and (21) were solvolyzing with considerable assistance, albeit, of different types. By comparison, (20)-OPNB was a factor of 10 slower than (8)-OPNB, according to Masamune. This he attributed to a geometrical distortion caused by the interaction of the hydrogens on C3 and C7.

It is entirely possible that (20)-OPNB is slower than (8)-OPNB due to a decrease in the internal angle in the bridge from 97' to 830. This would increase the strain energy associated with ionization thus destabilizing the transition state and the resulting trishomocyclopropenyl cation intermediate.

In comparing the benzo-analog (22) of (20), Tanida16 observed essentially the same result. He noted that the parent in this series

(23) solvolyzed to give products arising from ring cleavage, and











OTs


0


22-OTs


OH


O


22-oH


commented that krel. comparisons would appear to have no meaning. Nonetheless, he added the observation that k(22)/k(23) was 800 at 550C.


OTs


OAc


HOAc


23


Coates examined the related exo-tetracyclo[3.3.0.0436.02,]oct-4-yl p-toluenesulfonate (24)-OTs and found two products.17 Rate studies led him to conclude that while the solvolysis was still anchimerically assisted, the methylene bridge had geometrically distorted


OTs 24


OAc





S +


OAc


the angle between the cyclopropyl edge bond and the back lobe of the leaving group. Coates compared the solvolysis rate of (24) with that of (17). However, the two structures would differ. significantly in rigidity, and the strain energy required for a similar change in the


9





10


critical relative geometry of the incipient cation center and the cyclopropane ring in the process of ionization would not be of the same magnitude for the two systems. This points out the necessity for making comparisons between structures which are closely related and adopting as the parent-model for any series, a compound of similar skeletal structure.

Haywood-Farmer later organized the rate studies performed on
3h
the compounds of the basic bicyclo[2.1.1] system as shown in Table II. The krel. comparisons to parent (23) are usefu., those to the 7-norbornyl series of questionable value, beyond the comparison of the two parent

compounds (23) and (3).

One of the objectives of this research was to contribute to a better understanding of the effects of orbital topography on chemical reactivity of the strained a-bond as exemplified by the cyclopropane edge bond. Rigid, bridged, polycyclic systems meet the needs for precisely known substrate geometries. It was decided that tetracyclo[4.3.0.02,4.03,7]nonan-9-ol and its ester could provide a useful probe in meeting this objective.

X
9
3 ~8
7
2 4 6

4 5
26


From the data in Table I, it is clear that cyclopropyl edge

participation in (8)-OPNB can be more effective than double bond participation in (2)-OPNB. This has been rationalized on the basis of a difference in orientation of the participating orbitals relative to





11


TABLE II. Comparative Rates in the Basic anti-Bicyclo(2.1.1] Series



Compound k rel.1000 krel. 7-norbornyla




x



1 10

23


x

108 1011

25b





x

108 1011



108 loll



20 X



106 10



24


aAlso at 100*C. bReference 18.





12


the incipient carbonium ion site.13 Addition of the methylene bridge should increase the ground state strain thereby increasing the driving force for participation in (26). If relief of strain is the overriding factor in the rate determining step of the solvolysis of (26), then the rate should be greater than that observed in (8). Additional rate acceleration may also be expected from the methylene bridge itself since this alkyl substitution on the cyclopropane ring would have the same inductively stabilizing effect as would methyl substitution at C4. Conversely, the same methylene bridge also appears to distort the molecular symmetry and may thereby decrease the orbital lineup for participation, resulting in a trend toward decreased solvolytic reactivity such as is observed in (17).

A second objective was a thorough analysis of the solvolysis products of (26). Tanida's contention that the tricyclooctyl system

(8) and its rearranged counterpart (14) both provide 0.1% of (8) when subjected to identical solvolysis conditions leads one to look closely for similar effects in structurally related systems.lOb

Along this line of thought, it is then reasonable to assume that alcohol (27) should be one of the predominant products of aqueous solvolysis since it appears to be the least strained. However, the


X
9

4 5
6
2 8

3
7
27





13


proportion of unrearranged alcohol (26) derived from non-classical ion

(28) by attack at C2 and C9 is statistically favored by 2:1. This is the opposite of the situation in (8) and should provide an interesting comparison.


OPNB
+--a

b- '


c

26-OPNB 28


b



c


OH OH






OH


Finally, it was felt that an attempt should be made to unify the various pieces of information available in the field of rigid, bridged, polycyclics having basic bicyclo[2.2.1]carbon skeletons so as to point out the information required to fill in the gaps, thus making the theory of orbital topography a more complete tool for estimating chemical reactivity.















CHAPTER II


SYNTHESIS AND CHEMISTRY

The first challenge to be met in accomplishing the goals of this

research was the synthesis of the parent alcohol tetracyclo[4.3.0.02,4.03,8]nonan-9-ol (26). The initial synthetic route planned is given as a flow diagram in Scheme I. With minor modifications this was also the best route found to give the required product in the best overall yield.

Hexachlorocyclopentadiene reacted exothermally with methanolic potassium hydroxide to give 1,2,3,4-tetrachloro-5,5-dimethoxycyclopentadiene (29), a well-known compound.19 The Diels-Alder addition of acrolein diethylacetal with the diketal (29) gave the white crystalline 1,2,3,4-tetrachloro-5-endo-(1,1-diethoxymethyl)-7,7-dimethoxybicyclo[2.2.1]hept-2-ene (30) in good yield. The 1Hnmr spectrum of (30) contained the expected pairs of non-equivalent methoxyl and ethoxyl protons and the acetal proton, a doublet at 84.35 (J = 4.5 Hz). The dechlorination of (30) proceeded smoothly to give endo-5-(1,1-diethoxymethyl)-7,7-dimethoxybicyclo[2.2.1]hept-2-ene (31).20 The nmr spectrum of (31) displayed the characteristic pattern of non-equivalent methoxyl and ethoxyl protons, the acetal proton doublet (now shifted upfield to 64.00, J = 9.0 Hz), and the new vinyl protons at 66.14. The large change in the coupling constant of the acetal proton was attributed to the change in bulk upon going from chlorine to a hydrogen substituent on the bridgehead carbon atom. The bulky chlorine atom apparently


14





15


SCHEME I


Cl Cl

Cl C KOH
CH3OH


Cl Cl


70%


CH30 OCH3 Cl Cl


acrolein diethyl
acetal

80%


29


0
11 /1


33


1)Cr03 2) HR,0+I


COOH


63%


NaBH4 NaOH 95-100%


H OH





COOH


(Ac)20 NaOAc 37-55%


CH30 OCH3







32 CHO






OAc



/

COOHR


dil. HCl


(COCi) 2


CH30 OCH3




Cl4
30
H OC2H5
OC2H5
Na
t-BuOH 90%



CH30 OCH3






31 H OC2H5
OC 2H5



OAc



/


36 COdl


35


OAc







NNH O<


NH2CONHNH2


75%


39 2 OH



-KOH >
HOCH2CH2OH

75-80%


CuO

72-80%


- OA 38 0


CR2N2


OAc








37 0 CHN2


34


?6





16


hindered free rotation of the equally bulky ethoxyl substituted carbon thus constraining the acetal proton to a position which was less than the optimal dihedral angle for maximum coupling. Relief of this steric constraint resulted in a more favorable dihedral angle and consequently a larger coupling. Isolation of (31) required careful distillation, to avoid all traces of acid, and to completely separate a low boiling contaminant. The conversion of ketal-acetal (31) into the key intermediate endo-bicyclo[2.2.1]hept-2-ene-7-one-5-carboxylic acid (33) was found to be best accomplished in stages, with minimal intermediate purification. The acetal group of (31) was hydrolyzed quantitatively by stirring for 1 hour at room temperature in dilute (5-10%) hydrochloric acid. The intermediate aldehyde (32) was isolated and oxidized with Jones reagent.21 This initial hydrolysis and isolation served to remove the bulk of the ethyl alcohol produced, which would otherwise have been oxidized in the succeeding step. It is possible to accomplish both hydrolysis and oxidation simultaneously with Jones reagent, but the workup is complicated by the products of the side reactions. The crude product of the oxidation was subjected to more acidic (25-30%) hydrolysis conditions to remove all remaining methoxyl groups. The infrared spectrum of (33) contained two carbonyl absorptions, the ketone at 1780 and the acid at 1710 cm~1. In the nmr, the vinyl protons were found shifted downfield to 66.55 and an acid proton now appeared at 10.67. The final purification of the carboxylic acid (33) was found to be very difficult if the distillation of (31) had not been conducted so as to remove all of the lower boiling contaminant. It was noted in passing that the rate of hydrolysis of the acetal and ketal groups was sufficiently different to allow isolation and further





17


purification of the ketal-aldehyde (32) if desired. This compound could then be oxidized under neutral or basic conditions which would allow isolation of a ketal-acid.

The intermediate keto-acid (33) was reduced with aqueous basic sodium borohydride22 to give alcohol (34) as a clear colorless syrup. The integrity of the product was established by confirming the loss of the 1780 cm~1 infrared band, and the concomitant appearance in the nmr of a resonance which disappeared upon addition of deuterium oxide. The vinyl protons shifted upfield to 66.05 while the proton at C7, syn to the double bond, appeared as a broadened singlet at 63.50. This is consistent with previously reported values for this stereochemistry.23

Three methods were examined for the preparation of 7-acetoxybicyclo[2.2.1]hept-2-ene-5-endo-carboxylic acid (35). Those methods utilizing acetyl chloride gave little or no product ester. Literature procedures22,24 which made use of acetic anhydride-sodium acetate mixtures gave the best yields, but recrystallization was complicated by the co-precipitation of a yellow, gummy, by-product. A recently published25 procedure employing acetic anhydride in pyridine, a modification of an older method, was found to give cleaner products but in somewhat reduced yield. The last method could possibly be improved by lengthening the reaction time or by employing higher reaction temperatures. The nmr spectrum of the acetate ester showed vinyl protons at 66.10, and, as expected, the appearance at 2.05 of the acetate methyl singlet accompanied by a downfield shift to 4.42 of the broad bridge proton singlet.

In retrospect, the mode of preparation of 7-acetoxy-bicyclo[2.2.1]hept-2-ene-5-endo-carboxylic acid chloride (36) was of critical





18


importance from an unusual standpoint. The compound was first prepared by addition of the acid, as a solution in pyridine, to the oxalyl chloride-ether reaction mixture. When the reaction was complete, the acid chloride was isolated by filtering off precipitated pyridinehydrochloride and evaporating the ethereal filtrates to constant weight. The product appeared pure enough for further reaction but was later found to contain about 10% of dissolved pyridine-hydrochloride. Attempts at purifying (36) by vacuum distillation resulted in decomposition and by column chromatography, in hydrolysis. Fortunately, acid chloride (36) could be prepared quantitatively and cleanly by addition of an ethereal solution of the acid to the reagent oxalyl chloride. The production of acid chloride (36) completely free of pyridinehydrochloride was vital to the success of later steps in the sequence. The infrared spectrum of the acid chloride showed a carbonyl adsorption at 1790 cm~1. The nmr of (36) differed from (35) only in that the vinyl protons were now observed at different chemical shift values, H2 at 66.00 and H at 6.25, due to the influence of the endo carbonyl chloride group. The proton on C5 shifted only slightly from 63.15 to

3.40 while H remained unchanged.

The diazoketone (37) was produced quantitatively by slow addition of an ethereal solution of the acid chloride (36) to an ethereal solution of diazomethane.23,25,26 Although the diazoketone appeared to be reasonably stable at room temperature, attempts to purify it by column chromatography resulted in loss of 30-40% of the material. When (37) was prepared from pyridine-hydrochloride free acid chloride, the need for further purification was removed. The diazoketone was characterized by the appearance of a sharp -C-N=N adsorption at 2100 cm~1 in the





19


infrared and a concurrent shift of the carboxyl carbonyl from 1690 to 1640 cm~1. The nmr was observed to differ-from that of (36) in the appearance of a new resonance at 64.0 for the diazomethyl proton, and a small upfield shift of the vinyl multiplets to 5.9 and 6.15, respectively.

The cyclization of diazoketone (3f) to form the keto-acetate (38) proved exasperatingly difficult. Many researchers , 8 have employed this pseudo-carbenoid insertion to produce polycyclic structures with varying degrees of success. The "catalyst" used to accomplish the intramolecular insertion has varied with the researcher. Ultimately the catalyst system suggested from the work of Ghatak et al.29 was found to give high, reproducible yields of (38). These workers used an activated CuO catalyst made from freshly prepared copper powder,30 whose CuO:Cu content did not vary appreciably from batch to batch. They also irradiated the reaction mixture during the decomposition with infrared heat lamps. The function of these lamps was not clear-but the authors clearly state that they found the yields to be greatly reduced when the lamps were not used. In retrospect, it appears that part of the irreproducible results obtained in the earlier attempts at cyclization was due to the interference of pyridine-hydrochloride dissolved in the acid chloride and carried through to the diazoketone.

In early work, any keto-acetate (38) produced could not be

separated by conventional means from a mass of polymeric by-products. The thick, sticky, yellow residue obtained was extracted with boiling ether and discarded. The residue obtained upon evaporation of the ether was very crude, but conventional treatment with semicarbazidehydrochloride and sodium acetate24 produced a solid which could be





20


recrystallized to a constant melting point. The cyclized product, when present, could only be isolated and purified as the semicarbazone (39) until an improved method of cyclization was developed. This proved to be advantageous in developing the modified Wolff-Kishner conditions employed later. Using the conditions as outlined by Ghatak et al., ketoacetate (38) was finally isolated in sufficient purity for spectral analysis. The infrared spectrum of (38) was very simple, consisting mainly of -C-H stretch adsorptions at 2880, 2980, and 3050 cm~1, broad carbonyl adsorption centered at 1730 cm~1 and the acetate ester band at 1240 cm1. The nmr spectrum, as expected, no longer contained vinyl resonances, but the acetate methyl singlet remained at 62.05 while the adjoining bridge proton moved downfield to 4.98. The semicarbazone derivative (39) was only slightly soluble in common deuterated solvents, but resonances were still observed at 62.05 and 5.00 even in the extremely dilute solutions. The correct elemental analysis, and the appearance of characteristic amide I and II bands in the infrared had to serve as adequate characterization for this derivative.

The overall yield of 9-acetoxytetracyclo[4.3.0.02,4.03,81nonan5-semicarbazone (39) from the Diels-Alder adduct (30) to this point was 0-13%. From a practical point of view, if the synthesis was begun with 194 grams of chlorinated adduct (30), the maximum yield of semicarbazone at this point would be 16 grams. Unfortunately, the yields in the cyclization reaction, prior to the discovery of the CuO:Cu catalyst, and the pyridine-hydrochloride contaminant, were nearer zero more often than they were near optimum (42%). In the interest of conserving both time and material, and recognizing that the Wolff-Kishner reduction, as yet untried, can often give poor yields, it was decided





21


to work in a model system. In the interest of expedience, the reaction sequence given in Scheme II was begun. This would allow us to optimize yields of the cyclization reaction and the Wolff-Kishner reduction while producing the hydrocarbon (44) which contains the same carbon skeleton as the parent alcohol. This was to prove valuable for comparative purposes.

The procedure outlined in Scheme II was first published by

Nickon et al.27d The pure endo-bicyclot2.2.1]hept-2-ene-5-carboxylic acid (41) was obtained in excellent yield from the Diels-Alder adduct

(40). Although using the less desirable procedure for preparation of the acid chloride and diazoketone, and the unreliable cyclization conditions, acid (41) was found to undergp-ring closure readily in 80% yield. The resulting ketone was characterized as the semicarbazone derivative (43) and was identical to the compound described by Nickon. The ease of cyclization of the diazoketone derivative of (41) was surprising as compared to the results obtained when the 9-position was acetoxy substituted. This could be partly explained as a deactivating influence exerted on the double bond by the bridge functionality.

Ketone (42) was purified by column chromatography and subjected to the conditions usually employed in the Huang-Minlon modification31 of the Wolff-Kishner reduction. No product was found in several attempts to produce the hydrocarbon (44) in this manner. A further modification32 employing the semicarbazone derivative produced a 70% yield of an azine. Hydrocarbon (44) was successfully produced by the method of Murray and Babiak.33 The hydrocarbon was quite volatile, disappearing in a few hours at room temperature unless tightly sealed. The infrared and nmr spectra were of little value in the characterization





22


SCHEME II


SMethyl Acrylate


1) aq. NaOH


/c 2)H 30+ C
40 COOCH3 COOR


1) 12,
2) Zn,


KI
HAc


H2NNHCONH2.HCl

NaOAc


43 NNHCONH2


HOCH2CH2OH HOCH2CH201
H2NNH2 KOR
KOH









44


1) (COCl)2 2) CH2H2 3) Cu, Heat


4\0


41 COOH


N N





23


of (44) and, for that matter, those of (26)-OH and its rearranged counterpart (27)-OH were later found to be similarly unrewarding. In addition to elemental analysis, the 13Cnmr provided additional information. All nine carbon atoms of (44) are unique and, indeed, nine carbon resonances were found in the 13C spectrum. The chemical shift values are listed with those of other similar carbon skeletons in Table III.

The conditions for the Wolff-Kishner reduction of semicarbazone

(43) were then successfully applied to semicarbazone (39). Further slight modifications in this reaction brought the yield up to acceptable levels. The 1Hnmr of (26)-OH consisted of a broad multiplet between 60.8 and 2.4 with H9 occurring as a broad singlet at 64.35. The alcohol sublimed readily, requiring the use of gc-mass spectral analysis. The fragmentation pattern is given in Table IV. The source of the base peak at m/e 79 and the second major fragment at m/e 70 are rationalized in Scheme LL Molecular ion (26)-M+ ring opens via a to give the bicyclo[3.2.1] radical ion (26)-b. The favored route to the base peak m/e 79 may be via loss of formaldehyde giving (26)-c which then looses hydrogen, or the processes may occur simultaneously to give m/e 105. Loss of acetylene to give the C H + ion is a facile process when compared to the competing pathway where hydrogen migration leads to radical ion (26)-d. Loss of cyclopentadiene (m/e 66) via a retro Diels-Alder leads to the second major fragment radical ion m/e 70.

The p-nitrobenzoate (PNB) and 3,5-dinitrobenzoate (DNB) derivatives of (26)-OH required for solvolytic studies were prepared by conventional procedures.24 The p-bromophenylurethane derivative (26)-PBPU








TABLE III.


13CNMR Chemical Shift Values (ppm)a


OH
i
h
f
b e
g e

C d


a
b
c
d
e
f
g
h
i


a h
f
b g e


c d


22.5 19.7
24.2 39.1
34.1 34.7 40.9 48.9 86.4


OH

a b




e c d
e d


a
b
c
d
e
i


a
b
c
d
e
f
g
h
i


a b
c (or i


29.5
42.2 39.7
38.4 31.3 78.2


a = 37.1 b = 40.3 c = 47.1


aValues relative to TMS, solvent CDC1,.
3


24


30.66
24.3 30.81 39.77 35.3
40.06 38.0
44.5
51.9





25


TABLE IV. Comparative Fragmentation Patterns (70 eV)


OH







26-OH



% of Base

10.0 16.5 19.7 10.0
19.4 14.7 13.5
22.4 22.4 100.0
22.4 27.7 52.9 5.3 26.5 26.5


Mass


(M+)
(-H20)


OH







27-OH



% of Base

19.7 10.6 10.5
4.4
20.3
8.4 5.5 16.1 11.6
23.4 13.1 17.2 100.0
14.4

21.9


136 118 117 108 105 95 92 91 80 79 78 77 70 69 68 67





26


SCHEME III


H OH







m/e 136 26(M+)



H OH






m/e 70


+


H OH

H


S


H OH


0 m/e 66


26-d


_ -- H in/e 79


H H







m/e 105


-H2



26


+


H OH

H





26-b




-I
-CH 20





27


was prepared for use in the x-ray structural determination which employed the heavy atom technique. Although the crystals formed by this derivative were always clustered, it was possible to dissect a single crystal of sufficient size for the structure determination. The compound (26)-PBPU crystallizes in the space group P21/c with unit cell dimensions, a = 12.067(4), b = 9.782(3), c = 12.060(4), and = 96.12(3), and four molecules per unit cell. The analysis was refined to an R factor of 0.065. The bond lengths and bond angles and the interatomic distances of importance are summarized in the Appendix. The computer generated drawing (ORTEP) which represents the data is given in Figure I. Three internal angles were defined:

a) the cyclopropane plane through atoms 2,3, and 4;

b) the skeletal plane through atoms 1,2,3, and 8;

c) the bridge plane through atoms 1,8, and 9. The calculated dihedral angles were

a-b = 69.70 b-c = 61.2' a-c = 9.10.

The consequences of the information revealed by the x-ray structural analysis is discussed in relation to kinetic and product studies in Chapter III.

Parent alcohol (26)-OH was cleanly oxidized -(Scheme IV) to

tetracyclo[4.3.0.02,4.03,8]nonan-9-one (45) with a methylene chloride solution of chromium trioxide dipyridine complex.34 As was characteristic of other members of this family, (26)-OH and (44), the ketone was quite volatile and sublimed readily. It was characterized by the appearance of a carbonyl adsorption in the infrared at 1765 cm~1. The nmr resonance for H9 at 64.35 had disappeared leaving only the broad multiplet from 60.8-2.5.

















C (3) C (113
C~ (121g C 14) an
C(1) C (91
0114


C (71
0121 N C (161
C I I1


C (51 6


FIGURE I. Molecular Structure of (26)-PBPU





29


SCHEME IV


ODNB


60% Acetone


OH







27


CrO3 Acetone CrO3-2 pyridine
CH2C12


47


LiAlH


OH







27


OH







26


DNBC1
Pyridine


I


0
11


/:


46


45





30


Ketone (45) decarbonylates readily (t = 5.2 hours at 185*C

giving bicyclo[3.2.1]octa-2,6-diene (46), a well-known hydrocarbon.35 The isolation and identification of (46) was convincing proof that alcohol (26) and ketone (45) possess the structures assigned.

The esters (26)-OPNB and (26)-ODNB were solvolyzed in aqueous acetone and rearranged cleanly to give tetracyclo[3.3.1.02,4.03,7]nonan-9-ol (27) in good yield. The rearranged alcohol was also quite volatile and was easily purified by sublimation. Characterization of this new symmetrical structure employed mass spectral analysis (see Table IV, fragmentation patterns). Alcohol (27) had a much larger molecular ion, twice that of (26)-OH. Other than loss of water, the fragmentation pattern of (27)-OH was sufficiently different from that of its isomeric parent (26)-OH to warrant further consideration. It was rationalized (Scheme V) that (27)-M+ would most likely undergo a series of ring openings similar to (26)-M+ via the analagous (27)-a to give (27)-b. Loss of formaldehyde and hydrogen lead to m/e 105 in about the same abundance as observed for (26). However, loss of acetylene from the homo-aromatic ion (27)-e is not as favored a process here as in (26)-OH. The base peak in (27)-OH arises from the alternative process, now more competitive, of hydrogen shift and loss of cyclopentadiene to give the radical ion m/e 70.

Alcohol (26) had melting point 145-146*C (sealed tube) while that of (27)-OH was found to be 236-237*C (sealed tube). The nmr spectrum of (27)-OH differed somewhat from that of (26)-OH in the region 61.2-2.4, the major difference being the H9 proton which occurred as a broadened triplet (J = 5.0 Hz) at 4.00. In models of the new carbon skeleton, H9 appears ideally situated for long range W coupling,





31


.SCHEME V


H








m/e 136 27-(M+)


H OH







m/e 70 m/e 66


HH

H





27-a


H OH 27-d


H

H /27-e
mle 105 nile 105




I7C 2H2




1+



H H

nile 79


H -:

H

>


27-b






- CH20


-H



27-c





32


which would account for the broadening. The 13Cnmr was consistent with a symmetrical structure containing six different types of carbon atoms. The chemical shift values and peak assignments are given in Table III.

Oxidation of (27)-OH using either Jones reagent21 or the chromium trioxide dipyridine complex in methylene chloride34 gave excellent yields of tetracyclo[3.3.1.0 214.O 3>,nonan-9-one (47). Both ketone

(45) and ketone (47) have a moist appearance when freshly sublimed but were homogeneous to capillary glpc. Ketone (47) had a strong infrared carbonyl band at 1725 cmd (vs. 1765 cm~1 for (45)), and its nmr spectrum compared to (27)-OH showed the absence of the Ha resonance. As a check on the stereochemistry of (27)-OH, ketone (47) was reduced with lithium aluminum hydride. It was hoped that a mixture of alcohols, -OH both syn and anti to the cyclopropane ring, would result. Glpc examination of the alcohol produced established that it was at least 99% a single component. Co-injection with authentic anti-(27)-OH produced a single peak, and the nmr and infrared spectra were superimposable with that of the original (27)-OH. Models show that the approach which would produce the syn epimer appears to be blocked by the axial-like hydrogens in the six-membered ring. The formation of only (27)-OH from (47) appears consistent with attack at the least hindered face of the carbonyl double bond. As proof of the structure of (27)-OH and consequently of (47), it was planned to decarbonylate (47) and identify the known cis-2,7-bicyclo[3.3.0]octadiene.36 However, ketone (47) proved very resistant to decarbonylation (estimated t! = 2800 hours at 240*C). Further evidence for the stable nature of this ketone is found in comparing the mass spectral





33


fragmentation pattern with that of (45) in Table V. The molecular ion of (47) is eight times the intensity of the ion from (45) and the base peak of (45) is the result of further skeletal rearrangement.

Since ketone (47) would not conveniently decarbonylate, the proof for the structure of alcohol (27) was sought via the reaction sequence given in Scheme VI.

Alcohol (27) was dissolved in thionyl chloride to give tetracyclo13.3.1.02'4.03,7]nonyl-9-chloride (48) as a white solid. Chloride *(48) was dechlorinated using the conventional procedure21 and gave the hydrocarbon tetracyclo[3.3.1.02 4.03 7]nonane (49) christened "triaxane" by Nickon and Pandit.37 Triaxane was extremely volatile and could only be kept in a sealed tube. The 13Cnmr shift values were consistent with those reported,38 and are listed with the structural assignments in Table III.

The production of the known diene (46) and the origin of (27)-OH by the routes shown in Scheme IV, coupled with the formation of hydrocarbon (49) as shown in Scheme VI, confirm the integrity of the structures of (26)-OH and (27)-OH as drawn. That the two carbon skeletons are different appeared evident from the comparison of the physical characteristics. Solvolytic studies were anticipated to produce marked differences in the chemical reactivity of the two systems.





34


TABLE V. Comparative Ketone Fragmentation Patterns


0







45

% of Base


2.3 66.6 16.6 100.0 33.3 66.6 25.0


0








47

% of Base

16.1
100.0
19.4 94.3
24.2 62.9 12.6


SCHEME VI


C1


SoC12


t-BuOH


48


49


III


Mass


(M+) (-CO) (C7H7+)


134 106 105 91
79 78 77


OH 27















CHAPTER III


KINETIC AND PRODUCT STUDIES

One of the goals of this research was to obtain the rate data and product information necessary to make comparisons between (8),

(17), and (26).

8 X 'XX
88
4 5 3 7 3 7
4 7 7

2 2 2 1 6

3 5 4
8 17 26


Work in the endo-anti-tricyclo[3.2.1.02,4]octyl system (8)

had been performed with p-nitrobenzoate (PNB) esters in 70% dioxane: water,7 and in 70% acetone:water.6 Work in the pentacyclo[4.3.0.02,4.03,8.05'7]nonan-9-ol (17) system was also performed with p-nitrobenzoate esters, but the solvent was 65% acetone:water.12 For comparative purposes it was deemed expedient to perform the rate studies on (26)-OPNB in 70% acetone:water.

The extent of reaction was followed by titration with standard sodium hydroxide to a phenolphthalein end point of the p-nitrobenzoic acid produced. These titrations generally gave normal first order plots for all esters examined except (26)-OPNB. The plots for this ester were linear to approximately one half-life and then curved upward, due to the formation of (27)-OPNB, a less reactive ester.


35





36


Representative plots for (26)-OPNB and (27)-OPNB are illustrated in Appendix Figures Ia, IIa, and IIIa. The observed infinity titers were within 3-5% of theory except in the case of (26)-OPNB. Careful examination of the contents of several ampules indicated that at 78-83% of theoretical infinity only (27)-OPNB remained. Solvolysis of (26)-OPNB is obviously accompanied by isomerization and internal return. Solvolysis reactions with concurrent rearrangement of the type observed in (26)-OPNB have often been encountered, and their kinetic treatment has been discussed.39 Roberts 39b considered the following system of which (26)-OPNB is a representative: ROH + HX

R'X k3 ) R'OH + HX (1)


He assumed in his derivation of the rate constant, that the apparent first order rate constant (k or k titi tric ) obtained using

an acid infinity titer which does not take account of rearrangement will be the sum of krearrangement plus ksolvolysis, i.e. k + k2' as long as k is much faster than k3. Straight line plots were obtained for (26)-OPNB when 80% of the theoretical infinity value (plus the blank) were used in the calculations. Eighty percent of theoretical infinity was found to correspond to observed, or actual infinity values. Since the value of (k + kr) most nearly represents the actual rate of ionization of (26)-OPNB, it is the value used hereafter in this work for comparative purposes.

For all compounds, duplicate runs were made at each temperature. Each run was treated by a standard linear least squares program and the average of the two values was employed in further calculations.





37


The enthalpy of activation (AN) was calculated using equation (2).

k, T -AHt 1 1
ln ( ) - ln( .) = (-- ---) (2)
k2 T2 R T, T2

When the average Ht had been determined, equation (2) was rearranged and used for extrapolations to 25*C. The entropy of activation AS* was calculated using equation (3).

k, k -AH*
R(lny) - R(ln h) = T + ASt (3)


The value for the constant term R(ln k/h) was found to be 47.18742.

The solvolytic kinetics of some previously investigated systems pertinent to our study were redetermined. These included the pnitrobenzoate ester of 7-norbornadienol (12)-OH, which was used to establish the kinetic relationships to other systems and to test the accuracy of our procedures, and the corresponding 3,5-dinitrobenzoate ester (12)-ODNB which was examined in 60% acetone:water in order to establish a relationship between our results and those of Klumpp40 in the same solvent. In addition, rates for anti-7-norbornenyl p-nitrobenzoate (2)-OPNB in aqueous acetone were carefully measured in order to more accurately assess the often quoted rate ratio k(2)/k(2) 103h Originally this comparative rate was obtained from the solvolytic rates of the respective chlorides (2)-Cl and

(12)-Cl in 80% acetone:water. 4d Haywood-Farmer determined the rate constant for (2)-OPNB in 70% dioxane:water, 10a but did not study

(12)-OPNB under the same conditions. The complete data obtained from the solvolysis studies are given in Chapter IV and summarized in Table VI. The log plots of the rate constants versus reciprocal temperature are shown in Appendix Figures IVa, Va, and VIa.





38


TABLE VI. Kinetic Resultsa


Compound


OPNB






26
OPNB







27

OPNB




'I
12

OPNB






1

ODNB




12


T*C

125.0 100.0 90.2 25b


150.0
140.0 125.0 25


125.0 100.0 90.2 25


k(sec.~1)


7.13 7.53 3.29
1.74


x
x
x
x


10~4 10~
10~5 10~


1.06 x 10~ 4.47 x 106 1.01 x 106
1.85 x 10-12


3.71
4.75 1.78 8.26


150.0
140.O 126.2 25



90.0 75.0 60.0 25


2.39
9.48 2.34 2.10


3.14 6.51 1.36 1.58


x
x
x
x


x
x
x
x


x
x
x
x


10
10~ 10~ 10~


10~5 106 10_6 10-12



10~4 10~ 10~5 10~ 7


AH4
(kcal/mole)

24.1 1.5


30.4 1.0


24.7 1.1


31.8 0.4


24.5 0.9


AS* (e.u.)

-12.9 3.8


-9.9 2.3


-12.7 3.1


-5.0 1.0


-7.5 2.5


70% acetone:water unless otherwise noted. All values at 25*C were extrapolated from higher temperatures. CValue taken from reference 19c. 60% acetone:water.





39


As Table VI shows, (26)-OPNB solvolyzes almost 10,000 times

(103'97) faster than (27)-OPNB. The use of the actual infinity titer in the calculation of the rate constant which best represents the rate of ionization was justifiable since Roberts' assumption was held to be valid when k was much greater than k3. The reevaluated rate acceleration at 25*C in 70% acetone:water for (12)-OPNB over (2)-OPNB was found to be 3900 (103.6). The rate of solvolysis of (12)-ODNB in 60% acetone: water was found to be 19 times faster than that of (12)-OPNB in 70% acetone:water at 2500. The data of other workers on compounds of interest are summarized in Table VII for comparative purposes.

Since it would be impossible to evaluate the rate constant for the solvolysis of (3)-OPNB in 70% acetone, it was necessary to fall back on the assumptioniOa that (2)-OPNB will be accelerated over (3)OPNB by the same amount that (2)-OBs is accelerated over (3)-OBs,

1.42 x 1011.1 The rate of (2)-OPNB at 25*C was obtained by extrapolation from data obtained at higher temperatures and, from this, the rate of (3)-OPNB was inferred. Although the bicyclohexyl system

(5) is conformationally mobile, its place in the overall picture was deemed of interest. The assumption was made that k /k
(5)-OTs (3)-OTs
would be the same as k /k . Since the rate of (3)-OPNB
(5)-OPNB (3)-OPNB 4
had been estimated as described above and the rates of (5)-OTs and

(3)-OTs1 are known, (5)-OPNB could be similarly estimated. The calculated absolute rates at 25*C for all systems of interest in this study are compared in Table VIII.

While the kinetic study has provided information on the formation of the transition state in the solvolysis of (26), and analysis of the products is needed to provide an insight into the nature of the





40


TABLE VII. Kinetic Values for Structurally Related Systems


Compound T("C)


OPNB 8 OPNB







17 ODNB 50
ODNB 51
OTs







5


125.0 25e





100.0 d
256


50.
25


k(sec.~I) 1.67 x 104
7.33 x 10-8









7.00 x 10~5 6.02 x 10~





4.93 x 10~4 2.81 x 10~







5.74 x 10-3 7.65 x 10~





9.61 x 10~5 3.8 x 10 6


AH+
(kcal/mole


24.8 26.8 26.2 23.7







24.1


a b
70% acetone:water. dAll values at 25*C were obtained by extrapolation. C65% acetone:water. 60% cetone:water. Corrected to 70% acetone: water and to OPNB ester. Value for acetolysis.


As+
(e.u.)

-7.8










-11.5







-3.8








-5.7


-5.0


Reference


10a 12b







40 40


4d









TABLE VIII.


Comparison of Relative Rates of P-Nitrobenzoate Esters
in 70% Acetone:Water at 25*C _


Compound
OPNB




3 OPNB







* OPNB 2 OPNB





12 PNB






27
OPNB





6


17


-k(sec )


1.48 x 10~




2 14
2.16 x 10







2. 10 x 10'



-9
8.26 x 10





- 12
1.85 x 10







6.02 x 10


orel.
7-Norbornyl


log k
-r el.


1


Reference


1


1.46 x 109







1.42 x 1011 5.58 x 1014 1.25 x 10" 4.07 x 10"


109- 2 10" *2 1011 2 10 1i 10 1 .







10 13.


1


4d







This work
and 1 This work This work







12b


OPNB


- 8
1. 74 x 10


1.17 x 1015


a65% acetone:water.


41


26


1015.1'


This work





42


TABLE VIII. (continued)


kr l
7-Norborny1


log krel.


Reference


OPNB 8 OPNB







51 OPNB







50
OPNB 20


7.33 x 10


-8
7. 65 x 10


2.80 x 10







6.31 x 10 10


4.95 x 10"5


5.17 x 1015 1.89 x 10" 4.26 x 1013


Compound


k(sec )


1015.7 10". 3 101 3.


10a 40 40







15b





43


intermediate carbonium ion. The esters of (26) and (27) were solvolyzed in the normal manner, i.e. with a steadily increasing concentration of acid (no buffer) and with the presence of a 2:1 excess of urea. Little difficulty was experienced in analyzing the products of the normal hydrolysis. Peak areas were obtained with the use of a digital integrator and were reproducible so that good accuracy was obtained in chromatographic analysis. The solvolysis of (26)-OPNB in the presence of urea required heating at 125*C for only 20 hours to achieve ten half-lives. The resulting solution was yellow, but interfering products appeared to be minimal. Attempted analysis of the solvolysis products of (27)-OPNB in the presence of urea was badly complicated by the formation of products from many side reactions and decomposition. Such a large number of by-products were formed that isolation of the pr duct alcohol(s) sufficiently pure and in good yield was virtually ;mpossible. The ester (27)-ODNB was prepared and its rate of solvolysis at 140*C determined in order that the overall reaction time could be reduced, thereby producing a cleaner product. Solvolysis of (27)-ODNB in the presence of urea at the shorter ten half-life time (46 hours) still resulted in extensive decomposition, but overall, in fewer by-products; therefore analysis of this product mixture was less complicated. The results of the product analyses are given in Table IX.

One aspect of Table IX bears further explanation here. After having established all of the product ratios except those for (27)OPNB or -ODNB in the presence of urea, a control experiment was run. A mixture of p-nitrobenzoic acid and urea was dissolved in 70% acetone at the same molar concentration as that of a solvolytic product run













TABLE IX. Product Analysis


Conditions

Normal Urea


Normal


Urea


Overall YLeid

97% 97%


97.5% 96%


(26) -OPNB

0.09%

0


0

0


(27) -OPNB

99.91% 100%


0

0


Products
% Ester (26)-OH

24.5% 0.11%

30.0% 0.04%a


0

0


(27)-OH

99.89%

99.96%


0 100%

0 100%


% Alcohol

72.5% 67.0%


97.5% 96.0%


a
The origini anid identification of this material is questionable.


Starting
Material


(26)-OPNB


(27) -ODNB





45


and treated to the solvolysis conditions. Chromatographic analysis after workup clearly indicated several minor products were eluting in the vicinity of the retention time for (26)-OH. Since none of that alcohol could have been in this sample, the analytical data acquired for the solvolysis of (27)-OPNB in the presence of urea were discarded as spurious. Analysis of (27)-ODNB solvolysis products, while less complex, was subject to residual doubt when a minor peak appeared which could have been 0.06% of (26)-OH. Accordingly, painstaking care was exercised in preparing a solution of (26)-OH which was sufficiently dilute to simulate the product sample. This solution gave only a very small peak at low attenuation (near maximum sensitivity of the detector). Coinjection of the two samples produced an extra peak in the recorder tracing for true (26)-OH which was only a few tenths of a minute different in retention time from the spurious peak which could easily have been mistaken for a trace of (26)-OH. This bit of evidence casts some doubt on the origin and identification of the minor alcohol product detected in the solvolysis of (26)-OPNB in the presence of urea. The 0.04% reported as (26)-OH could as easily have been the spurious peak which was present to the extent of 0.06% in the (27)-ODNB solvolysis product.

It can be seen from the analysis of the products that internal return and attack by solvent in the solvolysis of (26)-OPNB occurs only at the anti- or the hindered endo-positions. The solvolysis products of (27)-ODNB arise exclusively from anti-attack of solvent.

It was deemed the best approach, for overall understanding, to

examine the results of this study by segregating the individual facets of the picture and discussing them separately. Accordingly, the





46


remainder

a)



b) c)


of this discussion is divided into the following parts: stereochemistry of the solvolysis and the nature of the carbonium ion,

the geometrical considerations, and the effects of strain.


The Stereochemistry ofthe Solvolvsis_ nd theNature of the Carbonium Jon

The data in Table VIII show that the rate of solvolysis of

(26)-OPNB is only a little slower than that of (8)-OPNB (k /k = 4.2), but 29 times faster than that of (17)-OPNB. As previously described,


OPNB






17

k 1
rel.

log krel. (3) io13.6


OPNB






26

29

iol s. 1


OPNB







8

121 10 *5'


(8) rearranges stereospecifically to give anti-(8), or mainly endo-(id). The stercospecificity of iLhe nucleophilic attack on the -,ttriediate oerLved from (3)-Of'B at cnly the ;,gi and hinfrod ano sites Is ouite cosisr ent with the trishenocyclopropenyl cation (23). Only (13) co] d so completely p.e'clude attack to Form na-(24) and son-(8). One arrant conceive of a set of F:lvy roq1li r-st >g r2 c _ ions giving c>Clusively the ructs observed. The as:er of (:7)--OH has Elso been discussed. -lera i the accelerated rate and the scrmbling of a deterium label clary point to cation (8).





47


+ ,,+ ,






113




The ketone (47) has an sp2 hybridized carbon atom at C
9

0l endo


ero"



47


and could be considered a good steric model for ion (28). When ketone (47) is reduced with lithium aluminum hydride, attack is



+''





28

96% from the exo side giving back alcohol (27) and 4% of three minor products (ratio 1:4:1).

The rate studies on (6)-OPNB clearly show acceleration of

nearly the same tiagnitude as that of (8)-OPNB. Product analyses indicate that nucleophilic attack is only at the anti and hindered endo sites. This is in direct contrast to the stereochemistry of the reduction of the ketone. Here, as in (8) and (17), the interpretation of the results clearly requires the intervention of the trishomocyclopropenyl cation (28).





48


Geometrical Considerations

The cyclopropane ring presumably stabilizes a carbonium ion by overlap of the bent C-C bonds, which are high in p-character, with the vacant p-orbital of the cation. From extensive studies on the addition of electrophiles to cyclopropane rings, it appears that the preferred mechanism is the edge approach. The ability of the cyclopropane edge to interact through space with a carbonium ion will be affected by the interatomic angles and other relative distances within the molecule. Theoretical studies by Hoffman42 have shown that there is a greater stabilization for 3-bicyclo[3.1.0]hexyl cation than for 4-cyclopentenyl, 2-indanyl, or cyclopentyl cations, and that the potential energy minimum occurs when the charged carbon atom is raised from 760 to 80* above the plane of the five-membered ring toward a chair form.





C. 76-80'





A search of the literature reveals a paucity of actual x-ray structural data on compounds related to (8), (17), and (26). Structural information was found for the anti-7-norbornenyl (2) skeleton,43a exo-anti-tricyclo[3.2.1.02 ,4]-octyl (10) skeleton,43b and the benzonorbornenyl (53) skeleton,43c none of which are of any assistance in evaluating the data now available on (26). Only the x-ray structure of the p-bromobenzoate ester of (20)18 bears mention. Masamune claims to have shown that (20) possesses a geometry similar to the one





49


predicted by Hoffman for the maximum delocalization of the positive charge of the bicyclo[3.1.0]hexyl cation. In (20) the charged carbon atom is raised 71* above the plane of the five-membered ring. Solvolysis studies on (20)-OPNBl5b show it to be a factor of 100 slower than

(8)-OPNB (Table VIII). Masamune concludes that the ground state geoinetry of the system is a decisive factor contributing to the rate acceleration. It is probably fortuitous that the rate acceleration of

(20) relative to 7-norbornyl is virtually identical with that of (17). As previously shown (Table II), (20) is accelerated 108 over its saturated parent (23) at 100*C and is identical with the acceleration afforded by its unsaturated analog (25).

It would appear that the real angles 0 and c and the interatomic distance AB may be more significant in the diagnosis of effec0D B
A Lv. grp.





tive overlap than are Hoffman's predictions.44 Masamune calculated the angle 0 (340), but nothing more was published. The angles 0, $, and the AB distance were calculated for (26) and were found to be:

0 = 450

= 200

AB = 2.182 angstroms.

With nothing more available for comparison than IMasamune's value for 0 (340), which is 110 smaller and more favorable, no conclusions can be drawn from these numbers.





50


In order to better visualize the molecule, the angles between the planes of interest, previously defined, were calculated and illustrated as a cutaway in Figure II. The carbon bearing the leaving group is found to be raised only 610 above the plane of the fivemembered ring. Careful examination shows that the skeleton of (26) is twisted by the methylene bridge with the result that the calculated values of 0, c, and the AB distance are averages, and in fact each carbon atom of the cyclopropane edge bond in (26) will have individually different values for these angles.45 N /



/ ' 9

4 k 2
'4



Calculations gave the following results:

Distance C2 - C9 2.283 angstroms; C3 - C9 = 2.343 angstroms

0 640 0 = 68

p 290 = 30*

Finally, as formation of the cation (28) progresses, it appears from models that C2 is held rigid in its original position by the methylene bridge, while C3 can move outward more freely and even farther from

C .

It appears clear then, from the detailed analysis of the x-ray structure, that insertion of the methylene bridge has impaired the orbital geometry for effective overlap and may be responsible for the reduced rate acceleration observed. Moreover, because of the individual differences in the angle and distance between C2 and C3 with














c9

..- LEAVING GROUP 8r





3 ~610





C

-2 110*









4


FIGURE II. Internal Plane Angles





52


respect to C9, the resulting cation would not appear to be formed symmetrically and this may further impair participation and reduce acceleration.

Further support for the conclusion that the orbital topography of (26) is less favorable overall for stabilization and additionally less favorable due to distortion in the cyclopropane edge bond, all due to the methylene bridge, was sought by an examination of the rate of extrusion of carbon monoxide from the corresponding ketones. The rate of synchronous loss of carbon monoxide and diene formation could provide a more sensitive probe for the effect of precise orbital alignment. This hypothesis is supported by the work of Allred46 on the synchronous loss of nitrogen and the formation of diene in the series (52-56), which produced the rates of nitrogen extrusion listed in Table X.

Allred suggested that the introduction of the methano bridge

in (54) caused severe strain in the transition state leading to diene. Lengthening of the bridge, as in (55) and (56),lessens this strain leading to enhanced rates of extrusion. The low reactivity of (54) when compared with (53) indicated that relief of strain involved in opening of the cyclopropane ring is not a major factor. Mention was made of the less favorable orientation of the cyclopropane edge bond in (54); however, preference for the increased strain argument was clearly stated. It would appear from inspection of models and from our results on decarbonylation of ketones (45) and (57) that the cyclopropane edge bond of (55) and (56) would certainly be more favorably oriented for overlap as the C-N bonds break, and that this more favorable overlap is at least as satisfactory as an explanation of Allred's results, and cannot be set aside.





53


TABLE X. Comparative Rates of Nitrogen Extrusion


k(sec )


16
4. 7 x 10


krel.




1


N

.N
53




N
54/
N
54





N


55





N


56


-4
1.04 x 10


_15
4.34 x 10


11
2.2 x 10


9.2


2.46 x 10~ 5.2 x 108


-5
8. 9 x 10


11
1. 9 x 10


N


52





54


Ketones (45), (47), and (57) were readily available from their corresponding alcohols.

0 0 0







45 47 57

The rate of decarbonylation of (57) was obtained gas chromatographically using the technique ultimately intended for both (45) and

(47), and compared favorably with the published value which had been obtained using Hnmr techniques. The data obtained are summarized below:


TABLE XI. Rates of Ketone Decarbonylation

(57) k 5 = 1.96 x 10~4 sec.1.10_X 10~4 sec.~ (lit. value)47

(45) k1650 = 4.45 x 10-6 sec.~

k 850 = 3.72 x 10-5 sec.~'

(47) k1650 = 8 x 10-13 sec.~1a

k 0 = 7 x 10~8 sec.I b


aEstimated assuming a linear relationship where log k (57) 165*/log k rel (26)-OPNB 1000 =
(45) 7-norbornyl(8)-OPNB

b
log k (47) 165*/log krel. 7-norbornyl(27)-OPNB 1000. Calculated
rel.(45) (8)-OPNB

from a single sample heated 76 hours at 240*C.





55


The data show k 165(57)/k65,(45) = 44, compared to the rate acceleration of solvolysis at 250 where k /k is only a factor
(8)-OPNB. (26)-OPNB
of four. The rate of decarbonylation would appear to be a more sensitive measure of the effect of the ethylene bridge on the overall distortion of the cyclopropane edge bond and the reduction in overall orbital alignment.

It could be argued that a comparison of the rates of solvolysis of a given compound cannot be compared to the rate of decarbonylation of its corresponding ketone since the mechanisms and transition states for the two processes may be entirely unrelated.

That the rate of decarbonylation of rigid ketones will be responsive to the degree of p-character of the cyclopropane edge bond orbitals has been shown by the excellent linear correlation which has been found to exist between the '3C-H coupling constants of the endo small-ring protons and the free energy of activation for decarbonylation of the appropriate ketone.48 The results of the x-ray structural determination have demonstrated the difference in p-character which exists in the orbitals of the cyclopropane edge bond in (26)-OH due to the methylene bridge.

It has been shown that a linear free energy relationship exists between the rate of decarbonylation of a series of unsaturated tricyclic ketones (A), and the rates of ionization of the ester of the corresponding saturated alcohol having the appropriate stereochemistry (B).49 The mechanistic significance of this relationship is not clear but the fact of its existence has led us to search for a similar relationship between the rates of solvolysis of other bridged polycyclic systems and the free energy of decarbonylation of





56


0
X





A (CH2) B (CH2)
n-2 n-2
n = 3,4,5,6

their corresponding ketones. This correlation would be expected to be more sensitive since the double bond in the tricyclic ketones (A) described, must, to a certain extent, exert a leveling effect on the decarbonylation rates.

Such a new structure-reactivity correlation would be more

useful for predictive purposes. Although limited, the data available to date are certainly encouraging.

From the foregoing discussion, it is apparent that the rate acceleration of (26)-OPNB over the 7-norbornyl parent model is due to the neighboring group participation by the cyclopropane edge bond as exemplified by the intermediate non-classical ion (28), and the lack of acceleration over (8)-OPNB can be explained as due to less favorable orbital topography introduced by the methylene bridge.


Effects of Strain

The cis-bicyclo[3.1.0]hexyl system (5) shows moderate rate enhancement when compared with the conformationally more rigid structures

(2) and (3) (Table VIII). The bridging which locks (5) in the chair form giving (8) results in a dramatic increase in rate: log (kB/k.) = 106.3. This has been attributed in part to the more favorable geometry for potential overlap in (8). Such bridging is also accompanied by





57


two other features: an increase in the rigidity and ring strain of the molecule, and a decrease in the internal angle and flexibility of the cation center.3h

It has been argued that since the transition state involves sp3 to sp2 rehybridization at the cation center, an increase in the internal bridge angle would facilitate the formation of the transition state and lead to an increase in the solvolysis rate. All available x-ray data on carbon skeletons arising from the basic bicyclo[2.2.1] skeleton clearly show that the angle in question is virtually identical, while the log krel. solvolysis rates vary from 0.4 (10) to 10'-"'(26). There is no reason to expect that the corresponding angle in '(8) and (17) should vary significantly. The relief of the necessity for complete sp 3_sp2 rehybridization upon going to delocalized ions (13), (18), and (28) will be a constant component of the rate unless the geometry for participation by the cyclopropane edge bond is altered.

The role of strain and geometry in the overall rate acceleration picture in the case of (51), when compared with 7-norbornyl, is cloudy. Here, the insertion of an additional methylene gives a basic bicyclo[3.2.1] system where the internal angle of the bridge bearing the leaving group appears to be expanded. This should supposedly facilitate ionization leading to a rate enhancement. Compared to 7-norbornyl, (51) is as fast as (8) and four times faster than (26) (Table VIII). The bridge angle in (27) is also expanded. However,

(27) solvolyzes 40,000 times slower than (8). Models indicate that the geometry of the cyclopropyl ring in (51) with respect to the





58


bridge bearing the leaving group is remarkably similar to (26). Moreover, (51) solvolyzes with exclusive rearrangement where the new bond is formed with the cyclopropane carbon atom on the side of the methylene bridge. The work of Hess50 helps in placing Klumpp's40 work on (51) in proper perspective. Unfortunately, the solvolysis


X XX




58 59 51

log krel. 1 105.3 10"-4

data on the cyclopropane analog of (55) have never been published.

Certainly relief of ground state skeletal strain occurs since both (26), '(8), and (51) rearrange to less strained structures. The difference in the free energy of reactants and products in the case of

(8)-OPNB has been calculated by Tanida10b from the equation

AFt = -2.3(l.986)(T*K) log (RP)

where R, the reactivity ratio, is represented by the rate constants for solvolysis of starting material and rearranged product, and P, the partition factor, is represented by the product ratio of rearranged alcohol to starting alcohol obtained in the solvolysis reaction under buffered conditions. In the case of (8)-OPNB and (14)-OPNB, R = 3.2 x 107, and P = 103, making AFt = -11.95 kcal/mole. Using the same relationship, Klumpp40 found the change in AFt for reactant

(51)-ODNB and its products to be 8.6 kcal/mole.

While the reactivity ratio R in the case of (26)-OPNB is

easily obtained, the factor P is not. We observed no (26)-OH from the solvolysis of (27)-ODNB or (26)-OPNB under buffered solvolytic





59


conditions. The limits of detection of (26)-OH for the method used in our product analysis are 10-12 gsec. -151 as little as 1 x 10~4 weight percent.

Assuming that P in our system is as small as 104 (it could be 106), AFt = -11.67 kcal/mole. This minimally small difference in the free energy of reaction could at best reduce the ratio k /k(26) to 1.5 (75C). The uncertainty in the evaluation of P makes it difficult to truly assess the difference in free energy in our system. The fact that the change in geometry only produced a change in k(8)/k 26) of 4.2-4.5 (25*-75*C) does not decisively eliminate Tanidas' argument for effect of strain in the rate acceleration, but it certainly advances the argument for the importance of orbital topography in affecting chemical reactivity.

As part of the characterization of compounds (26)-OH and

(27)-OH and their respective hydrocarbons, 13Cnmr were recorded. The resonances and their assignments were given previously in Table III. A casual comparison of the chemical shift values between the two hydrocarbons revealed an interesting difference. Parent hydrocarbon

(26) had a range of chemical shift values three times that of (27). Since the hybridization of the C-C bonds joining a carbon atom to its neighbors can affect 13C chemical shift values significantly it was assumed, perhaps naively, that the (26)-H skeleton varied more widely in overall hybridization of its C-C bonds than did (27)-H. Carbon atoms joined in chemical bonds which are greatest in p-character are found at highest field relative to 1MS (tetramethylsilane). 37b,52 As the chemical shift value increases over the normal range for cyclic structures, then the p-character of the carbon to carbon bonds must





60


be decreasing. When the 1 rC spectra of the alcohols (26) and (27) were later obtained, the same correlation was found and Table XII was constructed. It would appear that the skeletal rearrangement of (26) to (27) occurs with a general decrease in the overall p-character of the skeletal bonds and a decrease in the range of carbon hybridizations. The changes observed in the 13Cnmr chemical values upon rearrangement may be indicative of an overall lower energy ground state for the rearranged skeleton.

In the discussion of the relevance of the angles 0 and $, the

third internal angle of the triangle was not mentioned. Later examination indicated that it, too, bears some consideration. Recall the diagram on page 50. The extension of the C4 to C2 bond meets the extension of the bond from the leaving group to C9 at an angle calculated to be 1130. The corresponding angle involving the C to C3 extension was found to be only 1080. The cyclopropane bent bonds (the Coulson-Moffitt model) are calculated to overlap at an angle of 1150.

Also available from the structural data are the angles around each carbon atom in the cyclopropane ring of (26). One index of the p-character in C-C bonds is the sum of the three internal angles.45 In a purely tetrahedral arrangement this would be 328.50. The sum of the angles around each of the atoms was calculated to be C2 = 2730

C3 = 2840

C9 = 3190.

In addition to being closer to C , and having a better overlap angle
9
for potential bond formation, C also has slightly more p-character than C
3





61


TABLE XII. Comparative 13Cnmr Chemical Shift Values (ppm)


OH
i
a h
f
b
g e

c d


26


a
b
C
h
g
f
e
d
i


OH

b




A c e d


27


22.5 19.7
24.2 48.9 40.9 39.1
34.1 34.7 78.2


a = 29.5 e = 31.3 b = 42.2 c = 39.7 d = 38.4 i = 86.4


Overall range of chemical shift
29.2 12.7


6
A c for
Rearrangement

7.0 9.8 7.1
-6.7
1.3 0.6 5.6 3.7 8.2


values (excluding i):





62


Finally, the internal C -C 3-C angle of (26)-OPNB which eventually becomes the bridge bearing the leaving group in (27) was already 1200.

The most logical conclusions that can be drawn from these bits of information give insight into the nature of the delocalized cation

(28). The ground state geometry of (26) is ideal to give with the




I'




28

least motion, ion (28), where there is likely to be more bonding between C2 and C than between C2 and C 3, or C3 and C . Cation (28) looks exactly like the delocalized ion one would expect from ionization of rearranged product (27). The rate of solvent attack on non-classical ion (28) to give (27)-OH or internal return to give (27)-OPNB, i.e., the product forming step, being much faster than the rate determining step, is also the step in which relief of the skeletal strain must be occurring. The relief of strain on going from (28) to (27) apparently increases the activation energy for the reverse process. This would account for the slower rate of solvolysis of (27)-OPNB. In addition to this, the energy barrier marked by the transition state from (26) to (28) is added making the production (26) from solvolysis of (27)OPNB very unlikely. All of this is compatible with the results of the product study in which no (26)-OH was ever detected from solvolysis of

(27)-ODNB, or -OPNB.





63


On the basis of the observed effect of the methylene bridge in

(26), reducing the favorable geometry for participation and hence reducing the rate acceleration with respect to (8), it seems very likely that the effect of the zero bridge between C5 and C, which changes (26) into (17) is even more detrimental to orbital alignment, leading to the further decrease in acceleration as had been observed.

The solvolytic behavior of (26)-OPNB has justified the effort

expended in developing an efficient synthetic route for its preparation and the thorough study of all aspects of its geometry. The title compound has suggested that the alignment of the cyclopropane edge orbitals with the site of developing positive charge could well be the major factor which governs the extent of anchimeric assistance and hence the observed rate acceleration. The need for further studies by way of 13Cnmr comparisons and x-ray structural determinations is apparent.















CHAPTER IV


EXPERIMENTAL

A. Synthesis

Infrared spectra were recorded on a Perkin-Elmer Model 621 or

137B spectrophotometer using sodium chloride optics. Solution spectra were recorded on a Beckman IR-10 using matched 0.1 mm sodium chloride cells. The absorption band positions reported hereafter are given in wave numbers (cm-1).

Nuclear magnetic resonance spectra were recorded on either a Varian Associates Model A-60, 60 MHz spectrophotometer, or a Varian Model XL-100, 100 MHz instrument. Chemical shift values for 1 Hnmr are reported in 6 units relative to tetramethylsilane (TMS) at 60.00. The chemical shift values of 13Cnmr are reported in ppm from TMS.

Melting points are uncorrected and were obtained with a Thomas Hoover melting point apparatus.

Mass spectra were recorded on an Associated Electronic Industries (AEI) Model MS-30 mass spectrometer at 70 eV. Accurate mass determinations were obtained using the same instrument linked with an auxiliary PDP-8 digital computer.

Microanalyses were performed by Atlantic Microlabs, Inc., Atlanta, Georgia.

Gas-liquid partition chromatography (glpc) was performed with a Varian Associates Model P1440 chromatograph utilizing flame


64





65


ionization detection. The columns employed were:

a) 100 ft. x 0.01 in. capillary coated with diethylene

glycol succinate (DEGS),

b) 100 ft. x 0.01 in. capillary coated with UCON LB-550, and

c) 100 ft. x 0.03 in. capillary coated with DEGS.

Integration of analytical glpc tracings was obtained with an Autolab 6300 digital integrator made by Spectra-Physics, Inc.

The x-ray data were provided by Dr. G. J. Palenik of the University of Florida Center for Molecular Structure. The x-ray diffraction data were collected on a Syntex automatic four-circle diffractometer using Cu Ka radiation (0< 2<1100) to give 1421 significant reflections.


Preparation of 1,2,3,4-Tetrachloro-5,5-dimethoxycyclopenta-l,3diene (29)

In a typical run, 640 grams (2.35 moles) of hexachlorocyclopentadiene (Aldrich H 600-2) were mixed with 475 ml of AR methanol in a three-liter three-necked flask fitted with a mechanical stirrer, addition funnel, and reflux condenser. To this mixture was added a solution containing 342 grams (6.05 moles) of potassium hydroxide in 1400 ml of methanol. The rate of addition was regulated so as to maintain gentle reflux. The mixture was stirred overnight at room temperature and diluted with 1500 ml of cold water to dissolve precipitated potassium chloride. The product was separated and the aqueous phase extracted with 3 x 150 ml of methylene chloride. The combined organic phases were washed with water and dried over anhydrous magnesium sulfate. The solution was filtered, the methylene chloride removed by distillation at atmospheric pressure, and the residue





66


vacuum distilled to yield 450 grams (70%) of a pale yellow liquid, bp 118-120'/8 torr. The iHnmr spectrum (CDCl3) contained only a sharp singlet at 63.33.19


Preparation of 1,2,3,4-Tetrachloro-5-endo-(1,1-diethoxymethyl)-7,7dimethoxybicyclo[2.2.l]hept-2-ene (30)

A solution of 180 grams (0.65 mole) of (29), 100 grams (0.77 mole) of acrolein diethyl acetal (Aldrich 2400-1), and 0.1 gram of diphenylamine was heated with stirring under reflux for 42-48 hours at 130-1350C. At the end of this time the reaction mixture was cooled until it could be handled comfortably and then poured into a beaker. Approximately 50 ml of methanol were added and the mixture cooled to ice bath temperature. In this manner the title compound was crystallized in sufficient purity for further reaction. Filtration was followed with washings consisting of 2 or 3, 15 ml portions of ice cold methanol. Highest purity material could be obtained by a single recrystallization from methanol. The average yield of pure solid, mp 78-78.5*C, was 315 grams (80%).

The IHnmr spectrum (CDCl3) exhibited signals for 64.35 (d,l; J = 4.5 Hz; HC(OC2H5)2); 3.62 (s,3; syn CH30); 3.55 (s,3; anti CH3 0); 3.55 (m,4; CH3 2 0); 2.92 (m,l; J = 5.0 and 9.0 Hz; 5-exo); 2.39 (d of d,l; J = 9.0 and 12.0 Hz; 6-exo); 2.01 (d of d,l; J = 12.0 and 5.0 Hz; 6-endo); and the non-equivalent ethoxyl methyls centered at 1.2 (d of t,6).

Anal. Calcd. for C14H20C1 404: C, 42.61; H, 5.08; Cl, 36.00.

Found: C, 42.75; H, 5.20; Cl, 35.98.





67


Preparation of endo-5-(1,1-diethoxymethyl)-7,7-dimethoxybicyclo[2.2.1]hept-2-ene (31)

In a typical dechlorination,20 194 grams (0.49 mole) of (30) were mixed with 240 grams tert-butanol, 1420 ml of tetrahydrofuran, and 150-175 grams of finely chopped sodium metal. The mixture was heated under reflux with efficient mechanical stirring for 8 hours, and then allowed to stir overnight at room temperature. The excess sodium metal was removed by pouring the thick reaction mixture through a stainless steel mesh screen. The sodium-free mixture was stirred with 400-500 grams of ice and enough additional water to cause a heavy granular mass of salts to settle out. Filtration, and removal of excess solvent yielded a thick slurry. Water was added until the organic layer separated cleanly. The product was separated, dried with anhydrous magnesium sulfate, and distilled. The average yield for the colorless liquid, bp 90*/0.3 torr, was 112 grams (90%). Chromatographic analysis (glpc, column a) indicated that the product was 98% pure. The IHnmr spectrum (CDCl3) exhibited signals for: 66.14 (m,2; HC=C); 4.00 (d,l; HC(OC2H5)2); 3.55 (m,5; CH3C2HO and H); 3.21 (s,3; syn CH30); 3.15 (s,3; anti CH30); 2.8 and 2.6 (m,2; 5-exo, H, and H4 overlapping); 2.15 (m,l; 6-exo); 1.18 (d of t,6; CH3CH20);

0.82 (q,l; 6-endo).

Anal. Calcd. for C H 240 : C, 65.60; H, 9.42.

Found: C, 65.67; H, 9.50.

Preparation of endo-bicyclo[2.2.l]hept-2-en-7-one-5-carboxylic
acid-(33)

The synthesis of (33) was accomplished in several steps without purification of the intermediate compounds. In a typical preparation, the acetal (65 grams, 0.25 mole) was hydrolyzed by stirring at room





68


temperature with about 2 liters of 10% hydrochloric acid for 1 hour, and then extracting with ether. The dried ether extract was evaporated to give a yellow oil. The infrared spectrum of this oil has a strong carbonyl absorption band at 1720 and a smaller one at 1780 cm1. The IHnmr spectrum (CDCl3) showed the absence of all ethoxyl proton resonances while a new one proton singlet was observed at 10.65. The ketal methoxyl protons were still present in the nmr spectrum although the infrared indicated a small amount of bridge hydrolysis had occurred. This oil was dissolved in dry acetone and oxidized at 0*C with Jones reagent21 until the calculated amount of reagent had been added (35 ml). At this point, the color change from orange to green was very slow (requiring more than 15 minutes). The acetone solution was decanted, diluted with an equal volume of water and extracted with ether. The ether solution was extracted with several portions of 10% sodium carbonate. The basic aqueous extracts were acidified with 30% sulfuric acid. The solution was stirred at room temperature for 2 hours, filtered, and extracted with methylene chloride. The organic extracts were dried with anhydrous sodium sulfate, filtered, and evaporated to yield a yellow solid. Recrystallization from hexane gives a pale yellow solid, mp 104-105%C, weighing 25 grams (63% yield from the acetal). The infrared spectrum shows carbonyl absorption bands at 1780 and 1710, and a 2500-3500 region showing the broad absorption typical of carboxylic acids. The lHnmr spectrum (CDCl3) exhibited signals for: 610.67 (s,l; COOH); 6.55 (m,2; HC=C); 3.3-2.9 (m,3; H1, H4, and 5-exo); 2.25 (m,1; 6-exo); 1.67 (d of d,1; 6-endo). Anal. Calcd. for CH 80 3: C, 63.15; H, 5.30.

Found: C, 63,13; H, 5.26.





69


Preparation of endo-anti--bicydlo[2.2Llhept-2-en-7-ol--5-carboxylic
acid *(34)

A solution containing 6.0 grams of sodium borohydride and 1.0 gram of sodium hydroxide in 50 ml of distilled water was stirred magnetically at room temperature while a solution of 57 grams of (33) and 24 grams of sodium hydroxide in 400 ml of distilled water was added over a period of 4 hours.22 The yellow solution was stirred overnight and then acidified with 10% hydrochloric acid. The acidic solution was extracted with 3 x 200 ml portions of ether. The ether solution was washed with water, dried over anhydrous sodium sulfate, filtered, and evaporated to give a colorless to pale yellow syrup (95-100%). The compound was considered pure enough for further reaction and was used without additional purification.. The infrared spectrum contained a broad carbonyl absorption at 1700 and typical carboxylic acid absorptions obscuring the 2500-3500 cm~1 region. The 2Hnmr spectrum (CDCl3) exhibited the following signals: 66.05 (m,2; HC=C); 3.50 (broad s,l; HCOH); 2.94 (m,l; H); 2.58 (m,l; HI); 2.5-1.0 (broad m,3; 5-exo, 6-exo, and 6-endo); and two resonances which varied their chemical shift value upon change in concentration and were removed upon shaking with D20.


Preparation of endo-anti-7-Acetoxybicyclo[2.2.1]hept-2-ene-5-carboxylic
acid (35)

Method A

A mixture containing 57 grams (0.37 mole) of the alcohol (34), 180 ml (1.7 moles) of acetic anhydride, and 60 grams (0.73 mole) of anhydrous sodium acetate was heated on the steam bath for 1 hour and cooled. To this mixture was added, cautiously, 54 ml of water, and the





70


resulting solution was carefully reheated.for 1 additional.hour.24 After cooling, the solution was diluted with an equal volume of water and extracted with 5 x 200 ml portions of methylene chloride. The extracts were washed free of excess acids and dried over anhydrous sodium sulfate. Evaporation of the solvent left a yellow semisolid which on recrystallization from cyclohexane gave 40 grams (55%) of white crystals, mp 103-104'C. The infrared spectrum (KBr) showed C=0 1735 (ester), 1690 (carboxyl), and C-0 1240 cm-2 (acetate). Y Y
The IHnmr spectrum (CDCl3) exhibited the following signals: 610.75 (broad s,l; COOH; exchanges with D20); 6.10 (m,2; HC=C); 4.42 (broad s,l; HCOCOCH3); 3.15 (m,2; H4 and 5-exo); 2.82 (broad m,1; H,); 2.05 (s,3; COCH3); 1.5 (d of d,l; 6-endo). Anal. Calcd. for C10 H 120 4: C, 61.21; H, 6.16.

Found: C, 61.34; H, 6.19.


Method B

A solution containing 58.0 grams (0.376 mole) of (34) in 100 ml of pyridine was stirred at room temperature while 100 grams of acetic anhydride was added dropwise over a 2-hour period.25 The solution was allowed to stir at room temperature overnight. An approximately equal volume of crushed ice was added and the semisolid so obtained was separated by filtration. Washing on the filter with 10 ml of ice cold ether gave 27 grams (37%) of a white solid, mp 102-103*C. This material was identical in every respect with that obtained by method A.





71


Preparation of endo-anti-7-Acetbxybicylo[2.2.l]hept-2-en-5-darboxylic
acid chloride (36)

Method A

A solution of 9 ml (0.1 mole) of oxalyl chloride in 50 ml of anhydrous ether was stirred magnetically under nitrogen at ice bath temperature while a solution containing 13 grams (0.08 mole) of (35) and 8 ml (0.1 mole) of pyridine in 100 ml of anhydrous ether was added dropwise over a period of 2 hours. The resulting mixture was allowed to warm to room.temperature with continued stirring and nitrogen flushing. The mixture was diluted to twice its volume with additional anhydrous ether and filtered. The filtrate was evaporated to yield a pale yellow oil which was used as obtained. The infrared spectrum showed acid chloride carbonyl at 1790, ester carbonyl at 1735, C-0- at 1240, and -C=C- at 3300 cm~1. The 'Hnmr spectrum (CDC13) exhibited signals for: 66.25 (apparent octet, 1; H3); 6.00 (apparent octet, 1; H2); 4.45 (broad s,l; H7); 3.5-3.3 (broad m,2; H partially overlapping 5-exo); 2.9 (m,l; H1); 2.10-2.00 (m,4; 6-exo overlapping COCHO3); 1.57 (d of d,l; 6-endo).


Method B

A solution containing 9.0 grams (0.046 mole) of (35) was added to a chilled solution of 10 ml of oxalyl chloride (14 grams, 0.11 mole) in ether. The addition required 2 hours, after which the ice bath was removed. The solution was stirred at room temperature for an additional 30 minutes to be sure all carbon monoxide had evolved. The solvent was removed on a rotary evaporator. The residue (9.8 grams, 100%) had IR and nmr spectra identical to that previously obtained. This preparation results in a pyridine-hydrochloride-free product.





72


Preparation of Diazoketone (37)

A solution containing 16-17 grams of the acid.chloride was

added with stirring at ice bath temperatures to an ethereal solution of diazomethane.22 The addition required about an hour and upon completion was allowed to stir an additional hour with warming to room temperature. The solution was filtered and concentrated on a rotary evaporator at room temperature or below. The crude diazoketone was used as prepared without further purification, but could be chromatographically purified on Alumina (neutral Woehlm) using 1:1 ether-hexane as the eluant. Purification was accompanied by loss of nitrogen and an average loss of material of about 30%. The infrared spectrum showed -C=N=N at 2100 and diazoketone carbonyl at 1640 cm1. The original ester absorptions at 1735 and 1240 cm~1 remained unchanged. The IHnmr spectrum (CDCl3) exhibited signals for: 66.15 (m,l; HC=C); 5.90 (m,l; HC=C); 5.30 (s,l; HC=N=N); 4.30 (m,l; H7); 3.05 (m,l; H); 2.90 (m,l; HI); 2.9-1.5 (m,6; 5-exo, 6-exo, 6-endo, and COCH3). Preparation of Diazomethane

For 16-17 grams of acid chloride (36), the following procedure26 provided a solution approximPtely 0.25 molar in diazomethane, a twofold excess.

A solution of 12 grams of sodium hydroxide in 30 ml of water was chilled in an ice bath and 40 ml of ethylene glycol monomethyl ether were added. The solution was stirred magnetically and 600 ml of technical grade ether added. To this mixture was added 65 grams of Dupont Nitrosan. The mixture was stirred for 30 minutes and the.ice bath was replaced by a water bath. The ethereal solution of diazomethane was





73


distilled using an efficient condenser until the distillates were no longer brilliant yellow in color.


Preparation of 9-Acetoxytetracyclo[4.3.0.02,4.03, 8]nonan-5-semicarbazone
-(39)

Diazoketone (37) was dissolved in 650 ml of anhydrous tetrahydrofuran and mixed with an equivalent weight of electrolytic copper dust. The mixture was stirred magnetically and heated slowly until refluxing began. A spontaneous reaction occurred which was allowed to continue without further heating. When this reaction subsided, heating was continued for another hour. After the observation of spontaneous reaction, 5 ml samples were withdrawn from the flask and evaporated on a sodium chloride window. The residue was examined for the presence of the 2100 cm~1 absorption band in the infrared spectrum. Refluxing was continued until the sample no longer contained this absorption band. Complete reaction was observed to require from 1 to 24 hours with the best yields obtained at the shortest reaction times. When the reaction was complete, the mixture was cooled, filtered, and the filtrate evaporated. The resulting residue was boiled with three successive portions (150 ml each) of ether and discarded. The ether extracts were evaporated, weighed, and mixed with 1.5 grams of semicarbazide hydrochloride and 2 grams of sodium acetate per gram of residue.24 The crude semicarbazone was obtained by diluting the reaction mixture with an equal volume of water and extracting with methylene chloride. The organic layers were dried and evaporated to yield a gummy white solid which was recrystallized from chloroform-hexane. The white solid so obtained retains solvent and melts about 162*C. After drying under





74


vacuum for 24 hours, the product melts 183-184*C. The yield overall from the acid (35) ranged from 0 to 42%. The infrared spectrum contained -NH at 3200, acetate -C-0 at 1240, carbonyl at 1735, and aide I
-1
and II absorptions at 1580 and 1670 cm . The poor solubility of the compound in common deuterated solvents gave extremely dilute solutions and poor IHnmr spectra. Those resonances observable (CDCl3) were 65.5 (-NH ); 5.0 (H9); and 2.05 (s; COCH3). Anat. Calcd. for C12HN15303: C, 57.81; H, 6.06; N, 16.86.

Found: C, 57.66; H, 6.13; N, 16.63.


Preparation of CuO-Cu Catalyst

A solution containing 100 grams of copper sulfate pentahydrate in 350 ml of distilled water was stirred vigorously while 35 grams of pure zinc dust were added over 30-45 minutes.30 The solid was washed by decantation until washings tested sulfate free to barium hydroxide solution. The red solid was stirred with 250 ml of 5% hydrochloric acid for 1 hour and decanted. A second 250 ml portion of acid was added and the misture left overnight. The copper powder was filtered off, washed with water until filtrates were neutral to pH-Hydrion paper, and drained an additional 30 minutes over suction. The solid was transferred to a quartz tube and heated in a tube furnace at 250C until no more water was observed, then to 500-600C for 15-20 hours.29 The black solid was cooled, ground, bottled, and kept in a desiccator until used. Average yield, 27 grams.





75


Preparation of anti-9-Acetoxytetracyclo[4.3.0.02i4 .038]nonan-5-one
(38)

In a typical run 13.5 grams (0.063 mole) of the pyridinehydrochloride free acid chloride (36) in 250 ml of anhydrous ether were added to an excess of freshly distilled.ethereal diazomethane at 0*C. When addition was complete (about 1 hour) the ice bath was removed and the solution was stirred for 1 additional hour. Precipitated polymeric materials were removed by suction filtration and the filtrate was reduced in volume on a rotary evaporator. The last half of the solvent was removed at temperatures no higher than 25*C. The resulting diazoketone in 200 ml of tetrahydrofuran was added to 800 ml of cyclohexane. Approximately 100 ml of additional tetrahydrofuran were added to completely dissolve the ketone in case it oiled out. The CuO-Cu catalyst (17 grams) was added and the flask fitted with a long efficient reflux condenser. Two 250 watt Sylvania Industrial Infrared lamps29 were positioned on opposite sides of the flask 2 to 3 inches from the surface of the glass and slightly above the liquid level. After an average of 2 to 4 hours from the time the lamps were illuminated, the 2100 cm~1 band had completely disappeared from the infrared spectrum of the reaction mixture. The lamps were removed, the solution allowed to cool until it could be handled, and the catalyst removed by suction filtration. The solvent was removed on the rotary evaporator and the resulting yellow residue chromatographed on Woehlm neutral alumina using a solution of 20% petroleum ether (20-40*) in diethyl ether. The second fraction to be eluted was found to contain 10.6 grams (80% yield based on (35)) of pure ketoacetate.. This was suitable for semicarbazone preparation. Two succeeding runs employing





76


the CuO-Cu catalyst and pyridine-hydrochloride free acid chloride produced 72% and 77% yields, respectively. The cyclization is-sensitive to the amount of "catalyst" employed and the lower yield was obtained when less catalyst was used. The infrared spectrum (neat).contained the following absorption bands: 3050(w), 2980(m), 2880(w), 1730(s), 1375(m), 1240(s), 1205(m), 1050(s), 910(m), 870(m), 860(m), 830(m), and 740(m) cm~1. The 1Hnmr spectrum (CDCl3) consisted of 64.98 (broad s,l; H ) 2.67-1.52 (broad m,8); and 2.03 (s, 3; COCH ). '9 I'3 Preparation of anti-Tetracyclo[4.3.0.02,4.03,8 ]nonan-9-ol (26)

A solution containing 5.0 grams (0.020 mole) of (39), 10 grams of potassium hydroxide, and 2.5 ml of anhydrous hydrazine in 50 ml of ethylene glycol was heated to 135-140C in an oil bath and maintained at that temperature for 1 hour. The reaction was carried out in a distilling apparatus which contained a 10 cm vigreux column. After heating for 1 hour the temperature of the reaction mixture was raised slowly to 185-190*C and the water and other lower boiling materials were removed. The temperature was held at 190*C until solid could be observed subliming into the condenser. Reaction was continued until substantial material had collected in the condenser and the reaction mixture was cooled to room temperature. The sublimed material was rinsed from the apparatus with pentane and the solution set aside. An additional 10 grams of potassium hydroxide and 2.5 ml of hydrazine were added to the reaction flask, and the heating cycle was repeated. When additional product had again filled the condenser, the reaction was cooled, diluted with an equal volume of water, and transferred to a continuous liquid extractor. The mixture was extracted for 48 hours





77


with pentane and returned to the distillation apparatus. The water was removed and the whole process described above was repeated. When product was no longer observed subliming into the condenser, the reaction mixture was extracted as before and discarded. The combined pentane extracts and washings were evaporated to yield a white crystalline solid mp 145-146%C (with sublimation). The material obtained as described is sometimes pale yellow, but can be purified by sublimation at 70'/760 torr, or recrystallized from small amounts of pentane. Overall yield in this manner from (39) was 2.0-2.2 grams (75-80%). This is an adaptation of the method of Babiak.33 The infrared spectrum (KBr) contained the following absorption bands: 3150(s), 2950(w), 2880(m), 2800(w), 1260(m), 1080(s), 1070(s), 930(w), 840(w), 825(w), 790(m), 740(m) cm1. The aHnmr spectrum (CDCl3) consisted of: 64.35 (broad s,l; H ); 2.4-0.8 (broad m,10; all other methylene and methine protons); and the concentration dependent hydroxyl proton 1.82-1.72 (broad s,l). The 13Cnmr spectrum (CDCl ) contained nine resonances, consistent with the asymmetric structure of the carbon skeleton. The chemical shift values and assignments are given in Table III. Mass spectrum: 70 eV, m/e 136 (M+). The fragmentation pattern is given in Table IV.

AnaZ. Calcd. for C92H 0: C, 79.36; H, 8.88.

Found: C, 79.60; H, 8.80.


Preparation of anti-Tetracyclo[4.3.0.02,4.03, 8jnonan-9-yl pbromophenylurethan (26)-PBPU

To a solution containing 205 mg (1.50 mole) of (26) dissolved in 5 ml of anhydrous benzene was added 327 mg of para-bromophenyl





78


isocyanate. The mixture was warmed-for 5 minutes (steam bath) and the benzene evaporated.24 The residue was partially dissolved in boiling hexane, filtered, and cooled. Three additional recrystallizations with hot filtrations were required to produce a constant melting product, mp 129-130*C; yield 270 mg (50%). The sample was dissolved in ether, diluted with an equal volume of hexane in a dust-free container with a loose fitting top, and the solution allowed to evaporate slowly. The crystal clusters so produced were suitable for use in x-ray structural analysis.

Anal. Calcd. for ClH16N02Br: C, 57.50; H, 4.82; N, 4.19; Br, 23.91.

Found: C, 57.43; H, 4.89; N, 4.19; Br, 23.81.


Preparation of anti-Tetracyclo[4.3.0.02,4.0 3',]nonan-w9-yl pnitrobenzoate (26)-OPNB
24
Using standard procedures, 200 mg (1.47 mmole) of alcohol

(26) were dissolved in dry pyridine and to the solution was added 220 mg of freshly recrystallized p-nitrobenzoyl chloride. The mixture was swirled to mix thoroughly and stored in the refrigerator overnight. The cold mixture was diluted with 10 grams of ice and the oil thus deposited was extracted with ether. The ether extract was washed successively with water, 5% aqueous hydrochloric acid, 10% sodium carbonate, water, and saturated aqueous ammonium chloride. Evaporation of the solvent gave a yellow solid which was recrystallized from ethanolwater in cream-colored plates '(95% yield), mp 86-87*C. Recrystallization from hexane gave the compound as fine needles isp 87-88*C. The IHnmr spectrum (CDCl3) was undistinguished except for the appearance of aromatic protons at 68.24 and a downfield shift of Hi9 to 65.30.





79


Mass spectrum: 70 eV, m/e 285 (M+). Anal. Calcd. for C1H NO-: C, 67.36; H, 5.30; N, 4.91.

Found: C, 67.16; H, 5.39; N, 4.91.


Preparation of ant-Tetracyclo[4.3.0 02,4 .03,8]nonan-9-yl 3,5dinitrobenzoate (26)-ODNB

In the same manner as described for (26)-OPNB, 2.6 grams

(0.019 mole) of alcohol (26) dissolved in 10 ml of anhydrous pyridine were treated with 5.9 grams (0.026 mole) of 3,5-dinitrobenzoyl chloride. Upon swirling, a spontaneous exothermic reaction occurred. The resulting mixture, when it had cooled, was treated as before and the product isolated as a pale yellow solid by filtration. Recrystallization from ethanol gave 5.8 grams (88%) of pale yellow crystals, mp 118-119*C. This derivative was insoluble in 60-80% acetone-water solutions at room temperature. The infrared spectrum (KBr) exhibited carbonyl stretch at 1720 cm-1 as well as bands for -C-0 at 1280, and NO2 at 1540 and 1340 cm~1. The 1Hlnmr spectrum (CDCl3) contained proton resonances at 69.16 (m,3; aromatic); 5.34 (broad s,1; H9); and 3.0-1.0 (m,10). Anal. Calcd. for ClsH,4N20,- C, 58.18; H, 4.27; N, 8.48.

Found: C, 58.15; H, 4.27; N, 8.56.


Preparation of Tetracyclo[4.3.0.02,4 .0 3, 'nonan-9-one (45)

A solution of 272 mg (2.00 mmole) of alcohol (26) dissolved in 5 ml of methylene chloride was added in one portion to a solution containing 3.1 grams of chromium trioxide-dipyridine complex in 60 ml of methylene chloride.34 The mixture changed from blood red to black immediately. The mixture was allowed to stir for 1 hour at room temperature and then filtered and evaporated. The residue was extracted with





80


boiling ether and filtered to remove last traces of chromium polymers. The ether was evaporated to yield a pale.yellow semi-solid which sublimed readily at 65*/760 torr. The sublimed material (185 mg, 70%) was still sticky as if wet, but had mp 96-97*C (sealed tube). Analysis by glpc on columns a and b showed only one substance present when the injection port temperature was held below 140*C. When the injection port temperature was increased to 200*C, a new peak was observed at shorter (10 minutes compared with the ketone at 37 minutes, column b) retention times due to decarbonylation of the ketone in the injection port. The infrared spectrum (CC1 ) contained the following absorption bands: 3020(m), 2960(m), 2880(m), 1765(s), 1470(w), 1360(w), 1315(m), 1150(m), 1110(m), 1080(m), 820(s), 740(w) cm- . The 1Hnmr spectrum (CDCl3) consisted of a broad complex multiplet 0.8 to 2.56. Calculated Accurate Mass: 134.0731 mass units.

Found: 134.0734 mass units.


Preparation of Chromium Trioxide-dipyridine Complex

A one-liter three-necked flask was thoroughly dried and equipped for mechanical stirring under nitrogen. Eighty grams of chromium trioxide, which had been dried over phosphorus pentoxide under vacuum for 24 hours, were added in very small portions to 600 ml of anhydrous pyridine maintained at 15-20*C. The addition required 5 to 7 hours and at the end of that time the complex had precipitated. The red solid was isolated by decantation and washed by stirring with 5 x 150 ml portions of anhydrous ether with subsequent decantation.. The wet complex was transferred to a vacuum desiccator over phosphorus pentoxide and evacuated to 20 torr for 8 hours.- The dried complex was stored in





81


the same manner. For oxidation, the complex was dissolved in methylene chloride as a 5% solution and used in a 6:1 mole ratio to the alcohol. Oxidations were usually complete in 5-15 minutes at room temperature.34 Preparation of Bicyclo[3.2.l]octa-2,6-diene (46)

A sample of ketone (45) weighing 125 mg was heated in a sealed

tube at 200*C for 15 hours. The tube was opened and the contents taken up in ether. The ether solution was passed through a short alumina column and evaporated to give the diene which eluted as a single peak on columns a and b (glpc). The infrared spectrum (CHC1 ) was identical with that reported by other workers.35 The IHnmr spectrum (CDCl3) was also identical with that previously reported, and contained the following resonances: 66.28 (d of d,l; J1 7 = 2.8 Hz; H7); 6.08 (broad d of d,l; J2 3 = 9.8 ,Hz; H2); 5.76 (d of d,l; J6 7 = 5.5 Hz, J5 =

2.6 Hz; H6); 5.25 (broad d,l; H3); 2.68 (broad m,2; H, and H5); 2.36 (H ); 2.00 (d of d,l; Jab 9.5 Hz, Jla = 4.0 Hz; H 8a)

1.90 (J =ndo 18.0 Hz; H nd); 1.76 (d of d,l; J b = 0.6 Hz; H 8b).


Preparation of Tetracyclo[4.3.0.02,4.03,8]nonan-5-semicarbazone (43)

This compound was prepared according to the method of Nickon et al.27d and gave the desired product in 16% yield overall in eight steps from methyl acrylate and cyclopentadiene. Semicarbazone (43) had mp 205-206*C (literature value 204-205*C). The infrared spectrum (KBr) contained absorptions for -N-H at 3300, -NH2 at 3150 (doublet), and amide I and II bands at 1670 and 1585 cm~1, respectively.





82


Preparation of Tetracyclo[4.3.0.02, 4.03,- 8nonane (44.). Method:A32

A solution containing 5.0 grams (0.026 mole) of.semicarbazone

(43) in 5 nl of ethylene glycol was added all at once-to a solution of

5 grams of potassium hydroxide in 50 ml of ethylene glycol maintained at 130*C. The reaction mixture was contained in a 100 ml round-bottomed flask which was attached to a 10 cm vigreaux column topped by a thermometer and connected to a water-cooled condenser. This same apparatus was employed for all subsequent reductions of this kind. After the addition, the solution was heated to 1600C for 30 minutes and then to 190*C for 2 hours. The reaction mixture was cooled to 120C and the more volatile material removed under vacuum (4 torr) and the syrupy plastic residue cooled to room temperature. The cold amorphous mass was stirred with 250 ml of cold water and the white solid which precipitated was filtered off, washed with additional cold water, and dried. The solid was recrystallized from methanol-hexane to give 1.8 grams (53%) of white needles, mp 189-190*. The infrared spectrum (KBr) contained a -C=N absorption at 1650 cm-1 as well as aliphatic C-H bending and stretching bands. The IHnmr spectrum (CDCl3) was a series of complex overlapping multiplets between 61.4 and 3.5. The mass spectrum (70 eV) showed m/e 264 (M+). The compound was identified as bis(tetracyclo[4.3.0.02,4.038*]nonan)-5-azine. Anal. Calcd. for C H N C, 81.77; H, 7.62; N, 10.59.
18 20 2
Found: C, 81.87; R, 7.70; N, 10.51.


Method B33

A mixture containing 1.0 gram (0.0052 mole) of (43), 2.0 grams of potassium hydroxide, and 0.5 ml of hydrazine in 20 ml of ethylene





83


glycol was placed in the apparatus previously described and.heated slowly to 160*C. As the temperature rose above 135C, the product hydrocarbon was observed slowly subliming into the condenser. After 2 hours the reaction mixture was.cooled and the product washed from

the apparatus with ether. The ether solution was washed with water and dried over anhydrous sodium sulfate. Evaporation of the solvent produced a white waxy solid which sublimed readily at room temperature and atmospheric pressure. The product, obtained in 60% yield, had mp 116-117* (completely submerged sealed tube). The iHnmr spectrum (CDCl3) exhibited resonances at 62.55 (broad s,l; HI); 2.35 (broad s,l; H1); and 0.8-2.0 (complex multiplet, 10). The 13Cnmr shift values and assignments are consistent with the expected hydrocarbon product and are listed in Table III. The mass spectrum (70 eV) had m/e 120 (M+, 17%), 105 (12t), 92 (15%), 91 (17%), 79 (10%), 78 (17%), 77 (16%), and 66 (100%).

Anal. Calcd. for C9H12: C, 89.93; H, 10.07.

Found: C, 89.91; H, 10.08.


Preparation of anti-Tetracyclo[3.3.1.02,4.03,7]nonan-9-ol (27)

A mixture containing 2.0 grams (0.0070 mole) of (26)-OPNB in

100 ml of 70% acetone-water was refluxed for 30 hours and cooled. The acetone was removed on the rotary evaporator and the resulting mixture diluted- with 70 ml of water in which 2.0 grams of sodium hydroxide were dissolved. The mixture was heated to-reflux for 1 hour, cooled to room temperature, and extracted with 5 x 50 ml portions of pentane. The pentane extracts were dried and evaporated to yield a.yellow.semisolid. Sublimation of this material gave 0.9 gram (94%) of white





84


crystalline product, mp 236-237* (completely submerged.sealed tube). Analysis by glpe of the product on column a revealed only one peak eluting after 22 minutes, with no evidence for alcohol.(26) which

normally eluted at 27 minutes. The infrared spectrum (KBr) contained the following absorption bands: 3200(m), 3040(w), 2940(m), 2920(w), 2840(w), 1340(m), 1060(m), and 1030(m) cm-1. The 1Hnmr spectrum (CDCl3) exhibited resonances at: 64.0 (t,l; J = 5.0 Hz; .H); 2.4 (broad m,2; Hi and H5); and 2.4-1.2 (complex m,9), The 13Cnmr contained six resonances, consistent with the symmetrical carbon skeleton. The chemical shift values and assignments are listed in Table III. The same alcohol was also prepared in greater than 90% yield by solvolysis of (26)-ODNB in 60% acetone-water in a sealed tube at 125*C for

6 hours.

Accurate Mass Anal.' Calcd. for C9H120: 136.0887 mass units.

Found: 136.0898 mass units.


Preparation of anti-Tetracyclo[3.3.1.02,4 .0,7 nonan-9-yl pnitrobenzoate (27)-OPNB

In the manner previously described, 200 mg (1.47 mmole) of

alcohol (27) were dissolved in dry pyridine, and 300 mg of p-nitrobenzoyl chloride were added. Standard workup and recrystallization from petroleum ether gave .350 mg (84%) of pale yellow plates, mp 143-144*C. Anal. Calcd. for C A NO 4: C, 67.35; H, 5.30; N, 4.91.

Found: C, 67.50; H, 5.38; N, 4.83.


Preparation of anti-Tetracyclo[3.3..0 ' .031,7]nonani9-yl 3,5dinitrobenzoate (27)-ODNB

In the manner previously described, 1.4 grams (0.010 mole) of alcohol (27) were dissolved in 10 ml of pyridine and 3.5 grams (0.016





85


mole) .of 3,5-dinitrobenzoyl.chloride were.added. Standard workup followed by recrystallization from ethanol-chloroform gave .very fine yellow needles, mp 173-174*C. The 1Hnmr spectrum (CDCl3) exhibited resonances at 69.17 (m,3; aromatic); 5.07 (t,1; H9); 2.75 (broad s,3; Hi, H., and H7); 2.4-1.6 (broad m,.7). Anal. Calcd. for C 1614 N206: C, 58.18; H, 4.37; N, 8.48.

Found: C, 58.17; H, 4.31; N, 8.47.


Preparation of Tetracyclo[3.3.1.02, 4.03,7]nonan-9-one (47)

A solution containing 200 mg (1.47 mmole) of alcohol (27) in

methylene chloride was oxidized with a 5% solution (60 ml) of chromium trioxide-dipyridine complex in methylene chloride. The reaction mixture was allowed to stand at room temperature for 10 minutes, filtered, and the solvent evaporated. The resulting semi-solid was sublimed to give 140 mg (70%) of colorless crystals, mp 161-162*C. Analysis by glpc on column a revealed only one peak at about 32 minutes. Injection port temperatures above 300*C were required to produce a tracing showing as much as 1-2% of the decarbonylation product, which eluted at 4 minutes. The infrared spectrum (CHCl ) contained absorption bands at 3055(w), 2945(m), 2860(m), 1725(s), 1440(w),,1315(w), 1060(w), 1020(w) cm~1. The 1Hnmr spectrum (CDCl3) resembled that of alcohol

(27) with the exception of the absence of the proton signal for H9 and the slight shift of the bridgehead multiplet from 62.4 to 2.6. Calculated accurate mass for C9H110: 134.0730 mass units.

Found: 134.0731 mass units.

A-ketone identical in all respects was prepared in 85% yield by treatment of alcohol (27) with Jones reagent.





86


Preparation of Tetracyclo[3.3.1.O',"' -0 7 ]nonane (49).

Alcohol.(27): (136 mg, 1.00 mmole) was dissolved in 0.5 ml of

thionyl chloride at room temperature and stirred for 3 hours. Excess thionyl chloride was removed on a rotary evaporator to give a white solid. Chloride (48) was dechlorinated without further purification by refluxing in 5 ml of tetrahydrofuran containing 0.2 ml of t-butanol and 0.1 gram of sodium metal for 6 hours. Workup of the dechlorination reaction gave a white semi-solid. Sublimation of the crude product gave 85 mg (70%) of the hydrocarbon "triaxane" (49). The hydrocarbon is very volatile and evaporates rapidly at room temperature. "Triaxane" had mp 180-181* (completely submerged sealed tube) (literature value 183-184.5*C).37 The 1Hnmr spectrum (CDCl3) exhibited




--b


c a



resonances at: 62.46 (broad s,3; W = 11 Hz; J = 1.5 Hz, J = bc be
5.0 Hz; H ); 2.01 (m,3; H ); 1.61 (m,3; J =10.5 Hz; R ); 1.28
b c ae e
(d,3; H ). The 13Cnmr spectrum (CDCl3) contained resonances at 47.1, 40.3, and 37.1 ppm relative to TMS (see Table III) which was consistent with those previously reported.37,38


Preparation of 7-Norbornadienyl p-nitrobenzoate (12)-OPNB
53
This ester was prepared using the literature procedure. . The product was recrystallized from petroleum-ether-carbon tetrachloride as long pale yellow needles, mp 104-105* (literature value 101-102*C) in 95% yield.





87


Anal. Called. for C H NO : C, 65.36; H, 4.31; N, 5.45.

Found: C, 65.27; H, 4.36; N, 5.45.


Preparation 6f 7-Norbornadienyl 3,5-dinitrobenzoate (12)-ODNB

In the manner previously described, 2.16 grams (0.020 mole) of norbornadienol were dissolved in 10 ml of dry pyridine and 5.9 grams (0.026 mole) of 3,5-dinitrobenzoyl chloride were added. Standard workup followed by recrystallization from petroleum ether-carbon tetrachloride gave 3.9 grams (65%) of yellow needles, mp 128-130* (dec.). The infrared spectrum (KBr) contained absorption bands at 3000(m), 1720(s), 1545(m), 1345(m), and 1280(s) cmuF. The 1Hnmr spectrum (CDCl3) exhibited resonances at: 69.16 (apparent t,3; aromatic); 6.80 (d of t,4; HC=C); 4.90 (broad s,l; H7); 3.83 (m,2; H1 and H4).

Anal. Calcd. for C14 H N 0 : C, 55.63; H, 3.34; N, 9.27.
4 02 6
Found: C, 55.55; H, 3.39; N, 9.33.


Preparation of anti-7-Norbornenyl p-nitrobenzoate (2)-OPNB

In the manner previously described, 1.0 gram (0.0091 mole) of

alcohol (2) was dissolved in 5 ml of dry pyridine and 2.8 grams (0.015 mole) of recrystallized p-nitrobenzoyl chloride were added. Standard workup followed by recrystallization from ethanol gave 2.1 grams (84%), of pale yellow crystals, mp 122-123* (literature value 118-119C). The 1Hnmr and infrared spectra were consistent with those reported in the literature.54





88


Preparation of anti-7-jQrbornenyl 3,5-dinitrobenzoate:(2)-ODNB

This ester was prepared in exactly the same manner-as-(2)-OPNB.

Recrystallization from ethanol gave 2.5 grams (85%) of very fine creamcolored needles, mp 134-135*C. The IHnmr spectrum (CDCl 3) exhibited resonances at: 69.18 (m,3; aromatic); 6.35 (t,2; HC=C); 4.67 (broad s,l; H7); 2.98 (broad m,2; Hi and H); 1.93 (m,2; H ); 1.23 (m,2; exo
H ).
endo
Anal. Calcd. for C,1 112N20 6: C, 55.26; H, 3.98; N, 9.21.

Found: C, 55.20; H, 4.02; N, 9.23.


B. Kinetic Studies

Preparation of Kinetic Solutions Preparation of standard sodium hydroxide

The standard sodium hydroxide titrant was prepared by diluting a 0.1 N Acculute solution to 1 liter. As supply demanded, 100 ml aliquots of this solution were diluted to 1 liter and standardized against primary standard grade potassium hydrogen phthalate using a phenolphthalein endpoint. The normality of the solutions so prepared ranged from a low of 0.01003 to 0.01025 N. Preparation of phenolphthalein indicator solution

The indicator solution for p-nitrobenzoate ester kinetics was prepared by dissolving 1 gram of phenolphthalein in 60 ml of ethanol and diluting with carbonate free water to 100 ml.


Preparation of acetone:water solutions

'Carbonate free deionized water was prepared by distilling the water and passing it through an Amberlite MB-3 ion exchange resin.55





89


Reagent grade.acetone was stirred with 1 gram of potassium permanganate and 10 grams of anhydrous calcium sulfate for 2 days. The solvent was filtered, distilled, and stored over molecular sieve.56

Appropriate volumes of acetone and water were mixed and stored under nitrogen pressure. The density of the solutions was evaluated using a Kimax pyncnometer. The 70% acetone:water mixture had a density of 0.9113 grams/ml. The 60% acetone:water mixture had a density of

0.9397 grams/ml.


Preparation 'of decarbonylation materials

The solvent used for ketone decarbonylation kinetics was chosen because of its long glpc elution time relative to the products under study. For this purpose 1,2,4-trimethylbenzene was distilled and the chromatographically pure fraction used as the solvent.

Marker compounds were chosen for retention times and molecular weights. Benzene, toluene, and m-xylene met these requirements. The reagent grade materials were distilled until a fraction showing only one glpc peak was obtained.


Kinetic Procedures

p-Nitrobenzoate and 3,5-dinitrobenzoate esters

A carefully weighed sample of the ester was dissolved in acetone: water solution and made up to 100 ml at 20*C. With a hypodermic syringe equipped with a 6-inch needle, 5.2 ml samples were divided among 18 medium wall-glass ampules which had been flushed with nitrogen. Each tube was frozen in liquid nitrogen, pumped to 0.5 torr, and sealed. The tubes.were divided into two sets. The nine tubes of a set were




Full Text

PAGE 1

THE SYNTHESIS, CHEMISTRY, AND SOLVOLYTIC BEHAVIOR OF TETRACYCLO[4-.3.0.02,4 .03 ,8]NONAN-9-0L, AND RELATED COMPOUNDS By DONNA DEMBAUGH McRITCHIE A DISS ERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE mUVERSITY OF FLORIDA IN PARTIAL FULFI LLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY .. UNIVERSITY OF FLORIDA 1975

PAGE 2

To Ruth

PAGE 3

ACKNOWLEDGMENTS The author wishes to acknowledge the creative contributions and helpful discussions of Dr. Merle A. Battiste during the course of this r esearch. She is especially grateful for his uncanny ability to be available when needed and yet nonintrusive when not needed. Sincerest thanks are due Dr. Rocco Fiato, Dr. Warren Nielsen, John Timberlake, Dick Galley, and Bob Posey for their suggestions, advice, humor, and most importantly for their friendship, freely given, without which the author's circumstances would have been far less enjoyable. A large measure of thanks to Judy Romanik and Judy Nielsen for being there when the author needed someone to giggle or cry with. Finally, and perhaps most importantly, the author acknowledges her husband, John, for his love, unflagging support, and unfailing assistance over all the little rough places. iii

PAGE 4

TABLE OF CONTENTS Page ACKNOWLEDGMENTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • iii L I ST OF TABLES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • vii LIST OF FIGURES........................ . . . . . . . . . . . . . . . . . . . . . . . . . . viii ABSTRACT. . . • • . • . . . . . . . . . • . • • . . . • . . . . . . . . • • . . . . . . . . . . . . . . . . . . . . . . • i x INTRODUCTION. . • . . . . . . . • . . . • . . . . • • . . . . • . . . • . • • • • . . . . • . . . • • • . . • • . • • 1 SYNTHESIS AND CHEMISTRy.......................................... 14 KINETIC AND PRODUCT STUDIES...................................... 35 The Stereoche mistry of the Solvolysis and the Nature of the Carbonium ,Ion. . • . • • . . . . . . . . . • . • . . • . . • . • . . . . . •• . . . . . . . • • 46 Geometrical Considerations. ..•......•............•.•...•...••. 48 Effects of Strain............................................. 56 EXPER IMENTAL. . . • . . . • . • . . • . . . . . . • . . . . . . . . . • • . . . . . . . . . . . • • . . . . . . . . . 64 A. Synthesis. . . . . • . . . • . . . . . . . . . . . . . . . . . . . . . . . . • . . . . • . . . . • . . . • 64 Preparation of 1,2,3,4-Tetrachloro-5,5-dimethoxycyclopenta-l,3-diene (29)........................ 65 Preparation of 1,2,3,4-Tetrachloro-5-endo-(1,1-diethoxymethyl)-7,7-dimethoxybicyclo[2.2.1]hept-2-ene (30).................................................. 66 Preparation of endo-5-(1,1-diethoxyrnethyl)-7,7-dimethoxybicyclo[2.2.l]hept-2-ene (31).................... 67 Preparation of endo-bicyclo[2.2.l]hept-2-en-7-one-5-carboxylic acid (33). ............................... 67 Preparation of endo-anti-bicyclo[2.2.1]hept-2-en-7-ol-5-carboxylic acid (34).......... ... ...... ............ 69 Preparation of endo-anti-7-Acetoxybicyclo[2.2.1]hept-2-ene-5-carboxylic acid (35)......................... 69 iv

PAGE 5

Page Preparation of endo-anti-7-Acetoxybicyc1o[2.2.1]hept-2-ene-5-carboxylic acid chloride (36)..................... 71 Preparation of Diazoketone (3?)........................... 72 Preparation of Diazomethane................... ........ .... 72 Preparation of 9-Acetoxytetracyc1o[4.3.0.02,4.03,B]nonan-5-semicarbazone (39)... .................... ..... .... 73 Preparation of CuO-Cu Catalyst............................ 74 Preparation of anti-9-Acetoxytetracyc1o [4.3.0.02,4.03,B]nonan-5-one (38)............... ........•. 75 Preparation of anti-Tetracyc1o[4.3.0.02,4.03,B]nonan-9-o1 (26)........................................... 76 Preparation of anti-Tetracyclo[4.3.0.02,4.03,8]nonan-9-y1 p-bromopheny1urethan (26)-PBPU.......... ... .... 77 Preparation of anti-Tetracyc1o[4.3.0.02,4.03,B]nonan-9-y1 p-nitrobenzoate (26)-OPNB... ....... ....... ..... 78 I Preparation of anti-Tetracyc1o[4.3.0.02,4.03,8]nonan-9-y1 3,5-dinitrobenzoate (26)-ODNB......... ......... 79 Preparation of Tetracyc1o[4.3.0.02,4.03,B]nonan-9-one (45).......................................... 79 Preparation of Chromium Trioxide-dipyridine Complex ...•............................................... 80 Preparation of Bicyclo[3.2.1]octa-2,6-diene (46). ... ... ... 81 Preparation of Tetracyc1o[4.3.0.02,4.03,B]nonan-5-semicarbazone (43)........ ........ .... .......•.... 81 Preparation of Tetracyc1o[4.3.0.02,4.03,8]nonane (44)............................................... 82 Preparation of anti-Tetracyc1o[3.3.1.02,4.03,7]nonan-9-o1 (27)........................................... 83 Preparation of anti-Tetracyc1o[3.3.1.02,4.03,7]nonan-9-y1 p-nitrobenzoate (2?)-OPNB.... .........•.......• 84 Preparation of anti-Tetracyc1o[3.3.1.02,4.03,7]nonan-9-y1 3,5-dinitrobenzoate (2?)-ODNB........... •...... 84 v

PAGE 6

Page Preparation of Tetracyclo[3.3.l.02,4.03,7]-nonan-9-one (47).......................................... 85 Preparation of Tetracyclo[3.3.l.02,4.03,7]-nonane ( 49) . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 Preparation of 7-Norbornadienyl p-nitrobenzoate (12)-OPNB................................................. 86 Preparation of 7-Norbornadienyl 3,5-dinitrobenzoate (12)-ODNB .............•...........•....................... 87 Preparation of anti-7-Norbornenyl p-nitrobenzoate (2)-OPNB.. .. . ... ... ... .... .. . .. .. ..... .......... ... .... . .. 87 Preparation of anti-7-Norbornenyl 3,5-dinitrobenzoate (2)-ODNB..... ............................• 88 B. Kinetic Studies........................................... 88 Preparation of Kinetic Solutions.......................... 88 Kinetic Procedures................ . • . . . . . . . . . . . . . • . . . . . . . • 89 Analysis of Data.......................................... 91 C. Solvolysis Product Studies ................................ 110 anti-Tetracyclo[4.3.0.02,4.03,s]nonan-9-yl p-nitrobenzoate (26)-OPNB................................... 110 anti-Tetracyclo [3.3.1. 02,4.03,7 ]nonan-9-yl esters (27)-OPNB and (27)-ODNB................................... 112 Auxiliary Experiments..................................... 114 APPENDIX. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . 116 BIBLIOGRAPHY. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 BIOGRAPHICAL SKETCH.............................................. 128 vi

PAGE 7

LIST OF TABLES Table Page I. Comparative Rates in the Bicyclo[2.2.l)heptan-7-yl Series, and the Tricyclo[3.2.l.02,4)-oct-8-yl Series....... 5 II. Comparative Rates in the Basic anti-Bicyclo[2.l.1) Series. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 III. 13CNMR Chemical Shift Values (ppm)................ ......... 24 IV. Comparative Fragmentation Patterns (70 eV).......... ....... 25 V. Comparative Ketone Fragmentation Patterns..... ...... ....... 34 VI. Kinetic Results............................................ 38 VII. Kinetic Values for Structurally Related Systems............ 40 I VIII. Comparison of Relative Rates of p-Nitrobenzoate Esters in 70% Acetone:Water at 25C................... ............ 41 IX. Product Analysis........................................... 44 X. Comparative Rates of Nitrogen Extrusion... .............. •.. 53 XI. Rates of Ketone Decarbonylation... ...... ....•.............. 54 XII. Comparative 13Cnmr Chemical Shift Values (ppm).... ......... 61 Ia. Bond Lengths............................................... 116 IIa. Interatomic Angles............. .. . . ...... ... .. . . .. . . .. ..... 117 IlIa. Interatomic Distances...................................... 117 vii

PAGE 8

LIST OF FIGURES Figure Page 1. Molecular Structure of (26)-PBPU ......................... 28 II. Internal Plane Angles ............ . ..................•.... 51 Ia. Plot of log[C] Versus Time to 10 x t for ( 26 )-OPNB ...... 118 Ila. Plot of ln [C] Versus Time to 1 x tl for ( 2 6) -OPNB ....•.. 119 IlIa. Plot of ln [C] Versus Time to 10 x t for ( 27 )-OPNB ...... 120 k 2 IVa. Plot of log k Versus liT for (26)-OPNB and ( 27)-OPNB .•... 121 Va. Plot of log k Versus liT for (12)-OPNB and (12)-ODNB ..... 122 VIa. Plot of log k Versus liT for (2) -OPNB .................•.. 123 viii

PAGE 9

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE SYN T HESIS, CHEMISTRY, AND SOLVOLYTIC BEHAVIOR OF T E TRACYCLO[4.3.0.02,4.03,8]NONAN-9-0L, AND RELATED COMPOUNDS By Donna Dembaugh McRitchie August, 1975 Chairman: Dr. Merle A. Battiste Major Department: Chemistry Solvolysis studies of endo-anti-tricyclo[3.2.l.02,4]octan-8-yl esters have revealed an enormous rate acceleration compared to 7-norbornyl due to the stabilization afforded by the tris-homocyclo-propenyl cation. The products of the solvolysis were greater than 99% rearranged, which has led some investigators to rationalize the large acceleration as partly due to strain relief on ionization lead-ing to a classical ion with charge concentrated alternately on C2 and Studies of pentacyclo[4.3.0.02,4.03,8.0s,7]nonan-9-yl esters have also demonstrated accelerated solvolysis rates, but with no re-arrangement possible. Here there could be no doubt of the intermediacy of a tris-homocyclopropenyl ion and there was no strain relief upon going from reactant to product. The alignment of the cyclopropane edge bond orbitals with the site of developing positive charge could be a more important influence ix

PAGE 10

on the rate of reactivity than relief of ground state strain in the rate determining ionization. To resolve this question the tetracyclo[4.3.0.02,4.03,8]nonan-9-yl esters, which appeared to be more highly strained than those of the tricyclo series, were prepared and studied. The synthesis of the title compound was accomplished through the key intermediate endo-bicyclo[2.2.1]hept-2-ene-7-one S-carboxylic acid, obtained by Diels-Alder addition of acrolein diethylacetal to 1,2,3,4-tetrachloro-S,S-dimethoxycyclopentadiene followed by hydrolysis and oxidation. Reduction of the key intermediate with sodium borohydride was followed by esterification to give endo-anti-7-acetoxybicyclo[2.2.1]hept-2-ene S-carboxylic acid. Treatment of the acetoxy-acid with oxalyl chloride followed by diazomethane, and then cupric oxide in boiling tetrahydrofuran gave anti-9-acetoxytetracyclo[4.3.0.02,4.03,8]nonan-S-one. Wolff-Kishner reduction of the semicarbazone derivative of the tetracyclic acetoxy-ketone gave the title compound. Solvolysis of the p-nitrobenzoate esters of the title compound yielded a less strained rearranged product, but the relative rate acceleration was only a factor of four slower than the endo-anti-tricyclo[3.2.1.02,4]octan-8-yl esters. Relief of strain in the rate determining ionization step of the tetracyclic system does not appear to be responsible for part of the observed acceleration. The methylene bridge which transformed tricyclo[3.2.1.02,4]octan-8-o1 into tetracyclo[4.3.0.02,4.03,8]nonan-9-01 has been shown by x-ray structural determination to have distorted the orbital geometry in two ways. The edge bond orbitals have been pulled out of alignment with the site of developing positive charge and their through space distance to that site has been made unequal. x

PAGE 11

T h e methylene bridge has made C2 -C9 overlap more favorable than c C 3 9 overlap, leading to a cation which resembles product. Solvent attack on this intermediate ion, after the rate determining step, gives r ise to the less strained product. Additional documentation for the effect of orbital topography on chemical reactivity was sought through a study of the 13Cnmr spectra of the reactants and products, and from a study of the relative rate of decarbonylation of the respective ketones. xi

PAGE 12

CHAPTER I INTRODUCTION The assistance to ionization by a neighboring S, y-double bond is most dramatically revealed by a comparison of the solvolytic rates of the syn and anti-7-norbornenyl tosylates (l)-OTs and (2)-OTs with the 7-norbornyl system (3).1 The acceleration factor, which was re-corded for posterity as lOll, was attributed to backside homoconjuga-tive stabilization to the developing carbonium ion center. This phenomenon was most succinctly expressed as the delocalized, or nonclassical, bishomocyclopropenyl ion (4);2a,b although others preferred the concept of rapidly equilibrating carbonium ions.2c,d x x x 6 1 2 4 More recently, research attention has been focused on the neigh-boring group reactivity of the cyclopropyl a-bond.3 The early research of Winstein and his coworkers in the ais and trans-bicyclo[3.1.0]hexyl systems (5) and (6) established the fact of assistance to ionization by the cyclopropyl bond.4 This, now historic, study of the solvolytic rates (k . . /k = 35), the demonstrated existence of a special salt trans effect in only the ais isomer,4 and a detailed product analysis, led to 1

PAGE 13

4 2 the postulation of the non-classical trishomocyclopropenyl ion (7) as the reactive Further support for (7) was obtained via H OTs H 6 s 4 5 6 deuterium labeling studies. Cis-tosylate (5) labeled at C3 was found to give products completely scrambled in the C l , C3 , and C s positions. There was no scrambling observed when the correspondingly deuterated trans-tosylate (6) was solvolyzed under the same conditions.4a,d,e + 7 Cyclopropyl participation was further probed in the 8-tricyclo[3.2.l.02,4]octyl systems (8_11).6,7 These systems were chosen because, x x 7 3 8 9 10 11 at this point, evidence for ion (7) was limited to the bicyclohexyl system. The maximum rate enhancement reported to that point was 922 8 which fell short of previously published calculations and theoretical predictions.9 In particular, if the C2 and C4 positions of the chair

PAGE 14

3 and boat conformations of both the cis and trans isomers were connected by an ethano bridge, the four possible combinations were effectively frozen, allowing an efficient probe of the influence of the cyclopropyl group. The 8-tricyclooctyl series would also demonstrate that the align-ment of the cyclopropyl edge bond orbitals with that of the developing cationic center was of critical importance. One further advantage of this series was the direct comparison which could now be made with the 7-norbornyl (3), anti-7-norbornenyl (2), and 7-norbornadienyl systems. x c6 12 This makes possible the evaluation, in a systematic manner, the stabiliz-ing effects of the normal a-bond, the cyclopropane a-bond, and the classic TI or double bond. The normal a-bonded compound (3) can now be designated as a reference, or "parent-model" for a series of structurally related compounds. Pincock reported that the p-bromobenzenesulfonate(10)-OBs underwent acetolysis at a rate 2.7 times slower than (3)-OBs.7 The observed retardation could be due to steric interference to solvation at the backside of the leaving group but, in any case, the cyclopropane ring clearly provided no rate acceleration, thus ruling out "face" partici-pation. The cyclopropane "edge" orbitals are oriented downward, away from the incipient carbonium ion site precluding edge participation. If orientation of the edge orbitals were of importance,then compound (B) should be expected to show a definite rate enhancement when compared

PAGE 15

to (3)-OBs. This was found to be the case in several laboratories simultaneously. 6 The results which were available at that time are summarized in Table I. 4 The rate enhancement of the endo-anti-p-nitrobenzoate ester (8)OPNB, due to direct participation in the ionization step, was, at that time, the largest reported in the literature. Product studieslOa,b which showed the overwhelming formation of rearranged product, e.g., (14)-OH, and only 0.1% of retained alcohol (8)-OH imply that most of the charge resides on C2 and C4 • > SOH 8-0PNB 13 > HO H 14-0H + l8-0H Particularly noteworthy here was the observation that both solvent capture and internal return occur only from the endo direction. This result was in marked contrast to results observed when the ketone (15) corresponding to (14) was reduced with lithium aluminum to yield a product of exclusively exo attack. lOb Internal return to a rearranged dJ ..... , endo Ir'exo HO H H OH 62% 15 14 16

PAGE 16

TABLE I. Comparative Rates in the Bicyc1o[2.2.1]heptan-7-y1 Series, and the Series Compound Reference x cb 1 ;) dS 1 2 1 1 rt) 7 10 li:; 11 ib 6 ; 6 8 k 25 reI. 1 lOll 0.4 15 1 109 103 1.7 . 37 1012 5

PAGE 17

6 ester requires the existence of a stable cation whose lifetime permits migration of the anion, and approach from another direction. lOa Equili-bration experiments were reported to produce a mixture containing 62% exo-alcohol which would thus appear to be both thermodynamically and kinetically preferred on steric grounds. This is analagous to the ion generated by the anti-norbornenyl esters and is supportive of the non-classical ion (13). The rate of acetolysis of (9)-OBs is only 15 times that of (3) OBs and gives a complex mixture of products.6,lOa,b In contrast, (11)OBs is 104 faster than (3)-OBs. Here the methylene group of the exo cyclopropane ring may sterically aid the departure of the brosylate . 10a,11 anlon. Alternative possibilities involving a concerted bond shift to C have been discussed. lOa 8 I While indeed (8)-OPNB was locked in position so as to allow ex-tensive backside participation via cation (13), it had been suggestedlO that relief of strain could account in part for the lower activation energy observed in the solvolysis of (8). Further research in cyclopropyl participation led Coatesi2 to synthesize the pentacyclic alcohol (17)-OH. Here, overall relief of + 2 > 17-0PNB 18 17-0H

PAGE 18

7 strain cannot be a major factor in the observed rate acceleration since the rearrangement is degenerate, i.e. produces a structure identical with the original. The sole product of hydrolysis in the presence of base was (17)-OH, parent alcohol. Deuterium labeling studies yielded products which pointed to a triply symmetric trishomocyclopropenyl cation (18). It was suggested that the greater reactivity of (8)-OPNB relative to (17)-OPNB could be attributed to some relief of strain in the solvolytic transition state coupled with a more favorable orientation of the cyclopropyl edge bond. It was Coates' opinion that anchimeric assistance to ionization, forming the non-classical, symmetrical ion .(18), supplied the main driving force in his compound (17). The concept that precise orientation would be required for maximum assistance to ionization by cyclopropane was advanced by Battistel3 and further corroborated by the investigations of Sargentl4 and Masamune.ISa 19 Sargent's work with 2-(trans-3-bicyclo[3.l.0]) ethyl p-nitrobenzenesulfonate (19) gave no evidence, either product or kinetic, for participation. Conversely, Masamune's study of the exo and endo-anti-tricyclo[3.1.1.02,'+]heptan-6-ols provided additional insight into the geometrical requirements for assistance. He showed that (20)-OPNB was but 56 times faster than (21)-OPNB. However, (20)-OPNB yielded only the two rearranged products expected from the trishomocyclopropenyl cation while (21)-OPNB produced an olefinic mixture via ring cleavage.

PAGE 19

8 OPNB It It:] > > + 2 OR OPNB R 20 OR > o + 6 + Q7 R OR 21 Gassman15b independently concluded that both (20) and (21) were f solvolyzing with considerable assistance, albeit, of different types. By comparison, (20)-OPNB was a factor of 10 slower than (8)-OPNB, accord-ing to Masamune. This he attributed to a geometrical distortion caused by the interaction of the hydrogens on C3 and C7 • It is entirely possible that (20)-OPNB is slower than (8)-OPNB due to a decrease in the internal angle in the bridge from 97 to 83. This would increase the strain energy associated with ionization thus destabilizing the transition state and the resulting trishomocyclo-propenyl cation intermediate. In comparing the benzo-analog (22) of (20), Tanida16 observed essentially the same result. He noted that the parent in this series (23) solvolyzed to give products arising from ring cleavage, and

PAGE 20

9 OTs OH > 22-0Ts 22-0H commented that k comparisons would appear to have no meaning. NonereI. theless, he added the observation that k(22)/k(23) was 800 at 55C. OTs OAc HOAc > 6 23 Coates examined the related exo-tetracyclo[3.3.0.03.,6.02,8]oct-4-yl p-toluenesulfonate (24)-OTs and found two products.17 Rate studies led him to conclude that while the solvolysis was still anchi-merically assisted, the methylene bridge had geometrically distorted OTs OAc > + 24 the angle between the cyclopropyl edge bond and the back lobe of the leaving group. Coates compared the solvolysis rate of (24) with that of (17). However, the two structures would differ: :significantly in rigidity, and the strain energy required for a similar change in the

PAGE 21

10 critical relative geometry of the incipient cation center and the cyclo-propane ring in the process of ionization would not be of the same magni-tude for the two systems. This points out the necessity for making comparisons between structures which are closely related and adopting as the parent-model for any series, a compound of similar skeletal structure. Raywood-Farmer later organized the rate studies performed on the compounds of the basic bicyclo[2.l.l] system as shown in Table The k comparisons to parent (23) are usefuJ, those to the 7-norbornyl rel. series of questionable value, beyond the compa rison of the two parent compounds (23) and (3). One of the objectives of this research was to contribute to a better understanding of the effects of orbital topography on chemical ( reactivity of the strained a-bond as exemplified by the cyclopropane edge bond. Rigid, bridged, polycyclic systems meet the needs for precisely known substrate geometries. It was decided that tetracycloand its ester could provide a useful probe in meeting this objective. 7 2 26 From the data in Table I, it is clear that cyclopropyl edge participation in (8)-OPNB can be more effective than double bond parti-cipation in (2)-OPNB. This has been rationalized on the basis ofa difference in orientation of the participating orbitals relative to

PAGE 22

TABLE II. Comparative Rates in the Basic anti-Bicyc16[2.1.1] Series Compound x 24 x k 1 100 re . a b Also at 100C. Reference 18. a k 1 7-norbornyl re • 11

PAGE 23

12 the incipient carbonium ion site.13 Addition of the methylene bridge should increase the ground state strain thereby increasing the driving force for participation in (26). If relief of strain is the overriding factor in the rate determining step of the solvolysis of (26), then the rate should be greater than that observed in (8). Additional rate acceleration may also be expected from the methylene bridge itself since this alkyl substitution on the cyclopropane ring would have the same inductively stabilizing effect as would methyl substitution at C4 • Conversely, the same methylene bridge also appears to distort the molecular symmetry and may thereby decrease the orbital lineup for participation, resulting in a trend toward decreased solvolytic reactivity such as is observed in (17). A second objective was a thorough analysis of the solvolysis products of (26). Tanida's contention that the tricyclooctyl system (8) and its rearranged counterpart (14) both provide 0.1% of (8) when subjected to identical solvolysis conditions leads one to look closely for similar effects in structurally related systems.10b Along this line of thought, it is then reasonable to assume that alcohol (27) should be one of the predominant products of aqueous solvolysis since it appears to be the least strained. However, the 6 2 7 27

PAGE 24

13 proportion of unrearranged alcohol (26) derived from non-classical ion (28) by attack at C2 and C9 is statistically favored by2:1. This is the opposite of the situation in (8) and should providE. an interesting com parison. OR OPNB / > b ! c 2 6-0PNB 28 Finally, it was felt that an attempt should be made to unify the various pieces of information available in the field of rigid, bridged, polycyclics having basic bicyclo[2.2.I]carbon skeletons so as to point out the information required to fill in the gaps, thus making the theory of o rbital topography a more complete tool for .estimating chemical reactivity.

PAGE 25

CHAPTER II SYNTHESIS AND CHEMISTRY The first challenge to be met in accomplishing the goals of this research was the synthesis of the parent alcohol tetracyclo[4.3.0.02,4.03,8]nonan-9-ol (26). The initial synthetic route planned is given as a flow diagram in Scheme I. With minor modifications this was also the best route found to give the required product in the best overall yield. Hexachlorocyclopentadiene reacted exothermally with methanolic potassium hydroxide to give l,2,3,4-tetrachloro-5,5-dimethoxycyclopentadiene (29), a well-known compound. 19 The Diels-Alder addition of acrolein diethylacetal with the dike tal (29) gave the white crystalline l,2,3,4-tetrachloro-5-endo-(1,1-diethoxymethyl)-7,7-dimethoxybi.cyclo[2.2.I]hept-2-ene (30) in good yield. The IHnmr spectrum of (30) .contained the expected pairs of non-equivalent methoxyl and ethoxyl protons and the acetal proton, a doublet at 84.35 (J = 4.5 Hz). The dechlorination of (30) proceeded smoothly to give methyl)-7,7-dimethoxybicyclo[2.2.l]hept-2-ene (31).20 The nmr spectrum of (31) displayed the characteristic pattern of non-equivalent methoxyl and e thoxyl protons, the acetal proton doublet (now shifted upfieJd to 84.00, J = 9.0 Hz), and. the new vinyl protons at 86.14. The large change in the coupling constant of the acetal proton was attributed to the change in bulk upon going from chlorine to a hydrogen substituent on the bridgehead carbon atom. The bulky chlorine atom apparently 14

PAGE 26

C1?i(Cl KOH > CH30H Cl Cl 70% \I d::; 1) Cr03 < 2) H30+ 33 COOH 63% NaOH 95-100% :j (AchO > NaOAc COOH 37-55% 34 NH2CONRNH2 < 39 NNHCONH2 KOH HOCH2CH20H 75-80% 75% SCHEME I CRiiCR, Cl !J Cl Cl Cl 29 32 CHO OAc COOH 35 38 o 15 acrolein diethyl acetal > 30 80% H OC2 H s 1 OC2RS Na t-BuOH 90% dil. < HCl 31 OC2 HS OC2 HS OAc (COCl)2 36 COCI Cuo < 72-80% 37 C /\ o CRN:;!

PAGE 27

16 hindered free rotation of the equally bulky ethoxyl substituted carbon thus constraining the acetal proton to a position which was less than the optimal dihedral angle for maximum coupling. Relief of this steric constraint resulted in a more favorable dihedral angle and consequently a larger coupling. Isolation of (31) required careful distillation, to avoid all traces of acid, and to completely separate a low boiling contaminant. The conversion of ketal-acetal (31) into the key intermediate endo-bicyclo[2.2.I]hept-2-ene-7-one-5-carboxylic acid (33) was found to be.best accomplished in stages, with minimal intermediate purification. The acetal group of (31) was hydrolyzed quantitatively by stirring for 1 hour at room temperature in dilute (5-10%) hydrochloric acid. The intermediate aldehyde (32) was isolated and oxidized with Jones reagent.2l This initial hydrolysis and isolation served to remove the bulk of the ethyl alcohol produced, which would otherwise have been oxidized in the succeeding step. It is possible to accomplish both hydrolysis and oxidation simultaneously with Jones reagent, but the workup is complicated by the products of the side reactions. The crude product of the oxidation was subjected to more acidic (25-30%) hydrolysis conditions to remove all remaining methoxyl groups. The infrared spectrum of (33) contained two carbonyl absorptions, the ketone at 1780 and the acid at 1710 em-I. In the nmr, the vinyl protons were found shifted downfie1d to 66.55 and an acid proton now appeared at 10.67. The final purification of the carboxylic acid (33) was found to be very difficult if the distillation of (31) had not been conducted so as to remove all of the lower boiling contaminant. It was noted in passing that the rate of hydrolysis of the acetal and ketal groups was sufficiently different to allow isolation and further

PAGE 28

purification of the ketal-aldehyde (32) if desired. This compound could then be oxidized under neutral or basic conditions which would allow isolation of a ketal-acid. The intermediate keto-acid (33) was reduced with aqueous basic sodium borohydride22 to give alcohol (34) as a clear colorless syrup. The integrity of the product was established by confirming the loss of the 1780 cm-1 infrared band, and the concomitant appearance in the 17 nmr of a resonance which disappeared upon addition of deuterium oxide. The vinyl protons shifted upfield to 66.0S while the proton at C7 ' syn to t h e double bond, appeared as a broadened singlet at 63.S0. This is consistent with previously reported values for this stereochemistry.23 Three methods were examined for the preparation of 7-acetoxybi-cyclo[2.2.l]hept-2-ene-S-endo-carboxylic acid (35). Those methods utilizing acetyl chloride gave little or no product ester. Literature 22 24 . procedures ' made use of acetic anhydride-sodium acetate mix-tures gave the best yields, but recrystallization was complicated by the co-precipitation of a yellow, gummy, by-product. A recently published2S procedure employing acetic anhydride in pyridine, a modifica-tion o f an older method, was found to give cleaner products but in somewhat reduced yield. The last method could possibly be improved by lengthening the reaction time or by employing higher reaction temperatures. The nmr spectrum of the acetate ester showed vinyl pro-tons a t 66.10, and, as expected, the appearance at 2.0S of the acetate methyl singlet accompanied by a downfield shift to 4.42 of the broad bridge proton singlet. In retrospect, the mode of preparation of 7-acetoxy-bicyclo-[2.2.1]hept-2-ene-S-endo-carboxylic acid chloride (36) was of critical

PAGE 29

18 importance from an unusual standpoint. The compound was first prepared by addition of the acid, as a solution in pyridine, to the oxalyl chloride-ether reaction mixture. When the reaction was complete, the acid chloride was isolated by filtering off precipitated pyridine-h y d rochloride and evaporating the ethereal filtrates to constant weight. The product appeared pure enough for further reaction but was later found to contain about 10% of dissolved pyridine-hydrochloride. At-tempts at purifying (36) by vacuum distillation resulted in decomposi-tion and by column chromatography, in hydrolysis. Fortunately, acid chloride (36) could be prepared quantitatively and cleanly by addition of an ethereal solution of the acid to the reagent oxalyl chloride. The production of acid chloride (36) completely free of pyridine-hydrochloride was vital to the success of later steps in the sequence. ( The infrared spectrum of the acid chloride showed a carbonyl adsorption at 1790 em-I. The nmr of (36) differed from (35) only in that the vinyl protons were now observed at different chemical shift values, H2 at 86.00 and Hs at 6.25, due to the influence of the endo carbonyl chloride group. The proton on Cs shifted only slightly from 83.15 to 3.40 while H7 remained unchanged. The diazoketone (37) was produced quantitatively by slow addition of a n ethereal solution of the acid chloride (36) to an ethereal solution of diazomethane.23,25,26 Although the diazoketone appeared to be reasonably stable at room temperature, attempts to purify it by column chromatography resulted in loss of 30-40% of the material. When (37) was prepared from pyridine-hydrochloride free acid chloride, the need for f urther purification was removed. The diazoketone was characterized by the appearance of a sharp -C-N=N adsorption at 2100 cm-I in the

PAGE 30

19 infrared and a concurrent shift of the carboxyl carbonyl from 1690 to 1640 em-I. The nmr was observed to differcfrom that of (36) in the appearance of a new resonance at 04.0 for the diazomethyl proton, and a small upfield shift of the vinyl multiplets to 5.9 and 6.15, re-spectively. The cyclization of diazoketone (S1) to form the keto-acetate (38) proved exasperatingly difficult. 22 27 28 Many researchers ' , have employed this pseudo-carbenoid insertion to produce polycyclic structures with varying degrees of success. The "catalyst" used to accomplish the intra-molecular insertion has varied with the researcher. Ultimately the catalyst system suggested from the work of Ghatak et a1.29 was found to give high, reproducible yields of (38). These workers used an 30 activated CuO catalyst made from freshly prepared copper powder, I whose CuO:Cu content did not vary appreciably from batch to batch. They also irradiated the reaction mixture during the decomposition with infrared heat lamps. The function of these lamps was not clear-but the authors clearly state that they found the yields to be greatly re-duced when the lamps were not used. In retrospect, it appears that part of the irreproducible results obtained in the earlier attempts at cyclization was due to the interference of pyridine-hydrochloride dissolved in the acid chloride and carried through to the diazoketone. In early work, any keto-acetate (38) produced could not be separated by conventional means from a mass of polymeric by-products. The thick, sticky, yeillow residue obtained was extracted with boiling ether and discarded. The residue obtained upon evaporation of the ether was very crude, but conventional treatment with semicarbazidehydrochloride and sodium acetate24 produced a solid which could be

PAGE 31

20 recrystallized to a constant melting point. The cyclized product, when present, could only be isolated and purified as the semicarbazone (39) until an improved method of cyclization was developed. This proved to be advantageous in developing the modified Wo1ff-Kishner conditions employed later. Using the conditions as outlined by Ghatak et al., ketoacetate (38) was finally isolated in sufficient purity for spectral analysis. The spectrum of (38) was very simple, consisting mainl y of -C-H stretch adsorptions at 2880, 2980, and 3050 cm-I , broad carbonyl adsorption centered at 1730 cm-I and the acetate ester band at 1 2 40 cm-I • The nmr spectrum, as expected, no longer contained vinyl resonances, but the acetate methyl singlet remained at 02.05 while the adjoining bridge proton moved downfield to 4.98. The semicarbazone derivative (39) was only slightly soluble in common deuterated solvents, but resonances were still observed at 02.05 and 5.00 even in the extremely dilute solutions. The correct elemental analysis, and the appearance of characteristic amide I and II bands in the infrared had t o serve as adequate characterization for this derivative. The overall yield of 5-semicarbazone (39) from the Diels-Alder adduct (30) to this point was 0-13%. From a practical point of view, if the synthesis was begun with 194 grams of chlorinated adduct (30), the maximum yield of semicarbazone at this point would be 16 grams. Unfortunately, the yields in the cyc1ization reaction, prior to the discovery of the CuO:Cu catalyst, and the pyridine-hydrochloride contaminant, .were nearer zero more often than they were near optimum (42%). In the interest of conserving both time and material, and recognizing that thB Wolff-Kishner reduction, as yet untried, can often give poor yields, it was decided

PAGE 32

21 to work in a model system. In the interest of expedience, the reaction sequence given in Scheme II was begun. This would allow us to optimize yields of the cyclization reaction and the Wolff-Kishner reduction while producing the hydrocarbon (44) which contains the same carbon skeleton as the parent alcohol. This was to prove valuable for comparative purposes. The procedure outlined in Scheme II was first published by Nickon et al.27d The pure endo-bicyclo[2.2.I]hept-2-ene-5-carboxylic acid (41) was obtained in excellent yield from the Diels-Alder adduct (40). Although using the less desirable procedure for preparation of the acid chloride and diazoketone, and the unreliable cyclization conditions, acid (41) was found to closure readily in 80% yield. The resulting ketone was characterized as the semicarbazone derivative (43) and was identical to the compound described by Nickon. The ease of cyclization of the diazoketone derivative of (41) was surprising as compared to the results obtained when the 9-position was acetoxy substituted. This could be partly explained as a deactivating influence exerted on the double bond by the bridge functionality. Ketone (42) was purified by column chromatography and subjected to the conditions usually employed in the Huang-MinIon modification31 of the Wolff-Kishner reduction. No product was found in several attempts to produce the hydrocarbon (44) in this manner. A further modification32 employing the semicarbazone derivative produced a 70% yield of an azine. Hydrocarbon (44) was successfully produced by the method of Murray and Babiak.33 The hydrocarbon was quite volatile, disappearing in a few hours at room temperature unless tightly sealed. The infrared and nmr spectra were of little value in the characterization

PAGE 33

o 43 44 22 SCHEME II Methyl > _ l _ ) _ a _ q _ o 2) H3Q+ Acrylate 40 COOCH 3 COOH HzNNHCONHz HCl < ! H2NNH2 KOH NaOAc HOCH2CHzOH KOH 1) < 2) 3) 1) 12 , KI 2) Zn, HAc (COCl)2 d:; CH2H2 Cu, Heat . 41 COOH

PAGE 34

23 of (44) and, for that matter, those of (26)-OH and its rearranged counterpart (2?)-OH were later found to be similarly unrewarding. In addition to elemental analysis, the 13Cnmr provided additional in-formation. All nine carbon atoms of (44) are unique and, indeed, nine carbon resonances were found in the 13C spectrum. The chemical shift values are listed with those of other similar carbon skeletons in Table III. The conditions for the Wolff-Kishner reduction of semicarbazone (43) were then successfully applied to semicarbazone (39). Further sligh t modifications in this reaction brought the yield up to accept-able levels. The IHnmr of (26)-OH consisted of a broad multiplet between 00.8 and 2.4 with H9 occurring as a broad singlet at 04.35. The alcohol sublimed readily, requiring the use of gc-mass spectral ( analysis. The fragmentation pattern is given in-Table IV. The source of the base peak at mle 79 and the second major fragment at mle 70 are rationalized in Scheme ILL Molecular ion (26)-M+ ring opens via a to give the bicyclo[3..2.l] radical ion (26)-b. The favored route to the base peak mle 79 may be via loss of formaldehyde giving (26)-c which then looses hydrogen, or the processes may occur simultaneously to give mle 105. Loss of acetylene to give the C6H7+ ion is a facile process when compared to the competing pathway where hydrogen migration leads to radical ion (26)-d. Loss of cyclopentadiene (m/e 66) via a retro Diels-Alder leads to the second major fragment radical ion mle 70. The p-nitrobenzoate (PNB) and 3,5-dinitrobenzoate (DNB) deriva-tives of (26)-OH required for solvolytic studies were prepared by con-24 ventional procedures. The p-bromophenylurethane derivative (26)-PBPU

PAGE 35

24 TABLE III. 13CNMR Chemical Shift Values (ppm)a f f b b e e c d c d a = 22.5 a 30.66 b 19.7 b 24.3 c = 24.2 c = 30.81 d 39.1 d 39.77 e 34.1 e 35.3 f 34.7 f 40.06 g 40.9 g 38.0 h = 48.9 h 44.5 i = 86.4 i 51.9 c c (or i) e d a = 29.5 a 37.1 b 42.2 b = 40.3 c = 39.7 c = 47.1 d 38.4 e = 31.3 i 78.2 aya1ues relative to TMS, solvent CDC13 •

PAGE 36

25 TABLE IV. Comparative Fragmentation Patterns (70 eV) 26-0H 27-0H Mass % of Base % of Base l36 (Mt-) 10.0 19.7 118 (-H2O) 16.5 10.6 117 19.7 10.5 108 10.0 4.4 105 19.4 20.3 95 14.7 8.4 92 l3.5 5.5 91 22.4 16.1 80 22.4 11.6 79 100.0 23.4 78 22.4 13.1 77 27.7 17.2 70 52.9 100.0 69 5.3 14.4 68 26.5 67 26.5 21.9

PAGE 37

+ m/e l36 26-(Mt) m/e 70 m/e 79 + SCHEME III + > 26-a m/e 66 m/e 105 26-d . > -H < 26-b + 26 + -CH 0 2 26

PAGE 38

27 was prepared for use in the x-ray structural determination which employed the heavy atom technique. Although the crystals formed by this derivative were always clustered, it was possible to dissect a single crystal."of sufficient size for the structure determination. The com pound crystallizes in the space group P21/c with unit cell dimensions, a = 12.067(4), b = 9.782(3), c = 12.060(4), and 8 = 96.12(3), and four molecules per unit cell. The analysis was refined to an R factor of 0.06S. The bond lengths and bond angles and the interatomic distances of importance are summarized in the Appendix. The computer generated drawing (ORTEP) which represents 'the data i s given in Figure I. Three internal angleswere.defined: a) the cyclopropane plane through atoms 2,3, and 4; b) the skeletal plane through atoms 1,2,3, and 8; c) the bridge plane through atoms 1,8, and 9. The calculated dihedral angles were a-b = 69.70 b-c = 61.20 a-c = 9 . 10 .. The consequences of the information revealed by the x-ray structural analysis is discussed in relation to kinetic and product studies in Chapter III. Parent alcohol ( 2 6)-OH was cleanly o xidized . (Scheme IV) to tetracyclo[4.3.0.02,4.03,B]nonan-9-one(45) with a methylene chloride solution of chromium trioxide dipyridine complex.34 As was characteristic of other members of this family, (26)-OH and ( 44), the ketone was quite volatile and sublimed readily. It was characterized by the appearance of a carbonyl adsorption in the infrared at176S cm-1 • The nmrresonance for Hg at04.3S had disappeared leaving only the broad multiplet from 00.8-2.S.

PAGE 39

c • 'J 28

PAGE 40

DNBCl Pyridine 1 CrO 3.2 pyridine . CH2C12 o 1/ 45 SCHEME IV > 46 29 60% Acetone> 27 Acetone cro31 o \I 47 27

PAGE 41

Ketone (45) decarbonylates readily = 5.2 hours at 185C giving bicyclo[3.2.1Jocta-2,6-diene (46), a well...;.known hydrocarbon.35 The isolation and identification of '(46) was convincing proof that alcohol (26) and ketone (45) possess the structures assigned. The esters (26)-OPNB and (26)-ODNB were solvolyzed in aqueous acetone and rearranged cleanly to give tetracyclo[3.3.l.02,4.03,7]-nonan-9-ol(27) in good yield. The rearranged alcohol was also quite volatile and was easily purified by sublimation. Characterization of this new symmetrical structure employed mass spectral analysis (see Table IV, fragmentation patterns). Alcohol (27) had a much larger molecular ion, twice that of Other than loss of water, the fragmentation pattern of (27)-OH was sufficiently different from that of its isomeric parent (26),...OH to warrant further consideration. It I 30 was rationalized (Scheme V) that, (27)-M+ would most likely undergo a series of ring openings similar to , (26)-M+ via the analagous (27)-a to give (27)-b. Loss of formaldehyde and hydrogen lead to mle 105 in about the same abundance as observed for (26). However, loss of acetylene from the homo-aromatic ion (27)-e is not as favored a process here as in (26)-OH, The base peak in (27)-OH arises from the alterna-tive p rocess, now more competitive, of hydrogen shift and loss of cyc10pentadiene to give the radical ion mle 70. Alcohol (26) had melting point l45-l46C (sealed tube) while that of (27)-OH was found tooe 236-237C (sealed tube). The nmr spectrum of (27),...OH differed somewhat from that of (26)-OH in the re-gion 61.2-2.4, the major difference being the H9 proton which occurred as a broadened triplet (J = 5.0 Hz) at 4.00., In models of the new car-bon skeleton, H9 appears ideally situated for long range W coupling,

PAGE 42

31 . SCHEME V H > m/e 136 2 ?-(M+) 2?-a H + XOH+ H 2?-b 0 < • -CH2O m/e 70 m/e 66 2?-d H H -H <: < 2?-e m/e 105 m/e 105 2?-c t -C,B, H m/e 79

PAGE 43

32 which would account for the broadening. The 13Cnmr was consistent with a symmetrical structure containing six different types of carbon atoms. The chemical shift values and peak assignments are given in Table III. Oxidation of (27 )-OHusing either Jones reagent2l or the chromium trioxide dipyridine complex in methylene chloride34 gave excellent yields of tetracyclo [3.3.1. 0 2" 4. 0 3, 1nonan-9-one (47). Both ketone (45) and ketone (47) have a moist appearance when freshly sublimed but were homogeneous to capillary glpc. Ketone (47) had a strong infrared carbonyl band at 1725 cm (vs. 1765 em-I for (45)), and its nmr spectrum compared to (27)-OH showed the absence of the H.9 resonance. As a check on the stereochemistry of (27)-OH, ketone (47) was reduced with lithium aluminum hydride. It was hoped that a mixI ture of alcohols, -OH both syn and anti to the cyclopropane ring, would result. Glpc e xamination of the alcohol produced established that it was at least 99% a single component. Co-injection with authentic anti-(27)-OH produced a single peak, and the nmr and infrared spectra were superimposable with that of the original (27)-OH. Models show that the approach which would produce the syn epimer appears to be blocked by the axial-like hydrogens in the six-membered ring. The formation of only (27)-OH from (47) appears consistent with attack at the least hindered face of the carbonyl double bond. As proof of the structure of (27)-OH and consequently of (47), it was planned to decarbonylate (47) and identify the known cis-2,7-bicycloI3.3.0]d " 36 octa l.ene. However, ketone (47) proved very resistant to.decarbonyla-tion (estimated t1 =2800 hours at 240C) • Further evidence for the . stable nature of this ketone is found in comparing the mass spectral

PAGE 44

33 fragmentation pattern with that of (45) . in Table V. The molecular ion of (47) is eight times the intensity of the ion from (45) and the base peak of (45) is the result of further skeletal rearrangement. Since ketone (47) would not conveniently decarbonylate, the proof for the structure of alcohol (27) was sought via the reaction sequence given in Scheme VI. AlcOhol (27) was dissolved in thionyl chloride to give tetracyclo13.3.1.02,4.03"]nonyl-9-chloride (48) as a white solid. Chloride (48) was dechlorinated using the conventional procedure2l and gave the hydrocarbon tetracyclo[3.3.l.02 4.03 ']nonane (49) christened "triaxane" by Nickon andPandit.37 Triaxane was extremely volatile and could only be kept in a sealed tube. The 13Cnmr shift values were consistent 38 with those reported, and are listed with the structural assignments in Table III. The production of the known diene (46) and the origin of (27)-OH by the routes shown in Scheme IV, coupled with the formation of hydro-carbon (49) as shown in Scheme VI, confirm the integrity of the struc-tures of (26)-OH and (27)-OH as drawn. That the two carbon skeletons are different appeared evident from the comparison of the physical characteristics. Solvolytic studies were anticipated to produce marked differences in the chemical reactivity of the two systems.

PAGE 45

TABLE V. Comparative Ketone Fragmentation Patterns Mass 134 (M+) 106 ( -CO) 105 91 (C7H7+) 79 78 77 27 > o /I fit; 45 % of Base 2.3 66.6 16.6 100.0 33.3 66.6 25.0 SCHEME VI Na t-BuOH 48 o II 47 % of Base 16.1 100.0 19.4 94.3 24.2 62.9 12.6 49 III 34

PAGE 46

CHAPTER III KINETIC AND PRODUCT STUDIES One of the goals of this research was to obtain the rate data and product information necessary to make comparisons between (8), (1?), and (26). 7 7 2 2 2 3 8 17 26 Work in the endo-anti-tricyclo[3.2.1.02,4]octyl system (8) had been performed with p-nitrobenzoate (PNB) esters in 70% dioxane: 7% 6 water, and in 70. acetone:water. Work in the pentacyclo.[4.3.0.02,4.03,8 .Os,7]nonan-9-ol(1?) system was also performed with p-nitrobenzoate esters, but the solvent was 65% acetone:water.12 For comparative purposes it was deemed expedient to perform the rate studie s on (26)-OPNB in 70% acetone:water. The extent of reaction was followed by titration with standard sodium hydroxide to a phenolphthalein end point of the p-nitrobenzoic acid produced. These titrations generally gave normal first order plots for all esters (26)-OPNB. The plots for this ester were linear to approximately one half-life and then curved due to.the formation of (27)-OPNB, a less reactive ester. 35 ' -

PAGE 47

36 Representative plots for (26)-OPNB and (27)-OPNB are illustrated in Appendix Figures la, IIa, and IlIa. The observed infinity titers were within 3-5% of theory except in the case of (26)-OPNB. Careful examina-tion of the contents of several ampules indicated that at 78-83% of theoretical infinity only (27)-OPNB .remained. Solvolysis of (26)-OPNB is obviously accompanied by isomerization and internal return. Solvolysis reactions with concurrent rearrangement of the type observed in ( 2 B)-OPNB have often been encountered, and their kinetic treatment has been discussed.39 Roberts39b considered the following system of which (26)-OPNB is a representative: ROH + HX RX R' X R' OH + HX ( He assumed in his derivation of the rate constant, that the apparent first order rate constant (k b d or k. . t') obtained using o serve t1tr1me r1C an acid infinity titer which does not take account of rearrangement will be the sum of krearrangement plus ksolvolysis' i.e. kj + k2 , as long as kl is much faster than k3 • Straight line plots were ob-(1) tained for (26)-OPNB when 80% of the theoretical infinity value (plus the blank) were used in the calculations. Eighty percent of theoreti-cal i nfinity was found to correspond to observed, or actual infinity values. Since the value of + k r ) most nearly represents the actual rate of ionization of it is the value used hereafter in this work for comparative purposes. For all compounds, duplicate runs were made at each temperature. Each run was treated by a standard linear.least squares program and the average of the two values was employed in further calculations.

PAGE 48

The enthalpy of activation (t.:al=) was calculated using equation .(2). -t.Hf11 =-(--._) R Tl Tz (2) When the average H::F had been determined, equation (2) was rearranged and used for extrapolations to 25C. The entropy of activation t.S:j:: was calculated using equation (3). kl k -R(ln h) -t.ItF --+ t.S::F T (3) The value for the constant term R(ln k/h) was found to be 47.18742. 37 The solvolytic kinetics of: some previously investigated systems pertinent to our study were redetermined. These included the p-nitrobenzoate ester of 7-norbornadieno1 (12)-OH, which was used to establish the kinetic relationships to other systems and to .test ( the accuracy of our procedures, and the corresponding 3,5-dinitrobenzoate ester (12)-ODNB which was examined in 60% acetone:water in order to 40 establish a relationship between our results and those of Klumpp in the same solvent. In addition, rates for anti-7-norbornenyl p-nitrobenzoate (2)-OPNB in aqueous acetone were carefully measured in order to more accurately assess the often quoted rate ratio / 3 3h k(12) k(2) = 10 • " , Originally this comparative rate was obtained from the solvolytic rates of the respective chlorides and (12)-Cl in 80% acetone:water.4d Haywood-Farmer determined the rate constan t for (2)-OPNB in 70% dioxane:water,lOa but did not study (12)-OPNB under the same conditions. The complete data obtained from the solvolysis studies are given in Chapter IV and summarized in Table VI. The log plots of the rate constants versus reciprocal temperature are shown in Appendix Figures IVa, Va, and VIa.

PAGE 49

Compound 26 125.0 100.0 90.2 2:):; TABLE VI. Kinetic Resultsa k(sec. -1) 7.13 x 10-'+ 7.53 X 10-5 3.29 X 10-5 1. 74 x 10-8 (kca1/mo1e) 24.1 1.5 38 (e.u.) -12.9 3.8

PAGE 50

39 As Table VI shows, (26)-OPNB solvolyzes almost 10,000 times (103 •97) faster than (2?)-OPNB. The use of the actual infinity titer in the calculation of the rate constant which best represents the rate of ionization was justifiable since Roberts' assumption was held to be valid when kl was much greater than k3 • The reevaluated rate acceleration at 25C in 70% acetone:water for (12)-OPNB over (2)-OPNB was found to be 3900 (103 • 6). The rate of solvolysis of (12)-ODNB in 60% acetone: water was found to be 19 times faster than that of (12)-OPNB in 70% acetone:water at 25C. The data of other workers on compounds of in-terest are summarized in Table VII for comparative purposes. Since it would be impossible to evaluate the rate constant for the solvolysis of (3)-OPNB in 70% acetone, it was necessary to fall back on the assumptionlOa that (2)-OPNB will be accelerated over (3)I OPNB by the same amount that (2)-OBs is accelerated over (3)-OBs, 1.42 x 1011•1 The rate of (2)-OPNB at 25C was obtained by extra-polation from data obtained at higher temperatures and, from this, the rate of (3)-OPNB was inferred. Although the bicyclohexyl system (5) is conformationally mobile, its place in the overall picture was deemed of interest. The assumption was made that k /k (5)-OTs (3)-OTs would be the same as k /k . (5)-OPNB (3)-OPNB Since the rate of (3)-OPNB 4d had been estimated as described above and the rates of (5)-OTs and 1 (3)-OTs are known, (5)-OPNB could be similarly estimated. The cal-culated absolute rates at 25C for all systems of interest in this study are compared in Table VIII. While the kinetic study has provided information on the forma-tion of the transition state in the solvolysis of (26), and analysis of the products is needed to provide an insight into the nature of the

PAGE 51

40 TABLE VII. Kinetic Values for Structurally Related Systems t,H+ t,S=F COIDEound T(Oc) -1 k(sec. ) (kcal/mole (e.u.) Reference 90.2a 1.67 x 10 -4 24.8 -7.8 loa 25b 7.33 x 10 -8 8 l25.0c 7.00 x 10-5 26.8 -11.5 l2b 25 6.02 x 10-10 17 100.Od 4.93 x 10-4 26.2 -3.8 40 25e 2.81 x 10-9 50 ODNB 100.Od 5.74 x 10-3 23.7 -5.7 40 25e 7.65 x 10-8 51 OTs 50.0f -5 24.1 -5.0 4d 9.61 x 10 25 -6 3.8 x 10 5 a70% acetone:water. values at were obtained by extrapolation. c65% acetone:water. 60% eCorrected to 70% acetone: water and to OPNB ester. JValue for acetolysis.

PAGE 52

.TABLE VII I. Comparison of Relative Rates of p-Nitrobenzoate in 70% Acetone:Water ComEound cE;NB OPNB 26 _1 .k(sec ) 1.48 x 10 -23 _14 2.16 x 10 -12 2.10 x 10 . _9 8.26 x 10 _12 1.85 x 10 _lOa 6.02 x 10 -8 1. 74 x 10 . . a65% acetone:water. k reI. 7-Norborny1 1 1. 46 x 109 , 1. 42 X lOll 5.58 X 1014 1. 25 X lOll 4.07 x 1013 1.17 X 1015 at 25C . log k reI. 1 109 • 2 1011• 1 41 Esters Reference 1 4d This work and 1 This work This work 12b This work

PAGE 53

TABLE VIII. (continued) Compound 50 OPNB 20 _1 k(sec ) _8 7.33 x 10 -8 7.65 x 10 -9 2.80 x 10 -10 6.31 x 10 k reI. 7-Norborny1 4.95 X 1015 5.17 X 1015 4.26 X 1013 log k 1 re . 42 Reference lOa 40 40 15b

PAGE 54

43 intermediate carbonium ion. The esters of (26) and (27) were solvolyzed in the normal manner, i.e. with a steadily increasing concentration of acid ( n o buffer) and with the presence of a 2:1 excess of urea. Little difficulty was experienced in analyzing the products of the normal hydrolysis. Peak areas were obtained with the use of a digital integrator and were reproducible so that good accuracy was obtained in chroluatographic analysis. The solvolysis of (26)-OPNB in the presence of urea required heating at 125C for only 20 hours to achieve t e n half-lives. The r esulting solution was yellow, but interfering products appeared to be minimal. Attempted analysis of the solvolysis products of in the presence of urea was badly complicated by the formation of products from many side reactions and decomposition. Such a large number of by-products were formed that isolation of the p r o duct alcohol(s) sufficiently pure and in good yield was virtuall; :mpossible. The ester (27)-ODNB was prepared and its rate of s o lvolysis at140C determined in order that the overall reaction time could be reduced, thereby producing a cleaner product. Solvolysis o f (27)-ODNB in the presence of urea at the shorter t e n half-life time (46 hours) still resulted in extensiv e d ecomposiUon, but overall, in fewe r by-products; therefore analysis of this product mixture wa s less complicated. The results of the product analyses are given i n Table IX. One aspect of Table IX bears further explanation here. After having established all of the product ratios except those for (27) OPNB or -ODNB in the presence of urea, a control experiment 'was run. A mLxture of p-nitrobenzoic acid and urea was dissolved in 70% acetone at the same molar concentration as that of a solvolytic product run

PAGE 55

TABLE IX. Product Analysis Starting Overall Material Conditions Yield (2 6)-OPNB (27)-OPNB % Ester Normal 97% 0.09% 99.91% 24.5% (26) -OPNE Urea 97% 0 100 % 30.0% Nor mal 97.5% 0 0 0 ( 2?)-ODNB Ure a 96% 0 0 0 a The origin and identification of this material is questionable. Products (26)-OH 0.11% 0.04%a 0 0 _(27) -OR 99.89% 99.96% 100% 100% % Al cohol 72.5% 67.0% 97.5% 96.0% .p .p-

PAGE 56

and treated to the solvolysis conditions. Chromatographic analysis after workup clearly indicated several minor products were eluting 45 in the vicinity of the retention time for (26)-OH. Since none of that alcohol could have been in this sample, the analytical data acquired for the solvolysis of (27)-OPNB in the presence of urea were discarded as spurious. Analysis of (2?)-ODNB solvolysis products, while less complex, was subject to residual doubt when a minor peak appeared which could have been 0.06% of (26)-OH. Accordingly, painstaking care was exercised in preparing a solution of (26)-OH which was sufficiently dilute to simulate the product sample. This solution gave only a very small peak at low attenuation (near maximum sensitivity of the detector). Coinjection of the two samples produced an extra peak in the recorder tracing for true (26)-OH which was only a few tenths of a minute different in retention time from the spurious peak which could easily have been mistaken for a trace of (26)-OH. This bit of evidence casts some doubt on the origin and identification of the m inor alcohol product detected in the solvolysis of (26)-OPNB in the p resence of urea. The 0.04% reported as (26)-OH could as easily havebE!en the spurious peak \vhich was present to the extent of 0.06% in the (27)-ODNB solvolysis product. It c a n be seen from the analysis of the products that internal return and attack by solvent in the solvolysis of (26)-OPNB occurs only at the anti-or the hindered endo-positions. The solvolysis product s of (27)-ODNB arise exc.lusively from anti-attac k of solvent. I t v,as deemed the best approach, for overall un d erstanding, to examine the r esults of this study by segregating the individual facets of the picture and discussing them separately. Accordingly, the

PAGE 57

remainder of this discussion is divided into the followin g parts: a ) stereochemistry of the solvolysis and the nature of the carboni um ion, b ) the geometrical considerations, a n d c) the effects of strain . Th er_e ocJ?emi.s So 8 the N a tu r e of the Carbonium Ion The data in Table VIII show that the rate of solvolysis of 4 6 (26)-OPNB is only a little slower than that of (8)-OPNB (k /k = 4.2), 8 2 6 but 2 9 times faster than that of (17)-OPNB. As previousl y describ ed, k reI. log k r eI. (3) 17 1 1013• 6 26 8 2 9 1 2 1 1015 . 7 ( 8 ) rcarr2Dges stereospecifically t o give anti-( 8 ) , or mainly endo-(14). The stere8specificity of the nucleophilic attack o n the intermediate derived from ( S )-OPNB a t only the CT./lti and hindEred endo sites i s quite consistent with the trishomocyclopropenyl cation (13). Only (1 3 ) couJd so completely attack to form exo-(14) and One ca"QDot concei\7e of a set o f fully equi.librab::1g r:J C:.ssical jans giving the products observed. The ESLer o f (17)-OH has elso been disc,:ssed. Here Cibain, the accelerat ed rate a1' o the scrc.mbling of a oec:teriuITl l abel clearly point to cation (18).

PAGE 58

47 18 1 3 The ketone ( 47 ) has an hybridized carbon atom at C9 o 11 exo 47 and could be considered a good steric model for ion (28). When ketone (47) is reduced with lithium aluminum hydride, attack is 28 96% from the exo side giving back alcohol (27) and 4% of three minor products (ratio 1:4:1). The rate studies on (26)-OPNB clearly show acceleration of nearly the same magnitude as that of (S)-OPNB. Product analyses in-die ate that nucleophilic attack is only at the a:n.ti and hindered endo sites. This is in direct contrast t o the stereochemistry of the reduc-tion of the ketone. Here, as in ( S ) and (17), the interpretation of the results clearly requires the intervent:i.o n of the trisoOiDocyclopropenyl cation (2S).

PAGE 59

48 Geometrical Consider ations The cyclopropane ring presumably stabilizes a carbonium ion by overlap of the bent C-C bonds, which are high in p-character, with the vacant p-orbital of the cation. From extensive studies on the addition of electrophiles to cyclopropane rings, it appears that the 41 preferred mechanism is the edge approach. The ability of the cyclo-prop ane edge to interact throu g h space with a carboniu m ion will be affected by the interatomic angles and other relative distances within the molecule. Theoretical studies by Hoffman42 have shown that there is a greater stabilization for 3-bicyclo[3.l.0]hexyl cation than for 4-cyclopentenyl, 2-indanyl, or cyclopentyl cations, and that the potential energy minimum occurs when the charged carbon atom is raised from 76 to 80 above the plane of the five-membered ring I toward a chair form. A search of the literature reveals a paucity of actual x-ray structural data on compounds related to (8), (1 7), and (26). Struc-43a tural information was found for the anti-7-norbornenyl ( 2 ) skeleton, (10) skeleton,43b and the benzonorbornenyl (53) skeleton,43c none of which are of any assistance in evaluating the data now available on (26). Only the structure of the p-bromobenzoate ester of (20)18 bears mention. Masamune claims t o have shown that (20) possesses a geometry similar to the one

PAGE 60

49 predicted by Hoffman for the maximum delocalization of the positive charge of the bicyclo[3.1.0Jhexyl cation. In ( 20 ) the charged carbon atom is raised 71 above the plane of the five-membered ring. Solvolysis studies on (20)-O PNBlSb show it to b e a factor of 100 slower than ( 8)-OPNB (T able VIII). M a s a mune concludes that the ground state geometry of the system is a decisive factor contributing to the rate acceleration. It is probably fortuitous that the rate acceleration of ( 20 ) relative to 7-norbornyl is virtually identical with that of (1?). As previously s hown (Table II), ( 20 ) is accelerated 108 over its saturated parent (23) at 100C and is identical with the acceleration afforded by its unsaturated analog (25). It would appear that the real angles 8 and and the intera tom i c distance AB may be more significant in the diagnosis of effec-B Lv. grp. tive o v erlap tha n are Hoffman's predictions.4 4 M asamune calculated the angle 8 (34), but nothing more was published. The a n g l e s 8 , , and the AB dista n c e "("ere calculated for ( 2 6 ) and were foun d to be: e 45 AB 2.182 a ngst rom s . With not hing mor e a v ailable for comparison than l1asam u ne's v alue for 8 (34), which is 11 smaller and mor e favorable, no con clus ions can be drawn f rom thes e n umbers.

PAGE 61

50 In order to better visualize the molecule, the angles between the planes of interest, previously defined, were calculated and il-lustrated as a cutaway in Figure II. The carbon bearing the leaving group is found to b e raised only 61 above the plane of the five-membered ring. Careful examination shows that the skel eton of (26) is twisted by the m ethylene bridge with the r esult that the calculated v alues of e , , and the AB distance are averages, and in fact each carbon atom of the cyclopropane edge bond in (26) will have individually 45 different values for these angles. " / / / It Calculations gave the following results: Distance C2 -C9 = 2.283 angstroms; C3 -C9 2.343 angstroms e Finally, as formation of the cation (28 ) progresses, it appears from models that C is held rigid in its original position by the methyle n e 2 . bridg e , while C can move outward more freely and even f arthe r from 3 It appears clear then, from the detailed analysis of the x-ray structure, that insertion of the methylene bridge has the orbital geometry for effective overlap and may be r esponsible for the reduced rate acceleration because of the in-dividual differences in the angl e and distance betw e e n C2 and C3 with

PAGE 62

Cg C3 F . IGURE II. Internal Plane Angles L E AVING GROUP VI I-'

PAGE 63

respect to C9 , the resulting cation would not appear to be formed symmetrically and this may further impair participation and reduce acceleration. 52 Further support for the conclusion that the orbital topography of (26 ) is less favorable overall for stabilization and additionally less favorable due to distortion in the cyclopropane edge bond, all due to the methylene bridge, wa s sought by a n examination of the rate ofextrusion of carbon monoxide from the corresponding ketones. The rate of synchronous loss of carbon monoxide and diene formation could provide a more sensitive probe for the effect of precise orbital alignment. This hypothesis is supported by the work of Allred46 on the synchronous loss of nitrogen and the formation of diene in the series (52-56), which produced the rates of nitrogen extrusion listed in Table X. Allred suggested that the introduction of the methano bridge in (5 4 ) caused severe strain in the transition state leading to diene. Lengthening of the bridge, as in (55) and (56), lessens this strain leading to enhanced rates of extrusion. The low reactivity of (54) when compared with (53) indicated that relief of strain involved in opening of the cyclopropane ring is not a major factor. Mention was mad e of the less favorable orientation of the cyclopropane edge bond in (5 4); however, preference for the increased strain argument was clearly stated. It would appear frOID inspection of models and from our results on decarbonylation of ketones (45) and (5 7 ) that the cyclopropane edge bond of (55) and (56) would certainly be more favorably oriented.foroverlap as the C-N bonds break, and that this more favorable overlap is at least as satisfactory as an explanation of Allred's results, and cannot be set aside.

PAGE 64

TABLE X . Comparative Rates of Nitrogen Extrusion ci::), N 5 2 td::" . N 53 0:://. N 54 G::l N: 5 5 56 _ 1 . k (sec ) _16 4.7 x 1 0 _ 4 1 . 0 4 x 10 _15 4.34 x 10 _7 2.46 x 1 0 -5 8.9 x 1 0 k rel. 1 2 . 2 x 10 1 1 9.2 5.2 x 1 0 8 1 1 l. 9 x 10 5 3

PAGE 65

54 Ketones (45), (47), and (57) were readily available from their corresponding alcohols. o o I 45 47 57 The rate of decarbonylation of (57) was obtained gas chromato-graphically using the technique ultimately intended for both (45) and (47), and compared favorably with the published value which had been b . d . IH h . o talne uSlng nmr tec nlques. The data obtained are summarized below: TABLE XI. Rates of Ketone Decatboriylation (57) (45) (47) k 1.96 X 10-4 165 = -1 sec. k1650 4.45 X 10-6 sec. -1 k1850 3.72 X 10-5 sec. " I k1650 8 X 10-13 sec. -la k 7 X 10-8 sec. -lb 240 aEstimated assuming a linear relationship where log k (57) 165/10g k (26)-OPNB 100 reI. (45) reI. 7-norbornyl(B)_OPNB log k 1 (47) 165 flog k 1 " " (27)-OPNB 100. re .(45 ) re . 7-norborny l(B)_OPNB from a single sample h eated 76 hours at 240C. b Calculated

PAGE 66

55 The data showk165o(5?)/k1650(45)::; 44., compared to the rate accelera-tion of solvolysis at 25 0 where k /k is only a factor (8)-OPNB (26)-OPNB of four. The rate of de carbonyl at ion would appear to be a more sensi-tive measure of the effect of the methylene bridge on the.overall dis-tortion of the cyclopropane edge bond and the reduction in overall orbital alignment. It could be argued that a comparison of the rates of solvoly-sis of a given compound cannot be compared to the rate of decarbonyla-tion of its corresponding ketone since the mechanisms and transition states for the two processes may be entirely unrelated. That the rate of decarbonylation of rigid ketones will be re-sponsive to the degree of p-character of the cyclopropane edge bond orbitals has been shown by the excellent linear correlation which has t been found to exist between the 13C _ H coupling constants of the endo small-ring protons and the free energy of activation fordecarbonylation of the appropriate ketone.48 The results of the x-ray structural determination have demonstrated the difference in p-character which exists in the orbitals of the cyclopropane edge bond in (26)-OH due to the methylene bridge. It has been shown that a linear fre e energy r elationship exi sts between the rate of decarbonylation of a series of unsaturated tri-cyclic ketones (A), and the rates of ionization of the ester of the corresponding saturated alcohol having the appropriate stereochemistry (B).49 The mechanistic significance of this relationship is not clear but the fact of its existence has l e d us to search for a similar relationship between the rates of solvolysis of other bridged polycyclic systems and the free energy of of

PAGE 67

56 o n n = 3,4,5,6 -thei r corresponding ketones. This correlation would be expected to be more sensitive since the double bond in the tricyclic ketones (A) described, must, to a certain extent, exert a leveling effect on the decarbonylation rates. Such a new structure-reactivity correlation would be more useful for predictive purposes. Although limited, the data avail-able to date are certainly encouraging. , From the foregoing discussion, it is apparent that the rate acceleration of (26)-OPNB over the 7-norbornyl parent model is due to t h e neighboring group participation by the cyclopropane edge bond as exemplified by the intermediate non-classical ion (2B), and the lack of acceleration over (B)-OPNB can be explained as due to less favorable orbital topography introduced by the methylene bridge. Effects of Strain The cis-bicyclo[3.l.0]hexy1 system (5) shows moderate rate en-hancement when compared with the conformationally more rigid structures (2) a n d (3) (Table VIII). The bridging which locks (5) in the chair form giving (8) results in a dramatic increase in rate: log (kB/ks) = 106 • 3 • This has been attributed in part to the more favorable geometry for potential overlap in (B). Such bridging is also accompanied by

PAGE 68

two other features: an increase in the rigidity and ring strain of theinolecule, and a decrease in the internal angle and flexibility of the cation center.3h It has been argued that since the transition state involves Sp3 to Sp2 rehybridization at the cation center, an increase in the internal bridge angle would facilitate the formation of the transi-57 tion state and lead to an increase in the solvolysis rate. All avail-able x-ray data on carbon skeletons arising from the basic bicyclo-[2.2.1J skeleton clearly show that the angle in question is virtually identical, .while the log k 1 solvolysis rates vary from 0.4 (10) to re • . 1015' 1 (2 6) . Th t h t th . d ere lS no reason 0 expect t a e correspon lng angl e in (8) and (17) should vary significantly. The relief of the necessity for complete sp3-sp2rehybridization upon going to delocalized ions (13), (18) , and (28) will be a constant component of the rate un-less the geometry for participation by the cyclopropane edge bond is altered. The role of strain and geometry in the overall rate accelera-tion picture in the case of 051), when compared with 7-norbornyl, is cloudy. Here, the insertion of an additional methylene gives a basic bicyclo[3.2.l] system where the internal angle of the bridge bearing the l eaving group appears to be expanded. This should supposedly facilitate ionization leading to a rate enhancement. Compared to 7-norbornyl, (51) is as fast as (8) and four times faster than (26) (Table VIII). The bridge angle in ( 27 ) is also However, (27) solvolyzes 40,000 times slower than (8). Models indicate that the geometry of the cyclopropyl ring in . (51) with respect to the

PAGE 69

58 bridge bearing. the leaving group . is remarkably similar to (26). Moreover, (51) solvolyzes with exclusive rearrangement where the new bond is formed with the cyclopropane carbon atom on the side of the methylene bridge. 5040 The work of Hess helps in placing Klumpp's work on (51) in proper perspective. Unfortunately, the solvolysis d)X 58 59 log k 1 105 • 3 1012• 4 reI. data on the cyclopropane analog of (55) have never .been published. Certainly relief of ground state skeletal.strain occurs since both (26).,(8), and (51) rearrange to less strained structures. The I difference in the free energy of reactants and products in the case of ' (8)-OPNB has been calculated by TanidalOb from the equation 6F:f = log (RP) where R, ,the reactivity ratio, is represented by the rate constants for solvolysis of starting material and rearranged product, and P, the partition factor, is represented by the product ratio of re-arranged alcohol to starting alcohol obtained in the solvolysis re-action under buffered conditions. In the case of (8)-OPNB and (14 )-OPNB, R = 3.2 X 107 and P = 103 , making 6p:f = -11. 95 kcal/mole. Using the same relationship, Klumpp40 found the change in 6Ff for reactant (51)-ODNB and its products to be 8.6 kcal/mole. ,While the reactivity ratio R in the case of (26)-OPNB is easily obtained, the factor P is not. We observed no (26)-OH from the solvolysis of (27)-ODNB or ( 26)-OPNB under buffered solvolytic

PAGE 70

59 conditions. The limits of detection of (26)-OH for the method used -12 -151 -4 in our product analysis are .10. . gsec. . as little as 1 x 10 weight percent. Assuming that P in our system is as small as 104 (it could be 106), 6Ft = -11.67 kca1/mo1e. This minimally small difference in the free' energy of reaction could at best reduce the ratio k(S)/k(26) to 1.5 (75C). The uncertainty in the evaluation of P makes it diffi-cult to truly assess the difference in free energy in our system. The fact that the change in geometry only produced a change in k(S/k(26) of 4.2-4.5 (25....:75 oC) does not decisively eliminate Tanidas' argument for effect of strain in the rate acceleration, but it certainly ad-vances the argument for the importance of orbital topography in affect-ingchemica1 reactivity. I As part of the characterization of compounds ,(26)-OH and (27)-OH and their respective hydrocarbons, 13Cnmr were recorded. The resonances and their assignments were given previously in Table III. A casual comparison of the chemical shift values between the two hydrocarbons revealed an interesting difference. Parent hydrocarbon (26) had a range of chemical shift values three times that of (27). Since the hybridization of the C-C bonds joining a carbon atom to its neighbors can affect 13C chemical shift values it was assumed, ,perhaps naively, that the (26)-H skeleton varied more widely in overall hybridization of its C-C bonds than did (27)-H. Carbon atoms joined in chemical bonds which are greatest in p-character are . 37b 52 found at highest field relative to TMS (tetramethylsllane). ' . As the chemical shift value increases over the normal r ange for cyclic structures, then the p-character of the carbon to carbon bonds must

PAGE 71

60 be decreasing. When the It spectra of. the alcohols ( 26) and (27) were later obtained, the same correlation was found and Table XII was constructed. It would appear that the skeletal rearrangement of (26) to (27) occurs with a general decrease in the overall p-character of the skeletal bonds and a decrease in the r a n g e of carbon hybridiza-tions. The changes observed in the 13Cnmr chemica l v alues upon re-arrangement may be indicative of an overall lower e n e r gy ground state for the rearranged skeleton. In the discussion of the relevance of the angles 0 and , the third internal angle of the triangle was not mentioned. Later e xamina-tion indicated that it, too, bears some consideration. Recall the diagram on page 50. The extension of the C4 to C2 bond meets the extension of the bond from the leaving group to C9 at an angle calcu-lated to be 113. The corresponding angle involving the C4 to C3 e xtension was found to be only 108. The cyclopropane bent bonds (the Coulson-Moffitt model) are calculated to overlap at an angle of 115. Also available from the structural data are the angles around each carbon atom in the cyclopropane ring of (26). One index of the 45 p-character in C-C bonds is the sum of the three internal angles. In a p urely tetrahedral arra n gement this would be 328.5. The sum of the angles around each of the atoms was calculated tobe C2 273 C 3 284 C9 319. In addition to being closer to C , and having a b etter overlap angle 9 for potential bond formation, C also has slightly morep-character 2 than C • 3

PAGE 72

61 TABLE XII. Comparative 13Cnmr Chemical Shift Values (ppm) f c b c d e d 0 6 c for 26 27 Rearrangement a = 22.5 a = 29.5 7.0 b 19.7 9.8 c 24.2 e 31. 3 7.1 h 48.9 b 42.2 -6.7 g 40.9 1.3 f 39.1 c = 39.7 0.6 e = 34.1 5.6 d 34.7 d 38.4 3.7 i 78.2 i 86.4 8.2 Overall range of chemical shift values (excluding i): 29.2 12.7

PAGE 73

Finally, the internal C -C -C angle of (26)-OPNB which even '+ 3 8 tually becomes the bridge bearing the leaving group in (27) was al-ready 1200 • 62 The most logical conclusions that can be drawn from these bits of information give insight into the nature of the delocalized cation (2 8). The. ground state geometry of (26) is ideal to give with the 28 least motion, ion (28), where there is likely to be more.bonding between C and C than between C and C , or C and C9 ' Cation (28) 2 9 2 3 3 looks exactly like the delocalized ion one would expect from.ionization of rearranged product (27). The rate of solvent attack on non-classical ion (28) to give (27)-OH or internal return to give (27)-OPNB, i.e., the product forming step, being much faster than the rate determining step, is also the step in which relief of the skeletal strain must be occurring. The relief of strain on going from (28) to. (27) apparently increases the activation energy for the reverse p rocess. This would account for the slower rate of solvolysis of (27)-OPNB. In addition to this, the energy barrier marked by the transition state from (26) to (28) is added making the production (26) from solvolysis of (27)OPNB very unlikely. All of this is compatible with the results of the product study in which no (26)-DH was ever detected from solvolysis of (27)-ODNB, or -DPNB.

PAGE 74

63 On the basis of the observed effect of the methylene bridge in (26), reducing the favorable geometry for participation and hence re-ducing the rate acceleration with respect to '(8), it seems very likely that the effect of the zero bridge between Cs and C7 which changes (26) into(1?) is even more detrimental to orbital alignment, leading to the further decrease in acceleration as had been observed. The solvolytic behavior of (26)-OPNB has justified the effort expended in developing an efficient synthetic route for its preparation and the thorough study of all aspects of its geometry. The title com-pound has suggested that the alignment of the cyclopropane edge or-bitals with the site of developing positive charge could well be the major factor which governs the extent of anchimeric assistance and hence the.observed rate acceleration. The need for.further studies I by way of 13Cnmr comparisons and x-ray structural determinations is apparent.

PAGE 75

CHAPTER IV EXPERIMENTAL Synthesis Infrared spectra were recorded on a Perkin-Elmer Model 621 or l37B spectrophotometer using sodium chloride optics. Solution spectra were recorded on a Beckman IR-IO using matched 0.1 mm sodium chloride cells. The absorption band positions reported hereafter are given in wave numbers (cm-l). Nuclear magnetic resonance spectra were recorded on either a Varian Associates Model A-60, 60 MHz spectrophotometer, or a Varian Model XL-lOO, 100 MHz instrument. Chemical shift values for lRnmr are reported in 0 units relative to tetramethylsilane (TMS) at 60.00. The chemical shift values of 13Cnmr are reported in ppm from TMS. Melting points are uncorrected and were obtained with a Thomas Hoover melting point apparatus. Mass spectra were recorded on an Associated Electronic Industries CAEI) Model MS-30 mass spectrometer at 70 eV. Accurate mass determinations were obtained using the same instrument linked with an auxiliary PDP-8 digital computer. Microanalyses were performed by Atlantic Microlabs, Inc., Atlanta, Georgia. Gas-liquid partition chromatography (glpc) was performed with a Varian Associates Model P1440 chromatograph utilizing flame 64

PAGE 76

65 ionization detection. The columns employed were: a) 100 ft. x 0.01 in. capillary coated with diethylene glycol succinate (DEGS), b) 100 ft. x 0.01 in. capillary coated with ueON LB-550, and c) 100 ft. x 0.03 in. capillary coated with DEGS. Integration of analytical glpc tracings was obtained with an Autolab 6300 digital integrator made by Spectra-Physics, Inc. The x-ray data were provided by Dr. G. J. Palenik of the University of Florida Center for Molecular Structure. The x-ray diffraction data were collected on a Syntex automatic four-circle diffractometer using eu Ka radiation (0< 20< 110) to give 1421 signi-ficantref1ections. dierte . (29.) In a typical run, 640 grams (2.35 moles) of hexach10ro-cyclopentadiene (Aldrich H 600-2) were mixed with 475 ml of AR methanol in a three-liter three-necked flask fitted with a mechanical stirrer, addition funnel, and reflux condenser. To this mi xture was added a solution containing 342 grams (6.05 moles) of potassium hydrox-ide in 1400 ml of methanol. The rate of addition was regulated so as to maintain gentle reflux . The mixture was stirred overnight at room temperature and diluted with 1500 m1 of cold water to dissolve pre-cipitated potassium chloride. The product was separated and the aqueous phase e xtracted with 3 x 150 m1 of methylene chloride. The combined organic phases were washed with water and dried.over anhydrous magnesium sulfate. The solution was filtered, the methylene chloride removed by distillation at atmospheric pressure, and the residue

PAGE 77

66 vacuum distilled to yield 450 grams (70%) of a pale yellow liquid, bp 118-120/8 torr. The IHnmr spectrum (CDC13 ) contained only a sharp singlet at 03.33.19 Preparation of 1,2,3,4-Tetrachloro-5-endo-(1,1-diethoxymethyl)-7,7-dimethoxybicyclo[2.2.l]hept-2-ene (30) A solution of 180 grams (0.65 mole) of (29), 100 grams (0.77 mole) of acrolein diethyl acetal (Aldrich 2400-1), and 0.1 gram of diphenylamine was heated with stirring under reflux for 42-48 hours at l30-l35C. At the end of this time the reaction mixture was cooled until it could be handled comfortably and then poured into a beaker. Approximately 50 ml of methanol were added and the mixture cooled to ice bath temperature. In this manner the title compound was crystal-lized in sufficient purity for further reaction. Filtration was , followed with washings consisting of 2 or 3, 15 ml portions of ice cold methanol. Highest purity material could be obtained by a single recrystallization from methanol. The average yield of pure solid, mp 78-78.5C, was 315 grams (80%). The IHnmr spectrum (CDC13 ) exhibited signals for 04.35 (d,l; J = 4.5 Hz; HC(OC2Hs)2); 3.62 (s,3; syn CH30); 3.55 (s,3; anti CH30); 3.55 (m,4; CH3CH20); 2.92 (m,l; J = 5.0 and 9.0 Hz; 5-exo); 2.39 (d of d,l; J = 9.0 and 12.0 Hz; 6-exo); 2.01 (d of d,l; J = 12.0 and 5.0 Hz; 6-endo); and the non-equivalent ethoxyl methyls centered at 1.2 (d of t,6). Found: C, 42.75; H, 5.20; Cl, 35.98.

PAGE 78

Preparation Of endo...:.5"':' (1; l...:.diethoxymethyl) -7, 7-dirilethoxybicyc1o In a typical dechlorination,20 194 grams (0.49 mole) of (30) were mixed with 240 grams tert-butanol, 1420 ml of tetrahydrofuran, and 150-175 grams of finely chopped sodium metal. The mixture was heated under reflux with efficient mechanical stirring for 8 hours, and then allowed to stir overnight at room temperature. The excess 67 sodium metal was removed by pouring the thick reaction mixture through a stainless steel mesh screen. The sodium-free mixture was stirred with 400-500 grams ofice and enough additional water to cause a heavy granular mass of salts to settle out. Filtration, and removal of excess solvent yielded a thick slurry. Water was added until the organic layer separated cleanly. The product was separated, dried with anhydrous magnesium sulfate, and distilled. The average yield . I for the colorless liquid, bp 900/0.3 torr, was 112 grams (90%). Chromatographic analysis (glpc, co1unm a) indicated.that the product was 98% pure. The 1Hnmr spectrum (CDCl) exhibited signals for: 3.21 (s,3; syn CH30); 3.15 (s,3; anti CH30); 2.8 and 2.6 (m .;2; 5-exo, H1 and H4 overlapping); 2.15 (m,l; 6-exo); 1.18 (d of t,6; CH3CH20); 0.82 (q,l; 6-endo). Anal. Calcd. for C H 0: C, 65.60; H, 9.42. 14 24 4 Found: C, 65.67; H, 9.50. Preparatiortof aCid . (33) The synthesis of (33) was accomplished in several steps without purification of the intermediate compounds. In a typical preparation, the acetal (65 grams, 0.25 mole) was hydrolyzed by stirring at room

PAGE 79

68 temperature with about 2 liters of 10% hydrochloric acid for 1 hour, and then extracting with ether. The dried ether extract was evaporated to give a yellow oil. The infrared spectrum of this oil has a strong carbonyl absorption band at 1720 and a smaller one at 1780 em-I. The IHnmr spectrum (CDCI 3 ) showed the absence of all ethoxyl proton resonances while a new one proton singlet was observed at 10.65. The ketal methoxyl protons were still present in the nmr spectrum although the infrared indicated a small amount of bridge hydrolysis had occurred. This oil 21 was dissolved in dry acetone and oxidized at OC with Jones reagent until the calculated amount of reagent had been added (35 ml). At this point, the color change from orange to green was .very slow (re-quiring more-than 15 minutes). The acetone solution was decanted, diluted with an equal volume of water and extracted with ether. The ether solution was extracted with several portions of 10% sodium carbonate. The basic aqueous e xtracts were acidified with 30% sul-furic acid. The solution was stirred at room temperature for 2 hours, filtered, and extracted with methylene chloride. The organic ex-tracts were dried,with anhydrous sodium sulfate, filtered, and evaporated to yield a yellow solid. Recrystallization from hexane gives a pale yellow solid, mp 104-l05C, weighing 25 grams (63% yield from the acetal). The infrared spectrum shows carbonyl absorption bands at 1780 and 1710, and a 2500-3500 region showing the broad absorption typical of carboxylic acids. The.1Hnmr spectrum (CDC13 ) e xhibited signals for: 010.67 (s,l; COOH); 6.55 (m,2; HC=C); 3.3-2.9 (m,3; Hl'HIt , and 5"";exo); 2.25 (m,l; 6-exo); 1.67 (d of d,l; 6-endo). Anal. Calcd. for CaHa03: C, 63.15; H, 5.30. Found: c, 63.13; H, 5.26.

PAGE 80

Preparation Merida -anti "':'bicyC1o [2. 21 Jhep t"':'2"':'en"':' 7",:,ol...:.5"':'carboxylic . acid (34) A solution containing 6.0 grams of sodium borohydride and 1.0 69 gram of sodium hydroxide in SO ml of distilled water was stirred mag-netically at room temperature while a solution of 57 grams of (33) and 24 grams of sodium hydroxide in 400 m1 of distilled water was added over a period of 4 hours.22 The yellow solution was stirred overnight and then acidified with 10% hydrochloric acid. The acidic solution was extracted with 3 x 200 ml portions of ether. The ether solution was washed with water, dried over anhydrous sodium sulfate, filtered, and evaporated to give a colorless to pale yellow syrup (9S"':100%). The compound was considered pure enough for further.reaction and was used without additional purification., The infrared spectrum contained a broad carbonyl absorption at 1700 and typical carboxylic acid absorp-tions obscuring the 2S00-3S00 cmI region. The lHnmr spectrum (CDC13 ) e xhibited the following signals: 86.0S (m,2; HC=C); 3.S0 (broad s,l; HCOH); 2.94 (m,l; H,+); 2.S8 (m,l; HI); 2.S-1.0 (broad m,3; S-exo, 6-exo, and 6-endo); and two resonances which varied their chemical shift value upon change in concentration and were removed upon shaking Preparation of endo-anti-7-Acetoxybicyc10[2.2.1Jhept...:.2"':'erie"':'S-carboXylic acid ( 35) Method A A mixture containing : S7 grams (0.37 mole) of the alcohol ('34), 180 ml (1.7 moles) of acetic anhydride, and 60 grams (0.73 mole) ,of anhydrous sodium acetate was heated on the steam bath,for 1 hour and cooled' . To this mixture was added, cautiously ,S4 ml of water , and the

PAGE 81

24 resulting solution was carefully reheated. for 1 additional.hour. After.cooling,the solution was diluted with an equal volume of water and extracted with 5 x 200 ml portions of methylene chloride. The extracts.were washedfree.of excess acids and dried over anhydrous sodium sulfate. Evaporation of the solvent left a yellow semisolid which on recrystallization from cyclohexane gave 40 grams (55%) of white crystals, mp l03-l04C. The infrared spectrum (KBr) showed C=01735 (ester), 1690 (carboxyl), and y C-O 1240 cm-1 (acetate). y The IHnmr spectrum (CDC13 ) exhibited the following signals: 010.75 (broad s,l; COOH; exchanges with D20); 6.10 (m,2; HC=C); 4.42 (broad s,l; HCOCOCH3); 3.15' (m,2;H4 and 5...;.exo); 2.82 (broad m,l; H1); 2.05 (s,3; COCH3); 1.5 (d of d;l; 6-endo). Anal. Calcd. for C H 0: C, 61.21; H, 6.16. 10 12 4 Found: Ci 61.34; H, 6.19. . Method B 70 A solution containing 58.0 grams (0.376 mole) of (34) in 100 ml of pyridine was stirred at room temperature while 100 grams of acetic anhydride was added dropwise over a 2-hour period.25 The solution was allowed to stir at room temperature overnight. An approximately equal volume of crushed ice was added and the semisolid so obtained was separated by filtration. Washing on the filter with 10 mlof ice cold ether gave .27 grams (37%) of a white solid, mpl02-l03C. This material , vas identical in every. respect with that obtained by.method A.

PAGE 82

71 Preparation . ,.;.7";'Acetoxybicydo [2.2 .1]hept,.;.2,,;,ene,,;,S,,;,carboxylic acid chloride (36) :. Method A A solution of 9 ml (0.1 mole) of oxalyl chloride in 50ml of anhydrous ether was stirred magnetically under nitrogen at:icebath temperature while a solution containing 13 grams (0.08 mole) of (.35) and 8 ml (0.1 mole) of pyridine in 100 mlof anhydrous ether was added dropwise over a period of 2 hours. The resulting mixture was allowed to warm to. room . temperature with continued stirring and nitrogen flush-ing. The mixture was diluted to twice its volume with additional anhy-drous ether and filtered. The filtrate was evaporated to yield a pale yellow oil which was used as obtained. The infrared spectrum showed acid chloride carbonyl at 1790, ester carbonyl at 1735, c-o-at 1240, and -C=C-:at 3300 cn.'-l. The IHnmr spectrum (CDC13 ) exhibited signals for: 06.25 (apparent octet, 1; H3); 6.00 (apparent octet, 1; H2); 4.45 (broad s,1;H7); 3.5-3; 3 (broad m,2; :Hq partially overlapping 5-exo); 2.9 (m,l; HI); 2.10-2.00 (ni,4; 6-exo overlapping COCH3); 1.57 (dof d,l; 6-endo). Method B A solution containing 9.0 grams (0.046 mole) of (35) was added to a chilled solution of 10 ml of oxalyl chloride (14 grams, 0.11 mole) in ether. The addition required 2 hours, after which the ice bath was removed. The solution was stirred at room temperature.for an additional 30 minutes.tobe sure all carbon monoxide had evolved. The solvent was removed on a .rotary evaporator • . The residue (9.8 grams, . 100%) had IR and nmr spectra identical to that previously obtained . . This preparation results in a pyridine-hydrochloride-free product.

PAGE 83

72 A solution containing , 16....:l7 grams of the-acid.chloride was added with_stirring at ice bath temperatures to an ethereal solution of diazomethane.22 The addition required about an hourandupon.comple-tion was allowed.to stir an additional hour with warming to. room tempera-ture. The solution was filtered and concentrated on a rotary evaporator at room temperature or below. The crude diazoketone was used as pre-pared without. further purification, but could be chromatographically purifled on Alumina (neutral Woehlm) using 1:1 as the eluant. .Purification was accompanied by loss of nitrogen and an average loss of material of about 30%. The infrared spectrum showed -C=N=N at 2100 and diazoketone carbonyl at 1640 cm-I • The original ester absorptions at 1735 and remained unchanged. The lHi:unr \ spectrum (CDC13 ) exhibited signals for: 06.15 (m,l; HC=C); 5.90 (m,l; HC=C); 5.30 (s,l; HC=N=N); 4.30(m,1;H7); 3.05(m;1;H4); 2.90 . (m:,l; Hi); 2.9-1.5 (m,6; 5-exo, 6-endo, and COCH3). Preparation of Diazomethane For l6-l7grams of acid chloride (36), the following procedure26 provided a solution approximately 0.25 molar in diazomethane, a twofold excess. A solution of 12 grams of sodium hydroxide in 30 ml of water was chilled in an ice bath and 40 ml of ethylene glycol monomethyl ether were added. The solution was .stirred magnetically and , 600 ml of techni-cal grade ether added. To this mixture was added 6S grams of Dupont Nitrosan.The mixture was stirred for 30 minutes and the ice bath was replaced by a water bath.. The ethereal. solution of diazomethanewas

PAGE 84

73 distilled.using an efficient condenser until .the distillates were no longer brilliant yellow in color. (39) Diazoketone(37) was dissolved in 650ml of anhydrous tetrahydro-furan and mixed with an equivalent weight of electrolytic copper dust. The mixture was stirred magnetically and heated slowly untilrefluxing began. A spontaneous reaction occurred which was allowed to continue without further heating. When this reaction subsided, heating was continued for another hour. After the opservation of spontaneous re-action, 5 ml samples were withdrawn from the flask and evaporated on a sodium chloride window. The residue was examined for the presence of the 2100 cm-1 absorption band in the infrared spectrum. Refluxing was continued until the sample no longer contained this absorption band. Complete reaction was observed to require from 1 to 24 hours with the best yields obtained at the shortest reaction times. When the reaction was complete, the mixture was cooled, filtered, and the filtrate evaporated. The resulting residue was boiled with three successive portions (150 m1 each) of ether and discarded. The ether extracts were evaporated, weighed, and mixed with 1.5 grams of semicarbazide hydrochloride and 2 grams of sodium acetate per gram of residue.24 The crude semicarbazone was obtained by diluting the reaction mixture with an equal volume of water and extracting withinethylene chloride. The organiclayers were dried and.evaporated to yield a gummy white solid which was recrystallized from chloroform-hexime.. The white solid so obtained retains solvent and melts about 162C. After drying under

PAGE 85

74 vacuum for 24 hours, the product melts l83-l84C. The yield overall from the acid (35) ranged from 0 to 42%. The infrared spectrum con-tained -NH at 3200, acetate -c-o at 1240, carbonyl at 1735, and amide I and II absorptions at 1580 and 1670 -1 cm The poor solubility of the compound in common deuterated solvents gave extremely dilute solutions and poor 1Hnmr spectra. Those resonances observable (CDC13 ) were 05.5 (-NE2); 5.0 (Hg); and 2.05 (s; COCE3). Anal. Calcd. for C12H1SN303: C, 57.81; H, 6.06; N, 16.86. Found: C, 57.66; H, 6.13; N, 16.63. Preparation of CuO-Cu Catalyst A solution containing 100 grams of copper sulfate pentahydrate in 350 ml of distilled water was stirred vigorously while 35 grams of . 30 pure zinc dust were added over 30-45 The solid was washed by decantation until washings tested sulfate free to barium hydroxide solution. The red solid was stirred with 250 ml of 5% hydrochloric acid for 1 hour and decanted. A second 250 ml portion of acid was added and the misture left overnight. The copper powder was filtered off, washed with water until filtrates were neutral to pH-Hydrion paper, and drained an additional 30 minutes over suction. The solid was transferred to a quartz tube and heated in a tube furnace at 250C 29 until no more water was observed, then to 500-600C for 15-20 hours. The black solid was cooled, ground, bottled, and kept in a desiccator until used. Average yield, 27 grams.

PAGE 86

Preparatiou' cifciriti [4; 3. 0 02 . ' (38) In a typical run 13.5 grams (0.063 mole) of the pyridine-hydrochloride' free acid chloride.' (36) in 250 ml of anhydrous ether were added to an excess of freshly distilled.ethereal diazoinethane at OC.When addition was complete (about 1 hour) the ice bath was 75 removed and the solution was stirred for 1 additional hour. Precipi-tated polymeric materials were removed by suction filtration and the filtrate was reduced in volume on a rotary evaporator. The last half of the solvent was removed at temperatures no higherthan.25C. The re-suIting diazoketone in 200 m1 of tetrahydrofuran was added to 800 ml of cyclohexane. Approximately 100 m1 of additional tetrahydrofuran were added to completely dissolve the ketone in case it oiled out. The CuO-Cu catalyst (17 grams) was added and the flask fitted with a long efficient reflux condenser. Two 250 watt Sylvania Industrial Infrared lamps29 were positioned on opposite sides of the flask 2 to 3 inches from the surface of the glass and slightly above the liquid level. After an average of 2 to 4 hours from the time the lamps were illuminated, the 2100 em-I band had completely disappeared from the infrared spectrum of the reaction mixture. The lamps were removed, the solution allowed to cool until it could be handled, and the catalyst removed by suction filtration. The solvent was removed on the rotary evaporator and the resulting yellow residue chromatographed on Woehlm neutral alumina using a solution of 20%.petroleum ether in diethylether • . The second fraction to be eluted was found.to contain 10.6 grams (80% yield based on. (35 of was suitable for semicarbazonepreparation. Two succeeding runs employing

PAGE 87

76 .the.CuO..,.Cucatalyst.andpyridine-:-hydrochloride freeacid.chloride produced72% and 77% yields, . respectively The eyclization is. sensitive to the amount of "catalyst" employed and the lower yield was obtained when less catalyst was used • . The infrared spectrum (neat) . contained the following absorption bands: . 3050(w), ,2980(m), ' 2880(w) ,1730(s), 1375{m), l240(s), l205(m),1050(s), 9lO(m),870(m), S60(m), 830(m), and 740(m) cmI • The IHnrnr spectrum (CDC13 ) consisted of 64.98 (broad s,l; Hg); 2.:67-1.52 (broad ni,8); and 2.03(s, 3; COCH3). Preparation of anti ..;.TetracyCl6 [4.3.0 ;02,4 .03,8 ]nonan..;,9..;.ol . (26) A solution containing 5.0 grams (0.020 mole) of.(39), 10 grams of potassium hydroxide, and 2.5 ml of anhydrous hydrazine iIi 50 rn1 of ethylene glycol was heated to.135-l40C in an oil bath and maintained at that temperature for 1 hour. The reaction was carried out in a distilling apparatus which contained a 10 cm vigreux column. After heating for 1 hour the temperature of the reaction mixture was raised slowly to185-l90C and the water and other lower boiling materials were removed. The temperature was held at 190C until solid could be observed subliming into the condenser. Reaction was continued until substantial material had collected in the condenser and the reaction mixture was cooled to room temperature. The sublimed material was rinsed from the apparatus with pentane and the solution set aside. An additional 10 grams of potassium hydroxide and 2.5 ml of hydrazine were.added.to the.reaction flask, and the.heating cycle ' was repeated. When additional product had again filled the condenser, the.reaetion was cooled, diluted with an equal volume of water, and transferred to a continuous liquid extractor • . The mixture was extracted for 48 hours

PAGE 88

77 with pentane and returned to the distillation apparatus. The water was removed and the whole process described above was repeated. When product was no longer observed subliming into the condenser, the re-action mixture was extracted as before and discarded. The combined pentane extracts and washings were evaporated to yield a white crystal-line solid mp l45-l46C (with sublimation). The material obtained as described is sometimes pale yellow, but can be purified by sublimation at 70/760 torr, or recrystallized from small amounts of pentane. Overall yield in this manner from (39) was 2.0-2.2 grams (75-80%). This is an adaptation of the method of Babiak.33 The infrared spectrum (KBr) contained the following absorption bands: 3l50(s), 2950(w), 2880(m), 2800(w), l260(m), 1080(s), 1070(s), 930(w), 840(w), 825(w), 790(m), 740(m) cm-l. The IHnmr spectrum (CDC13 ) consisted of: 04.35 I (broad s,l; H ); 2.4-0.8 (broad m,lO; all other methylene and methine protons); and the concentration dependent hydroxyl proton 1.82-1.72 (broad s,l). The l3Cnmr spectrum (CDC13 ) contained nine resonances, consistent with the asymmetric structure of the carbon skeleton. The chemical shift values and assignments are given in Table III. Mass spectrum: 70 eV, m/e 136 (M+). The fragmentation pattern is given in Table IV. Anal. Calcd. for C9Hl20: C, 79.36; H, 8.88. Found: C, 79.60; H, 8.80. Preparation of anti-Tetracyclo[4.3.0.02,4.03,s]nonan-9-yl pbromophenylurethan (26)-PBPU To a solution containing 205 mg (1.50 mmole) of (26) dissolved in 5 ml of anhydrous benzene was added 327 mg of para-bromophenyl

PAGE 89

78 isocyanate ... The .mixture was warmed .for 5 minutes. (steam bath) and the benzene evaporated;24 . . The residue was partially dissolved. in. boil-ing hexane, filtered, and cooled . . Three additional recrystallizations with hot filtrations were required to produce a constant melting pro-duct, mp. 129'7"130C; yield. 270 mg(50%).Thesamplewas dissolved in ether, diluted with an equal volume of hexane in a dust-free container with a loose fitting top, and the solution allowed to evaporate slowly. The crystal clusters so produced were suitable for use inx-ray struc-tural analysis. Anal. Calcd. forC16H16N02Br: C, 57.50; H, 4.82; N, 4.l9;Br, 23.91. Found: C,57.43; H, 4.89; N,4.l9; Br, 23.81. Preparation of anti "';'TetracyC16 [4 3 0 It 03,8 ]nonan"';'9-yl prtitrobenzoate (26) "';'OPNB ... . . . ( .. . : 24 Using standard procedures, . . 200 mg (1. 47 mIDole) of alcohol (26). were dissolved in dry pyridine and to.the solution was added 220 mg of freshly recrystallized p -nitrobenzoyl chloride. The mixture was swirled to mix thoroughly and stored in the refrigerator overnight. The cold mixture was diluted with 10 grams of ice and the oil thus deposited was extracted with ether. The ether extract was washed successively with water, 5% aqueous hydrochloric acid, 10% sodium car-bonate, water, and saturated aqueous ammonium chloride. Evaporation of the solvent'gave a yellow solid which was recrystallizedfrom.ethanol-water in cream-colored plates (95% yield), mp tionfromheXane gave the compound as fineneedlesmp87....:88C . . The 1 Hnmr spectrum (CDCl ) was undistinguished except for the appearance . 3 of aromatic protons at 08.'24 and a downfield shift of Hg-to05.30.

PAGE 90

Mass spectrum: 70 eV, m/e285 (M+). Anal. Calcd. C, .67.36; H, 4.91. Found: C,67.l6; H, 5.39; N, 4.91. Preparafioriofci:riti -Tetracyclo 3. 0 cf-,4 ] rionan":9":yl 3; 5. diriitroberizoate (26}":ODNB In the same manner as described for' (26) -OPNB, 2.6 grams 79 (0.019 mole) of alcohol (26) dissolved in 10 ml of anhydrous pyridine were treated with 5.9 grams (0.026. mole) of 3,5-dinitrobenzoyl chloride. Upon swirling, a spontaneous exothermic reaction occurred. The result-ing mixture, when it had cooled, was treated as before and the product isolated as a pale yellow solid by filtration.' Recrystallization from ethanol gave 5.8 grams (88%) of pale yellow crystals, mp 118-ll9C. This derivative was insoluble in 60-80% acetone-water solutions at \ room temperature. The infrared spectrum (KBr) exhibited carbonyl stretch at 1720 cmI as well as bands for -C-O at 1280, and N02 at 1540 and 1340 cmI . The IHnmr spectrum (CDC13 ) contained proton resonances at 09.16 (m,3; aromatic); 5.34 (broad $,1; H9); and 3.0-1.0 (m,lO). Anal. Calcd. for C16H14N206.! C, 58.18; 4.27; N, 8.48. Found: C, 58.15; H, 4.27; N, 8.56. A solution of 272 mg . (2 .00 mmole) of alcohol. (26) dissolved in 5 m1 of methylene chloride was added in one portion to a solution con-taining 3.1' grams of _ chromd.um trioxide-dipyridine complex in 60 ml of 1 . hI' "d 34 methy ene c orl e. The mixture changed from blood red to black immediately. .Themixture was -a1lowedto.stir for 1 hour at room tempera-ture and then filtered and evaporated. The residue was extracted with

PAGE 91

80 boiling ether. and filtered_to remove last traces of chromium .polymers. The ether was evaporated to. yield a pale yellow semi-solid-whichsub-limed readily at 65 /760 torr ... The sublimed material (185 mg, 70%) wasstill.sticky as ifwet,but had mp96"":97C (sealed Analysis byglpc on columns a and b showed only one substance present when the injection port temperature was held below 140C. When the injection port.temperature was increased to 200C, a new peak was observed at shorter (10 minutes compared with the ketone at 37 minutes, column:b) .retention times due to decarbonylation of the ketone in the injection port. The infrared spectrum (CC14 ) contained the following absorption bands: 3020.(m) , 2960(m), 2880(m),l765(s),1470(w), l360(w), l3l5(m), l150(m), lllO(m), 1080(m),820.(s), 740(w) cm-I • The IHnmr spectrum (CDC13 ) consisted of a broad complex multiplet 0.8 to 2.50. ( Calculated Accurate Mass: 134.0731 mass units. Found: 134.0734 mass units. PreparationofChromitim Trioxide-dipyridineComplex A one-liter three-necked flask was thoroughly dried and equipped for mechanical stirring under nitrogen. Eighty grams of chromium tri-oxide, which had been dried over phosphorus pentoxide under vacuum for 24 hours, were added in very small portions to 600 ml of anhydrous pyridine maintained at l5-20C. The addition required 5 to 7 hours and at the end of that time the complex had precipitated. The .red solid was isolated by decantation and washed by stirring with 5 x 150 ml portions of anhydrous ether with subsequent decantation • . The .wet complex was transferred to a vacuum desiccator over phosphorus pentoxide and.evacuated.to 20 torr.for 8 hours. The dried. complex was stored in

PAGE 92

81 the same manner. For oxidation, the complex was dissolved in methylene chloride as a 5% solution and used in a 6:1 mole ratio to the alcohol. Oxidations were usually complete in 5-15 minutes at room temperature.34 Preparation of Bicyclo[3.2.l]octa-2,6-diene (46) A sample of ketone (45) weighing 125 mg was heated in a sealed tube at 200C for 15 hours. The tube was opened and the contents taken up i n ether. The ether solution was passed through a short alumina column and evaporated to give the diene which eluted as a single peak on columns a and b (glpc). The infrared spectrum (CHCl ) was identical with that reported by other workers.35 The IHnmr spectrum (CDC13 ) was also identical with that previously reported, and contained the follow-ing resonances: 86.28 (d of d,l; J I 7 = 2.8 Hz; H7); 6.08 (broad d of d,l; J2 3 = 9.8 1Hz; H2); 5.76 (d of d,l; J6 7 = 5.5 Hz, J5 6 = 2.6 Hz; H6); 5.25 (broad d,l; H3); 2.68 (broad m,2; HI and Hs); 2.36 (H ); 2.00 (d of d,l; J = 9.5 Hz, JIa = 4.0 Hz; Hsa); 4_exo ab 1.90 (J = 18.0 Hz; H4 d); 1.76 (d of d,l; JIb = 0.6 Hz; exo-endo -en a Hsb ) • Preparation of Tetracyclo[4.3.0.02,4.03,s]nonan-5-semicarbazone (43) This compound was prepared according to the method of Nickon 27d et al. and gave the desired product in 16% yield overall in eight steps from methyl acrylate and cyc1opentadiene. Semicarbazone ( 43) had mp 205-206C (literature value 204-205C). The infrared spectrum (KBr) contained absorptions for -N-H at 3300, -NH2 at 3150 (doublet), and amide I and II bands at 1670 and 1585 em-I, respectively.

PAGE 93

82 'Preparation Of TetracyClo [4 3.0.02 , 3,H 8 ]nonane ' &4). MethodA32 A solution containingS.O grams (0.026 mole) '.of ,semicarbazone in S.ml, of ethylene glycol was , added, all, at once to, a solution of S grams of ,potassium hydroxide in SOml of ethylene glycol maintained at 130C. The reaction mixture Was contained in a 100ml round-bottomed flask which was attached to a 10 cmvigreaux column topped by a thermometer and connected to a water-cooled condenser. This same apparatus was employed for all subsequent reductions of this kind. After the " addition, ,the solution was heated ,to 160C for 30 minutes and then to 190C for 2 hours. The reaction mixture was cooled ,to 120C ,and the more volatile material removedundervacuum:(4 torr) and the syrupy plastic residue cooled to room ,temperature.' The cold ' amorphous mass I was stirred with2S0 ml of cold water and the white solid which precipi-tated was filtered off, washed with additional cold water, and dried. The solid was recrystallized from methanol-hexane to give 1.8 grams (S3%) of white needles; mp ,189-190. The infrared spectrum (KBr) contained a -C=N absorption at16S0cm-1 as well as aliphaticC-H bending and stretching bands. The IHnmr spectrum (CDC13 ) was a series of complex overlapping multipletsbetween 01.4 and 3.S. The mass spectrum (70 eV) showed mle 264 (M+). The compound was identified as bis(tetracyclo[4.3.0.02'''.Os,sJnonan)-S-azine. Anal. Calcd. ,for C.1SH20N2: C,81.77; Ii, 7.62; N,10.S9., 'Found:, C,81.87; Ii, 7.70; N, 10.51. Method ' B33 ' A mixture, containing 1.0gram,(O.OOS2mole),of(43)" 2.0 grams of potassium hydroxide, and O.S mlofhydrazine in 20ml of ethylene

PAGE 94

glycol was placed in the apparatus.' previously described . and heated . slowly to 160C • ' .. As. the temperature. rose above .135C, . ' the product' hydrocarbon was . obserVed slowly subliming into the condenser • . After 2 hoursthe.reaction mixture wasc06led and the product washed from 83 the apparatus with ether • . ' The . ether' s .olution was washed' with water and dried over anhydrous sodium.sulfate. Evaporation of.the solvent produced a white waxy solid which sublimed readily at room temperature and atmospheric pressure. The product, obtained in 60% yield, had mp 116-117 . (completely.submerged sealed tube). The IHnmr spectrum (CDC13 ) exhibited resonances at 02.55 (broad s,l; H1); 2.35 (broad s,l;Ha); and 0.8-2.0 (complex multiplet, 10). The 13Cnmrshift values and assignments are consistent with the expected hydrocarbon product and are listed in Table IILThe mass spectium(70 eV)had .m/e120 (M+, 17%), 105 (12:7.), 92 (15%), 91 (17%), 79 (10%), 78 (17%),. 77 (16%), and 66 (100%). Anal. Calcd. for C9H12: C,89.93; H, 10.07. Found: C, 89.91; H, 10.08. Preparation Of anti-TetracyClo [3 3. L 02 , 4 ] nonan":' 9":'0 1 . (27) A mixture containing 2.0 grams (0.0070 mole) of (26)-OPNB in 100 m1 of 70% acetone-water was refluxed for 30 hours and cooled. The acetonewas removed on the rotary evaporator and the resulting mixture dilu with 70 ml of water in which 2.0 grams of sodium hydroxide were dissolved • . The .mixture washeated,to.reflux for 1 hour, cooled to room temperature; and extracted with 5 x 50 m1 portions of pentane • . The .pentane extracts were dried and.evaporated to yield ayellow.semisolid. Sublimation of thismaterial'gave 0.9 gram (94%) of white .

PAGE 95

84 crystalline product, mp 236-2370 (completely submerged sealed tube). Analysis by glpc of the product on. column a revealed. only one. peak eluting after .22 minutes, with no evidence for. alcohol (26)whi.ch. normally eluted at 27 minutes • . The infrared spectrum (KBr) . contained the following absorption bands:'. 3200 (m), 3040 (w) , . 2940 (m), 29 20.(w) , 2840{w), l340(m), 1060(m), arid 1030(m) cm-1.The IHnmr 'spectrum (CDC13 ) exhibited resonances at: 04.0 J = 5.0 Hz; . Hg); 2.4 (broad m,2; HI and H s); and 2.4-L2 (complex m,9), The 13Cnmr tained six resonances, consistent with the symmetrical carbon skeleton. The cheinicalshift values and assignments are listed.in Table III. The same alcohol was also prepared in greater than 90% yield by solvo1y-sis of. (26)-ODNB in 60% acetone-water in a sealed tube at 125C for 6 hours. ACcu:r'ate Mass Anal. Calcd. for C9H120: 136.0887 mass units. Found: 136.0898 mass Preparation'Of p nitrobenzoate . (27) ..;.OPNB In the manner previously described, 200 mg (1.47nunole) of alcohol (27) were dissolved in dry pyridine, and 300 mg of p-nitrobenzoyl chloride were added. Standard workup and recrystallization from pe-troleum ether gave . 350 mg (84%) of pale yellow plates, mp :l43-l44C. Anal. for C16HISN04: C,67:35; H, 5.30; N, 4.91. Found: C,67.50; H, 5.38; N, 4.83. Preparation' of 'anti";'TetrcicyC!o [3 3 .1.02,4 ]nonciri,.;.9";'yl 3 ;5-diriitrobenzoate •. (27)';'ODNB . . In the. manner previously described, 1. 4 grams : (0; 010 mole) of alcohol .(27). were dissolved in :10 m1 of pyridine and 3.5 grams (0.016

PAGE 96

mole).of 3;5-dinitrobenzoylchloride.were.added.Standardworkup followed by.recrystallizatioQ. from gave .yery ;fine ye11ow .needles, mp l73-174C • . The . 1Ihunr spectrum' (CDC13 ) . exhibited resonances at 09 (m, 3 ; aromatic); 5 . : 07 . H 9); 2.75 (broad s,3; HI' Hs ; ' andH7 ) ; 2.4-1.6. (hroadm,7). Anal. Calcd • . for C, 58.18; H ' , 4.:37; N," $.48. Found: C, 58.17 ; H, 4.31; N, 8.47. Preparation Of TetracyClo [3 3; 1.02,4.03,7 ]ri.onan...:.9"':'cine(47) A solution containing 200 mg (1.47 mmole) of alcohol (27) in 85 methylene chloride was oxidized with a .5% solution (60 ml) of chromium trioxide-dipyridine complex in.methy1ene chloride. The reaction mix-ture was allowed to stand at room temperature forlO minutes, filtered, and the solvent evaporated. The resulting semi-solid was sublimed to give 140 mg (70%) of colorless crystals; mp l6l-l62C. Analysis by glpc on column a revealed only one peak at about 32 minutes. Injec-tion port temperatures above 300C were required to produce a tracing showing as much as 1-2% of the decarbonylation product, which eluted at 4 minutes. The infrared spectrum (CHC13 ) contained absorption bands at 3055(w), 2945(m), 1060(w), 1020(w)cm-l • The IHnmr spectrum (CDC13 ) resembled that of alcohol (27). with the exception of the absence of the proton signal for H9 and the slight shift of the bridgehead multiplet from 02.4 to 2.6. Calculated accurate mass forC9H100: 134.0730 mass units • . . ;Found: . ,134.07 31 mass . units. A . ketone. identical in. all. respects was ' prepared iri '85% yield by. treat, ment of alcohol (27) with Jones. reagent.

PAGE 97

86 Preparation of Tetracydo [3 . ' 3 1:02:,4 ]rioriarie(49}. , , . Alcohol. (27.).: (136 mg; r.OOmmole): was dissolved:in O.Smlof thionyl chloride at room temperature and stirred for 3 ,hours. :Excess thionyl chloride was removed on a rotary evaporator. to give a white solid • . Chloride (48) was dechlorinated without further purification by refluxing in 5 mlof tetrahydrofuran containing O. 2ml of t-butanol and 0.1 gram of sodium metal for 6 .hours. Workup of the dechlorina-tion reaction gave a white semi-solid. Sublimation of the crude product gave 85 mg . (70%) of the hydrocarbon "triaxane" (49) • . The hydro-carbon is very volatile and evaporates rapidly at room temperature. "Triaxane" had mp180"':18lo (completely submerged sealed tube) (literature value 183"':184. 5C) .37 The IHnmr spectrum (CDC13 ) exhibited =: c a resonances at: 02.46 (broad s,3; = 11 Hz; J = 1.5 Hz, J bc be 10.5 Hz; H ); 1.28 e 5.0 Hz; H ); 2.01 (m,3; H ); 1.61 (m,3; J = b c ae (d,3; H). , The 13Cnmr spectrum (CDC13 ) contained resonances at 47.1, a 40.3, and 37.1 ppm relative to TMS (see Table III) which was consistent with those previously reported.37,38 Preparation of 7"';'Norbornadieriyl p"';'riitroberiioa te (12) "';'OPNB . ' . .' ' 53 This ester was prepared using the product was recrystallized from petroleum..,.ether-carbon tetrachloride as long pale yellow needles; mpl04:"'105 (lfterature'valuelOl:";102C) in 95% yield.

PAGE 98

Anal.'Calcd. C;:65.36; H, 4.31; N, 5,"45. Found: , C,:65.27;H,'4.36; N,,5.45. 'Preparation of -7":'Norbornadieriy13; 5..:.dirii troberiioate ' ,(12) ":'ODNB In-the manner previously described, 2.16 grams ,(0.020, mole) ,of norbornadienol were dissolved in 10 ml of dry pyridine and 5.9 grams (0.026 mole) of 3,5-dinitrobenzoyl chloride were added. Standard workup followed by recrystallization from petroleum tetrachloride gave 3.9 grams (65%) of yellow needles; -mp 128-1301) , (dec.). The infrared spectrum (KBr) contained absorption bands at 3000(m),1720(s), l545(m), l345.(m), and l280(s) cm-1 • The IHnmr spectrum (CDC13 ) exhibited resonances at: 09.16 (apparent t,3;aro-matic); 6.80 (d of t,4; HC=C); 4.90 (broad s,1;' H7); 3.83 (m,2; HI Anal. Calcd. for C H NO: C, 55.63; H,3.34; N, 9.27. 14 10 2 6 Found: C,55.55; H, 3.39; N, 9.33. Preparation of anti..:.7":'Norbornenylp-rtitrobenzoate (2)":'OPNB In the manner previously described, 1.0 gram (0.0091 mole) of 87 alcohol (2) was dissolved in 5 ml of dry pyridine and 2.8 grams (0.015 mole) of recrystallized p-nitrobenzoyl chloride were added. Standard workup followed by recrystallization from ethanol gave 2.1 grams (84%), of pale yellow crystals, mp122-123_ (literature value l18-1l9C). The IHnmr and infrared spectra-were consistent with those ,reported in , S4 the literature.

PAGE 99

88 . This ester was prepared ineXactly the samemanner.as. (2)-OPNB. Recrystallization from etharlOl gave. 2.5 grams (85%) ofver' fine creamcolored needles; mp134:....l35C. .The1lliunrspedrum (CDC 1 ) exhibited . . 3 . resonances.at: 09.18 (ni;3;aromatic); 6.35 {t,2;. HC=C); 4.67. (broad s,l; H7); 2.98 (broad m;2; Hl and HIf.); 1.93 {m;2; H); 1.23 (m,2; exo H . ) endo Anal. Calcd. for CHH1ZNz06: C,55.26; H,3.98; N, 9.21Found: C, :55.20; H, : 4.02; N, 9.23. Studies P:teparcitioIi of Kinetic Solutions The standard ,sodium hydroxide titrant was prepared by diluting a 0.1 N Acculute solution to 1 liter. As supply demanded, 100 m1 aliquots of this solution were diluted to 1 liter and standardized against primary standard grade potassium hydrogen phthalate using a phenolphthalein endpoint. The normality of the solutions so prepared ranged from a low of 0.01003 to 0.01025 N. The indicator solution for p-nitrobenzoate ester kinetics was prepared by dissolving 1 gram of phenolphthalein in 60 mlof ethanol and diluting with carbonate free water to 100 ml. Carbonate free deionized water .wasprepared by distilling the water and passing. it. through an Amberlite MB-3 ion resh.5S

PAGE 100

89 grade.acetone was .stirred with 1 gram of potassium permanganateand 10 grams of .anhydrousccilciumsulfate for 2 days. The solvent was ' filtered, . and stored.over molecular Appropriate volumes of acetone and water were mixed and stored under nitrogen pressure.' The ' density of the solutions was evaluated using a Kimax pyncnometer • . The 70% acetone:watermixturehad a density of 0.9113 grams/ml. The 60%acetorie:water mixture had a density of 0.9397 gramS/ml. Preparatioriof decarbonylationmaterials The solvent used for ketone.decarbonylation kinetics was chosen because of its long glpcelution time relative to the products under study. For this purpose 1,2,4-trimethylbenzene was distilled and the chromatographically pure fraction used as the solvent. Marker compounds were chosen for retention times and molecular weights. Benzene, . toluene, and m-xylene met these requirements. The reagent grade materials were distilled until a fraction showing only one glpc peak was obtained. Kinetic Procedures A carefully weighed sample of the ester was dissolved in acetone: water solution and made up to 100 ml at 20C. With a hypodermic syringe equipped with a 6-inchneedle, 5.2 mlsamples.were dividedamongl8 medium wall: glass" ampules flushed with:nitr:ogen. Each tube was frozen. in . liquid nitrogen" , . pumped ' to O. 5 torr;. and. sealed. The tubes were . divided. into two . .The nine "tubes. of a set . were

PAGE 101

90 innnersed in a preheated. constant. temperature silicone oil. bath.. After an equilibration time of from 5 . tol5 minutes, . the first sample was removed to.an ice bath, this being the zero time. One tube was left in the bath for 10 .x t and was used to obtain the actual infinity . titer. Prior to opening, the.tubes.wereequilibrated.to 20C in a water sample was removed by. means of a constant delivery pipet calibrated with acetone:water solutions. The sample was delivered into a flask containing 10 m! of carbonatefree water and 2 dropsof phenolphthalein solution. The samples were titrated to the firstap-pearance of a permanent pink color. Ketones A carefully weighed sample ofthe.ketone was dissolved in 2m! of l,2,4-trimethylbenzene and the marker compound (weighingIO% that I of the ketone) was added. A hypodermic syringe was employed in trans-ferring 0.1 ml of the sample to each of 18 car ius tubes which had been flushed with nitrogen. The samples were frozen in liquid nitrogen, pumped to.O.5 torr, and sealed. Zero and infinity samples were taken as for the esters. All ampules were stored until one run was complete. In preparation for analysis, ampules were frozen acetone-dry ice until all material had condensed in the bottom of the tube. The tube was opened and the contents transferred to a tightly stoppered vial. A typical sample was inj ected. into the gas chromatograph five times in succession.and the diene-marker ratio measured by the digital integrator. The .ayerage of: :the five values was . used. to calculate the rate constanL

PAGE 102

91 AJialysisOfData .Theraw datawere.plottedmanually.to.insure linearity of . the resul ts ',and then processed by computer. using a FOCAL 8 program writ ten by Dr. Roy King • . This program was easily adjusted either-milliliters of sodium hydroxide; milligram or microvolt ratio of standard to product, versus time • . This least squares treatment pro-duced and printed the In obserVed,whlchcould also.be.used for graphic plots, along with the standard deviation and the rate constant. The data are presented for each run in.thetables below, along with.the molarity of the solution and the.temperature of the bath. Thepercent of theoretical infinity is given in parentheses after the actual in-finity titer. The half life was obtained by dividing In 2 by the rate constant. I For all compounds studied, rates were determined.at three temperatures. When data were available on the same compounds from other sources, rates were recomputed through the FOCAL 8 program to insure compatibility with our own rate constants for the same compound.

PAGE 103

. Solvolysis of 7..,.Norbornadieny1p -nitrobenzoate . . . . '(zg,)"';'OPNR' in 70% Aee tone" Time , . (seen 360 1460 1860 2060 2260 2460 2660 Inf. . Run .111 Acid Titer '(ml) , 0.315 1.425 1.645 ' 1.910 1.960 2.280 2.280 3.715 (97%) Temperature: 125 0.2e Molarity: 0.008M k = 10-4 0.14 sec.-1 t = 1883 sec. . . Time (sec.) 360 1060 1260 1460 1660 1860 Inf. Run 112 Acid Titer , (nil) , 0.210 1.010 1.245 1.385 1.630 1. 725 3.790 (98%) Temperature: 125 0.2e Molarity: ' O. 008M k = X 10-4 0.11 sec.-1 t , = '1853 sec. , In . Observed -0.4838 -0.5848 ' -0.7218 -0.7499, -0.8550 -0.9512 -0.0570 -0.3099 -0.3982 -0.5623 -0.6072 92 Published rates.for this ester.in 65% acetone:wat:er .. at125e were 5.8 X 10-4 ' 0.01 in :70%, acetone:wateratI25e, 3.71 x 10-4 •19c

PAGE 104

Time (sec. ) 300 3800 7800 10700 14500 18030 In. Run 113 Acid Titer (m1) 0.290 0.930 1.545 1.960 2.380 2.750 4.610 (102%) Temperature: 100 O.le Molarity: 0.0076M k = 4.74 X 10-5 0.045 sec.-1 t = 14,619 sec. Time (sec.) 560 3750 7400 10800 14600 20000 In. Run #4 Acid Titer (m1) 0.240 0.810 1.400 1.890 2.340 2.880 4.610 (102%) Temperature: 100 O.le Molarity: 0.0076M k = 4.77 X 10-5 0.053 sec.-1 tl = 14,513 sec. 93 In Observed -0.0650 -0.2253 -0.4082 -0.5537 -0.7264 -0.9079 In Observed -0.0535 -0.1932 -0.3620 -0.5276 -0.7087 -0.9803 The data of other workers19c were analyzed using the same focal program. Temperature: 90.2 0.05e Molarity: approximately 0.007M k = 1.79 X 10-5 0.01 sec.-1 (Run #1) = 38,715 sec. k = 1.77 X 10-5 0.01 sec.-1 (Run 112) = 39,152 sec.

PAGE 105

Solvolysis of 7-Norbornadieny1 3,5-dinitobenzoate (12)-ODNB in 60% Acetone Time (sec.) 300 7400 13600 27400 35700 58200 72000 85700 94000 Inf. Run 111 Acid Titer (m1) 0.175 0.500 0.730 1.255 1.560 2.100 2.360 2.625 2.820 3.790 (96%) Temperature: 60 0.05e Molarity: 0.0067M k = 1.36 X 10-5 0.03 sec.-1 t = 51,Q05 sec. Time (sec.) 300 8700 15100 28200 44200 59600 72400 87800 102900 Inf. Run 112 Acid Titer (m1) 0.175 0.535 0.790 1.260 1.650 2.060 2.430 2.624 2.870 3.745 (98%) Temperature: 60:<>. o.osoe Molarity: 0.0065M k = 1.36 X 10-5 0.03 t = 50,869 sec. -1 sec. In Observed -0.0473 -0.1415 -0.2140 -0.4022 -0.5304 -0.8078 -0.9749 -1.1799 -1. 3630 In Observed -0.0479 -0.1542 -0.2369 -0.4101 -0.5808 -0.7989 -1. 0468 -1. 2073 -1. 4539 94

PAGE 106

Time (sec.) 360 2000 3760 5400 7460 9000 10860 14760 Inf. Run 113 Acid Titer (ml) 0.200 0.515 0.875 1.140 1.445 1.650 1.885 2.215 3.525 (99%) Temperature: 75 0.05e Molarity: 0.0060M k = 6.52 X 10-5 0.05 sec.-1 t =10,628 sec. Time (sec.) 360 2000 3600 5400 7320 10840 14400 20400 Inf. Run 114 Acid Titer (ml) 0.205 0.540 0.855 1.160 1. 440 1.880 2.210 2.620 3.525 (99%) Temperature: 75 0.05e Molarity: 0.0060M k = 6.51 X 10-5 0.05 sec.-1 t = 10,639 sec. 95 In Observed -0.0584 -0.1579 -0. 2853 -0.3907 -0.5275 -0.6313 -0.7473 -0.9901 In Observed -0.0599 -0.1663 -0.2778 -0.3991 -0.5251 -0.7624 -0.9863 -1. 3599

PAGE 107

Run 115 Time Acid Titer (sec.) (ml) 240 0.190 540 0.505 1040 0.920 1540 1.310 2040 1.660 2540 1.910 3030 2.185 3560 2.425 Inf. 3.610 (98%) Temperature: 90 0.05e Molarity: 0.0062M k = 3.16 X 10-4 0.04 sec.-1 tl = 2192 sec. Time (sec. ) 240 600 1000 1500 2020 2500 3000 3500 Inf. Run 116 Acid Titer (m1) 0.210 0.580 0.930 1.310 1.675 1.915 2.165 2.400 3.620 (98%) Temperature: 90 0.05e Molarity: 0.0062M k = 3.12 X 10-4 0.02 sec.-1 t = 2222 sec. 96 In Observed -0.0541 -0.1507 -0.2942 -0.4508 -0.6159 -0.7533 -0.9297 -1.1142 In Observed -0.0598 -0.1746 -0.2969 -0.4492 -0.6212 -0.7531 -0.9117 -1. 0878

PAGE 108

Initial Solvolysis of nonan-9-y1 p-nitrobenzoate (26)-OPNB at 125C Time Acid Titer (sec.) (ml) In[C]a 320 0.223 -0.066 610 0.680 -0.217 910 1.130 -0.392 1210 1.500 -0.563 1800 1.895 -0.785 2410 2.095 -0.919 3610 2.555 -1. 320 5410 2.760 -1.570 7210 2.810 -1. 640 9010 2.850 -1.700 12610 2.855 -1. 710 18010 2.885 -1. 760 Actual Infinity 2.900 (82.8%) Theoretical Infinity 3.405 Blank 0.080 Calculated Infinity 3.485 on the calculated infinity. 97 The graphical presentations of log C versus time for this initial run are presented in Figure la, Appendix. The graphical presentations of subsequent runs using 80% of calculated infinity (which had been shown to be equivalent to the actual infinity) are shown contrasted with those using the calculated infinity in Figure IIa, Appendix.

PAGE 109

Solvolysis of anti-Tetracyc1o[4.3.0.02,4.03,8]nonan-9-y1 p-nitrobenzoate (26)-OPNB Time (sec. ) 360 8660 12660 15660 18060 24660 25660 27660 Inf. Run 111 Acid Titer (m1) 0.140 0.690 0.925 1.110 1.215 1.475 1.500 1.585 3.226 (calculated) Temperature: 90.2 0.2e k = 2.30 x 10-5 0.04 sec.-l t = 30,130 sec. Time (sec. ) 360 9660 12660 15660 21090 23660 25660 Inf. Run /12 Acid Titer (m1) 0.145 0.760 0.920 1.100 1.310 1.440 1.530 3.226 (calculated) Temperature: 90.2 0.2e Molarity: 6.71 x 10-1M k = 2.33 X 10-5 0.03 sec.-l t = 29,742 sec. In Observed -0.0444 -0.2407 -0.3379 -0.4217 -0.4726 -0.6111 -0.6254 -0.6759 In Observed -0.0460 -0.2686 -0.3357 -0.4170 -0.5210 -0.5913 -0.6430 98 Rate constant calculated using 80% of calculated infinity (actual infinity) and the data from Run 112. k =3.29 X 10-5 0.07 sec.-l = 21,063 sec.

PAGE 110

Time (sec.) 360 2360 2660 3660 4660 Inf. Run 113 Acid Titer (ml) 0.155 0.480 0.535 0.670 0.805 3.353 (calculated) Temperature: 100 0.2C Molarity: 6.98 x 10-3M k = 5.28 X 10-5 0.07 sec.-1 = 13,125 sec. Time (sec. ) 350 2650 4650 6650 8650 10650 11650 12650 Inf. Run 114 Acid Titer (m1) 0.170 0.540 0.815 1.080 1.340 1.520 1.600 1. 670 3.353 (calculated) Temperature: 100 0.2C Molarity: 6.98 x 10-3M k = 5.26 X 10-5 0.08 sec.-1 tl = 13,175 sec. 99 1n Observed -0.0473 -0.1545 -0.1738 -0.2229 -0.2745 1n Observed -0.0520 -0.1756 -0.2785 -0.3888 -0.5102 -0.6039 -0.6485 -0.6893 Rate constant calculated using 80% of calculated infinity (actual infinity) and the data from Run #4. k = 7.53 X 10'-5 0.1 sec.-1 (at 80% of infinity) = 9203 sec.

PAGE 111

Time (sec.) 300 800 1000 1200 1400 1600 Inf. Run 115 Acid Titer (m1) 0.240 0.965 1.265 1.440 1.668 1.860 3.640 (calculated) Temperature: 125.2 0.2e Molarity: 7.43 x 10-3M k = 4.97 X 10-4 0.08 sec.-1 = 1394 sec. Time (sec. ) 300 800 1000 1100 1500 1600 Inf. Run 116 Acid Titer (m1) 0.275 1.015 1.220 1.280 1.700 1.825 3.478 (calculated) Temperature: 125.2 0.2e Molarity: 7.09 x 10-3M k = 4.97 X 10-4 0.14 sec.-1 tl = 1394 sec. In Observed ':"0.0682 -0.3080 -0.4270 -0.5035 -0.6129 -0.7155 In Observed -0.0824 -0.3451 -0.4320 -0.4589 -0.6710 -0.7441 100 Rate constant calculated using 80% of calculated infinity {actual infinity) and the data from Run 116. k (at 80% of infinity) = 7.13 X 10-4 t 972 sec. -1 0.22 sec.

PAGE 112

Solvolysis of anti-Tetracyc1o[3.3.1.02,4.03,7]nonan-9-y1 p-nitrobenzoate (2?)-OPNB Time (sec.) 3740 28750 44454 88000 175400 255400 Inf. Run 111 Acid Titer (m1) 0.215 0.295 0.340 0.450 0.665 0.855 3.100 (96%) Molarity: 6.76 x 10-3M k = 9.86 X 10-7 0.086 sec.-1 t = 703,553 sec. Run 112 Time (sec. ) Acid Titer (m1) 3750 83400 171730 264620 351660 428950 517090 Inf. 0.185 0.365 0.540 0.735 0.875 0.985 1.140 2 . 488 (100%) Temperature: 125 0.2e k = 1.03 X 10-6 0.014 sec.-1 = 672,816 sec. In Observed -0.0719 -0.1000 -0.1162 -0.1568 -0.2415 -0.3227 In Observed -0.0773 -0.1586 -0.2447 -0.3502. -0.4334 -0.5040 -0.6129 101

PAGE 113

Time (sec. ) 3000 15200 32630 85400 101400 171100 190500 Inf. Run 113 Acid Titer (m1) 0.215 0.500 0.815 1.500 1. 735 2.170 2.220 2.566 (97%) Temperature: 150 0.2e Molarity: 5.48 x 10-3M k = 1.04 X 10-5 0.02 sec.-1 = 66,635 sec. Time (sec. ) 3000 14190 29800 41860 86200 100900 173600 Inf. Run 114 Acid Titer (m1) 0.215 0.485 0.800 0.980 1.580 1. 720 2.185 2.560 (96%) Temperature: 150 0.2e Molarity: 5.46 x 10-3M k = 1.07 X 10-5 0.01 sec.-1 t = 64,776 sec. 102 1n Observed -0.0875 -0.2167 -0.3822 -0.8784 -1.1275 -1. 8687 -2.0037 1n Observed -0.0877 -0.2100 -0.3747 -0.4826 -0.9602 -1.1144 -1. 9208

PAGE 114

Time (sec.) 1100 78550 122350 163600 208130 251350 Inf. Run 115 Acid Titer (m1) 0.295 0.870 1.135 1.335 1.525 1.700 2.360 (100%) Temperature: 1400 o.ooe Molarity 5.04 x 10-3M k = 4.52 X 10-6 0.09 sec.-1 = 153,351 sec. Time (sec.) 600 44330 88450 129250 173300 263100 289370 Inf. Run 116 Acid Titer (ml) 0.185 0.500 0.755 0.980 1.170 1.460 1.530 2.050 (99%) Temperature: 1400 O.le Molarity: 4.38 x 10-3M k = 4.43 x 10-6 0.03 sec.-1 t = 156,467 sec. 103 In Observed -0.1335 -0.4599 -0.6557 -0.8342 -1.0392 -1. 2744 In Observed -0.0946 -0.2796 -0.4593 -0.6502 -0.8459 -1. 2457 -1.3720

PAGE 115

Solvolysis of 9-yl 3,5-dinitrobenzoate (2?)-ODNB Time (sec.) 360 6960 11100 15100 21200 27900 Inf. Run 111 Acid Titer (m1) 0.090 0.565 0.800 0.940 1.140 1.335 1. 970 (97%) Temperature: 140 O.lC Molarity: 4.16 x 10-3M k = 3.87 X 10-5 0.1 sec.-1 = 17,911 sec. Time (sec.) 300 5800 9360 14700 21770 24670 28700 InL Run 112 Acid Titer (ml) 0.090 0.485 0.670 0.880 1.190 1.230 1. 330 1. 930 (95%) Temperature: 140 O.lC k = 3.97 X 10-5 0.1 sec.-1 = 17,460 sec. In Observed -0.0468 -0.3380 -0.5210 -0.6485 -0.8646 -1.1324 In Observed -0.0480 -0.2894 -0.4264 -0.6087 -0.9588 -1. 0144 -1.1686 104

PAGE 116

Solvolysis of anti-7-Norborneny1 p-nitrobenzoate (2)-OPNB Time (sec. ) 360 28300 57800 77600 93000 109500 115300 Inf. Run III Acid Titer (m1) 0.205 0.985 1. 630 1.955 2.165 2.365 2.480 3.585 (99%) Temperature: 140 0 .05e Molarity: 7.37 x 10-3M k = 9.52 X 10-6 0.15 sec.-1 = 72,794 sec. Time (sec. ) 360 22300 60100 84000 36800 101700 116200 Inf. Run 112 Acid Titer (m1) 0 . 175 0.810 1.640 2.045 1.155 2.260 2.460 3.595 (99%) Temperature: 140 0.05e Molarity:. 7.37 x 10-3M k = 9.45 x 10-6 0.08 sec.-1 t = 73,333 sec. In Observed -0.0589 -0.3212 -0.6064 -0. 7884 -0.9263 -1.0781 -1.1771 In Observed -0.0499 -0.2553 -0.6092 -0.8415 -0.3875 -0.9908 -1.1531 105

PAGE 117

Time (sec.) 300 . 4900 10500 15600 22300 28500 33600 lnf. Run 113 Acid Titer (m1) 0.220 0.580 0.965 1.275 1.575 1.845 2.060 3.525 (99%) Temperature: 150 O.lC Molarity: 7.36 x 10-3M k = 2.41 X 10-5 0.03 sec.-1 = 28,755 sec. Time (sec. ) 300 4900 11100 16200 21800 27600 32100 37040 lnf. Run 114 Acid Titer (m1) 0.215 0.560 0.955 1. 295 1.565 1.840 1.945 2.140 3.525 (99%) Temperature: 150 O.lC Molarity: 7.36 x 10-3M k = 2.37 X 10-5 0.04 sec.-1 1:J. = 29,240 sec. 106 In Observed -0.0644 -0.1798 -0.3199 -0.4489 -0.5921 -0.7413 -0.8782 In Observed -0.0629 -0.1730 -0.3160 -0.4579 -0.5869 -0.7383 -0.8027 -0.9344 19c From the data of other workers, the rate constant was obtained by FOCAL 8 treatment of their values for acid and infinity titers. Temperature: 126.2 0.05C Molarity: 7.5 x 10-3M k = 2.17 X 10-6 0.05 sec.-1 (Run Ill) = 5322 min. k = 2.51 X 10-6 0.05 sec.-1 (Run 112) = 4601 min.

PAGE 118

Decarbonylation of endo-Tricyclo[3.2.1.02,4]octan-8-one (52) Time (sec. ) 601 1806 3605 7356 10820 InL Run 111 Ratio Peak Areas 0.486 1.587 2.752 4.386 5.099 5.857 (actual) Temperature: 165 0.2C Molarity: 0.867M k = 1.92 X 10-4 0.016 sec.-1 tl = 3609 sec. Time (sec. ) 600 1800 2700 3660 5400 7300 9000 10800 InL Run 112 Ratio Peak Areas 0.378 1. 253 1. 782 2.254 2.910 3.409 3.747 3.935 4.485. (actual) Temperature: 165 0.2C Molarity: 0.785M k = 1.99 X 10-4 0.02 sec.-1 tl = 3482 sec. In Observed -0.0866 -0.3160 -0.6346 -1. 3817 -2.0447 In Observed -0.0880 -0.3276 -0.5064 -0.6983 -1. 0465 -1. 4275 -1. 8046 -2.0986 107

PAGE 119

Decarbony1ation of 9-one (45) Time (sec. ) 1900 72075 107856 146876 184395 234002 267340 Inf. Run III Ratio Peak Areas 0.064 2.397 3.262 4.001 4.678 5.298 5.664 8.351 (actual) Temperature: 165 0.2e Molarity: 0.514M k = 4.22 X 10-6 0.082 sec.-1 = 164,218 sec. Run 112 Time (sec. ) Ratio Peak Areas 2060 66500 99000 153160 180800 Inf. 0.073 . 2.522 3.484 4.756 5.357 9.363 (actual) Temperature: 165 0.2e Molarity: 0.448M k = 4.67 x 10-6 0.036 sec.-1 t = 148,394 sec. In Observed -0.0077 -0.3383 -0.4953 -0.6522 -0.8214 -1. 0063 -1.1340 In Observed -0.0078 -0.3138 -0.4654 -0.7092 -0.8490 108

PAGE 120

Time (sec.) 600 16400 26600 33900 52210 Inf. Run 113 Ratio Peak Areas 0.150 4.142 5.782 6.592 8.067 9.363 (actual) Temperature: 185 0.2e Molarity: 0.447M k = 3.78 X 10-5 0.083 sec.-1 = 18,333 sec. Time (sec.) 600 7170 14400 22700 29200 34080 Inf. Run 114 Ratio Peak Areas 0.150 2.150 3.820 5.223 6.115 6.685 9.363 (actual) Temperature: 185 0.2e Molarity: 0.447M k = 3.66 X 10-5 0.026 sec.-1 tl = 18,934 sec. 109 In Observed -0.0162 -0.5841 -0.9611 -1.2525 -1. 9455 In Observed -0.0162 -0.2609 -0.5242 -0.8161 -1.0587 -1. 2517

PAGE 121

no , . c .... Solvolysis Product Studies' [4 3 0 . ..: <26}":'OPNB . " . -. " . ' . ." Normalcoriditions A sample of <26 ):-OPNB weighing 0.708 gram was dissolved in SO ml of 70% acetone-water in a glass ampule which had been flushed with dry nitrogen. The contents were frozen in liquid nitrogen, pumped to O.S torr, and sealed. The tube was heated in a constant temperature sili-cone oil bath for 10 x t and cooled. The tube was diluted with2S0ml of water, made basic with 10% sodium carbonate, and extracted with S x 100 m1 portions of pentane. The pentane was dried over anhydrous magnesium sulfate overnight, filtered, and evaporated. The resulting white solid weighed 0.390 gram (97%). Thin layer chtoma-tography showed only. two spots. The alcohol fraction was separated by subliming it out of the mixture and washing the solid into a tared vial with a . few drops of ether. After evaporation of the ether, the alcohol fraction weighed 0.280 mp 232-23SoC (completely submerged sealed tube). The ester residue weighed 0.100 gram (24.S%) and had mp140-l42C. The ester fraction was dissolved in ether and reduced with an ethereal solution of lithium aluminum hydride. The reduced ester in.ether was treated with saturated aqueous ammonium chloride to destroy unreacted lithium aluminum hydride and then with anhydrous. sodium carbonate.'. The clear dry supernatant. ether solution was analyzed for.its.alcohol ratio without furtherpu1;'ification • . Both alcohol fractions were examined.chromatographically, tirstat attenua-tion32 to.64,detectorsensitivity at 10...,..11 mv,whereonlyonepeak on scalewasobserVed;thenat.attenuation 2 to 4, detector.sensitivity

PAGE 122

111 where . . the main. peak was ' . qff . scale. and minor. components. could be. evaluated.. The . alcohol isolated direc tly . from the reac tiori. mixture contained 0.112% of the unrearrC3:nged .alcohol (26) eluting.at20.2 minutesrelative.to alcoho1,(27).at,l6.2 minutes, when run at 2 to 4 x highsensitivities usually showed a minor peak and broad shoulder from 12 to 16 minutesrelativeto.(27) atl6.2 minutes. The origin ,of this trace material is unknown, but it is thought to be the.result of acid catalyzed decomposition of alcohols produced on solvolysis. The pattern can also be observed in analytical samples of both alcohols and thus may ori-ginate in the chromatograph itself. The alcohol fraction obtained by lithium aluminum hydride reduction of recovered p-nitrobenzoate ester was found to contain 0.094% of unrearranged alcohol.(26) and three minor I components .of unknown origin. These peaks were characteristic of nearly all p-nitrobenzoate esters reduced with lithium aluminum hydride and eluted at 13.4,14.2, and 15.5 minutes (ratio 1:8:5) relative to . (27)-OH at 16.2 minutes. These peaks may have constituted as much as 2% of this alcohol fraction but may have been the result of lithium aluminum hydride reduction of the p-nitrobenzoate portion of the ester • . Basiccortditions A weighing 0.693 gram (2.43 IDmole) was dis-solved in .50ml of 70% acetone-water in a flushed ampule and o .325 gram . ' (5. 42.mmole) of urea was added.. The tube was . sealed as pre-viouslydescribed and innnersed.in a constant temperature bath. for :10 x t. extracted, . the ampule yielded 0.496 , gram of sticky yellow solid. . The materialwaschfomatographed on .

PAGE 123

112 Florisil. using 1: 1 . benzene...,.petroleum ether. as the. eluant. _ . The . re-arranged ester eluted ;first, product. alcohol second, o. (67%) . representing a .97% yield over-all. . . The ester fraction was reduced to. alcohol with. an, ethereal solu-tion of lithium aluminum hydride as previously descrihed.andthe alcohol fractions analyzed chromatographically. Product alcohol appeared to contain 0.04% of unrearranged alcohol .(26), c;lnd the same ,extraneous peaks previously ,described. The alcohol obtained from ,reduction ,of residual ester contained no unrearranged alcohol withi:i1the limits of detection of ,the analytical gas chromatograph,' and traces.of the three . peaks 'previously described, which were always found. cinti";'Tetracyclo [3 3 ]rionari..;,9-yl esters' (27) ";'OPNB . ' arid ' (27) Normalcoriditions A sample of (27)-OPNB weighing 0.0755 gram was dissolved in 5 m1 of 70% acetone-water and sealed in an ampule as previously described. The solution was heated for 10 x t (7.5 days'at 150C). The tube was cooled,opened, and the contents made basic to pH Hydrion paper with 10% sodium carbonate solution. The solution was extracted with pentane and the extracts were dried and evaporated. The white solid so ob-tainedwelghed 0.033 gram (92%) and ,consisted primarily ofalcoho1.(2?). No aromatic protons were found in the IHnmr, and thin layerchromatography showed only one major spot,indicatingthatlLttleor no unreacted ester was present.' Chro.matographi.c analysis of the product alcohol with no further purification indicated that it was 86% (27)";'OH and 6% of im-purities with, retention times of , 12"';16 .minutes relative, to (2?).,.,.OH at There Was no trace of (26)";'OH

PAGE 124

113 to. the limits of .detection • . This . analysis was repeated with (2?)-ODNB. Here 10 x t . was , onlY 46 .hours.at :140Cj:andthereactionwas,expected . ' . . ..' ....... . Alcohol.(27): was . recovered in 97.5% yield.andwaspure . to the limits .. ofdetection,. except for. theusual 12"';16 lllinutebroad peak,.as of an authentic sample of . (27),.,OR. Basic conditions Solvolysis of a sample of (27):-:-OPNB in 70%' acetone-water in the presence of urea for 10 x 11.' (7.5 days at 150C) afforded a black tar containing so many by-products that a reasonable product was virtually impossible. Therefore, a sample of C:27)-ODNB weighing 1.009 grams (3.05 mmole) in'75 ml of 70% acetone-water containing 0.3614 gram (6.02mmole) of urea was sealed as previously described. After heating 10 x t (46 hours at , 140C) a black oil was obtained. The material, I which smelled strongly of ammonia, was diluted with 400 llli of water and extracted with pentane. The pentane extracts were washed successively with. 2 x 100 mlof 2% hydrochloric acid, 2 xlOO ml of 5% sodium carbonate, and 2 xlOO ml of water. After drying, the solvent was evaporated to give 550 mg of solid which was sublimed to give 400 mg (96%) of alcohol (27), whose 1 Hnmr spectrum was identical with that of an authentic sample • . The alcohol was carefully analyzed on the gas chromatograph both before and .after.sublimation.The sample was found to contain a small amount (0.06%) of a substance which elutes but.near to . the retention time of(26):-OH. An ether solution of au then-tic, (26):-OH sufficiently dilute. to give a . relatively weak; ' . but. observable glpcpeakat . the low vas . prepared' . A-few'drops' of . this. solution was mixed with. the' . solvolysis sample. 'Injection of

PAGE 125

114 . this mixture. showed tha t . the minor. peak. observed in. produc t : samples fiom.solvolysis.containingureamay,be.anartifact_resultingfiom the decomposition of urea, or interaction of urea with the : acid produced during.the_solvolysis. Theglpctracing obtained for.this -mixture clearly showed the previously.detected small unknown peak to elute .. seconds short of the added (26),..OR . . ltwas concluded that the product of basic solvolysis of . (2?)-ODNB in the presence of urea contains no (26)-OH • . Auxiliary Experiments Reduction of esters The integrity of the esters . employed in the product.studies was verified in the following manner •.. Samples of the esters weighing at least 25mgwere di13solved in ether and reduced with an ethereal solution of lithium aluminum hydride. The resulting mixture was quenched with saturated ammonium chloride solution and shaken with anhydrous sodium carbonate. The supernatant was decanted and reduced in volume to approximately 0.2ml with a stream of dry nitrogen. After this treatment, esters of . (26)-OH gave a solution which was found to contain as . much . as 1.5%0. (27)..,.OH and a pattern_consisting of abroad peak 12"":16 minutes, as previously described. The same solution obtained from esters o (27)""OH was free of . (26),..OH to.the limits of.detectionof the flame ionization detector, .but.also contained.theusual pattern.of-trace contaminants • . Decompositibnproductsfiom-urea A solution containing 0.086 gram (1. 4 .x 10-3:nnnole) of urea and o .093giam . (5.6 x lO-41nmol e ) of p-nitrobenzoic acid in 5:ml of 70%

PAGE 126

115 acetone-water was sealed in an ampule as described for a regular.solvoly-sis sample., This washeated.at150C.for 7.5 days ClO.x t ... of(2?)-. . . . OPNB) .'. After. cooling, the thick. blackinass was worked up . as though. it were a . regular. solvolysis' sample.' Evaporation. of. the pentane extracts' left a black oil. As with actual' samples, '. the oil was placed. in a sublimation apparatus and held .at 75C/760 torr for 4 hours.' . The . sublimed material was washed ftomthe cold finger and eXaminedchtomatographically. A small sample of .the oil before sublimation was also examined. Before sublimationthesample contained 46 elutablepeaks ranging in.retention time from 1.5 minutes to 80 minutes. The alcohols . (26), and. (27) normally were eluted at 32 and 24 minutes, respec-tively, under the conditions employed (column b).' The sample after sublimation contained only 20 elutable peaks extending t644 minutes' I retention time. Several of the peaksin.this sample occurred' in the region betWeen 30 and 34 minutes where (26)-DH normally eluted. When the experiment was repeated with 3,5-dinitrobenzoic acid, essentially the same result was obtained.

PAGE 127

' APPENDIX ' TABLE 'Ia; . Bond 'Lengths C(l) C(2) 1.538 0.012 C(l) -C(6) 1.602 C(l) C(9). 1.550 0.011 C(2). C(3) 1.534 ' 0.012 C(2) 'C(4) 1.529 0.012 CO) -C(4) 1.543 ' 0.012 C(3) C(8) 1.543 0.012 C(4) C(5) 1.522 0.013 ' C(5) C(6) 1.560 0.012 C(6) C(l) 1.552 0.013 C(l) C(8) 1.514 0.012 C(8) C(9) 1.542 0.011 C(9) I 0(2), 1.454 0.009 C(10) 0(1) 1.216 0.011 C(IO) 0 (2). 1.377 0.011 C(IO) -N 1.363 0.011 C(ll) -N 1.443 0.010 C(ll) -e(12). 1.385 0.011 C(ll) C(16) 1.406 0.011 C(12) -e(13) 1.403 0.011 e(13) -e(14) 1.390 0.012 e(14) . ""'Br 1.893 0.009 C(14) -e(lS) 1.376 0.012 C(IS) -e(16) 1.413 0.012 116

PAGE 128

. . Interatomic Angles Central . . Atom C 2 , C.1 C 6 102.-7 (.6)' 'C2 C.l C9. Cs. c . oJ.. Cg' 103.4(.6) C; .' C2 .' ; :Clt 109.1(.7) C..1 'C2. C3 103.2(..7) . C 3 . ' C 2 ,C,+ 60 5 (.'5) C2. . C3 C,+ 59.6(.5) C z C a C e 104.8(.7) 'C,+ C1 C e 119 .8(.7) C z C,+ C3 59.9(.5) C z ,C,+ Cs 105.9 (.7) 'C3 'C,+ Cs 120.2(.7) C,+ Cs ,C6 101. 9 (.7) Cl C6 C7 103.8(.6) C.l : C 6 C s 104.9(.7) C s C 6 C 7 105.6 (.7) :C6 C7 Ce 101.1(.7) C 3 C e C 7 110.7 (. 7) C3 C a Cg 98.8(.6) C 7 C e C9 101.2(.. 7) C I . \ C9 C e 96.1 (.6) Cl ; Cg z 108.6(.6) C e C9 'Oz 114.6 (. 6) 01 C.l 0 'Oz 124.1(.8) 01 C.l 0 N 128.3(.8) C12 ,ell N 123.0(.7) GlZ Cll C 16 121.6 (.7) C16 Cll N 115 .3(.6) Cll C12. C 13 120.3 (. 7) ClZ C13 GIlt 117.7 (. 8) C13 GI,+ CIS 123.0(.8) CIa C I'+ Br 118.1(.6) CIS C1I+ Br 118.9(.7) C.l'+ CIS C16 119.5(.8) CII CI6 CIS 117.9 (.7) C9 z C.lO 116.2(.6) TABLE IlIa. InteratomicDistances C(9) C(3) C(9) -C(2) C(9) -C(7) C(9) -C(6) o 2.343 A 2.283 2.361 2.475 117

PAGE 129

G -O.H G -0.31 u 00 o r-l -0. -0. Q G G o 0.2 T = 12SoC G Q (!) o G) 0.4 0.6 0.8 1.2 1.4 TIME (sec) x 10 FIGURE Ia. Plot of log[C] Versus Time to 10 x tl/2 for (26)-OPNB o 1.6 f-" f-" co

PAGE 130

u ....... ..-t -0.2,. -0.4 -0.6 -0.8 .Actual Infinity 2 FIGURE Ira. 4 Theoretical ? Infinity 6 8 TIME (sec) x 10 10 12 Plot of In [C] .Versus . Time . to . 1 . x tl12r T = 125C 14 16 f-' f-' \0

PAGE 131

u r":f -0.2 -0.6 -1.0 -1. 4,. -1. 8 20 40 60 80 100 120 . -3 TIME (sec) x 10 . FIGURE IlIa. Plot of In [C] Versus Time to 10 x tl;2 for (27)-OPNB T 1S0C 140 160 f-' N o

PAGE 132

-3.0" -4.0. eo 0 r-i -5. -6. 2.3 2.4 liT x 103 2.5 eo 0 r-i -3.0 "-4.0,. -5.0 -6.0 2.5 FIGURE IVa • . P1dtof log k Versus1/T for (26)-OPNBarid . (2?)-OPNB 2.6 liT x 103 2.7 t-' N t-'

PAGE 133

-3.0 -4.0 ,bO 0 r.-I -5.0 -6.J cr;ffiB 2.5 2.6 liT x 103 2.7 -3.0 -4.0 bo a r-i " -5.0 -6: 0 J)NB 2.7 2.8 2.9 liT x 103 'FIGURE ' Va. Plot of log k Versus liT for (12)";'OPNBsnd '(12)..;.ODNB 3.0 t-' N N

PAGE 134

00 0 rl -3.0 -4.0 -5.0 -6.0 2.3 2.4 liT x 103 2.5 FIGURE VIa. Plot of log kVersusl/T for . (2)..:.QPNB t-' N W

PAGE 135

BIBLIOGRAPHY 1. S. Winstein, M. Shatavsky, C. J. Norton, and R. B. Woodward, J. Amer. Chem. Soc., 77, 4183 (1955). 2. For review see: a) S. Winstein, Chem. Soc. spec. Publ., No. 215 (1965). b) S. Winstein, Quart. Rev.,23, 141 (1969). c) H. C. Brown and H. M. Bell, J. Amer. Chem. Soc., 85, 2324 (1963). d) N. C. Deno in Progress in Physical Organic Chemistry Vol. 2, S. G. Cohen, A. Streitwieser, Jr., and R. W. Taft, Eds.,Interscience Publishers, Inc., New York, N. Y., 1956, p. 479. 3. a) S. Sare1, J. Yove11, and M. Sare1-Imber, Angew. Chem., .Int. Ed. Engl., 7, 577 (1968). b) M. Hanack and H. J. Schneider, Angew. Chem., Int. Ed. Engl., 6, 666 (1967). c) B. Capon, Quart. Rev., Chem. Soc., 45 (1964). d) E. Vogel, Aneew. Chem., 72, 4 (1960). e) N. C. Deno, Chem. and Eng. News, 42, No. 40, 88 (1964). f) N. C. Deno, Prog. Phys. Org. Chem., 2, 129 (1964). g) P . . D. Bartlett in Nonclassical Ions, W. A. Benj amin, Inc., New York, N. Y., h) J. Haywood-Farmer, Chem. Rev., 74, 315 (1974). 4. a) S. Winstein, J. Sonnenberg, and L. De Vries, J. Amer. Chem. Soc., 81, 6526 (1959). b) S. Winstein, J. Amer. Chem. Soc., 81-, 6528 (1959). c) S. Winstein and J. Sonnenberg, J. Amer. Chem. Soc., 83, 3244 (1961). d) S. Winstein and J. Sonnenberg, J. Amer. Chem. Soc., 83, 3235 (1961) . e) S. Winstein, E. C. Friedrich, R. Baker, and Yan-i Lin, Tetrahedron, Suppl., 8 (II), 621 (1966). 5. R. L. Picco1ini and S. Winstein, Tetrahedron, Suppl., 2, 423 (1963). 6. a) M. A. Battiste, C. L. Deyrup, R. E. Pincock, and J. Haywood-Farmer, J. Amer. Chem. Soc., 89, 1954 (1967). b) H. Tanida, T. Tsuji, and T. Irie, J. Amer. Chem. Soc., 89, 1953 (1967). 7. J. Haywood-Farmer, R. E. Pincock, and J. I. Wells, Tetrahedron, 22, 2007 (1966). 8. T. Norin, Tetrahedron Letters, 37 (1964). 124

PAGE 136

125 9. S. Winstein, P. Bruck, P. Radlick, and R. Baker, J. Amer. Chern. Soc., 86, 1867 (1964). 10. a) J. Haywood-Farmer and R. E. Pincock, J. Amer. Chern. Soc., 91, 3020 (1969). b) H. Tanida, Accts. Chern. Res. 1, 239 (1968). c) M. J. S. Dewar, and J. M. Harris, J. Amer. Chern. Soc., 90, 4468 (1968), and 92, 6557 (1970). d) Y. E. Rhodes, and T. Takino, J. Amer. Chern. Soc., 90, 4469 (1968). 11. M. A. Battiste, P. F. Rankin, and R. Edelman, J. Amer. Chern. Soc., 93, 6276 (1971). 12. a) R. M. Coates, and J. L. Kirkpatrick, J. Amer. Chern. Soc., 90, 4162 (1968). b) R. M. Coates, and J. L. Kirkpatrick, J. Amer. Chern. Soc., 92, 4883 (1970). 13. M. A. Battiste, J. Haywood-Farmer, H. Ma1kus, P. Seidl, and S. Winstein, J. Amer. Chern. Soc., 92, 2144 (1970). 14. G. D. Sargent, R. L. Taylor, and W. H. Demisch, Tetrahedron Letters, 2275 (1968). 15. a) S. Masamune, R. Vukov, M. J. Bennett, and J. T. Purdham, J. Amer. Chern. Soc., (94, 8239 (1972). b) P. G. Gassman, and X. Creary, J. Amer. Chern. Soc., 95, 2729 (1973). 16. Y. Hata and H. Tanida, J. Amer. Chern. Soc., 91, 1170 (1969). 17. R. M. Coates and K. Yano, Tetrahedron Letters, 2289 (1972). 18. S. Masamune, S. Takada, N. Nakatsuka, R. Vukov, and E. N. Cain, J. Amer. Chern. Soc., 91, 4322 (1969). 19. a) J. S. Newcomer and E. T. McBee, J. Amer. Chern. Soc., 71, 946 (1949). b) R. Yates and P. Eaton, Tetrahedron, 12, 13 (1961). c) C. L. Deyrup, Ph.D. Dissertation, Boston University (1970). 20. P. G. Gassman and P. G. Pape, J. Org. Chern., 29, 160 (1964). 21. A. G. Bowers, T. G. Ha1sa11, E. R. H. Jones, and A. J. Lemin, J. Chern. Soc., 2548 (1953). 22. D. J. Beames, J. A. and L. N. Mander, Austr. J. Chern., 25, 137 (1971). 23. P. Story, J. Org. Chern., 26, 287 (1961); E. I. Snyder and B. Franzus, J. Amer. Chern. Soc., 86, . 1166 (1964). 24. R. L. Shriner, R. C. Fuson, and D. Y. Curtin in The Systematic Identification of Organic Compounds, Fifth Edition, John Wiley & Sons, Inc., . New York, N. Y., 1964.

PAGE 137

126 25. P. K. Freeman, B. K. Stevenson, D. M. Balls, and D. H. Jones, J. Org. Chern., 39, 546 (1974). 26. L. E. Fieser and M. Fieser in Reagents for Organic Synthesis, Vol. I, John Wiley & Sons, Inc., New York, N. Y., 1967. Diazomethane Preparation, p. 191. 27. a) W. Kirmse in Carbene Chemistry, Second Edition, Academic Press, London, 1971, p. 338. b) S. A. Monti, D. J. Bucheck, and J. C. Shepard, J. Org. Chern., 34., 3080 (1969). c) M. M. Fawzi and C. D. Gutsche, J. Org. Chern., 31, 1390 (1966). d) A. Nickon, H. Kwasnik, T. Swartz, R. o. Williams, and J. B. DiGiorgio, J. Amer. Chern. Soc., 87, 1615 (1965). 28. G. W. Glumpp, G. Ellen, J. Japenge, and Miss G. M. de Hoog, Tetra hedron Letters, 1741 (1972). 29. U. R. Ghatak, P. C. Chakraborti, B. C. Ranu, and B. Sanyal, J. Chern. Soc. Chern. Commun., 548 (1973). 30. R. Q. Brewster and T. Groening, Org. Syn. Coll. Vol. 2, 445 (1943). 31. Huang-MinIon, J. Amer. Chern. Soc., 68, 2487 (1946); D. J. Durham, D. J. McLeod, and J. Cason, Org. Syn.Coll. Vol. 4, 510 (1963); S. Hunig, E. Lucke, and W. Brenninger, Org. Syn., 43, 34 (1963). , 32. J. P. John, S. Swaminathan, and P. S. Venkaramani, Org. Syn. Coll. Vol. 5, 749 (1973). 33. R. K. Murray, Jr., and K. A. Babiak, J. Org. Chern., 38, 2556 (1973). 34. J. C. Collins, W. W. Hess, and F. J. Frank, Tetrahedron Letters, 3363 (1968). 35. J. M. Brown, and J. L. Occo1owitz, J. Chern. Soc. (B), 411 (1968); S. Winstein, M. Ogliaruso, M. Sakai, and J. M. Nicholson, J. Amer. Chern. Soc., 89, 3656 (1967); W. R. Moore, W. R. Moser, and J. E. LaPrade, J. Org. Chern., 28, 2200 (1963). 36. P. K. Freeman and T. D. Ziebarth, J. Org. Chern., 38, 3635 (1973). 37. A. Nickon and G. Pandit, Tetrahedron Letters, 3663 (1968). 38. J. B. Stothers in Carbon-13 NMR Spectroscopy, Academic Press, New York, N. Y., 1972, p. 397. 39. a) C. F. Wilcox, and M. E. Mesirov, J. Amer. Chern. Soc., 84, 2757 (1962). b) K. L. Servis and J. D. Roberts, Tetrahedron Letters, 1369 (1967). c) R. S. Macomber, J. Org. Chern., 38, 2568 (1973).

PAGE 138

127 40. a) G. Ellen and G. W. Klumpp, Tetrahedron 2995 (1974). p) G. Ellen and G. W. Klumpp, Tetrahedron 3637 (1974). 41. C. J. Collins, Chern. 543 (1969); C • . C. Lee, Progr. Phys. Org. 129 (1970); G. E. Schenk, and F. A. L. Anet, Tetra hedron 2779 (1971); L. Radom, J. A. V. Buss, and P. V. R. Schleyer, J. Amer. Chern. 311 (1972); N. Bodor, M. J. S. Dewar, and D. H. Lo, 5303 (1972); C. C. Lee, S. Vassie, and E. C. F. Ko, 8931 (1972). 42. R. Hoffman, Tetrahedron 3819 (1965). 43. a) A. C. McDonald, and J. Trotter, Acta CrystaZZogr. 243 (1965). b) A. C. McDonald, and J. Trotter, Acta CrystaZZogr. 456 (1965). c) H. Kayama, and K. Ohada, J. Chern. Soc. 940 (1969). 44. T. F. W. McKillop and B. C. Webster, 1879 (1970). 45. L. N. Ferguson in Organic Molecular Structure, Willard Grant Press, Boston, 1975, p. 492. 46. E. L. Allred, and A. L. Johnson, J. Amer. Chern. 1300 (1971) . 47. S. C. Clarke, and B. L. Johnson, Tetrahedron 3555 (1971) . 48. M. A. Battiste, Personal Communication. I 49. J. W. Nebzydoski, Ph.D. Dissertation, University of Florida, 1969, p. 120. 50. B. A. Hess, Jr., J. Amer. Chern. (1969). 51. G. W. Ewing in Instrumental Methods of Chemical Analysis, McGrawHill, Inc., .New York, N. Y., 1969, p. 465. 52. G. C. Levy and G. L. Nelson in Carbon-13 Nuclear Magnetic Resonance for Organic Chemists, John Wiley & Sons, Inc., New York, N. Y., 1972, p. 24, and see also reference 38. 53. S. Winstein and C. Ordronneance, J. Amer. Chern. 2884 {1960). 54. J. Haywood-Farmer, Ph.D. Dissertation, University of British Columbia, 1967, p. 135. 55. R. A. Sneen and A. L. Baron, J. Amer. Chern. 614 (1961). 56. J. K. Kochi and G. S. Hammond, J. Amer. Chern. 75, 3445 (1953).

PAGE 139

128 BIOGRAPHICAL SKETCH Donna Dembaugh McRitchie was born on July 16, 1934 in Superior, Pennsylvania to Albert and Helen Dembaugh. The family moved to Erie, Pennsylvania where she attended public schools, graduating from Erie East High School in 1952. Upon completion of her undergraduate studies at The College of Wooster, Wooster, Ohio in 1956 she was employed as a chemist by Battelle Memorial Institute at Columbus, Ohio. In December of 1955 she married John J. McRitchie and upon completion of his Ph.D. in 1961 was employed as a research chemist by Arthur D. Little of Cambridge, Massachusetts. After moving to Washington in 1967, she , taught chemistry at Wenatchee Valley College in Wenatchee until enter-ing graduate school at Oregon State University in the fall of 1970. She completed her studies toward the M.S. degree in September of 1971 and entered the University of Florida in the spring of 1972.

PAGE 140

I certify that I have read this study and that in my op1n10n it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. erIe A. Battiste, Chairman Professor of Chemistry I certify that I have read this study and that in my op1n1on it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. William M. Jones /' Professor of Chemis ry I certify that I have read this study and that in my op1n10n it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. I certify that I have read this study and that in my op1n10n it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Gus J. P lenik Professor of Chemistry

PAGE 141

I certify that I have read this study and that in my oplnlon it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Charles M. Allen, Jr. Associate Professor of Blochemistry This dissertation was submitted to the Graduate Faculty of the Department of Chemistry in the College of Arts and Sciences and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August, 1975 Dean, Graduate School