INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.
Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6 " x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.
University Microfilms International A Bell & Howell Information Company 3 0 0 North Z eeb R oad. Ann Arbor. Ml 48 1 0 6 -1 3 4 6 USA 313/761-4700 800/521-0600
Order Number 9201695
Synthetic studies on dodecahedrane. Generation and reactions of its radical and anion
Lagerwall, Dean Robert, Ph.D.
The Ohio State University, 1991
UMI 300 N. ZeebRd. Ann Arbor, MI 48106
SYNTHETIC STUDIES ON DODECAHEDRANE. GENERATION AND REACTIONS OF ITS RADICAL AND ANION
DISSERTATION
Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University
by
Dean Robert Lagerwall, B.S.
* * * * *
The Ohio State University 1991
Dissertation Committee: Approved by Professor Leo A. Paquette Professor Matthew Callstrom Professor Viresh Rawal Adviser Department of Chemistry To my parents ACKNOWLEDGEMENTS
My sincere thanks are extended to Professor Leo A. Paquette for the opportunity to perforin this research in an atmosphere allowing for productivity and creativity. The enthusiasm which he demonstrated in all aspects of chemistry was both contagious and appreciated. I would like to thank the many members of the Paquette group with whom I have had contact throughout my studies at OSU. The friendship, encouragement, time and efforts which they have contributed to the group made the research environment not only fruitful but also enjoyable. I would also like to thank Professor Stuart Kulp of Moravian College for introducing me to Organic Chemistry. The fire he began in my heart has continued to grow and allowed me (with the help of many) to truely enjoy my chosen career. Finally, I thank my parents for the freedom, support and encouragement they gave me in all aspects of my life during my academic experience. VITA
June 18,1963 ...... Born - Sacramento, CA
June, 1985 ...... Bachelor of Science cum laude. Moravian College, Bethlehem, PA
September 1985-September 1986 Graduate Teaching Associate, The Ohio State University
Sepetmber 1986-June 1991 ...... Graduate Research Associate, The Ohio State University
PUBLICATIONS
Kiplinger, J. P.; Marshall, A. G.; Kobayashi, T.; Lagerwall, D. R.; Paquette, L. A.; Bartmess, J. E. "Measurement of the Double-Bond Strain Energy in Dodecahedrene and Adamantene by Thermochemical Bracketing of Gas-Phase Ion-Molecule Reactions." lournal of the American Chemical Society. 1989, J_LL, 6914.
FIELD OF STUDY
Major Field: Chemistry Studies in Organic Chemistry TABLE OF CONTENTS
DEDICATION...... ii
ACKNOWLEDGEMENTS...... iii
VITA ...... iv
LIST OF TABLES...... viii
LIST OF FIGURES...... ix
LIST OF SCHEMES...... x
CHAPTER PAGE
I. INTRODUCTION...... 1 A. Background ...... 1 B. Goal of this Research...... 3 C. Organization of this T ex t...... 4
II. SYNTHESIS OF DODECAHEDRANE...... 6 A. Synthesis of Diketodiester 8 ...... 6 B. Trost Sequence Applied to 1 7 ...... 11 C. Synthesis of Dichlorodiester 2 0 ...... 14 D. Closing the Front of the Molecule ...... 17 E. Closing the Back of the M olecule ...... 19 F. Improved Access to Secododecahedrane ...... 23 G. Improved Access to Dodecahedrane ...... 26
III. DODECAHEDRYL RADICAL STUDY...... 33 A. Background ...... 33 1. Targeting Radical Precursors ...... 33 2. Derivatization Methods Available ...... 37 B. Radical Precursors ...... 39 1. Iodododecahedrane ...... 39 2. (Phenylseleno)dodecahedrane 42 v C. Dodecahedryl Radical Generation ...... 46 1. Reaction Design ...... 46 2. Reaction With Acrylonitrile ...... 47 3. Trapping of 56 With Acrylates ...... 50 4. Attem pted Trapping of 56 With 6 land 62 ...... 56 5. Trapping With 2-Cyclopentenone ...... 58 6 . Addition-Elimination Methodology ...... 60 D. Conclusion ...... 69 References ...... 72
IV. DODECAHEDRYL ANION STUDY...... 77 A. Background ...... 77 B. Radical Anion Methodology ...... 78 1. Alternate Anion Precursor ...... 81 2. Reaction Design ...... 83 C. Transmetallation Methodology ...... 86 1. Reaction Design ...... 86 2. Single Electron Transfer Exploration 94 D. Conclusion ...... 96 References ...... 98
V. EXTENSIONS TO DODECAHEDRYL CATION CHEMISTRY 102 A. Background ...... 102 B. Utilization of Zinc Iodide ...... 103 C. Dodecahedryl Alcohol: Improved Synthesis 107 D. l,4-Di(dodecahedryl)benzene ...... 108 E. Conclusion ...... 112 References ...... 114
VI. DODECAHEDRENE...... 116 A. Introduction ...... 116 B. Measurement of the Double-Bond Strain Energy of Dodecahedrene by Thermochemical Bracketing of Gas-Phase Ion-Molecule Reactions ...... 117 1. Background ...... 117 vi 2. T heory ...... 118 3. R esults ...... 119 C. Dodecahedrene: Laboratory Synthesis and Trapping ...... 127 1. Background ...... 127 2. R esults ...... 127 D. S u m m ary ...... 132 References ...... 133
VII ATTEMPTED REGIOSPECIFIC DISUBSTITUTION 135 A. Background ...... 135 B. R esults ...... 139 C. Conclusion ...... 144 References ...... 146
EXPERIMENTAL SECTION...... 148
APPENDIX A...... 165
APPENDIX B...... 167
APPENDIX C...... 169
APPENDIX D...... 174
BIBLIOGRAPHY...... 184
vii LIST OF TABLES
TABLES PAGE
1. Some Dodecahedryl GC Retention Times ...... 39
2. Heats of Hydrogenation of 94 and 95 by Several Methods (kj/mol) ...... 123
3. Summary of Data Collection and Structure Refinement Parameters for 17 ...... 168
4. Summary of Data Collection and Structure Refinement Parameters for 2 1 ...... 171
5. Bond Lengths for 21 ...... 172
6 . Intramolecular Bond Angles Involving the Non- Hydrogen Atoms for 21 ...... 173
viii LIST OF FIGURES
FIGURES PAGE
1. X-Ray Structure of 17 ...... 13
2. X-Ray Structure of 2 1 ...... 16
3. Origins of Dodecahedryl C arbons...... 166
4. Numbering Scheme for 2 1 ...... 170
ix LIST OF SCHEMES
SCHEMES PAGE
1. Conversion of Pagodane to Dodecahedrane ...... 2
2. Domino Diels-Alder Products ...... 8
3. Selective Hydrolysis and Iodolactonization ...... 9
4. Conversion of Iodolactone 10 to 13 ...... 10
5. Trost Sequence ...... 12
6 . Conversion of 17 to Dichlorodiester 2 0 ...... 15
7. Birch Reduction of 2 0 ...... 18
8 . Conversion of 25 to 2 8 ...... 19
9. Conversion of 28 to the Low Temperature Photolysis Product ...... 20
10. Preparation of Keto Aldehyde 3 4 ...... 22
11. Preparation of Secododecahedryl Alcohol ...... 22
12. Improved Access to 36 and Conversion to Secododecahedrane ...... 26
13. Dehydrogenation to Dodecahedrane ...... 28
14. Attempted Allylation of 38 ...... 33
15. General Reaction Studied by Bechwith and Pigou on Sjj2 Radical Reactivity ...... 34 16. Formation of 40 Upon Attempted Nitration of 3... 35
17. Unsuccessful Nitration of 3 ...... 35
18. Route to Aminododecahedrane ...... 36
19. Unsuccessful Direct Animation Attempt on 3 8 ...... 36
20. Cationic Derivatization Techniques ...... 38
21. Reaction of 38 with AII 3 (Possible Intermediacy of 46?) ...... ,. 40
22. Reactions of 38 with Znl 2 ...... 41
23. Previous Unsuccessful Attempts to Obtain 4 7 ...... 43
24. Preparation of TMSSePh ...... 44
25. Reactions of 38 with TMSSePh and ZnX 2 ...... 45
26. Generation and Trapping of 56 with Acrylonitrile . 48
27. Assumed Mechanism and Possible Side Reactions in the Formation of 57 ...... 49
28. Attempted Acrylate Trapping of 5 6 ...... 51
29. Catalytic Tin Hydride Generation ...... 53
30. Formation of Trace Acrylate Trapping Product 54
31. Tin Pinacolate Methodology ...... 55
32. Attempted Trapping of 56 with 61 and 62 ...... 57
xi Trapping of 56 with 2-Cyclopentenone ...... 60
Examples of Addition/Elimination Methodologies. . . 61
Mechanism of Allylation Using 65 ...... 61
Successful Allylation of 56 with 6 5 ...... 63
Preparation of 6 8 by Sasaki ...... 66
Synthesis of Allyldodecahedrane ...... 67
Mechanism of Allylation with Allyltrimethylsilane . 67
Preparation and Attempted Trapping with 70 ... . 68
Generic Representation of Possibilities in Typical Radical Reactions ...... 71
Previously Attempted Metallations ...... 78
Preparation of DBB and LiDBB ...... 80
LiDBB Model Study ...... 80
Attempted Anion Generation with LiDBB and 38. . . 81
Preparation of TMSSPh and 7 9 ...... 82
Successful Generation and Trapping of the Dodecahedryl Anion from 7 9 ...... 85
Attempted Anion Generation with NaN ...... 86
Successful Transmetalation of 38 ...... 87
xii Competing Single Electron Transfer Products During Transmetalation ...... 89
Lack of Deuterium Incorporation in THF-dg During Transmetalation ...... 90
Options Available During Transmetalation ...... 91
Cross-Coupling Conditions of Eguchi ...... 95
Possible Mechanism for the Generation of Coupled Products Using Eguchi's Conditions ...... 95
Possible Effect of 46 on Rate of Electrophilic Trapping...... 106
Original Preparation of 5 3 ...... 107
Improved Synthesis of Dodecahedryl Alcohol ...... 108
Possible Precursor to Dumbellane ...... 1 10
Synthesis of l,4-Di(dodecahedryl)benzene ...... 112
Generation of Dodecahedrene in the MS ...... 1 19
Unstrained Analog of Dodecahedrene ...... 121
Unstrained Adamantene Analog ...... 122
Elimination in the Adamantane System ...... 122
1,2- and 1,4-Elimination Products of 100 ...... 124
Generation and Trapping of Dodecahedrene ...... 131
xiii 6 6 . Mechanism of 1-Adamantanyl Alcohol Formation Noted by Perkins and Turner ...... 132
67. Trapping of the Dodecahedryl Dication ...... 136
6 8 . Norrish Type I Photoreaction ...... 137
69. Norrish Type II Photoreaction ...... 138
70. Proposed Regioselective Disubstitution ...... 138
71. Preparation of 112 ...... 139
72. Options Available to Diradical 113 ...... 141
73. Possible Origin of Photolysis By Products of 112 . . 142
74. Preparation of 115 ...... 143
xiv CHAPTER I INTRODUCTION
A. Background Dodecahedrane (3) is the most complex member of the structurally intriguing CnHn hydrocarbons whose carbon frameworks define regular polyhedra (1-3). The //, symmetry (120 symmetry operations) dodecahedrane possesses qualifies it as the most highly symmetric of all organic molecules. Adding to its allure is the presence of a cavity incapable of solvation (trans-cavity distance of 4.31 A) within this highly rigid, topologically spherical molecule and the near ideal tetrahedral character attained by each sp3-hybridized carbon atom.
^ 1
V
3
1 2
Much speculation and theorization about the properties of dodecahedrane and its derivatives appeared before the first successful synthesis was achieved by the Paquette group in 1982. The recently reported conversion of pagodane (4) to dodecahedrane by the Prinzbach group is impressive despite the low yield ( 8%) (Scheme 1) and demonstrates the continued fascination with this molecule. The emergence of dodecahedrane from the theoretically possible to the synthetically accessible served to increase interest in this aesthetically pleasing system.
catalyst
4 3
Scheme 1. Conversion of Pagodane to Dodecahedrane
The utilization of the dodecahedryl nucleus as a potential carrier for functional groups of pharmaceutical interest was a major driving force in the derivatization of 3. Of particular interest were various amino derivatives that were to be compared to the biologically active 1-aminoadamantane hydrochloride (approved in 1966 by the FDA for use against Asian influenza). To establish if 3 more symmetrical, suitably derivatized condensed spherical alicyclics of dimensions larger than adamantane would exhibit antiviral effects, an intense effort was mounted to obtain the desired dodecahedryl targets. Studies directed toward the functionalization of the dodecahedryl nucleus were initiated. Unfortunately, the amount of material available has always hampered the acquisition of information about this system. Nonetheless, after the preparation of the monobrominated derivative, the remarkable ease with which dodecahedrane underwent electrophilic substitution became readily apparent. These investigations, however, showed no successful trapping of either the dodecahedryl radical or anion. The scarcity of material thus led the vast majority of available dodecahedrane through the proven cationic route rather than the unknown radical or anionic paths in attempted derivatizations. The desired amino derivatives were prepared via a cationic pathway and their antiviral properties evaluated. While these molecules exhibit antiviral activity, the corresponding hydrochloride salts were found to be at least 5- to 10-fold less active than 1- aminoadamantane hydrochloride and to be more toxic.
B. Coal of this Research Although a plethora of derivatives have become accessible via the cationic pathway, the prior neglect to the radical and anion 4 chemistry of dodecahedrane could no longer be ignored. It was the goal of this researcher to examine the radical and anion chemistry of dodecahedrane and to evaluate their potential for further derivatization of the dodecahedryl nucleus. Additionally, interesting side studies were to be explored as access to new products became available.
C. Organization of this Text Chapter II details the synthesis of dodecahedrane. The 22-step linear synthesis is briefly explained with extra attention given to improvements that I have contributed. Assembly of the twelve polyfused cyclopentane rings to yield a molecule containing 30 carbon-carbon bonds between 20 methine units starting from cyclopentadiene is impressive. Due to the amount of starting material needed for the desired studies, I was obliged to complete this synthesis twice. Chapter III addresses the potential for radical chemistry to deliver new derivatives and explores the reactivity characteristics of this structurally unusual radical. Chapter IV details the first successful generation and trapping of the dodecahedryl anion and discusses the limits of the anionic methodology of dodecahedrane. Chapter V briefly summarizes the extensions made to the cationic chemistry of dodecahedrane uncovered and developed in the 5 previous two chapters. An improved synthesis of dodecahedryl alcohol exemplifies this improved methodology. Additionally, l,4-di(dodecahedryl)benzene was prepared by making use of the cationic reactivity of dodecahedrane. This molecule represents the first of a new class of compounds containing two dodecahedryl nuclei in one molecule. The synthesis and characteristics of this molecule will be discussed. Dodecahedrene, the mono-olefin of 3 is explored in Chapter VI. The double-bond strain energy of this highly pyramidalized olefin has been measured by thermochemical bracketing of gas-phase ion- molecule reactions. Additionally, a derivative made available during the radical study allowed in situ generation and trapping of dodecahedrene as the epoxide. In Chapter VII, possible 1,2-disubstitution of the dodecahedryl nucleus by photochemical means was investigated. Unfortunately, the characteristics of the products obtained upon irradiation did not allow evaluation of this methodology. Throughout this text, the reader will note that the unusual characteristics of dodecahedrane and its derivatives both help and hinder the exploration of this exquisite system. CHAPTER II SYNTHESIS OF DODECAHEDRANE
A. Synthesis of Diketodiester 8 The cornerstone to the synthesis of dodecahedrane is the so- called domino Diels-Alder reaction . 1 When originally modified to deliver potential dodecahedryl intermediates ,2 recourse was made to finely divided sodium on alumina 3 to generate the needed sodium cyclopentadienide (5). This method of anion generation unfortunately mandated laborious filtration and was later abandoned in favor of a more recently developed procedure 4 involving oil-free sodium hydride .5 The freshly prepared sodium cyclopentadienide is then oxidatively coupled on exposure to 0.5 mol equivalent of iodine. The reaction is carried out at low temperature to avoid isomerization of the 9,10-dihydrofulvalene ( 6 ) thus formed. Addition of dimethyl acetylenedicarboxylate (DMAD )6 at -78°C leads to a pair of Diels- Alder products each of which is capable of a second or domino Diels- Alder reaction. The completed reaction leads to a one-to-one mixture of C 2 *symmetric diesters, the desired 8 possessing C 2 V symmetry (Scheme 2)7. During the second Diels-Alder reaction, the mixture is not cooled but allowed to exotherm. Depending on the reaction, if left on 6 7 its own, the final temperature may vary from 10°C to 67°C. The exothermicity of this reaction was often a problem, for as the final temperature increased, so did the amount of polymeric by-products with a corresponding decrease in yield. To avoid this problem, I included in the procedure the need of controlling the reaction mixture's final temperature (0 - 10°C, ice bath if necessary). This temperature control is not mentioned in the Organic Synthesis preparation ,5 but is vital if consistent yields are desired.
Diester 8 is readily separated from the unwanted isomer by making use of the hindered nature of it's ester functions. Thus, selective hydrolysis at -10 to + 10°C in methanol using a cold solution of sodium hydroxide not only solubilizes the unwanted isomer, but also any polymeric material from the domino Diels-Alder reaction.
Extraction and recrystallization deliver the crystalline diester 8 .
More forcing hydrolysis of 8 leads to the internal diacid 9 in 15-20% yield from sodium cyclopentadienide (Scheme 3). Although this yield may seem low, it should be noted that this structure already contains four of the requisite cyclopentane rings with six cis-locked methine hydrogens, fourteen of the final twenty carbons, and appropriate functionality to allow continued use of the symmetry of the intermediate. Under standard iodolactonization conditions, the diacid could be "cross-corner" oxygenated to deliver the bis- iodolactone 10 efficiently. This methodology circumvents the isomer complications involved in the direct hydroboration-oxidation of the olefins .8 DOMINO DIELS-ALDER REACTION \
CH3OOC-CSC-COOCH3
XOOCH3 \ ^ C O O C H
COOCH;
COOCH;
COOCH ICOOCH COOCH3
Scheme 2. Domino Diels-Alder Products cooch3 /^L2jty~C00H coocHa ^NjOOH
selective hydrolysis
COOCH3 COOCH3 COOCH3 COOCH3 8
separation; hydrolysis
l2. Kl, NaH C03
(82 %) |COOH COOH
10
Scheme 3. Selective Hydrolysis and Iodolactonization
Opening of 10 with catalytic sodium ethoxide 9 in ethanol at room temperature forms bis(iodohydrin) 11. The mild reaction conditions are the result of the steric stress on the system and allows
ll's functional groups to survive .10 Jones oxidation of 11 followed by reduction of the resultant bis(iodo-ketone) 12 by zinc/copper couple 11 in methanol yields diketodiester 13, possessing a C 2 axis of 1 0 symmetry. At this stage, a single recrystallization assures the p u rity of 13 (Scheme 4).
NaOEt EtOH 84%
10
NgfcCrjOy, HzSO* acetone, 0° C 92%
Zn-Cu NH4CI MeOH 6 8 % C 0 2Et O C 0 2Et
1 3 12
Scheme 4. Conversion of Iodolactone 10 to 13 11
Hi Trost Sequence Applied to 17 Through the use of Trost's diphenylsulfonium cyclopropylide, bisketone 13 was converted to biscyclopentenone 17 according to Scheme 5 in 55% overall yield. In order to shift the reversible formation of the ylide in the desired direction and still avoid alkaline hydrolysis of the esters, an eight-fold excess of the diphenylsulfonium cyclopropyl tetrafluoroborate (14) was used with only a slight excess of potassium hydroxide. On treatment of the initially formed epoxycyclopropanes with fluoroboric acid, an acid- catalyzed rearrangement leads to a mixture of spirocyclobutanones (15, major isomer shown). The excess Trost Reagent (14) may be recovered while purifying these cyclobutanones. Baeyer-Villiger oxidation with 30% hydrogen peroxide in methanol provides the corresponding dilactones which, upon treatment with phosphorous pentoxide in methanesulfonic acid at 50°C for 36 hours, undergo intramolecular Friedel-Crafts acylation to give the desired biscyclopentenone 17. Note that 17 possesses all twenty of the needed carbons of dodecahedrane while maintaining an axis of symmetry.
A crystal of 17 suitable for X-ray analysis was grown . 1 2 Shown in Figure 1 is an ORTEP drawing of this structure (notice the lack of an X-ray crystallographic axis of symmetry). 1 2
1 . [ > - S P h 2 BF4
KOH, DMSO
2. HBF4 jC 0 2Et 1 3 C 0 2Et 1 5 C 0 2EtO
H2 0 2> 55 % overall NaOH MeOH
p2o 5 CH3 SO 3 H 50 °C
1 7 1 6
Scheme 5. Trost Sequence ORTEP drawing of biscyclopentenone 17
Figure 1. X-Ray Structure of 17 1 4
C. Synthesis of Pichlorodiester 20 A key transformation in the synthesis of dodecahedrane is the hydrogenation of 17 over 10% palladium on carbon. As one would expect from the topology of 17, the four additional hydrogens are delivered from the sterically unencumbered convex surface, forcing the cyclopentanones to the inner regions of the developing sphere. When treated with sodium cyanoborohydride, followed by lactonization, the highly crystalline bislactone 19 results. The geometry of this precursor is now beginning to resemble closely that of dodecahedrane. Treatment of 19 with methanol saturated with HC1 in the presence of additional chloride ion sources allows the
formation of dichlorodiester 20 (Scheme 6 ). During the formation of 20, a small amount of a by-product also resulted. Examination of this compound indicated the presence of a hydroxyl group and a remaining ethyl ester. Recrystallization of this by-product produced a crystal suitable for X-ray analysis, which established the substance to be 21 (Figure 2). This compound resulted from reaction of a small portion of diketone 13 with only one equivalent of the cyclopropyl ylide. This "mono-Trost" adduct was then carried throughout the reaction sequence as an impurity until its reactivity difference was noted .13 This mono-lactone was, fortunately, unreactive to the conditions used to open 19, conveniently allowing separation of this by-product and continued progress to dodecahedrane. 1 5
■OEt
Pd-C 95% .OEt
1 7 1 8 NaCNBHg CH3OH 73% pH *4; CeHe.hT -HzO c u « W s > 0
HCI, EUN*Cr LiCI, MeOH 39-44%
20 19
Scheme 6 . Conversion of 17 to Dichlorodiester 20 ORTEP Drawing of By-Product 21 Viewing Through the Top, Central Bond
Figure 2. X-ray Structure of 21 17
D. Closing the Front of the Molecule At this stage, all twenty carbons of the dodecahedryl framework are present and a C 2 axis of symmetry has been maintained, thus allowing for rapid construction of a fairly complex molecule. The synthesis must now deviate from this pattern while the frontside of the molecule is closed .14 Dissolving metal reduction of dichlorodiester 20 in liquid ammonia generates a radical anion at one of the ester carbonyls (Scheme 7). This species undergoes intramolecular S^2 displacement of chloride ion, acceptance of a second electron, and loss of methoxide to deliver ketone 22. The central bond of this 1,4-dicarbonyl system is now cleaved and reduction of the remaining chloride yields the intermediate dianion 23. Upon addition of one equivalent of chloromethyl phenyl ether, a 2:1 ratio of the desired adduct 25 to the transannular species 24 is obtained. This reaction is especially subject to the formation of by products. Unless exact amounts of all reagents are used , 15 the yield of desired keto ester drops drastically. Despite these limitations, I was able to increase the scale of this reaction to a level more commensurate with the amount of material available. 1 8
2 0
1. -C l-
NH.
22 23 JPhOCHzCI
25(43% ) PhOCHj 24(22% >
Scheme 7. Birch Reduction of 20
Irradiation of 25 continues construction of the sphere by generating tertiary alcohol 26 via homo-Norrish closure. This reaction proceeds cleanly, suggesting that the requirements indicated by Scheffer and Dzakpasu 16 for this type of reaction are well met by 25. Dehydration occurs smoothly without skeletal rearrangement and subsequent reduction of olefin 27 with diimide 17 yields 28.
Once again, the product contains a plane of symmetry (Scheme 8). 1 9
CHaO PhOCH2 PhOCH,
hv/pyrex (ilter t -BuOH-benzene ho: Et3N 8 5 -9 0 %
25 2 6 p -TsOH b en zen e 90%
CH30 q P hOCH2 V s
HjNNHfc, H20 2 CuS04, MeOH 95%
2 8 2 7
Scheme 8. Conversion of 25 to 28
E. Closing the Back of the Molecule At this point, the stage was set for closing the back of the molecule. Recourse was again made to homo-Norrish chemistry. The reduction of 28 with diisobutyl aluminium hydride (Dibal-H) to alcohol 29 followed by oxidation yielded aldehyde 30. On small scale, PCC oxidation was the most convenient method. For more reasonable quantities, Swern oxidation proved superior. Although the yields were similar, only the latter was capable of consistent scale-up (Scheme 9). 20
CHgO PhOCH;
Dibal-H QbHb 95%
2 8 2 9 PCC, CH2Clz 67 - 80 % or Swern
OH
hv / pyrex filter toluene: ethanol -78 °C 32%
31 3 0
Scheme 9. Conversion of 28 to the Low Temperature Photolysis Product
The utilization of chloromethyl phenyl ether as the trap for dianion 23 was for the purpose of preventing enolization of this aldehyde. In its absence, oxidation of the corresponding alcohol to the aldehyde was unsuccessful . 14 Photolysis of tertiary aldehyde 30 does indeed result in a second homo-Norrish reaction, but decarbonylation and photoreduction are also possible. In the last 21 instance, product 29 may be readily recycled; the decarbonylated products are, however, not useful to the synthesis of dodecahedrane . 18 To reduce the extent of competing decarbonylation, the reaction mixture was irradiated at a low temperature, but even then only a 32% yield of 31 could be isolated. With the enolization problem solved, the phenoxymethyl appendage could be removed to continue towards dodecahedrane (Scheme 10). Birch reduction of 31 delivered dihydrobenzene 32, which afforded diol 33 upon acidic hydrolysis. Again, PCC oxidation was limited to small scale. Swem oxidation conditions overcame the problematic scale-up of this oxidation, bringing the yield of aldehyde 34 back to an acceptable level. In the past, the aldehyde carbonyl of 34 was removed by performing a retro-Claisen reaction with potassium hydroxide, but only a 37% yield resulted. Recently, good use was made 19 of the point driven home so well in the previous photoreaction, namely the strong tendency of tertiary aldehydes to decarbonylate. Indeed, direct photolysis of keto-aldehyde 3 4 initially forms the decarbonylated ketone which undergoes the desired homo-Norrish reaction upon continued irradiation to yield the seco-alcohol of dodecahedrane (35, Scheme 11). OH PH
Li, NHs, THF CjHgOH 95%
31 3 2 10% HQ THF, r.t. 3 hours
OH pHO
(COCIfe, DMSO CHjCfe, -78° C; EtaN
3 4 3 3
Scheme 10. Preparation of Keto Aldehyde 34
OH PHO
hvl pyrex fitter f-BuOH: CeHe
3 4 35
Scheme 11. Preparation of Secododecahedryl Alcohol 23 f. Improved Access to Secododecahedrene The amount of work that has been accomplished on the synthesis of dodecahedrane by previous researchers has led to the optimization of many of the reactions prior to my involvement with the project. However, I was able to contribute to the synthesis by modifying reaction conditions to make yields consistent (i.e. domino Diels-Alder) or overcoming scale-up problems (i.e. dichlorodiester scale-up or use of Swern methodology). Improvements in the beginning of a linear synthesis are good, but become more appealing as the molecule nears completion. Of the three laboratory manipulations remaining to complete the synthesis of dodecahedrane, two of them had always proven problematic. An improvement in the overall yield would translate directly into an increase in the resultant quantity of dodecahedrane. With this in mind, I sought to improve this final set of reactions. Historically, the dehydration of seco alcohol 35 was accomplished by stirring 35 in the presence of p-TsOH with warming to 70°C. Although seco olefin 36 resulted and could be obtained pure after chromatography, the yield was low and very unpredictable (generally 30 to 50%). In addition to the desired olefin, additional compounds began to appear as the reaction progressed. I reasoned that the olefin was unstable to these conditions and a shorter reaction time was desired. In fact, the reaction could be accomplished in a fraction of the time, without evidence of decomposition, if 35 was stirred with silica gel at 70°C 24 prior to the addition of a catalytic amount of p-TsOH .20 Instead of the previous 3-hour reaction period, a mere 30 minutes was sufficient to consume 35. The yield due to this modification did not increase relative to that previously realized, but the average yield (when inconsistency was accounted for) was increased. At this stage of the synthesis, a 50% yield is still quite painful. Although the improvement was significant, I continued to consider alternative reaction variations. The possibility of ionic hydrogenation 21 of the seco alcohol to afford secododecahedrane (37) in one step attracted my interest. This methodology has been shown to convert tertiary alcohols and substituted olefins into the corresponding saturated systems. The alcohol was often found to react through the intermediate dehydrated olefin which underwent protonation to yield the reactive carbocation .22 Since elimination of water from 35 necessitated the intermediacy of a cation and any seco olefin formed was also expected to deliver a cation readily, the ability of this carbocationic species to participate in an ionic hydrogenation reaction was explored. When seco alcohol 35 dissolved in dichloromethane containing triethylsilyl hydride was treated with one drop of trifluoroacetic acid at room temperature, very rapid conversion (within 15 minutes) to 36 resulted. As the conversion to the saturated system was expected to take longer, the reaction was allowed to continue. 25
Unfortunately, even after three days, only a small quantity of secododecahedrane was observed. Although the one-step conversion of 35 to 37 was thus not effected, facile dehydration conditions were discovered that do not contribute to the decomposition of the seco olefin. Additionally, the lack of by-products not only resulted in increased yields, but also allowed a more readily isolated and purified product. Thus, stirring 35 in dry dichloromethane at room temperature, addition of one drop of trifluoroacetic acid and stirring for 15 minutes resulted in the complete consumption of alcohol and formation of 36 (easily monitored by TLC). Passage of the crude mixture through a silica gel plug followed by solvent removal (in vacuo, no heat required) afforded the seco olefin. Short column chromotography gave pure olefin 36 ( 8 6 %, Scheme 12). This improvement over the historical method greatly increases the amount of dodecahedrane that can be produced from a given amount of starting material. Not only did the average yield of this step (relative to the previous 30 to 50%) nearly double, but the ease with which the product is obtained was increased .23 Historically, the diimide reduction of the pure seco olefin was never a problem, consistently returning 37 in high yield. 26
TsOH, C6H6 Silica Gel, A 50%
CFaCCfeH. CH2Cfe,86 %
THF. HjjNNH j . Nal04> EtaSiH. C uS04 CFaCOjH, CHjCfe
3 7
Scheme 12. Improved Access to 36 and Conversion to Secododecahedrane
G. Improved Access to Dodecahedrane The final step of the dodecahedrane synthesis has always been difficult. A great deal of work by previous researchers has revealed the use of the catalyst system 5% Pt/Al 2C>3 to be the most fruitful in terms of returned dodecahedrane. Heating of an intimate mixture of 37 with this catalyst produced 3 in 35 to 50% yield after purification. This reaction was particularly sensitive to shaking, for 27
the lack of external motion resulted in abominable yields. An operating temperature of 220 to 250°C was also required for high yields. This is easier said than done when you realize the mixture must be shaken quite often (2 to 3 times an hour for the first 8 to 10 hours, 3 to 10 additional times until 24 hours is up). The temperature had been previously maintained by the use of a molten salt bath (KN 0 3 /N a N 0 2 : 10/7) and kept at 220°C by means of a heating mantle. With the amount of material to be made, this meant the heating mantle was often on for weeks at a time. Since heating mantles were not built for this type of abuse, they often "burned out," resulting in solidification of the salt, breaking of the thermometer 24 and occasional loss of 3 by breaking the reaction tubes. Add to this the cost of replacing a two-liter heating mantle every 2 to 3 weeks and the method gets to be quite troublesome. Although this method allowed shaking of the reaction, the cumbersome safety shield and molten salt proved awkward. The inconvenience and cost of this step warranted the development of an alternate heating method. The ideal method should meet certain criteria. It should not involve expensive heating mantles, it should allow ready temperature control while permitting easy shaking, and should (as an added bonus) not be space consuming. A method exhibiting these characteristics was found! Through modification of a "hot tube" equipped with a thermometer 24 and attachment of a clamp to the insulated tube, the temperature could 28 be accurately controlled while allowing ready shaking. Since the metal sheath of the tube was its own shield, no space consuming shield was necessary. Additionally, since the integrity of the heating mantle no longer had to be considered, the temperature could be raised to 250°C. The combination of all these factors also increased the isolated yield of dodecahedrane by close to 10%. This method routinely returned 60 to 65% of pure dodecahedrane (Scheme 13).
5 % Pt/AIA
250°C sealed tube 60-65%
3 7 3
1HNMR(CDCI3) S 3.38 (s) 13C NMR (CDCI3) 66.93 ppm
Scheme 13. Dehydrogenation to Dodecahedrane
I completed this synthesis twice. In general, diacid 9 was converted totally to bislactone 19 or dichlorodiester 20. At this stage, batches of the remaining intermediate were taken through the sequence to dodecahedrane. In this way, although a large amount of dodecahedrane (>50 mg) was never available, a continuing supply emerged from the "pipeline." The first run through began with just 29 shy of 1/2 kg of diacid 9; the second one began with just over 1 kg of 9 (approximately one third of bislactone 19 remains). A large portion (about one quarter to one third) of the initially formed dodecahedrane was allotted for heat of combustion studies presently being performed. With the synthesis of 3 detailed, Appendix A illustrates pictorially that all twenty carbons originate from three molecules and depicts their relative locations. Additionally, the formal nomenclature of dodecahedrane is included there. 30
References
1. McNeil, D.; Vogt, B. R.; Sudol, J. J.; Theodoropoulos, S. S.; Hedaya, E. J. Am. Chem. Soc. 1974, 2£, 4673.
2. Paquette, L. A.; Wyvratt, M. J. J. Am. Chem. Soc. 1974, 2£, 4671.
3. These conditions were developed and refined by Matzner. See: a) Matzner, E. A. Ph. D. Thesis, Y^le University, 1958; b) Doering, W. v. W. In Theoretical Organic Chemistry - The Kekule Symposium: Butterworths: London: 1969 p 45.
4. This innovation was pioneered by R. T. Taylor and M. W. Pelter (Miami University).
5. Taylor, R. T.; Pelter, M. W.; Paquette, L. A. Org. Syn. 6 8 , 198, (1989).
6 . Huntress, E. H.; Lesslie, T. E.; Bomstein, J. Org. Svn. Coll Vol 4, 329 (1963).
7. Paquette, L. A.; Wyvratt, M. J.; Berk, H. C.; Moerck, R. E. J. Am. Chem. Soc. 1978. 100. 5845.
8. Paquette, L. A.; Balogh, D. W. J. Am. Chem. Soc. 1982, 104. 774.
9. Originally sodium methoxide in methanol was used, but the ethyl ester groups proved more tolerant of the conditions needed in the upcoming "Trost Series." See: Paquette, L. A.; Miyahara, Y. J. Org. Chem.. 1987, 52., 1265.
10. Epoxide formation is deterred by the molecular superstructure which maintains these groups in a geometric relationship approaching 120°. 31 11. a) Moriarty, R. M.; Adams, T. Tetrahedron Lett.. 1969, 3715; b) Grewe, R.; Heinke, A.; Sommor, C. Chem. Ber., 1956, ££, 1978.
12. This compound was used an an introductory probe to see how the carbonyl group's effect on the electron density of the central olefin distorts the geometry of this norborenone-type system. Lagerwall, D. R.; Galucci, J. C.; Paquette, L. A. Acta Cryst. in press.
13. The large quantities of material handled at this stage of the project did not warrant the removal of all trace impurities. When a convenient purification point was reached, the impurities were removed, reducing the overall time of the sequence. Also, as noted by Dr. J. King, the mono-Trost impurity is not distinguishable from the desired isomers until the reaction sequence is over.
14. Paquette, L. A.; Balogh, D. W.; Temansky, R. J.; Begley, W. J.; Banwell, M. G. J. Org. Chem.. 1983,4&, 3282.
15. Any excess reagent would cause a change in concentration of other reagents and thus make the reaction's yield lower. For example, too much trap will consume the desired product by forming the dialkylation product while too little trap is just as detrimental.
16. Scheffer, J. R.; Dzakpasu, A. A. J. Am. Chem. Soc. 1978, 100. 2163.
17. Diimide reduction of similar twisted olefins was found superior to a Parr hydrogenation apparatus. See relevant references in: Paquette, L. A.; Temansky, R. J.; Balogh, D. W.; Taylor, W. J. L_ Am. Chem. Soc. 1983. 105. 5441.
18. The use of these products as intermediates to azadodecahedrane is currently underway. Dr. A. Gazit, unpublished results.
19. Dr. T. Kobayashi was the first to notice and utilize this improvement. 32
20. The silica gel was found to be important, for without it the yields again became unpredictable. No reaction was observed before the addition of TsOH. Prior to solvent removal, it was beneficial to add triethylamine and filter the crude mixture.
21. For a review, see: Kursanov, D. N.; Parnes, Z. N.; Loim, N. M. Synthesis. 1974, 633.
22. Carey, F. R.; Tremper, H. S. J. Org. Chem.. 1971, 26, 758.
23. Unfortunately, although the first improvement of this reaction was discovered early in my involvement with the project, the later was found as I neared the end of my research and thus will help the future members of the project more than it did me.
24. Used in conjunction with a "thermowatch" to control the temperature. CHAPTER III DODECAHEDRYL RADICAL STUDY
A. Background 1. Targeting Radical Precursors Previous researchers found bromododecahedrane (38) to be an unacceptable radical source, for when it was reacted with allyl tri-n- butylstannane under standard radical conditions, no desired allyl- dodecahedrane was formed (Scheme 14).1
Initiate + Bu3S n '
3 8 3 9
Scheme 14. Attempted Allylation of 38
With this precedent, it was deemed necessary to explore alternate radical precursors. Beckwith and Pigou have published a quantitative study on the order of reactivity of various groups X
33 34 toward Sh 2 attack by tributyltin radicals when R remains constant
(Scheme 15).2 They found I > Br > PhSe > Cl > ArS > MeS, noting a general trend towards increasing reactivity down and to the right of the periodic table.
R-X + Bu3Sn» ------► R» + Bu3Sn-X
Scheme 15. General Reaction Studied by Beckwith and Pigou
on Sh 2 Radical Reactivity
Although the slightly less reactive3 xanthate esters 4 and aliphatic nitro compounds 5 have also been shown to be effective sources of alkyl radicals ,6 the corresponding dodecahedryl derivatives were not seriously considered. The unusually high insolubility of dodecahedryl alcohol (53) would make this xanthate precursor difficult to handle. Direct nitration of the dodecahedryl nucleus has been previously attempted, yielding the nitrate (40) instead of the desired nitro derivative 41 (Scheme 16).7 More recently, collaboration with the Olah group to utilize their expertise in this field again failed to return any 41 (Scheme 17).8 Oxidation of the known aminododecahedrane (45) was not considered practical due to the length of the reaction sequence (Scheme 18).9 35
ONO.
n o 2b f 4 CHjCIj
3 4 0
Scheme 16. Formation of 40 Upon Attempted Nitration of 3
NO.
n o 2bf4
■ - X - " EtN02 O 0tort, 6 days
3 4 1
Scheme 17. Unsuccessful Nitration of 3
The direct chloramination of 3 with trichloroamine/aluminum trichloride10, unlike the adamantane case11, returned no desired amino derivative . 1 Recently, the similar amination of bromoadamantane 12 rekindled interest in the direct preparation of 45 from 38. Unfortunately this promising route produced no desired amine (Scheme 19). 36
CONH,
MejAINkfe
,NHCOzf-Bu
Phl(OjCCFJj f-BuOH, 50°C
Scheme 18. Route to Aminododecahedrane
NH.
NCIs, A(CI3
-X-
38 4 5
Scheme 19. Unsuccessful Direct Amination Attempt on 38
The next two intermediates targeted were iodododecahedrane (46) and (phenylseleno)dodecahedrane (47) because of their 37 anticipated ease of homolytic cleavage and high potential as radical precursors.
SePh
4 6
2. Derivatization Methods Available With these targets in mind, an overview of the type of chemistry available to effect derivatization of the dodecahedryl nucleus is first warrented. When dodecahedrane is stirred in bromine at room temperature overnight, 3 8 is formed quantitatively. No polybrominated dodecahedranes are formed under these conditions indicating the ionic nature of this process. Bromide 38 has played a key role in the development of dodecahedryl chemistry , 13 but the only useful derivatization technique involved cationic chemistry. Just as adamantyl bridgehead substitution relies heavily upon polar Sn 1 substitution reactions under electrophilic conditions14, so does that of dodecahedrane. For only in the presence of a Lewis acid and an electrophile was successful derivatization accomplished. Scheme 20 illustrates some of the derivatives available via this cationic route. 38
Cation Ravlaw
4 8 .X .C I 51
AgOTf
XH 3 OH CHaClj
FeClj, C6Hb
3 8 5 0 AgOTf. CH3 CN. then KfeO heat
NHCOCH3 OR
52.R-COCF 3 5 4 4 1 ,R - N 0 2 53.R-H
Scheme 20. Cationic Derivatization Techniques
Halogen exchange (to F or Cl) may be achieved with the appropriate Lewis acid in an inert solvent such as dichloromethane. When benzene is used as the solvent, a Friedel-Crafts reaction conveniently delivers phenyldodecahedrane (50). Preparation of the 39 methyl ether takes advantage of the ability of silver(I) ion to induce ionization of bromododecahedrane. Addition of silver triflate to a
solution of the bromide in methanol-dichloromethane ( 1 :1) at room temperature gave 51. In a similar manner, the trifluoroacetate 5 2 was prepared upon exposure of 38 to silver trifluoroacetate and trifluoroacetic acid. Finally, formation of the N - dodecahedrylacetamide was achieved by again making recourse to silver triflate. Suspension of 38 and AgOTf in hot acetonitrile followed by hydration delivered 54.
B. Radical Precursors 1. Iodododecahedrane Because iodo compound 46 was determined to be a valuable intermediate in the radical investigation, a variety of methods were examined to prepare this elusive species. Literature precedent 15 indicated that 38 should be converted to the iodide with aluminium triiodide. However, when freshly prepared AII 3 was reacted with 38, no iodo derivative was isolated. Instead, dodecahedryl alcohol 53 was obtained (Scheme 21). Formation of dodecahedryl iodide may have been accomplished under these conditions, with conversion to the alcohol occurring upon work-up. Evidence for this was the appearance of a product peak with the expected retention time (see TABLE 1) when the crude mixture was analyzed by capillary GC, but the disappearance of this peak and production of 53 40 upon work-up. The aqueous work-up used to remove the aluminum salts must hydrolyze this highly polarizable species, the cation being trapped by water to return the alcohol.
TABLE 1: Some Dodecahedryl GC Retention Times
GC Retention Time Compound (min.) DDH-H (3) 8.6 DDH-C1 (48) 11.9 DDH-Br (38) 15.2 possibly DDH-I (46) 19.0
ch 3c n Al,(foil) Alla
OH
All, Workup
3 8 4 6 5 3
Scheme 21. Reaction of 38 with AII 3 (Possible Intermediacy of 46 ?) 4 1
To avoid an aqueous work-up, a halogen exchange was attempted in light of the fact that both the chloro- and fluoro- derivatives had proven accessible by this methodology. Anhydrous zinc iodide was used as a convenient Lewis acid. Although Znl 2 is not as polarizing as the Lewis acids used previously to effect 1 substitution, longer reaction times were expected to accommodate the desired reaction. When 38 was stirred with excess Znl 2 in dichloromethane, the surprising emergence of dodecahedrane resulted! What sheds a little more light on this curious reaction is the formation of dodecahedryl alcohol if a little moisture is present (Scheme 22).
Znlj.CHjClg
3 8 3
OH
Znlg,CHgCI2 trace H20
3 8 5 3
Scheme 22. Reactions of 38 with Znl 2 42
A possible explanation for these observations is the intermediacy of the desired iodododecahedrane. Alkyl iodides are known for their ease of homolytic cleavage (thus their use as radical precursors). If iodododecahedrane was an intermediate, its decomposition to the dodecahedryl radical followed by hydrogen abstraction could explain the presence of dodecahedrane (a trace of chlorododecahedrane was also observed in some reactions). Although the presence of the alcohol is an indication of the cation and not the corresponding halide, any iodododecahedrane present is expected to be more readily converted to the cation relative to the bromide due to the more easily polarizable carbon-iodine bond. In any case, iodododecahedrane was deemed too unstable to work with as a radical precursor.
2. (Phenylseleno)dodecahedrane Parallel to our quest of the iodo derivative, efforts were directed toward the synthesis of the second of the desired radical precursors, (phenylseleno)dodecahedrane. Attempted synthesis of this molecule by other researchers had met with no success (Scheme 2 3 ) .16
Although, in theory, 47 was available 17 by quenching lithiododecahedrane with a phenylselenenyl halide, the lithio derivative of dodecahedrane had not yet been successfully generated and thus a cationic pathway was examined. The utility of (trimethylsilyl)phenylselenide (TMSSePh) had recently been 43 demonstrated in the formation of alkyl phenylselenides under electrophilic conditions18. This reagent was, however, not stable in the presence of these Lewis acids needed to bring about Sn 1 substitution of the dodecahedryl framework19. The lack of success in the previous attempts to prepare 47 using these harsh Lewis acids initiated an effort to apply milder catalysts. The stability of zinc salts (ZnCl 2, ZnBr 2 , and Znl 2>20 to TMSSePh made possible the desired study.
_ _ PhSeH .. 3 8 ------. — ► No Reaction AgOTf
Me2AISePh 38 ► No Reaction
52 -H ► No Reaction 60°C
Scheme 23. Previous Unsuccessful Attempts to Obtain 4 7
TMSSePh was conveniently prepared from diphenyldiselenide
(55) according to the procedure of Sonoda .21 Reduction to sodium phenylselenide followed by trapping with trimethylsilyl chloride and distillation afforded the desired product (Scheme 24). 44
Na° TMSCI PhSeSePh 2 PhSeNa PhSeSiMe3
55 56 57
Scheme 24. Preparation of TMSSePh
The low solubility of ZnCl 2 in dichloromethane coupled with the competing transhalogenation leading to chlorododecahedrane made it an unacceptable choice. A small amount of the desired product was produced upon exposure of 38 to TMSSePh in the presence of ZnBr 2 - Although halogen exchange no longer complicated the mixture, the low solubility of the salt contributed to a prohibitively long reaction time. Znl 2 , however, worked exceedingly well! Not only was it expected to be more soluble than the other two salts, but as noted earlier, any halogen exchange would lead to a proven highly reactive intermediate, readily convertible to the desired product (Scheme 25). Thus, admixture of bromododecahedrane in dichloromethane with excess Znl 2 and excess TMSSePh in the dark led to complete consumption of 38 after three days at room temperature. The isolation of 47 from this reaction mixture illustrates two important points about working with dodecahedryl derivatives. First, due to the scale generally used, the functionalized dodecahedrane is almost always the limiting reagent and separation of the excess reaction components is often a problem. In this case, the highly toxic selenium compounds offered a challenging purification obstacle. This was overcome by making use of the ease of air oxidation of 45 phenylselenol to diphenyldiselenide. Thus, after washing the zinc salts from the reaction mixture, the phenylselenol was readily converted to the more easily handled diphenyldiselenide under a
TMSSePh ZnCfe, CH-jCIa
3 8 4 8 (trace)
.SePh
TMSSePh ZnBr2l CH2CI2
3 8 4 7 (trace)
.SePh
TMSSePh Znlg, CHjCIj
3 8 4 7
Scheme 25. Reactions of 38 with TMSSePh and ZnX 2 46 stream of air. Originally this compound was removed by sublimation followed by preparative TLC to provide approximately 65% yield of the desired (phenylseleno)dodecahedrane. However, the general insolubility of dodecahedryl derivatives in diethyl ether permitted direct trituration of the crude mixture of diphenyldiselenides with a minimum amount of diethyl ether as a means of removing the diselenide and leaving pure 47 (62% isolated yield).
C. Dodecahedryl Radical Generation 1. R eaction Design With a potential radical precursor in hand, generation of the dodecahedryl radical was pursued. Since no successful radical trapping had yet been achieved, it was decided not to complicate the reaction analysis with a complex product. In the choice of an appropriate radical trap, careful thought had to be given to assure a readily identifiable product. The high molecular weight and low volatility of dodecahedryl derivatives makes certain compounds not emerge from a capillary gas chromatography column. Without this valuable tool, determination of the reaction mixture's contents becomes very time consuming. Alternatively, a chromophore in the product facilitates purification and identification by allowing preparative thin layer chromatography to be simplified. 47
2. Reaction with Acrylonitrile With these criteria in mind, acrylonitrile was chosen as the radical trap. The successful trapping of 1-adamantyl radical by this olefin had been demonstrated by Eguchi 22 using the tributyltin hydride (donor) - azobis(isobutyronitrile) (AIBN, initiator) method 23 to bring about reaction (This will be referred to as the "tin hydride method" throughout this manuscript). Their success in preparing bridgehead substituted adamantane derivatives (via the free radical pathway) was encouraging. To our delight, when (phenylseleno)dodecahedrane was treated under the above conditions, the desired nitrile was formed as the major product (by GC and GC/MS). These extremely encouraging preliminary results initiated a parallel investigation to answer a puzzling question: namely, what was the possibility of generating radical 56 from the more abundant and theoretically more reactive bromododecahedrane? As mentioned earlier, previous attempts by others to utilize 38 in this manner were fruitless even though there are countless literature examples of alkyl bromides serving as radical precursors. Now that the radical had been generated and successfully trapped, the product could now indicate the utility of alternate radical precursors. When bromododecahedrane was treated in the same manner as phenylselenide 47, a similar reaction profile resulted! Indeed, the radical could be generated from the bromide, contrary to previous reports. In fact, from the bromide, the desired nitrile 57 could be 48 isolated in 65% yield with 3 formed as the major byproduct (Scheme 26).
Formation of tha Dodocahodryl Radical
.SePh ,CN
[•••
Major
4 7 5 7 AIBN, CjHe, A BusSnH Bu3SnH
5 6
3 8
Scheme 26. Generation and Trapping of 56 with Acrylonitrile
As in the adamantyl case, the reaction course is depicted as the well-known cyclic radical chain process (Scheme 27)22. Possible side reactions are reduction of 56 to 3 and competitive addition of tin radical to the olefin leading to a tin adduct. Alternatively, the secondary radical could react with additional trap to form polymeric material (telomerization). ^ C N BuaSrv B u g S n ^ ^ ' CN [H]
^ C N polymeric 58 material
Scheme 27. Assumed Mechanism and Possible Side Reactions in the Formation of 57 50
A choice now had to be made: from which radical precursor, the bromide or the phenylselenide, should the radical chemistry be explored? The use of the bromide had some obvious advantages: 1) the loss of material that resulted from the preparation of the phenylselenide was avoided, 2 ) the purification of bromododecahedrane was more convenient than that of 47, and 3) the bromide was much easier to observe by capillary GC (38 at 15.2 min; 47 at 77 min), resulting in simplified reaction analyses. It was thus decided to investigate the bulk of the dodecahedryl radical chemistry from the bromide, knowing that if initiation of the reaction became a problem, we could fall back upon the available (phenylseleno)dodecahedrane.
3. Trapping of 56 with Acrylates When methyl acrylate was used as the radical trap utilizing the tin hydride method, only a complex mixture of polymeric materials resulted. Ethyl acrylate behaved similarly, yielding a mixture impossible to analyze due to the polymers present. Even from 47, the completed reaction mixture's components could not be identified. To gain access to the reaction mixture's components, the reaction time was shortened. Although there would still be bromide present, the degree of polymerization might not prevent product identification. Upon analysis (GC/MS), the only dodecahedryl product observed was dodecahedrane, resulting from hydrogen atom transfer to the dodecahedryl radical (Scheme 28). X - Br. SePh R - Me, Et
Conditions Attempted
Vary radical precursor X - Br, SePh Vary time of reaction: to monitor reaction Vary initiation method: AIBN, A; AIBN,hv, (RgSn^, A; (RgSn^, hv Vary conditions: trap concentration, hydride concentration
Scheme 28. Attempted Acrylate Trapping of 56
Since the acrylates are similar in radical reactivity to acrylonitrile ,24 it seemed plausible that if the hydride concentration could be lowered, the desired product would be formed. Thus to hinder the competing reduction, the tin hydride was formed in situ, in catalytic amounts. The use of catalytic tin hydride has several benefits. Since the concentration of the available hydride is diminished, the trap may compete more effectively for the dodecahedryl radical, lowering the amount of reduction that occurs. Also, since less tin compound is needed, less has to be removed from the completed reaction, simplifying isolation of the desired product. Finally, from a toxicity 52 standpoint, small quantities of tin compounds are preferred over larger amounts. In general, a catalytic tin hydride method makes use of the conversion of a tin halide to the desired tin hydride in the presence of a hydride source (UAIH 425, NaBH426, NaCNBH 327). Ethanol or tert- butanol have been used as solvents to trap the resulting diborane. Bergbreiter and Blanton further extended this methodology by avoiding the use of alcoholic solvent when including a phase transfer catalyst to transport the NaBH 4 to toluene28. Since I desired not to use an alcoholic solvent (due to the general insolubility characteristics of dodecahedryl derivatives) and felt no need to complicate the reaction mixture by including a toxic crown ether, I developed a method suited to my needs. Envisioning the in situ generation of tributyltin bromide from hexabutylditin and bromododecahedrane after initiation with AIBN, I reasoned the slight solubility of NaBH 4 in benzene would be adequate to allow the desired reactions to occur (Scheme 29). When heated to 75°C under these conditions, both methyl and ethyl acrylate produced a trace amount of the desired products. By GC/MS analysis, neither compound gave a parent peak but both exhibited the same fragmentation pattern (consistent with a mono substituted dodecahedryl derivative with a CH 2 CH 2CO appendage: m/z of 258, 259, 260, 286, 314, 315). The 258, 259, and 260 fragments are fairly common for the dodecahedryl nucleus with 259 generally preferred; the m/z of 286 corresponds to M'+-HC 0 2 R; m/z 53 of 314 to M-+-HOR; and m/z of 315 to M'+-OR. Only the relative intensities of the m/z 314 and 315 peaks differed between the two products. When R = Me, the retention time was 54 minutes and with
R = Et, the retention time was 66 minutes. These factors led to the assignment of the two minor components as the desired esters. Unfortunately, the major components were still dodecahedrane and remaining dodecahedryl bromide. Polymerization of the acrylates was still a complication (Scheme 30).
(BuaSnfc — t s t * — - 2B usSn
BitaSn. + R-Br ------► BuaSn-Br + R •
2 Bu3Sn-Br + 2 NaBH4 2 BugSnH + 2 NaCI + B2H6
R • +
+ BuaSn-H + BuaSn <
Scheme 29. Catalytic Tin Hydride Generation 54
It is possible that the dodecahedryl radical may have added to the acrylates better than the results seem to indicate. If the rate of telomerization 29 occurred faster than conversion to the desired product, the polymeric adduct may not be observable by GC/MS analysis, resulting in observation only of the monoadduct. The GC/MS analyses were run for extremely long (hours) periods of time to check this possibility, but no additional peaks emerged.
,c o 2r
+
3 8 59a + b 3 (trace) R .M e.E t
Scheme 30. Formation of Trace Acrylate Trapping Product
It has been noted 22 that very reactive alkenes lose their utility as radical traps due to the many competing reactions of which they are capable. Whether it be polymerization without involvement of the bromide, tin radical addition to the trap, or polymerization of the intermediate radical, their inability to deliver the desired product renders them useless. Two of the many problems to overcome if the acrylates were to be used successfully are the elimination of the excess acrylate (due 55 to the complicated reaction analysis) and the lowering of competing reduction to dodecahedrane. Methodology recently developed by the
Hart group 30 utilizing bis(trimethylstannyl)benzopinacolate (60) addresses both of these problems. "Tin Pin," as 60 is affectionately called, has been shown to allow good yields of the trapping adduct to be obtained even when only one equivalent of the acrylate is present. Additionally, the production of an intermediate tin enolate avoids the need for a hydrogen atom source during reaction (Scheme 31). This precedent encouraged investigation with the dodecahedryl system .
Me3SnO OSnMe 3 Ph—|— |—Ph Ph Ph
60
OMe R*X + 60 ^ J ------*^C02Me R^^^OSnMea H2°
Scheme 31. Tin Pinacolate Methodology
Unfortunately, this methodology did not solve all of the problems associated with the addition of dodecahedryl radical to the acrylates. Irradiation decomposed the Tin Pin as expected, but there was inefficient conversion of 38 to the radical. Upon workup, no 56 desired adducts were observed and the mixture contained predominantly unreacted 38. A trace of dodecahedrane was observed. Perhaps under these conditions, the tin radical may compete successfully for the scarce acrylate, leaving hydrogen atom abstraction as the only reasonable option available to any dodecahedryl radical formed. Alternatively, the rate of decomposition of the dodecahedryl radical by hydrogen atom abstraction may compete with the addition of this trap. This may be unlikely due to the observance of successful radical trapping when utilizing the tin hydride method, (although there was a significant difference in trap concentrations). In any case, the quest for more suitable radical traps continued.
4. Attempted Trapping of 56 with 61 and 62
The successful trapping of maleic anhydride 31 (61) by 1- adamantyl radical 32 suggested its use as a less reactive, more stable radical trap. Unfortunately, when 38 was submitted to a wide variety of radical forming conditions ,33*34 the only identifiable components of the resulting mixture (by GC/MS and TLC;MS) was dodecahedrane and starting bromide. In the hope that a chromophore containing product would ensure its observation, N- phenylmalimide (62) was used as an alternative radical trap.3S The desired product should be UV active and allow even small quantities to be located and identified. Again, under a variety of reaction conditions, only 3 and 38 were identified (Scheme 32). 57
initiation Starting [H] Material
38
initiation Starting NPh [H] + Material
X - Br, SePh (38.47)
Scheme 32. Attempted Trapping of 56 with 61 and 6 2
As in the comparison of the cation chemistry of adamantane with dodecahedrane, wide differences in their radical chemistry is becoming obvious. The 1-adamantyl radical has been shown to be nucleophilic . 36 To compare the relative nucleophilicities of bridgehead radicals, their relative rates of solvolysis has been u s e d . 37 As found earlier1, conditions used to hydrolyze 1- bromoadamantane (10% K 2 CO 3) returned 38 unchanged, which suggests the dodecahedryl radical is less nucleophilic than the 1- adamantyl radical and reactions successful in one case may not operate in the second. This lower nucleophilicity is given credit for allowing the reduction of the radical to compete successfully with the desired trapping reaction. 58
5. Trapping of 56 with 2>Cyclopentenone Generation of the dodecahedryl radical (56) was not considered a problem. What was thought to be troublesome was the lack of nucleophilic character exhibited by the radical. With very reactive traps, polymerization becomes a complication and with less reactive traps, competing reduction becomes predominant. What was needed was a radical trap between these extremes. One which like acrylonitrile was reactive enough to encourage the dodecahedryl radical to add, but not polymerize or telomerize prior to conversion to the desired adduct. 2-Cyclopentenone (63) had been used as a radical trap 31 and was thought to fit these criteria. Once again, utilization of the "tin pin method" returned only 38 after consumption of the tin pinacolate. It was becoming obvious that this methodology (although impressive) was unsuitable in the present context. Since thermal initiation of this radical reaction was unsatisfactory due to the instability of the cyclopentenone, a photolytic initiation method was employed .38 Using the tin hydride method (Bu 3 SnH, AIBN, 38, hv), only a trace of the desired adduct was seen (GC/MS), with the major product being dodecahedrane. To hinder the competing reduction, the catalytic tin hydride methodology developed earlier was again deployed. When 38 dissolved in deoxygenated benzene containing cyclopentenone, hexabutylditin, and solid NaBH 4 was irradiated, the amount of reduction product decreased relative to the desired adduct. Unfortunately, as the reaction time was extended to ensure 59 complete consumption of the starting bromide, the yield of desired adduct decreased. The ketone product may not be stable under prolonged exposure to these conditions, as indicated by the appearance of a more polar product lost in the baseline impurities.
By shortening the length of photolysis, an optimized yield of 6 4 (35%) could be realized. Dodecahedrane and starting bromide accounted for an additional 30% (eluted together by prep TLC, approximately 1:1 by GC )39 (Scheme 33). Isolation of this ketone was at first quite troublesome. The molecule contained no chromophore and was thus UV inactive. The use of iodine as a reversible stain failed to indicate the location of the product on TLC plates. And finally, the reaction mixture was complex due to the side reactions of the cyclopentenone .40 The approximate location of the product was readily identified by staining a small portion of the plate, but the tendency of preparative TLC plates to elute unevenly necessitated the acquisition of additional information to ensure consistent isolated yield and purity. On small scale, impurities often overlooked on "normal" scales become obvious. Such was the case with a "phthalate" impurity introduced during work-up (i.e. from a rubber stopper, septa, or squirt bottle). I noticed the desired ketone eluted just above this UV active phthalate impurity. When included in the reaction mixture, even if the plate eluted unevenly, the product could still be obtained pure. Thus by making use of this "indicating" band, the desired adduct could be readily isolated. 60
(Bu3Sn)2, NaBH4 64 hv, C6^ 35% c5
38 63
+
3 ' v------' 3 8 30%
Scheme 33. Trapping of 56 with 2-Cyclopentenone
6 . Addition-Elimination Methodology
Tin hydride is capable of very rapid H-atom transfer ,4 1 generating a chain transfer agent (I^Sn*)- Unfortunately, this allows reduction of the initially generated radical to be a serious side reaction, limiting the usefulness of certain traps. A very powerful alternative method involves generation of the chain transfer agent by a fragmentation (rather than atom abstraction) reaction .42 Here the rapid fragmentation of a C-Y bond produces Y* and a neutral olefin. Most commonly, radical addition to allyl or vinyl derivatives with the appropriate substituents leads to these intermediates. The net result of this type of methodology is either allylation or vinylation of a suitable radical precursor (Scheme 34).
Scheme 34. Examples of Addition/Elimination Methodologies
Keck and co-workers have shown through systematic studies that free radical allylation with allylstannanes is a very powerful
method for the functionalization of organic molecules .43 The propagation sequence for allyl tri-n-butylstannane (65) is outlined in Scheme 35.
R-Br + Bu,Sn»
R* + ^ \ ^ SnBu3 ------► R^^^^SnBu,
r x ^ ^ S n B u a ------r / / \ ^ + BuaSn*
Scheme 35. Mechanism of Allylation using 65 62
After formation of a carbon-centered radical, addition to the allylstannane yields the adduct radical. Rapid P-scission of the tri-/i- butylstannyl radical delivers the olefin and the chain transfer agent. As mentioned earlier, this methodology was not successful in the hands of previous researchers. But since I now knew generation of the dodecahedryl radical could be accomplished from either the bromide or the phenylselenide, I reexamined this reaction. The characteristic advantages of this methodology (over the tin hydride method) stem from the absence of tin hydride in the system. The lifetime of the initial radical is now not limited by the. rate of hydrogen atom transfer. This ability to provide long lifetimes for intermediate radicals (to permit relatively slow reactions), yet rapidly transfer the chain was particularly appealing when considering this methodology. The only requirement is that addition of the radical to the allyltin be more rapid than decomposition of the radical. The use of only two equivalents of 65 in certain reactions studied by Keck43b suggested that only a small excess of this trap is adequate for successful reaction (unlike the large excess generally needed when utilizing the tin hydride method). No reaction was observed under these conditions and even when 20 equivalents of 65 was present, the conversion of 38 to the radical was inefficient
(after 8 hours irradiation with AIBN present, starting bromide was still the major component of the reaction). By GC/MS, 63 allyldodecahedrane had been formed, but surprisingly, so had dodecahedrane! In a study to survey the scope and limitation of this chain reaction, Migita subjected various organic halides to the reaction of allylic tin compounds .44 He found the efficiency of the free radical chain reaction between allylic tin compounds and organic halides depends on the nature of the halide and is affected by the ease of halogen abstraction by the stannyl radical. Electrophilic radicals were found to attack the n-bond of allyltin easily, whereas aralkyl and alkyl radicals were less reactive, requiring excess allyltin to be employed for a successful chain reaction .45 When a 200-fold excess of 65 was used, consumption of the bromide was complete. The major product (by GC/MS and GC) was the desired allyldodecahedrane, but there was still a significant amount of 3 present (30%, Scheme 36).
.Br
initiation
38 39 3
Scheme 36. Successful Allylation of 56 with 65 64
The large excess of tin reagent needed for sufficient conversion was problematic. Although a variety of methods are commonly used to remove tin compounds ,46*47 many of them utilize the insolubility of various tin derivatives in diethyl ether and are not suitable with the dodecahedryl system. The result was an inability to isolate the allyldodecahedrane sufficiently free of tin impurities. A quick note about the mass spectra of dodecahedryl derivatives is warrented. Parent hydrocarbon 3 has a molecular weight of 260 mass units and exhibits only the parent peak in its mass spectrum. The characteristic feature of 3 and its derivatives is the lack of fragmentation. The MS of a monosubstituted derivative will depend on the substituent, but will most commonly exhibit fragmentation with a m/z between 259 (cleavage of DDH-X) and the parent peak. Certain derivatives such as the allyl compound also show a peak at m/z = 258 due to a McLafferty rearrangement .4 8 When coupled with the parent peak, this allows for easy identification of allyldodecahedrane by GC/MS. The production of 3 as a side-product must now be addressed. Even without a formal hydride source, reduction still occurs. The dodecahedryl radical must be reacting via hydrogen atom abstraction from one or more sources in the reaction mixture (possibilities include tin compounds, solvent, or dodecahedrane). Whatever the source, the decomposition of this radical was occurring at a rate similar to its addition to the allyltin. This is rationalized knowing that the addition of an unactivated allylstannane is not considered a 65 particularly facile process ,42 especially for an unactivated radical such as that derived from dodecahedrane. Despite the shortcomings of this reaction, it does suggest additional reactivity characteristics of the dodecahedryl radical. First of all, it does deliver the desired allyl adduct as the major identifiable product, and hence a successful reaction from this point of view. The fact that dodecahedrane was formed suggests that Sh ' reaction of dodecahedryl radical with the allyltin is not fast enough to construct an effective chain process; some of the dodecahedryl radical may be consumed via a non-chain by-pass. Finally, the prolonged irradiation necessary to realize satisfactory conversion suggests a small kinetic chain length44. Because allyldodecahedrane was desired for other aspects of this project49 and we wished to characterize this species, an alternate route was developed. Once again, cation chemistry was utilized.
The preparation of 1-allyladamantane ( 6 8 ) under electrophilic conditions from the corresponding chloride developed by Sasaki 50 was used as precedent. An 85% yield of 68 resulted when chloride
6 6 was treated with allyltrimethylsilane (67) at room temperature in the presence of titanium tetrachloride (Scheme 37).
The known cation-stabilizing ability of silicon 51 increases the nucleophilicity of an unsaturated bond containing a p-silyl group.
Accordingly, it was expected that the dodecahedryl system would take advantage of such an enhancement in an electrophilic substitution reaction. Although TiC ]4 has been used successfully for 66 generation of the intermediate cation, I wished to avoid the transhalogenation by-product. To this end, Znl 2 was utilized as the Lewis acid in an attempt to extend its utility as a catalyst in electrophilic reactions with the dodecahedryl system.
(67)
66 68
Scheme 37. Preparation of 68 by Sasaki
In actuality, this reaction worked exceedingly well! When bromododecahedrane dissolved in dry dichloromethane containing excess allyltrimethylsilane and some solid Znl 2 was stirred for 3 days in the dark at room temperature, allyldodecahedrane was produced in 65 % yield (Scheme 38). The general representation of this reaction, in mechanistic terms, is outlined in Scheme 39 .52 From a preparative viewpoint, these reaction conditions were far superior than those used with allyltin, for there were no impurities to speak of. Further purification, if necessary, was conveniently carried out by preparative TLC. Luckily, no dodecahedryl by-products complicated the reaction, for had this been the case, preparative TLC would have been ineffective .5 3 67
Me3Si' Znl2, CH2CI2
3 8 3 9
Scheme 38. Synthesis of Allyldodecahedrane
Me3Si* Me3Si< 3> ^ s^ 'C>\ •TMSNu 1 \ Nu J
Scheme 39. Mechanism of Allylation with Allytrimethylsilane
P-Stannylstyrene 70 has been shown to undergo addition- elimination reactions with alkyl radicals to yield styrene derivatives.54*55 Hydrostannylation of phenylacetylene (69) according to the procedure of Saihi and Pereyre56 afforded the desired styrene as a mixture of isomers (>75% E). In the hope that this addition-elimination methodology could be extended to the dodecahedryl system, 70 and bromododecahedrane were reacted under dodecahedryl radical-forming conditions (AIBN, CgHg, hv). Unfortunately, other than starting bromide, the only dodecahedryl product identified was dodecahedrane (Scheme 40). Apparently, the dodecahedryl radical is again decomposing by H-abstraction faster 68 than the desired addition to the styrene (rapid p-scission of any intermediate formed would deliver the desired olefin and prevent any telomerization). The more reactive P-stannyl acrylate23* 72 was initially expected to be an alternative addition-elimination trap. Like the acrylates used in the tin hydride method, however, this substituted acrylate was not expected to survive the dodecahedryl radical- forming conditions. Similar considerations halted the search for yet other addition-elimination radical traps.
AIBN — Ph + HSnBuj Bu3Sn 6 9 7 0
7 0
initiation
3 8 71
Scheme 40. Preparation and Attempted Trapping with 7 0 69
D. Conclusion Derivatizatiop of the dodecahedryl system has been achieved by free-radical methodology. However, the derivatives were obtained quite painstakingly. The utility of free radical methodology on the dodecahedryl system is, in general, not of synthetic value. This is due, in part, to the lack of flexibility of radical traps. The low yields, care needed in the development of reaction conditions, co production of by-products and resulting difficulties in purification of the desired targets are some of the negative aspects of this methodology. We do, however, now know a great deal more about the reactivity of the dodecahedryl radical. When I first began this study, there were no examples of the site specific generation and trapping of the dodecahedryl radical. We can now generate the radical from either the bromide or phenyl selenide and, in the presence of a suitable radical trap, achieve the desired radical trapping product. The key is the suitability of this trap to the reactivity characteristics of the dodecahedryl radical. Unfortunately, very few standard radical traps accommodate the dodecahedryl radical's reactivity characteristics. As in all radical reactions, it is the relative rates of competing reactions that dictate the outcome. Although the yields may be optimized by careful design of reaction conditions to disfavor competing side reactions, the outcome is often beyond the control of the chemist. 70
Illustrated in Scheme 41 is a generic representation of some of the options available to the dodecahedryl radical (DDH has been used as a representation of dodecahedrane in this Scheme) and other components of a typical radical reaction. Based on solvolytic properties, the dodecahedryl radical was predicted to be less nucleophilic than the 1-adamantyl radical. This lower nucleophilicity (i.e. ka) has been used to explain the extent of competing reduction (ka vs. kn) when utilizing the tin hydride method. When addition- elimination methodology was explored (no tin hydride present), reduction still occurred to some extent. A low nucleophilicity (and lack of electrophilicity) would account for slowing addition of the radical to the allyltin and allow decomposition by H-atom abstraction to become possible (ka vs. kn')* Also, addition-elimination methodology results suggest that the rate of halogen abstraction by the tin radical (kx) is also sluggish, leading to inefficient conversion to the dodecahedryl radical. This assumption is supported by related studies involving the tin hydride method. The length of time needed to consume the bromide was such that the radical trap was often unstable (even though the same radical trap has been used successfully in other systems). Finally, the inability to achieve significant radical formation when utilizing tin pinacolate methodology may also support these assumptions. The inability of the tin radical to generate the dodecahedryl radical efficiently may be linked to tin's competing reaction with the trap (kx vs. ka"). Decomposition of the dodecahedryl radical could be possible in the low concentrations of trap available (Ich' vs ka). The combination of inefficient dodecahedryl radical formation and low nucleophilicity would certainly be disastrous to the synthetic utility (flexibility) of the radical chemistry of dodecahedrane. The observed reactivity is consistent with these assumptions.
DDH-X + *SnR3 DDH • + X-SnR3
DDH* + Trap DDH-Trap
DDH* + H-SnRg DDH-H + *SnR3
SH DDH* DDH-H kw
DDH-Trap*
DDH-Trap* DDH-Trap’
•SnR3 + Trap SnR3-Trap
R’-Trap* + Trap R’-Trap-Trap*
Scheme 41. Generic Representation of Possibilities in Typical Radical Reactions 72
References
1. Jeff Weber, Ph. D. Dissertation, The Ohio State University, 1987.
2. Beckwith, A. L. J.; Pigou, P. E. Aust. J. Chem. 1986, 2£, 77.
3. Curran D. P. Synthesis 1988, 417.
4. a) Barton, D. H. R.; Crombie, S. W. J. Chem. Soc.. Perkin Trans. 1 1975, 1574. b) Review: Hartwig. W. Tetrahedron 1983. 39. 2609.
5. Ono, N.; Miyake, H.; Kaji, A. Chem. Lett. 1985, 635.
6 . See also: Giese, B. Angew. Chem.. Int. Ed. Engl. 1985,24. 553 and relevant references cited therein.
7. a) Impurities in the nitronium tetrafluoroborate are credited for returning the nitrate. See reference 1. b) For the preparation of nitronium tetrafluoroborate free of nitrosonium ions, see: Elsenbaumer, R. L. J. Org. Chem. 1988, 52., 437.
8. Personal communication between: Olah, G. A.; Paquette, L. A.; King, J. K.; Trivedi, N. J.; Lagerwall, D. R.
9. Weber, J. C.; Paquette, L. A. J. Org. Chem. 1988. 53. 5315.
10. a) Kovacic, P.; Roskos, P. D. J. Am. Chem. Soc. 1969,21, 6457. b) Kovacic, P.; Goralski, C. T.; Hiller, J. J., Jr.; Levisky, J. A.; Lange, R. M. J. Am. Chem. Soc. 1965, £1, 1262.
11. Kovacic, P.; Chaudhary, S. S. Org. Svn. 1969. 48. 4.
12. Cahill, P. A. Tetrahedron Lett. 1990. 31. 5417.
13. Paquette, L. A.; Weber, J. C.; Kobayashi, T. J. Am. Chem. Soc. 1988,1111, 1303. 73 14. Reviews: a) Fort, R. C. Adamantane-The Chemistry of Diamond Molecules: Marcel Dekker: New York, 1976. b) Engler, E. M.; Schleyer, P. v. R. In MTP International Review of Science. Alicvclic Compounds: Parker, W., Ed.; Butterworths, London, 1973; Vol. 5. c) Bingham, R. C.; Schleyer, P. v. R. Fortschr. Chem. Forsch. 1971. 18. 1. d) Sevast'yanova, V. V.; Krayuskin, K. M.; Yurchenko, A. G. Russ. Chem. Rev. (Engl. Transl.) 1970, 39. 817. e) Fort, R. C., Jr.; Schleyer, P. v. R. Chem. Rev. 1964, M , 277.
15. Bhatt, M. B.; Baba, J. R. Tetrahedron Lett. 1984, 2 5 .. 3497.
16. Unpublished results, Dr. T. Kobayashi.
17. The preparation of l-(Phenylseleno)adamantane has also been achieved via a Sr n I mechanism, see: a) Palacios, S. M.; Alonso, R. A.; Rossi, R. A. Tetrahedron 1985,41, 4147. b) Review: Rossi, R. A.; Pierini, A. B.; Palaaios, S. M. in Advances in Free Radical Chemistry. Vol 1. JAI Press Inc.; Greenwich, Conn., 1990; p 193 - 252.
18. Detty, M. R.; Seidler, M. D. J. Org. Chem. 1981. 46. 1283.
19. The following were found unstable with TMSSePh: AgOTf, SnClj, TiCl 4, Q 1CI2, CuCN.
20. For the use of the Znl 2/TMSSePh system to open tetrahydrofurans, see: Miyoshi, N.; Hatayama, Y.; Ryu, I.; Kambe, N.; Murai, S.; Sonoda, N. Synthesis 1988. 175.
21. Miyoshi, N.; Ishii, H.; Kondo, K.; Murai, S.; Sonoda, N. Synthesis 1979, 300.
22. Ohno, M.; Ishizaki, K.; Eguchi, S. J. Org. Chem. 1988. 53. 1285.
23. a) Giese, B. Angew. Chem.. Int. Ed. Engl. 1983, 22 , 753. b) Giese, B.: Radicals in Organic Synthesis: Formation of Carbon- Carbon Bonds: Pergamon Press: Oxford, 1986.
24. Giese, B. Angew. Chem.. Int. Ed. Engl. 1984,22, 69. 74 25. a) Kuivila, H. G.; Menapace, L. W. J. Org. Chem. 1963,2&, 2165. b) Kuivila, H. Acc. Chem. Res. 1968 ,1, 299.
26. Corey, E. J.; Suggs, J. W. J. Org. Chem. 1 9 7 5 , 2 5 5 4 .
27. Stork, G.; Sher, P. M. J. Am. Chem. Soc. 1986, 108. 303.
28. Bergbreiter, D. E.; Blanton, J. R. J. Org. Chem. 1987, 52, 473.
29. For an example of methyl acrylate giving noticeable (10%) double adduct in a radical reaction, see: Adlington, R. M.; Baldwin, J. E.; Basak, A.; Kozyrod, R. P. J. Chem. Soc.. Chem. Comm. 1983. 944.
30. Hart, D. J.; Seely, F. L. J. Am. Chem. Soc. 1988. 110. 1631.
31. For additional examples of anhydrides as radical traps, see: Ramaiah, M. Tetrahedron 1989.43. 3541.
32. Fukunishi, K.; Tabushi, I. Synthesis 1988, 826.
33. Including varying conditions of the tin hydride method, Tin Pin methodology and catalytic tin hydride techniques.
34. The "Tin Pin" conditions returned only starting material.
35. N-Methylphthalimide has been successfully used as a radical trap, see reference 31.
36. Testaferri, L.; Tieeco, M.; Tingoli, M.; Fiorentino, M.; Troisi, L. Chem. Soc.. Chem. Comm. 1978, 93.
37. Lefort, O.; Fossey, J.; Gruselle, M.; Nedelic, J-Y.; Sorba, J. Tetrahedron 1985.41. 4237.
38. 450 Watt Hanovia Lamp.
39. Neither dodecahedrane or bromododecahedrane stain by any method. The "best" way to find them on the prep TLC plate is to hold the plate up to the light and observe a "shadow" in the expected region. Alternatively, there is a slight difference in 75 the texture of the surface of the plate where the compounds are located.
40. Irradiation of enones is known to produce head-to-head and head-to-tail dimers.
41. k~104 - 106 M^sec*1, see reference 25b.
42. Curran, D. P. Synthesis 1988, 489.
43. a) Keck, G. E.; Enholm, E. J.; Yates, J. B.; Wiley, M. R. Tetrahedron 1985, 4JL, 4079; b) Keck, G. E.; Yates, J. B. J. Am. Chem. Soc. 1982. 104. 5829.
44. Migita, T.; Nagai, K.; Kosugi, M. Bull. Chem. Soc. Jpn. 1983, 50. 2480.
45. The previous researcher used conditions similar to Keck's, possibly explaining his lack of success.
46. Several work-up procedures for the removal of tin halides are commonly used, the most popular work-up involves treatment of the reaction mixture with fluoride, see: a) Milstein, D.; Stille, J. K. J. Am. Chem. Soc. 1978,100. 3636; b) Leibner, J. E.; Jacobus, J. J. Org. Chem. 1979. 44. 449.
47. A method for the removal of hexaalkylditins and tin hydrides has been developed, see: Curran, D. P.; Chang, C-T. J. Org. Chem. 1989. 54. 3140.
48. a) McLafferty, F. W. Anal. Chem. 1956, 21 , 306. b) McLafferty, F. W. Anal. Chem. 1959, H , 82.
49. See photochemistry section of Chapter VII.
50. a) Sasaki, T.; Usuki, A.; Ohno, M. J. Ore. Chem. 1980,4£, 3559; b) Sasaki, T.; Usuki, A.; Ohno, M. Tetrahedron Lett. 1978, 19. 4925.
51. Fleming, I. in Comprehensive Organic Chemistry, ed. D. H. R. Barton and W. D. Ollis, Pergamon, Oxford, 1979, Vol. Ill, p. 541. 76
52. Magnus, P. Aldrichimica Acta 1980,12.(3) 43.
53. Due to the similarity in Rf values for many of the non-polar dodecahedryl derivatives, preparative TLC is an ineffective separation method.
54. a) Russell, G. A.; Tashtoush, H .; Ngoviwatchai, P. J. Am. Chem. Soc. 1984.106. 4622; b) See also reference 31.
55. Baldwin, J. E.; Kelly, D. R. ^ Chem. Soc.. Chem. Comm. 1985, 682.
56. Saihi, M. L.; Pereyre, M. p ^ ll. Soc. Chim.. Fr, 1977. 1251. CHAPTER IV DODECAHEDRYL ANION STUDY
A. Background The difficulty in forming bridgehead anions is well docum ented . 1 To address this problem, researchers have spent much time and effort developing alternate methods for their synthesis, resulting in some improved procedures for generating these species.
For example, 1-adamantyllithium 2 was not synthetically available until the work of Dubois appeared in 1983 after he extended the methodology he developed to prepare the corresponding Grignard reag en t.3 The potential utility of a well behaved dodecahedryl anion is obvious. The derivatives available from such a species would be limited only by the chemist's imagination. Unfortunately, the dodecahedryl anion falls into the bridgehead anion category and its formation had not been realized prior to this study .4 It was the goal of this researcher to generate the lithio derivative of dodecahedrane and evaluate its synthetic potential for further derivatization of the dodecahedryl framework. Prior to my involvement with this aspect of the project, previous researchers attempted the generation of the dodecahedryl anion by standard means. Unfortunately, all of their efforts met with 77 78 no success (Scheme 42). Various attempts to generate the corresponding Grignard reagent followed by trapping returned only dodecahedrane. Similarly, metalation attempts with sodium or lithium followed by trapping again produced only dodecahedrane. Transmetalation procedures also failed. Under a variety of conditions, trapping after treating the bromide with either n-BuLi or t-BuLi returned only dodecahedrane.
1. UP 2. trap
M°-U°. Na°.Mg°
38
1. R-Li 2. trap
R - n-Bu, f-Bu
38 3
Scheme 42. Previously Attempted Metalations
B. Radical Anion Methodology Upon reviewing these results, an alternate anion generation methodology was sought. Aromatic radical anions have long been 79 known to react with alkyl and aryl halides to generate the corresponding anions .5 Sodium naphthalide (NaN) has been studied very extensively, but Freeman has found lithium di-f-butylbiphenyl (75, LiDBB) to be a superior radical anion, allowing excellent yields of
alkyllithiums to be obtained .6 To test the applicability of this methodology and gain synthetic experience with this system, a model study utilizing 1-bromoadamantane (76) was initiated. p, p ’-Di-/e/7-butylbiphenyl (74, DBB) was prepared according
to the method of Curtis 7 by Friedel-Crafts reaction of /-butyl chloride with biphenyl (Scheme 43), prior to its becoming commercially available. Formation of the radical anion was achieved using the conditions suggested by Freeman 8 by adding finely cut lithium metal to a solution of DBB in THF at 0°C. After 3 to 5 hours, the dark blue solution was cooled to -78°C, 76 was added, and the mixture was stirred for 15 minutes prior to quenching with dimethyl sulfate. After work-up, the ratio of anion trapping to reduction was found to be 4:1 by GC (Scheme 44). If this ratio carried to the dodecahedryl system, we would be pleased. However, as we have found out in both the cation and radical studies, the adamantyl system does not always parallel the reactivity of the dodecahedryl system. To obtain some relative information between the two systems, we compared the corresponding reduction potentials. Dubois has shown that by utilizing reduction potentials of cage structure halogen derivatives, the yield of the corresponding organolithium compound may be predicted .9 He noted that when the reduction potentials of 80 two tertiary alkyl bromides are compared, the more negative the reduction potential, the higher the expected yield of the lithium anion. As the reduction potential becomes more positive, the more by-products result from a radical pathway (i.e. H-atom abstraction or coupling). He further concluded that the yield is inversely proportional to the stability of the corresponding cage-structure radical, supporting this with a correlation involving the thermolysis rates of the corresponding cage-structure peresters (commonly associated with the relative stability of the radical).
7 3 DBB, 7 4
.+ 7 4 Li THF, 0°C o o
LiDBB, 75
Scheme 43. Preparation of DBB and LiDBB
LiDBB, -78°C; M62OSO3
76 77 78
Scheme 44. LiDBB Model Study 81
When comparing the reduction potentials of the 1-adamantyl system with the dodecahedryl system, a profound difference was observed. The Ej /2 for electrochemical reduction of bromododecahedrane in anhydrous HMPA was found to be 0.39 V more positive than for 1-bromoadamantane (-2.158 and -2.547 V vs
SCE ) , 10 suggesting a lower degree of anion formation and higher radical stability in the dodecahedryl system relative to the 1- adamantyl system. Indeed, as predicted, when attempts were made to prepare the dodecahedryl anion from the bromide, even using LiDBB, followed by trapping, only reduction was observed (Scheme 45). It was now obvious that an alternate anion precursor was needed.
LiDBB, -789C; Mel
38 3
Scheme 45. Attempted Anion Generation with LiDBB and 38
1. Alternate Anion Precursor Screttas and co-workers have shown that sulfides may be cleaved by reagents such as lithium naphthalide (LiN) to form the 82 organolithium species .11 LiDBB has since been found to return higher yields and faster rates at lower temperatures than LiN .12 The anion precursor thus targeted was dodecahedryl phenylsulfide (79), expected to be readily obtained via similar methodology developed in the preparation of the dodecahedryl phenylselenide. Phenylthiotrimethylsilane (81, TMSSPh) was conveniently prepared from thiophenol, hexamethyldisilane and a catalytic amount of imidazole .13 The treatment of 38 in dichloromethane with excess 81 in the presence of anhydrous Znl 2 returned the desired phenylsulfide 79 in 93% yield after 10 hours at S0°C (Scheme 46). Again, purification was most conveniently achieved by trituration with diethyl ether. Sulfide 79 was very insoluble in ether, possibly accounting for the increased yield relative to the (phenylseleno)dodecahedrane.
[imidazole] SH + (TMSfeNH SSiM0s o o 80 81
SPh
/ 1 1 I \ PhSSiMe* Znl2 CHjCfe, 50°C *
38 79
Scheme 46. Preparation of TMSSPh and 79 83
2. Reaction Design At this point, the design and goal of the ensuing study should be addressed. First of all, the trap choice must be considered. Since the successful trapping of the dodecahedryl anion had never been achieved, it was deemed imperative that whatever the trap, the resulting product would be readily identified. Two possibilities immediately come to mind: methyl- and allyldodecahedrane, both known compounds. Two disadvantages of preparing the methyl derivative were realized. Methyldodecahedrane (82) and 3 have similar retention times on GC columns , 14 complicating the interpretation of results. Also, 82 is occasionally a by-product in certain batches of 3 , 15 making the results more ambiguous. Allyl dodecahedrane presented neither of these complications and was expected to be the product of anion trapping with allyl bromide. The immediate goal of the anion study was to identify promising reaction conditions capable of delivering the anion in synthetically useful quantities. If found, these avenues would then be explored, thus avoiding the unnecessary loss of valuable starting material. From this point of view, it was the relative amount of reduction to anion trapping product that most concerned us. Also, a quantitative relationship was not necessary until such routes were identified and the conditions optimized. What was necessary was the development of methodology allowing the dodecahedryl products not to be lost during work-up, and the lack of interference of any impurities with reaction analysis. The trap chosen must allow 84 convenient monitoring of the reaction by GC/MS and capillary GC. Helpfully, dodecahedryl derivatives tend to be non-volatile so impurities generally do not hamper the analysis. The scarcity of the starting material dictated that the smallest amount of compound be used that could still deliver useful information. In this way, the above goals could still be met without a great loss of material. Typically, 2 to 3 mg ("0.005 mmol) of the dodecahedryl compound was used. Obviously, great care and attention to detail was necessary on this scale. Trace moisture would almost assuredly deliver only dodecahedrane, resulting from quenching of the anion . 16 The reaction was designed so that excess radical anion was present to help avoid this possibility. With the target anion precursor in hand, study of the possible dodecahedryl anion formation was initiated. Standard conditions involved the addition of 79 in THF 17 to LiDBB solution at -78°C, stirring briefly and quenching with excess allyl bromide (immediate dispersion of color), followed by warming to room temperature. After dilution with benzene, passage through a short silica gel plug and solvent removal, a white solid remained (mostly DBBs). A minimum amount of diethyl ether permitted the non-dodecahedryl compounds to be removed allowing GC/MS analysis. Allyldodecahedrane was present, but relative to dodecahedrane, only to the extent of 2-3:8-7. No starting phenylsulfide was present. The major reaction component was dodecahedrane (Scheme 47). 85
.SPh
LiDBB, -78°C;
7 9 3 3 9
Scheme 47. Successful Generation and Trapping of the Dodecahedryl Anion from 7 9
Unfortunately, when the material removed by trituration was subjected to sublimation (house vacuum, 55°C overnight; conditions that would not remove any dodecahedryl species), a small quantity of dodecahedryl compounds was found. Although the relative I amounts of 39 and 3 indicated that neither compound was removed preferentially, the trituration of future reactions was avoided to ensure procurement of consistent relative product ratios. This problem was readily overcome by sublimation of the crude solid prior to GC/MS analysis. In this way, the relative amounts of 39 and 3 could be monitored. Although NaN has been shown to be less effective than LiDBB in its ability to generate anions from alkyl bromides, the corresponding reaction was carried out on (phenylthio)dodecahedrane. Treatment of excess NaN with a solution of 79 in THF at -78°C returned only dodecahedrane and starting material. No allyldodecahedrane was observed (Scheme 48). 79 79 3
Scheme 48. Attempted Anion Generation with NaN
The inability of (phenylthio)dodecahedrane to deliver synthetically useful ratios of anion to reduction products was disappointing and caused its dismissal as a useful anion precursor. However, for the first time evidence of the dodecahedryl anion had been obtained.
C. Transmetalation Methodology 1. Reaction Design As mentioned earlier, previous attempts to generate the anion by transmetalation met with no success. I wished to reexamine this methodology with the hope of gathering more information about the dodecahedryl anion. Contrary to previous results, treatment of bromododecahedrane in THF at -78°C with excess r-BuLi produces lithiododecahedrane (83, Scheme 49)! Thus trapping with allyl bromide delivered allyldodecahedrane in approximately a 3:7 ratio relative to dodecahedrane. Conveniently, unlike the radical anion 87 methodology, extraction, filtration, and solvent removal afforded a product mixture consisting of almost solely dodecahedryl derivatives. In general, the dodecahedryl compounds were obtained in >90% combined yield, indicating the efficiency of recovery.
/ -BuLi, -78°C; + THF
38 3 8 3
Scheme 49. Successful Transmetalation of 38
The ability to generate the anion by a transmetalation reaction offered an opportunity to learn more about this reaction and the lithiododecahedryl anion by varying the reaction conditions. Consumption of the starting bromide was very rapid. The passage of only 10 seconds between the addition of f-BuLi and quenching with allyl bromide at -78° C was adequate to guarantee that no bromododecahedrane remained. To ascertain the stability of lithiododecahedrane, the quench was postponed for 2 hours (*78°C). Even then, the trapping product was obtained in the typical 3:7 ratio relative to dodecahedrane. Finally, the temperature of the reaction mixture was varied from -95 to -42°C with similar product ratios. Since all variations led to the same anion trapping to reduction ratio, 88 convenience led to the typical conditions consisting of addition of excess f-BuLi to 38 in THF at -78°C. Following a 20 to 30 second delay, the trap was added, stirred an additional 15 minutes and allowed to warm to room temperature prior to the usual work-up. As the reaction was studied, a minor component (~5%) was almost always observed. From the GC/MS , 18 and exact mass 19 it was determined to be (f-butyl)dodecahedrane (84). Unfortunately, due to the small amount produced and the inability to separate it from the other reaction components, it could not be examined further. Supporting this assignment was the production of a similar by product assigned as (s-butyl)dodecahedrane (85) (again by GC/MS )20 when 5-BuLi was utilized to generate the anion. When s-BuLi was used to generate the anion, no anion trapping product was observed, instead only dodecahedrane and 85 were identified (no starting bromide however). When n-BuLi was used in an analogous manner, only dodecahedrane was formed ,21 no other dodecahedryl products were observed (some starting bromide remained however). Formation of the minor components 84 and 85 is rationalized in terms of a competing single electron transfer (SET) pathway available to the dodecahedryl bromide .22 Coupling of the radicals would lead to the observed products (Scheme 50). 89
R-Li + R* (SET) -LiBr
R« t -Bu R « s -Bu
3 8 86
8 4 8 5
Scheme 50. Competing Single Electron Transfer Products During Transmetalation
One way to optimize a reaction is to develop conditions that disfavor competing side reactions. With this in mind, we explored some possible paths to dodecahedrane. An alternate pathway available to 8 6 could account for the formation of some of the dodecahedrane in the anion generation attempts. If 86 abstracted a hydrogen from the solvent, dodecahedrane would result. To test this possibility, THF-dg was used as the solvent. Upon analysis, no deuterium incorporation was observed, eliminating this pathway as a possible cause of dodecahedrane production (Scheme 51). 90
f-BuLi, -78°C, THF-de; -X- ,Br
3 8 86
(Only 3 and 38)
Scheme 51. Lack of Deuterium Incorporation in THF-dg During Transmetalation
Scheme 52 outlines some of the possibilities that may occur under these conditions to account for the reaction products. When excess f-BuLi is added to bromododecahedrane, single electron transfer or transmetalation may occur. Both pathways must be operating since products arising from both pathways have been observed. Decomposition of radical pair 86 by solvent-H abstraction has been eliminated 23 as a source of dodecahedrane. The small fraction of coupling observed also suggests dodecahedrane may arise from an anionic pathway. 8 6
f-BuLi +
Transmetalation
f-BuBr +
f-BuLi
Trap X'X
Scheme 52. Options Available During Transmetalation 92
The transmetalation reaction of bromododecahedrane and t-
BuLi is expected to be reversible ,24 but the excess /-BuLi should react with the formed f-BuBr to produce isobutane and isobutylene, forcing the reaction in the desired direction. The dodecahedryl anion has a few options once it is generated. Since 83 may be in close proximity to the f-BuBr as it is formed, 83 may compete with the excess f-BuLi and react with /-BuBr, returning dodecahedrane .25*26 Alternately, any moisture present would similarly return 3. Once these possibilities have been exhausted, the relatively stable anion could be trapped in the desired fashion when the appropriate reagent is added. The consistency with which approximately a 3:7 ratio of anion to reduction product is seen under a variety of conditions, when considering the sensitivity of the scale to moisture, suggests this ratio to be a reactivity characteristic and not due to extraneous moisture. For any moisture would surely lead to widely varied ratios. The exact determination of the pathway to dodecahedrane was beyond the scope of this investigation, but one of the more suitable explanations of this ratio is the competing reaction of lithiododecahedane with the acidic proton of the f-BuBr generated (although no experiments were performed to investigate this possibility). The lack of sensitivity of the product ratio to the varying conditions dictated the inability to obtain synthetically useful anion formation via transmetalation methodology. 93
Even though a consistent 3:7 anion trapping vs reduction was unacceptable from a synthetic point of view, the desire for an isolatable and purifable derivative was great. Unfortunately, dodecahedrane and allyldodecahedrane are inseparable by preparative TLC, the most widely used purification method available on this scale. More time consuming separation techniques were unwarranted for this known compound. What was needed was an anion trap that yielded a product with a difference in polarity relative to dodecahedrane (and t-butyldodecahedrane), preferably was UV active to facilitate detection and added a significant amount of mass to the dodecahedryl system (to make up for the low yield). Unfortunately, the most useful instrument when analyzing the reaction mixture (the GC/MS) did not respond well to the last preference. Prohibitively long retention times resulted (if the compound emerged at all) as the mass of the derivative increased. And, in general, the ability of a compound to be UV active often results in an increased molecular weight (i.e. aromatic systems). Thus the ability to monitor a reaction meant the product could probably not be readily isolated in pure form and when isolation would be facilitated, the determination of the product's existence was hindered. On top of these problems was an expected yield of no greater than 30%. This would require larger amounts of starting bromide to be dedicated to the isolation of any anion trapping products than was desired. For these reasons, efforts to obtain anion trapping products was abandoned.?? 94
2. Single Electron Transfer Exploration Although only a small amount of SET pathway was chosen by the dodecahedryl bromide under the conditions explored, it was realized that if the system could be coaxed into preferring this pathway, a useful derivatization technique would be available. The cross-coupling of tertiary alkyl bromides with organometals is frequently difficult due to the inability of an Sn 2 mechanism (two-electron-transfer process) to play a significant role. The result is generally the dominance of side reactions such as elimination . 28 In cases where elimination is not a viable alternative ,29 a radical recombination mechanism (single-electron- transfer process) is often involved (frequently more than one mechanism may operate concurrently30). Eguchi developed a successful cross-coupling of 1-halo- adamantanes with a variety of Grignard reagents to yield the corresponding bridgehead substituted products .31 He found that when a non-Lewis basic solvent such as dichloromethane was used, moderate yields of the cross-coupled product resulted with some reduction also taking place (Scheme S3).
Unlike literature precedent involving alkynyl allanes ,32*3 3 Eguchi found that the mode of bond formation was not arising from an ionic pathway, but was actually due to participation of a radical- type intermediate. This encouraging information initiated an investigation of the utility of this methodology on the dodecahedryl system. Unfortunately, when methyl and vinyl Grignard reagents 95
RMgX CH2CI2
X - Cl, Br, I major minor R -1 °, 2°, 3° alkyl; allyl, aryl.
Scheme 53. Cross-Coupling Conditions of Eguchi
.Brl
R-MgX R* + MgX • MgXBr
3 8 5 6
MgXBr S-H
3
Scheme 54. Possible Mechanism for the Generation of Coupled Products Using Eguchi's Conditions 96 were reacted with bromododecahedrane in CH 2 CI2 solution, the reduction product and starting bromide were the major components of the mixture. Only trace amounts of cross-coupling products could be identified (Scheme 54)34 Under these conditions, reduction to dodecahedrane, like the pathway assumed for formation of adamantane ,31 may be considered to be the result of a radical process .35 Alternatively, had transmetalation occurred, dodecahedrane would have also been returned upon work-up. No experiments were performed to address this possibility since, although transmetalation with Grignard reagents is possible,24 the tendency of the more electropositive metals (such as Na, K and Mg) to undergo side reactions, makes transmetalation a less desirable alternative. Whatever the source of dodecahedrane, it was obvious that this coupling methodology would not deliver useful quantities of the desired coupling products and was thus abandoned .36
H. Conclusion This study showed that the generation of dodecahedryllithium could be accomplished, but not in synthetically useful amounts. This anion was found to be stable (-78°C for at least 2 hours) and readily add to allyl bromide to produce allyldodecahedrane. Two methods for the preparation of this species were developed. Whether formed from (phenylthio)dodecahedrane and the radical anion LiDBB or from 97 bromododecahedrane by transmetalation with /-BuLi, approximately a 3:7 ratio of anion trapping to reduction was obtained. In the later case, a small degree of SET pathway was taken by the bromide, evident in the identification of cross-coupled products. SET methodology was briefly examined as a possible route to additional derivatives, but abandoned when competing reduction was found to predominate. 98
References
1. Fort, R. C. Jr.; Schleyer, P. v. R. Adv. Alicvclic Chem. 1966,1, 283.
2. Molle, G.; Bauer, P.; Dubois J. E. J. Org. Chem. 1983, 18., 2975.
3. Molle, G.; Bauer, P.; Dubois J. E. J. Org. Chem. 1982,42, 4120.
4. Weber, J. and Kobayashi, T. unpublished results.
5. For a review of radical anions and dianions, see: Holy, N. L. Chem. Rev. 1974, 21, 243.
6 . Freeman, P. K. Hutchinson, L. L. Tetrahedron Lett. 1976, 22 , 1849.
7. Curtis, M. D.; Allred, A. L. J. Am. Chem. Soc. 1965,82, 2554.
8. a) Freeman, P. K. Hutchinson, L. L. J. Org. Chem. 1983, 4&, 4 705; b) Freeman, P. K. Hutchinson, L. L. J. Org. Chem. 1980, 4£, 1924.
9. Dubois, J.-E.; Bauer, P.; Kaddani.B. Tetrahedron Lett. 1985. 26. 57.
10. The reduction potentials were determined by Amy E. Clough in HMPA vs. AgC104 and converted to SCE.
11. Screttas, C. S.; Micha-Screllas, M. J. Org. Chem. 1978.43. 1064.
12. Rucker, C. Tetrahedron Lett. 1984, 25.. 4349.
13. Glass, R. G. J. Organomet. Chem. 1973, £1, 83.
14. Methyldodecahedrane is actually more volatile than dodecahedrane. 99
15. This occurs if 31 remains after Birch reduction. The corresponding phenyl ether is carried through the sequence until it is cleaved in the final dehydrogenation step.
16. Glassware was oven dried, cooled in a desiccator, assembled, and flame-dried under a vigorous stream of nitrogen prior to use. Any additions to the reaction vessel were through a septa or past vigorously exiting nitrogen.
17. Sulfide 79 was difficult to dissolve in THF, making the exact amount added often unknown. This does not interfere with the goal of this investigation.
18. GC/MS Parent m/z =316; P-Me m/z = 301; 260, 259, 258.
19. Exact Mass DDH-C(CH3)3: calc 316.2190; obsd 316.2164 (the sample had a DDH-H impurity present that did not interfere with this result).
20. GC/MS Parent m/z =316; P-Et m/z - 287; 260, 259, 258.
21. Non-dodecahedryl impurities complicated the reaction analysis.
22. For evidence of SET of alkyl bromides with n-BuLi, see: Bailey, W. F.; Ovaska, T. V.; Leipert, T. K. Tetrahedron Lett. 1989, 2Q. 390 1 .
23. There are other hydrogen atom sources present in the reaction, but the solvent is the most abundant.
24. Negishi, E. Organometallics in Organic Synthesis: Wiley: New York, 1980; p 38.
25. The formation of hydrocarbon formally derived from the reduction of halide substrate appears to be a general occurrence in the lithium-halogen interchange reactions that involve /-BuBr, see: Bailey, W. F.; Nurmi, T. T.; Patricia, J. J.; Wang, W. J. Am. Chem. Soc. 1987,1Q2, 2442. 100
26. The reaction of f-BuLi with f-butyl halide generated from the halogen-metal exchange is well known, see: Corey, E. J.; Beames, D. J. J. Am. Chem. Soc. 1972. 94. 7210.
27. Some traps fitting at least some of the above criteria were attempted but no desired isolatable or assignable products were obtained. In all cases, however, dodecahedrane was formed. Attempted traps included: TMS-C1; PhCH 2-Br; PhOCH2- Cl; PhSe-Cl and PhCHO. As when trapping with allyl bromide, the f-butyl dodecahedrane byproduct was usually observed.
28. See examples for Grignard case in: Kharasch, M. S.; Reinmuth, 0. Grignard Reactions of Nonmetallic Substances: Prentice-Hall: New York, 1954; pp 1067 - 1125.
29. Elimination would lead to a violation of Bredt's rule in the adamantyl case and lead to the highly pyramidalized dodecahedrene in the dodecahedrane case.
30. Muraoka, K.; Najima, M.; Kusabayashi, S. J. Chem. Soc.. Perkin Iia.nL. 21986, 761.
31. Ohno, M.; Shimizu, K.; Ishizaki, K.; Sasaki, T.; Eguchi, S. J. Org. £Jh££L. 1988, £2., 729.
32. Negishi, E.; Baba, S. J. Am. Chem. Soc. 1975,22, 7386.
33. Trialkynylaluminium reagents have been reported (See Reference 32) to undergo a substitution reaction at the adamantane bridgehead position via a carbocation intermediate. When an attempt was made to couple trihexynylalane with bromododecahedrane (utilizing suggested conditions), no substitution was observed.
34. Methyldodecahedrane was readily identified (GC/MS comparison) with an authentic sample. Vinyldodecahedrane was the assumed trace component, assigned by its retention time and MS fragmentation pattern.
35. Bauer has demonstrated that the Barbier synthesis of alcohols does not necessarily involve intermediate formation of an 101
organometallic compound. In certain cases, there was observed a radical pathway in which the anion radical (R-X* ) resulting from the attack by a halogenated derivative on lithium is directly trapped by the ketone or ketyl radical on the metal surface before the organometallic compound forms. See: Bauer, P.; Molle, B. J. Am. Chem. Soc. 1982. 104. 3481. Because any dodecahedryl alcohols formed by this reaction were expected to be unobservable by methodology readily available and due to the negative results of the above mentioned coupling attempts, the cost of material needed to pursue this route was considered prohibitive..
36. The possibility of forming a dodecahedryl trimethyltin species was attempted using Me 3 SnLi, methodology known to proceed via a SET pathway in other bridgehead systems, but the large amount of reduction again forced abandonment of this methodology. See: Adcock, W.; Iyer, V. S.; Kitching, W.; Young, G. J. J. Org. Chem. 1985, 3706. CHAPTER V EXTENSIONS TO DODECAHEDRYL CATION CHEMISTRY
A. Rack ground When the cation chemistry of dodecahedrane was originally investigated , 1 the 1-adamantyl cation (87) was used as a model system. However, time and time again it was noted that reactions successful with 1 -bromoadamantane failed with bromododecahedrane. There are marked differences between the dodecahedryl cation
( 8 8 ) and 87.2 While neither cation is expected to achieve p la n a rity ,2*3 8 8 lacks stabilizing factors available to 87. Since no carbon-carbon or carbon-hydrogen bond can be located trans to the vacant, strongly directed orbital of 8 8, no normal stabilizing effect of alkyl substituents is possible (unlike with 87). Solvolytic rate data are not available for 8 8, but similar factors are attributed to a 1 0 19- fold solvolytic rate retardation of the tosylate precursors of 8 9 relative to 87.4 The inability of bromododecahedrane to be hydrolyzed under conditions used to hydrolyze 1-bromoadamantane 5 expose this reactivity difference .6 In fact, it is the exception rather than the rule that reactions will be successful on both systems .7 Despite this
102 103 reactivity difference, the catio n ic chem istry has provided the m ost accessible route to substituted dodecahedranes. Although the transient generation and trapping of the dodecahedryl cation is quite useful, strong and highly polarizing Lewis acids were always needed for successful reaction. Unfortunately, many electropfriles were intolerant of these conditions and thus limited the variety of derivatives available using this methodology.
87 88 89
JL Utilization of Zinr If^ide Throughout the previous two chapters, the use of anhydrous zinc iodide has resulted in the production of dodecahedryl derivatives not previously available. Whether to a llo w electrophiles unstable to the more harsh Lewis acids to survive during the formation of 8 8 or simply to avoid the often encountered 104 transhalogenation, Znl 2 has extended the range of electrophiles suitable for 8 8 to trap. These newly obtained derivatives have directly resulted in the exploration of additional areas of dodecahedrane chemistry [viz. dodecahedrene (Chapter 6 ) and photochemistry (Chapter 7)].
SePh ,SPh
4 7 7 9 3 9
Although the formation of 8 8 is slower than with the more polarizing Lewis acids, the rate at which it is formed remains on a "convenient" time scale, generally allowing the product to be obtained within three days. This is a very small price to pay for the
extension to carbocation methodology which Znl 2 has provided.
All of the zinc salts utilized (ZnCl 2, ZnBr 2 , Znl 2) were shown to
be capable of generating 8 8 , but Znl 2 was far superior to the others.
The enhancement in reactivity is attributed to the ability of Znl 2 to produce a higher concentration of the intermediate cation, which was then trapped by the electrophile present. Two possibilities that exist
for the differences are the increased solubility of Znl 2 in
dichloromethane relative to ZnCl 2 or ZnBr 2 and/or the generation of a 105 dodecahedryl intermediate formed from a transhalogenation reaction. Both of these possibilities will be addressed. In general, due to the decreased ionic character of the metal- iodine bond, iodo salts are more soluble in organic solvents relative to the corresponding chloro- and bromo- salts (i.e. use of this is made in the Finkelstein reaction to prepare alkyl iodides: the insolubility of the NaCl resulting from the halogen exchange forces the equilibrium in the desired direction). A higher concentration of Zn+ + in solution would be expected to increase the concentration of 8 8 present and thus lead to an increased rate of electrophilic trapping. When one considers the possible intermediacy of iodododecahedrane (46) and its effect on the rate, things become more complicated. First of all, since 46 could not be isolated (or even positively identified), we can at most suggest it is formed in situ. This seems reasonable since transhalogenation has often been observed when utilizing Lewis acids containing chlorine or fluorine. Recall the appearance of the unknown, unstable compound that resulted from the reaction of AII 3 with 38 and returned only dodecahedryl alcohol upon aqueous work-up and the emergence of the same type of peak if Znl 2 was stirred with 38 in CH 2 CI2 (no moisture present, prior to work-up) with a GC/MS fragment of m/z = 259. Couple this with the unexpected reduction of bromododecahedrane to dodecahedrane with Znl 2 and evidence for this transhalogenation begins to mount. 106
As noted earlier, the presence of 46 may have no effect on the increased rate. All that can be reasonably stated is that 46 should be converted to cation 8 8 faster than 38 (k.|>k_Br)- But since there are additional unknown rates involved, no definite effect on the overall rate of electrophilic trapping may be stated (Scheme 55).
k-ar k,Br
38 88 4 6
‘Trap
,Trap
Scheme 55. Possible Effect of 46 on Rate of Electrophilic Trapping
Whatever the exact reasons for this rate enhancement, it was successfully utilized to derivatize dodecahedrane. To further extend the usefulness of Znl 2 , the previously reported problem of dodecahedryl bromide hydrolysis was reexamined. 107
C. Dodecahedryl Alcohol: ___ Improved -Synthesis The inability to replace the bromine atom of 38 with a hydroxyl group to give 53 under a wide variety of conditions was demonstrated by previous researchers . 1 At the time, the preferred method for obtaining the desired alcohol was via the direct functionalization of the parent system (Scheme 56). Treatment of 3 with lead tetraacetate and trifluoroacetic acid furnished the trifluoroacetate 52. Direct hydrolysis with aqueous sodium hydroxide gave the alcohol 53 in 76% yield. This radical reaction (3 to 52) was found to proceed erratically, often returning products assumed to arise from polyfunctionalization.
OCOCF; OH
10% NaOH
3 5 2 5 3
Scheme 56. Original Preparation of 53
An opportunity to avoid this sequence was noted. By making use of the newly developed Znl 2 methodology, a direct one step preparation of 53 was envisioned. By modification of the conditions noted earlier during the attempted preparation of dodecahedryl iodide, a highly efficient synthesis of 53 was developed. 108
When moist dichloromethane 8*9 containing 38 and Zn ^ 10*11 was allowed to stir in the dark, 53 was the only product. After 2 days, only a trace of the bromide was present. It could be removed by trituration (ether) or more conveniently by resubmission to reaction conditions . 12 In this way, near quantitative yields of the alcohol can be produced (Scheme 57).
.OH
Znl2, ChfcClg trace H20
3 8 5 3
Scheme 57. Improved Synthesis of Dodecahedryl Alcohol
D. 1.4-Di(dodecahedrvlIbenzene The ability to include in one molecule, two dodecahedryl nuclei has been envisioned since the molecule first became available. Unfortunately standard coupling procedures did not accommodate the characteristic reactivity patterns of 38. Of particular interest was the parent didodecahedrane 90, affectionately called "dumbellane." Although seen arising from the coupling of two dodecahedryl radicals ,13 the requirement of generating the radicals in close proximity to each other necessitated the intermediacy of a 109 precursor already containing two dodecahedryl moities. A promising precursor was 93 envisioned obtainable from 45 (Scheme 58) by known methodology . 14 Solid state photolysis of 93 is expected to yield 90 but the ability to arrive at the desired intermediates has not been realized . 15
92 110
NH.
4 5
9 3
Scheme 58. Possible Precursor to Dumbellane
The homologated versions of type 91 were hopefed to arise via the dodecahedryl anion and suitably substituted dodecahedranes but the lack of useful anion chemistry shut down these possibilities. However, the known ease of cationic substitution made the acquisition of 92 a likely canidate. The Friedel-Crafts reaction of bromododecahedrane with phenyldodecahedrane (50) was expected to return 92. The original preparation of 50 from 38 utilizes benzene as a solvent. Consequently the C 2o H i 9+ intermediate is guaranteed exposure to a high concentration of the aromatic coreactant. The result is a rather brief reaction time (3 h at 32°C). A comparable concentration gradient is not available when considering a related electrophilic route to 92. Nonetheless, admixture of equimolar amounts of 38 and 50 in dry dichloromethane followed by addition of one crystal of AICI 3 resulted in the gradual precipitation of an off-white solid during 5 days at room temperature (Scheme 59). Following the addition of water and trituration with CH 2CI2 and ether, a 51% yield of the highly insoluble 92 results ((M+) m/z calc. 594.3224, obsd. 594.3286). The insolubility of this species made it particularly difficult to work with. However, because it was insoluble in solvents the starting materials were soluble in, purification was readily achieved. Additionally, these impurities were easily monitored by GC/MS and thus ensured the sample's purity. The slight amount of 92 soluble in CDCI 3 necessitated long-term acquisition to obtain a 13C NMR spectrum.
This C4 6 H 42 hydrocarbon exhibited six of the anticipated eight signals, the two singly unique dodecahedryl carbons not making their appearance under these circumstances. The two benzenoid carbon absorptions rigorously corroborate the para substitution pattern (as does the singlet in the *H NMR ).16 112
Scheme 59. Synthesis of l,4-Di(dodecahedryl)benzene
E. Conclusion The use of cation chemistry has proven to be the most valuable route to the derivatization of the dodecahedryl nucleus. By expanding the electrophiles available to trap the transient carbocation, I was able to open avenues to new aspects of dodecahedrane chemistry. Additionally, the reactivity of the dodecahedryl cation allowed synthesis of the first tethered dodecahedrane to be realized. l,4-Di(dodecahedryl)benzene belongs to a special class of compounds containing two dodecahedryl nuclei per molecule and was found to exhibit high insolubility and be highly 113 symmetrical (8 symmetrically non-equivalent carbons expected for this 46-carbon compound). 114
References
1. Weber, J. Ph. D. Dissertation, The Ohio State University, 1987.
2. Paquette, L. A. Chem. Rev. 1989. 89. 1051.
3. Fort, R. C. Jr.; Schleyer, P. von R. Adv. Alicvlic Chem. 1966,1, 283.
4. Bingham, R. C.; Schleyer, P. von R. J. Am. Chem. Soc. 1971. 93. 3189.
5. Sasaki, T.; Eguchi, S.; Kateda, T. J. Org. Chem. 1974. 39. 1239.
6 . Attempted hydrolyses included: HCONH 2, A; AgOTf, NH3; Me2AlNH2; HN(TMS)2 ,A.
7. The most helpful similarity in adamantane and dodecahedrane chemistry is the ability of both systems to undergo mono- bromination selectively.
8. One drop of water is shaken with 5 ml CH 2C12 and allowed to settle. The organic layer is used as the reaction solvent.
9. The reaction does not work in THF; only starting material rem ains.
10. Solid Znl 2 must remain in the reaction.
11. When ZnB r 2 is substituted for Znl2, only a trace of the alcohol (>5%) is formed after 3 days.
12. Allowing the reaction to stir for longer periods of time ensures the absence of starting bromide and simplifies the product purification.
13. The inability to generate the lithio derivative in adequate quantities would not allow the known lithium cuprate coupling methodology to be feasible, see: Schafer, J.; Polbom, K.; Szeimies, G. Chem Ber. 1988. 121. 2263. 115
14. a) Prochazka, M.; Ryba, O.; Lim, D. Coll. Czech. Chem. Comm. 1968,22, 3387; b) Engel, P. S.; Lee, W.-K.; Marschke, G. E.; Shine, H. J. J. Org. Chem. 1987, £2., 2813; c) Adam, W.; Mazenod, F.; Nishizawa, Y.; Engel, P. S.; Baughman, S. A.; Chae, W.-K.; Horsey, D. W.; Quast, H.; Seiferling, B. J. Am. Chem. Soc. 1983. 105. 6141: d) Ohme, v. R.; Preuschhof, H. Liebigs Ann. Chem. 1968. 713. 74; e) Engel, P. S.; Chae, W.-K.; Baughman, S. A.; Marschke, G. E.; Lewis, E. S.; Timberlake, J. W.; Luedtke, A. E. J. Am. Chem. Soc. 1983. 105. 5030.
15. Skerlj, R., personal communication.
16. Paquette, L. A.; Lagerwall, D. R.; King, J. L.; Niwayama, S.; Skerlj, R. Tetrahedron Lett, in press. CHAPTER VI DODECAHEDRENE
A. Introduction Dodecahedrene (94), the unsaturated analog of the parent dodecahedrane, holds a great deal of interest as a highly strained olefin1. It should, when examined crystallographically, exhibit a high degree of pyramidalization about the olefinic carbon atoms ,2 probably with near tetrahedral geometry at these centers. The rigidly locked skeleton of cage systems such as adamantane and dodecahedrane causes a high degree of distortion at any olefinic bonds present, thus creating problems in their synthesis. This type of challenge has always appealed to the scientific community as evident by the intense interest in adamantene (95). This severely distorted anti-Bredt olefin has recently been prepared, matrix isolated, examined spectroscopically ,3 and shown to possess a double bond rather than existing as a 1,2-biradical. Previously reported force-field calculations 4 suggest that adamantene is highly pyramidalized and twisted at the olefinic carbons, although it is not pyramidalized to as high a degree as would be required in the dodecahedrene system. We have successfully measured the "double bond strain energy" of dodecahedrene 5 and although we could not isolate it, dodecahedrene was prepared and trapped as the epoxide. 116 117 S3
9 4 9 5
B. Measurement of the Double-Bond Strain. Energy of Dodecahedrene bv Thermochemical Bracketing of _GasJhase Ion-Molecule Reactions^ 1. Background Electron ionization of various dodecahedrane derivatives produces a positive ion at m/z 258, corresponding to a C 2oHig*+, a species isobaric with the radical cation of dodecahedrene. In particular, compound 52 and phenyldodecahedrane show m/z 258 as the dominant peak in their mass spectra under high-energy electron impact. However, C20H 1 g*+ may represent a ring-opened structure because of the high degree of strain in the system. The possibility of forming the neutral dodecahedrene from 52 intensified our interest in this system. Our indirect evidence for the gas-phase production of dodecahedrene is based on the observation of a ^-elim ination reaction when strong gas-phase anionic bases react with trifluoroacetoxydodecahedrane 52. The ionic product of this reaction is trifluoroacetate ion at m/z 113. In order to infer the structure of a neutral molecule that cannot be directly observed, however, the 118 appearance of an expected ionic product is not a sufficiently strong argument. Consideration of the alternate pathways is necessary to support this postulated reaction mechanism. The arguments disfavoring these alternate pathways are discussed elsewhere 5 and only the accepted pathway will be discussed here.
5 2 5 0
2. T heory A brief theoretical view of the experiment is warranted. The gas-phase ion-molecule reactions were carried out in the trapped-ion cell of a Nicolet FTMS-2000 mass spectrometer7 in collaboration with the Marshall group at The Ohio State University. Gas-phase synthesis of olefins as neutral molecules may be accomplished by a |3- elimination reaction. However, since only exothermic processes are generally observable in ion cyclotron resonance owing to the near- thermal conditions and limited ion-trapping period, incorporation of a strained double bond into a molecule requires a very strong gas- phase base and a very weakly basic anionic leaving group. 119
If the extra amount of enthalpy required to introduce a double bond into a strained system is greater than the heat of formation for a corresponding unstrained system, then the reaction will be endothermic and not observed. Thus, by varying the strength of the gas-phase bases and knowing the corresponding heats of formation, we can bracket where the reaction is observed and where no reaction occurs. The difference in the resulting heats of formation can then be related to the amount of extra energy needed to incorporate the olefin in the strained system. Scheme 60 pictorially describes this process with 52.
+ "0CCFa + CH3 ° H
FTICRMS
No Reaction
5 2
Scheme 60. Generation of Dodecahedrene in the MS
3. R esults Under the ion trapping conditions in the FT/ICR mass spectrometer, elimination of trifluoroacetate from 52 is observed with methoxide but not with ethoxide. We therefore begin by 120 comparing the thermochemistry of the observed elimination of 96 to that of the open-chain analog shown in Scheme 61. This reaction also produces a tetrasubstituted alkene, in this case a planar unstrained
one. From ionic and neutral heats of formation ,8 the reaction is calculated to be 173 kJ/mol exothermic for RO* = methoxide and 159 kJ/mol exothermic for RO* = ethoxide. Because methoxide induces elimination in trifluoroacetoxydodecahedrane, whereas ethoxide does not, the amount of extra energy to produce the strained olefin, dodecahedrene, compared to the unstrained olefin 2,3-dimethyl-2- butene (97), must therefore lie between the two values for the formation of 97, namely, at 166±>12 kJ/mol. The advantage of this comparative method in obtaining relative heats of reaction is that no absolute thermochemistry for the dodecahedryl derivative is needed. Roughly speaking, when the heat of hydrogenation of 97, 100 kJ/mol, is added to the relative enthalpy of elimination, a net heat of hydrogenation of dodecahedrene of 266±-l7 kJ/mol 9 is obtained. In the elimination reactions designed to produce adamantene (model study for this methodology), the leaving group is the same as for the dodecahedrene-forming reactions, but the bases necessary to promote reaction were weaker. Similar to the unstrained analog 9 7 in the dodecahedrene case, comparison is made with the unstrained olefin 99 in the adamantene elimination (Scheme 62). AHrx for the formation of 99 is calculated 10 to be -129 kJ/mol with PhCH2 0 ' as the base and -118 kJ/mol with (/-(^H ^C H O *. If we assume that adamantene is produced in the elimination reaction, these two bases 121 would bracket the thermochemical threshold for the elimination reaction producing adamantene (Scheme 63). The amount of extra energy, relative to 99, necessary to produce this strained olefin is 124±6 kJ/mol. Again the heat of hydrogenation of adamantene can be obtained as the sum of this value and the heat of hydrogenation of 99. In a manner similar to 94, we obtain an apparent AHhytj of
228 kJ/mol for adamantene.
+ - II + BH OCCF3 0 1 9 7 OCCFa B or
> J <
9 6 No Reaction
With 5 2 Elimination with OMe Comparison of its ionic and 173 kJ/mol neutral heats of formation No Reaction with OEt (2,3KJimethyl-2-butene model) 159 kJ/mol
166 ± 12 kJ/mol
Adding the heat of hydrogenation of 2,3-dimethyl-2-butene of 100 kJ/mol, 266 kJ/mol is obtained for the heat of formation of dodecahedrene
Scheme 61. Unstrained Analog of Dodecahedrene Scheme 62. Unstrained Adamantene Analog
PhCHjO Elimination Observed
FTICRMS
100 (/-Bu )2CHO No Reaction
Elimination product was 40 kJ/mol more stable than dodecahedrene This was contrary to adamantene's predicted stability.
Alternate elimination mechanism available
Scheme 63. Elimination in the Adamantane System
Although published calculations 4 of the heats of hydrogenation of adamantene and dodecahedrene predict that adamantene should be the more strained olefin, our results imply that 40 kJ/mol less energy should be required to incorporate a double bond into the adamantane skeleton. This inconsistency forces consideration of a more complex possibility, namely a 1,4-elimination reaction (Scheme 64). In an unstrained system, a 1,4-elimination would be unlikely; 123
the energy gained from the formation of a second double bond would be offset by the breaking of a carbon-carbon single bond. For example, the heat of hydrogenation of two ethene molecules to produce butane is approximately -230 kJ/mol . 11 In the case of the adamantene system, however, this cost is offset by the relief of
strain upon opening of the ring. AMI calculations 12 indicate that diene 101 produced by the 1,4-elimination should be "163 kJ/mol lower in energy than adamantene. In contrast, in the case of dodecahedrene, there is no energy advantage in a 1,4-elimination.
The diene produced is calculated 12 to be 22S kJ/mol higher in energy relative to dodecahedrene, presumably because the rigid system prevents much relief of strain in the diene (unlike the previous case). Also, the 1,4-elimination in this system would necessarily proceed via a grossly misaligned orbital construct. Therefore, we can state
with reasonable certainty that dodecahedrene is produced in a 1,2- elimination whereas adamantene is not. The heat of hydrogenation of adamantene is better estimated as the sum of the threshold value (-116 kJ/mol), the heat of hydrogenation of 99, and the calculated -163 kJ/mol difference
between the diene 1 0 1 and adamantene energies, for AHhyd=-391 kJ/m ol. Recently reported calculated heats of hydrogenation (by force-
field molecular mechanics methods 13*14 and by molecular orbital calculations at the AMI level12) have been compared to the experimentally determined values obtained by the bracketing 124 method described for dodecahedrene and adamantene and are listed in Table 2.
OCCF3 1,2-elimination —X—
1 0 0 9 5
OCCF3
1 ^-elimination
H 100 101 163 kJ/mol lower in energy relative to adamantene (AM1)
Scheme 64. 1,2- and 1,4 -Elimination products of 100 125
TABLE 2
Heats of Hydrogenation of 94 and 95 by Several Methods (kJ/Mol)
-AH -AH -AH (AMI)* (MMl)b (expt)c
Dodecahedrene (94) 304, 529<* 185 266 Adamantene (95) 373, 2104 275 228
aCalculated by the method of Reference 12. bForce Field, Reference 14 cExperimental, this work ^For double hydrogenation (1.4-elimination product)
Although force-field calculations have been used successfully for estimating the relative energies of conformational isomers ,14 the method has never proven reliable in providing absolute heats of formation or heats of reaction. Adamantene and dodecahedrene are different types of strained olefins , 15 namely bridgehead versus nonplanar zero-bridged; thus it is unreasonable to expect an accurate theoretical prediction from such calculations on two such different models. On the other hand, the semiempirical molecular orbital calculations returned a better understanding of these systems. In the adamantene system, geometry optimization by AMI shows significant pyramidalization of the double bond at both olefinic carbons (more at the bridgehead), but a significant amount of 126 distortion of the system is in the twisting of the double bond necessary to accommodate a trans configuration in a six-membered ring. The torsion angle across the double bond is nearly 17° twisted, instead of completely pyramidalized. In dodecahedrene, as expected, all the distortion is in pyramidalization, with a torsion angle between the a-carbons of 133° (versus 180° in a planar bond). The experimental results shown here indicate that although adamantene is not so rigidly locked into a pyramidalized configuration as dodecahedrene, the twisting distortion of its double bond contributes significantly to its instability. In their evaluation of the stability of a variety of bridgehead olefins, Maier and Schleyer13 discuss a parameter they call "olefinic strain," which is closely related to the heat of hydrogenation of a bridgehead ofefin. Rough rules of thumb may be used with this parameter to predict whether a given olefin may be isolatable. Baised on their results, adamantene should be diffucult to isolate, even in low temperature matrices, whereas dodecahedrene should be possibly stable at very low temperatures. Since adamantene has recently been isolated, our results indicate. that dodecahedrene may be synthetically accessible, at least at low temperature.16 127
£ Dodecahedrene: Laboratory Preparation and Trapping 1. Background Encouraged by the predicted ability to isolate dodecahedrene under the right conditions, efforts were turned to the acquisition of this highly pyramidalized ofefin by standard "wet chemistry" operations. Elimination of phenylseleninic acid from organic molecules is a well established method for the introduction of unsaturation into a wide variety of systems. The phenyl selenoxides that undergo this elimination upon pyrolysis are generally prepared from the oxidation of the corresponding phenyl selenides, which themselves can be prepared via nucleophilic or electrophilic m ethods.
2. Results Thus, (phenylseleno)dodecahedrane (4 7), available by electrophilic methodology developed for the radical study, now promised to be a potential precursor to dodecahedrene. The plan was simple: oxidation of 47 to the corresponding phenylselenoxide (102); isolation followed by pyrolysis to yield the olefin under conditions suitable for its isolation. Unfortunately, the unpredicted instability of the dodecahedryl phenyl selenoxide made this plan impossible. As expected, the oxidation of 47 with a slight excess of meta- chloroperbenzoic acid (mCPBA) rapidly caused consumption of the starting phenyselenide. Unfortunately, contrary to our expectations, 128 the phenylselenoxide 102 was not very stable. At or near room temperature this species decomposed to dodecahedrene and dodecahedryl alcohol (their formation will be explained later), resulting in an impure product after work-up. This unexpected reactivity is attributed to the favorable alignment (perfectly eclipsed) of the phenylselenoxide and the a-protons. It thus became impossible to isolate 102 so that dodecahedrene could be formed under more favorable conditions. The initial strategy to gain access to this olefin had to be rethought, in particular, conditions that would allow dodecahedrene to survive long enoungh to confirm its existence were desired. Since the by-product of dodecahedrene formation is acidic (phenylseleninic acid) and the by-product of the oxidant (mCBA) is an acid, the reaction had to be buffered. If not, dodecahedrene would be protonated to yield the dodecahedryl cation and undergo electrophilic substitution. Under these conditions, the highly pyramidalized olefin was expected to have only a finite lifetime and would have to be trapped to facilitate determination of its presence. Although Diels-Alder reactions are often used to trap highly reactive olefins formed in situ, the large amount (relative to dodecahedrene) of trap needed and the presence of oxidant in the reaction mixture were expected to complicate reaction analysis. The most convenient trap was thought to be mCPBA,17 with resultant epoxide 103 formation. Since the reaction mixture had to be buffered, the acid- sensitive epoxide was expected to survive. 129
Excess mCPBA was necessary. Not only did phenylselenide 47 have to be converted to 102, but the resulting phenylselenic acid could consume the oxidant. On top of this was the need for a large mCPBA concentration to ensure successful trapping of the olefin. Due to the excess oxidant, the buffer chosen had to be stable in the system and be readily separated from the sensitive epoxide.
Anhydrous Na2HP 0 4 in dichloromethane was found to be suitable. When considering the reaction conditions, care again had to be given to making sure the excess reagents and by-products were readily separable from the desired epoxide. An acidic work-up was out of the question, as the epoxide was expected to open to the substituted cation and undergo electrophilic substitution. Removal of the acidic components and buffer were not a problem, a mere aqueous basic wash would satisfactorily remove them. The excess mCPBA could be sufficiently removed by converting it to mCBA with sodium sulfite in the basic wash. Any remaining non-dodecahedryl components were readily removed by trituration with ether. When 47 was allowed to react in this buffered environment containing excess mCPBA, GC/MS analysis showed the presence of the desired epoxide and dodecahedryl alcohol (Scheme 65). The epoxide exhibited a parent peak of m/z 274 (C 2 0 H 18O) and eluted faster than the alcohol (m/z 276, C2oH2oO)- The ratios of these compounds varied from only a trace of the epoxide to a predominance of the epoxide as the conditions were optimized. After standard work-up, these dodecahedryl compounds could be obtained free from the non- 130 dodecahedryl species, but not free from each other. Attempted chromatography served only to decompose the epoxide and returned only the alcohol. After a variety of separation attempts, I found the less polar epoxide could be separated from the alcohol on basic alumina when eluting with a solvent containing 3% triethylamine. The epoxide was stable to these conditions and was isolated in 46% yield from phenylselenide 47. Alcohol 53 was also isolated upon continued elution in 27% yield. The instability of the epoxide is evident by the eventual appearance of additional peaks in the !H NMR as time passes. The carbon and proton NMRs of this species exhibit the epoxide's symmetry and shielding/deshielding characteristics. In particular, a 7-line 13C NMR is seen as the result of the two symmetry planes in 103. The presence of dodecahedryl alcohol as the major by-product will now be addressed. Two pathways that exist for its formation are: (a) acid-catalyzed hydration of dodecahedrene, or more probably (b) a competing free radical decomposition as noted by Perkins and Turner18 in the adamantyl series. Although the former would deliver the observed alcohol, the system was buffered and anhydrous. The consistency with which 53 was always found suggests it to be a reaction characteristic rather than due to extraneous moisture. In an attempt to prepare adamantene from 1- adamantyl phenylselenoxide (104), Perkins and Turner found that upon heating 104, 1-adamantanol (105) was formed exclusively. 131
The mechanism they proposed is given in Scheme 66. They supported this mechanism by successful spin trapping of 1-AdO- radicals (although the origin of the hydrogen was not clear). A similar mechanism is expected for the conversion of 102 to 53.
.SePh ,SePh "1
-PhSeOH
4 7 102 9 4 .■ S mCPBA
OH
5 3 1 0 3
Scheme 65. Generation And Trapping of Dodecahedrene 132
■SePh OSePh
j- PhSe •
1 0 5
Scheme 66. Mechanism of 1-Adamantyl Alcohol Formation Noted by Perkins and Turner
D. Summary
The highly pyramidalized dodecahedrene was generated in the laboratory and trapped as its epoxide (46%). The unexpected ease in which the precursor phenylselenoxide underwent elimination forced the trapping of 94 rather than its direct isolation. Additionally, the double-bond strain energy of dodecahedrene was measured by thermochemical bracketing of gas-phase ion-molecule reactions. When applied to measuring the double-bond strain energy of adamantene, however, a different mechanism was found to operate. 133
R e fe re n c e s
1. a) Szeimies, G. In Reactive Intermediates: Abramovitch, R. A., Ed.; Plenum: New York, 1983; Vol 3, pp 299-366. b) Eaton, P. E. Tetrahedron 1979.35. 2189.
2. a) Watson, W. H.; Galloy, J.; Bartlett, P. D.; Roof, A. A. M. J. Am. Chem. Soc. 1981, 103. 2022. b) Bohm, M. C.; Carr, R. V. C ; Gleiter, R.; Paquette, L. A. J. Am. Chem. Soc. 1980. 102. 7218. c) Volland, W. V.; Davidson, E. R.; Borden, W. T. J. Am. Chem. Soc. 1979.101. 533. d) Radzizewski, J. G.; Yin, T.-K.; Miyake, F.; Ranzoni, G. E.; Borden, W. T.; Michl, J. J. Am. Chem. Soc. 1986, 108. 3544. e) Ranzoni, G. E.; Yin, T.-K.; Borden, W. T. J. Am. Chem. Soc. 1986. 108. 7121. 0 Paquette, L. A.; Shen, C.-C.; Krause, J. A.J. Am. Chem. Soc. 1989. 111. 2351.
3. a) Conlin, R. T.; Miller, R. D.; Michl, J. J. Am. Chem. Soc. 1979, 101. 7937. b) Michl, J.; Radziszewski, G. J.; Downing, J. W. Pure Appl. Chem. 1983. 55. 315.
4. a) Ermer, O. Aspekte von Kraftfeldrechnungen: Wolfgand Baur Verlag: Munich, W. Germany, 1981. b) Matella, C. J.; Jones, M.; Schleyer, P. v. R.; Maier, W. F. J. Am. Chem. Soc. 1979, 101. 7634. c) Allinger, N. L.; Sprague, J. T. J. Am. Chem. Soc. 1972, 24, 5734.
5. Kiplinger, J. P.; Tollens, F. R.; Marshall, A. G.; Kobayashi, T.; Lagerwall, D. R.; Paquette, L. A.; Bartmess J. E. J. Am. Chem. Soc. 1989,111, 6914.
6. This section contains a brief summary of the material contained in reference 5. For a more detailed discussion, the reader is encouraged to see this reference.
7. Nicolet Instrument Corp., 5225-1 Verona Rd., Madison, WI 53711-0508. 134
8. Santos, I.; Balogh, D. W.; Doeke, C. W.; Marshall, A. G.; Paquette, L. A. J. Am. Chem. Soc. 1986. 108. 8183.
9. The larger error is due to a symmetry correction, see reference 5.
10. Neutral molecule gas-phase heats of formation are available from; Pedley, J. B.; Naylor, R. D.; Kirby, S. P. Thermochemical Data of Organic Compounds. 2nd ed.; Chapman and Hall; London, 1986, or may be calculated (trifluoroacetate esters) from group additivity rules. See: Benson, S. W. Thermochemical Kinetics. 2nd ed.; Wiley; New York, 1976. Anion gas-phase heats of formation are also available, see reference S and relevant references cited therein.
11. See reference 10
12. Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P.J. Am. Chem. Soc. 1985, 107. 3902. Calculations performed using AMPAC version 4.0 (QCPE Program No. 455) implemented on an IBM RT computer by J. E. Bloor, M. Spotswood, and J. E. Bartmess at the University of Tennessee, Knoxville.
13. Maier, W. F.; Schleyer, P. v. R. J. Am. Chem. Soc. 1981,101, 1891.
14. a) Wertz, D. H.; Allinger, N. L. Tetrahedron 1974, 30, 1579. b) Allinger, N. L.; Sprague, J. T. Ibid. 1975. 31. 21. c) Allinger, N. L. Adv. Phvs. Org. Chem. 1976, H , 1.
15. Fawcett, F. S. Chem. Rev. 1950, 42. 219.
16. The Prinzbach group has recently reported the synthesis of a tetrasubstituted dodecahedrene, see Melder, J.-P.; Pinkos, R.; Fritz, H.; Prinzbach, H. Aneew. Chem.. Int. Ed. Enel. 1989, 28. 3 0 5 -3 1 0 .
17. Oxidation of seco-dodecahedrene (36) has been shown to deliver the corresponding seco-epoxide with mCPBA, see: Paquette, L. A.; Kobayashi, T. Tetrahedron Lett. 1987, 3531.
18. Perkins, M. J.; Turner, E. S. J. C. S. Chem. Comm. 1981, 139. CHAPTER VII ATTEMPTED REGIOSPECIFIC DISUBSTITUTION
A. Background The regiospecific disubstitution of dodecahedrane has always been hampered by the inability of the group first introduced to influence significantly the reactivity of any of the remaining five positions over the others. The result (i.e. polybromination or radical oxygenations) is an inseparable mixture of isomers. Of the five possible isomers of a disubstituted dodecahedrane, methodology only exists to obtain one regiospecifically. By making use of the positive charges' desire to become as distal as possible, the trapping of dodecahedryl dication 10 61 with methanol yields only the 1,16-disubstituted ether 107. The demonstrated ability to convert 107 to the corresponding dibromide 108 upon exposure to bromine has opened the door to 1,16 disubstitution of the dodecahedryl framework (Scheme 67). Although two 1,2-disubstituted dodecahedranes have been prepared (109 and 110),2*3 the desire for access to additional 1,2- disubstituted derivatives was still high. With cationic chemistry returning the 1,16-isomer, radical substitution being non- regiospecific, and anion chemistry unavailable from an unfunctionalized derivative, there remained few alternatives 135 136 available to meet our goals. Of the most promising, a photochemical means was held in high regard.
MeOH
Scheme 67. Trapping of the Dodecahedryl Dication
1 0 9 110
The ability of a photoexcited carbonyl group to efficiently functionalize a remote position on an alkyl chain is well known.4 In fact, some of the key steps in the synthesis of dodecahedrane rely on the photochemistry of the carbonyl group. Of particular interest was 137 the ability of a photoexcited carbonyl to abstract a y-hydrogen followed by coupling of the diradical formed to deliver a photocyclized product. This is, unfortunately, not the only path available to the carbonyl and our enthusiasm had to be bridled. Many of the photochemical processes available to aldehyde and ketone carbonyls may be classified as Norrish Type I or Norrish Type II reactions. In the Type I case (Scheme 68), bond rupture occurs a to the carbonyl group to deliver an initially formed acyl and alkyl radical. The acyl radical often decarbonylates and couples with the alkyl radical (or undergoes other typical radical chemistry).
O O hv A CO + R’—R
Scheme 68. Norrish Type I Photoreaction
When the Norrish Type II path is followed, the photoexcited carbonyl can abstract a y-proton5 to deliver intermediate biradical 111. Commonly, 111 has two choices: collapse to form cyclic alcohols (photocyclization) or bond-scission (photoelimination) (see Scheme 69). The path chosen is greatly dependent on the structure of the molecule and the reaction conditions utilized. 138
OH R------
OH 111 A +
Scheme 69. Norrish Type II Photoreaction
If a carbonyl group could be appropriately positioned relative to the dodecahedryl nucleus, it was expected that the corresponding photoexcited carbonyl could undergo the Norrish Type II photocyclization to yield a 1,2-disubstituted dodecahedryl derivative.6 The preference for an oxidizable alcohol made aldehyde 112 the initially desired target for the exploration of this possible regiospecific disubstitution methodology.
R .R
Scheme 70. Proposed regioselective disubstitution 139
IL Results The availability of 112 was seen originating from the readily accessible allyldodecahedrane. Initial attempts to form 112 relied upon the ozonolysis of 39 in dichloromethane at -78°C followed by reductive work-up. Although the aldehyde was formed, purification and characterization was hampered by the presence of a dodecahedryl impurity (assumed due to the oxidation of the dodecahedryl nucleus7). Unfortunately, this aldehyde was not very stable and decomposed upon attempted purification. To avoid this, the in situ generation of a diol with catalytic OSO 4 in the presence of sodium periodate 8 conveniently returned aldehyde 112 in 63% isolated yield (Scheme 71).
H
0 s 0 4 N al04
3 9 112
Scheme 71. Preparation of 112
When 112 was irradiated 9 in C 6 D 5 through pyrex, no reaction was observed. However, irradiation through quartz caused consumption of the aldehyde and formation of a highly insoluble tan 140 solid. The solubility characteristics of this material made NMR analysis impossible. Additionally, a mass spectrum of this highly insoluble material returned no helpful information to aid in this material's identification.10*u The inability to work with the major product of the reaction intensified our interest in the CgDg-soluble products of the reaction. Methyldodecahedrane is known to exhibit its methyl protons at 1.14 ppm as a singlet. Since decarbonylation of 112 was considered a possibility (albeit remote), evidence for the appearance of this singlet was closely monitored. After irradiation, this singlet was not observed. When a GC/MS analysis was performed on the soluble portion of the reaction, there were indeed dodecahedryl products present. The observed products included methyldodecahedrane and starting material as w ell as dodecahedryl alcohol and a phenyl-d 5 substituted dodecahedrane (50-d5>. Since no methyl signal was observed in the NMR, the first two compounds are explained as the result of the partial decarbonylation of 112 in the injection port of the GC/MS (as observed when analyzing pure 112.12) The latter two products are readily rationalized if the formation of the desired 1,4-biradical is invoked. Had y-hydrogen abstraction resulted, 113 could have undergone the competing photoelimination to extrude the enol of acetaldehyde and the formation of dodecahedrene (Scheme 72). Since dodecahedrene is highly reactive, it is not expected to be stable under these conditions for any length of time. Any trace of acid 141 could readily deliver the dodecahedryl cation. This cation, if formed, could react with the solvent in a Friedel-Crafts reaction or be trapped by a trace of moisture and account for the latter two products observed (Scheme 73).
Scheme 72. Options Available to Diradical 113 142
Scheme 73. Possible Origin of Photolysis Byproducts of 112
The appearance of these products suggests that at least some of the desired 1,4-biradical is being formed. Unfortunately, the only identifiable products of the reaction are minor components and most likely due to competing photoelimination. The inability to identify the major component of the reaction (which was still considered to possibly be the desired cyclobutanol) coupled with the relative instability of the starting aldehyde caused consideration of an alternate photocyclization precursor. The corresponding methyl ketone 115 was chosen. 143
Access to 115 originated form allyldodecahedrane as the result
of a Wacker-Type oxidation 13 (Scheme 74). This palladium-catalyzed oxidation efficiently delivered the methyl ketone in 95% isolated yield. When 39, disolved in a DMF/H 2 O mixture, was stirred in the presence of PdCl 2 , O2 , and a trace of HC1, efficient delivery of the desired methyl ketone resulted.
PdCI2, Og DMF, H P , H*
30 1 1 5
Scheme 74. Preparation of 115
Irradiation of this hopeful cyclobutanol precursor (in C 5H 6 or CH3CN) through quartz caused the consumption of starting material
but again resulted in the production of another very insoluble tan solid. Unfortunately, this material proved no more cooperative than its predecessor and resisted all attempts at useful NMR or MS analysis. A possibility still existed for obtaining useful information on this material. If it was the desired cyclobutanol, it might be dehydrated to yield an olefin more readily analyzed. Unfortunately, 144 all attempts to dehydrate this material failed to deliver any additional products (possibly due to the highly insoluble nature of this material). When attention was turned to the soluble portions of the reaction mixture, unlike the previous case, no dodecahedryl species was observed. This suggests that if 113 was formed, photoelimination was not as prevalent as from aldehyde 1 1 2 . The lack of information available from these two photochemical reactions was particularly discouraging. A photoinduced reaction was occurring, but the product's physical characteristics were complicating analysis. Phenyl ketones are known to be quite accommodating when prompted to undergo photocyclization . 14 However, since the most practical route to the corresponding dodecahedryl derivative would involve the relatively unstable aldehyde 1 1 2 and the system was not expected to be capable of being analyzed by GC/MS due to its molecular weight, hope was not high. Finally, the lack of utility of the product in a general 1 ,2 -disubstitution methodology and the expected similarity in product characteristics caused the abandonment of this plan before it had begun.
£. Conclusion Although a photochemical reaction is occuring when 112 and 115 are irradiated, the corresponding products' characteristics prohibited their identification. From the by-products observable 145 from the photolysis of 112, some of 1,4-biradical 113 is expected to have been generated, suggesting the possible formation of the desired cyclobutanol, but no evidence could be gathered to confirm or deny this expectation. 146
R e fe re n c e s
1. a) Olah, G. A.; Prakash, G. K. S.; Kobayashi, T.; Paquette, L. A. L_ Am. Chem. Soc. 1988, 110. 1304. b) Olah, G. A.; Prakash, G. K. S.; Fressner, W.-D.; Kobayashi, T.; Paquette, L. A. J. Am. Chem. Sfifix 1988, m , 8599.
2. Paquette, L. A.; Miyahara, Y. J. Org. Chem. 1987. 52. 1265.
3. Paquette, L. A.; Miyahara, Y.; Doecke, C. W. J. Am. Chem. Soc. 1986, IQS., 1716.
4. a) Evans, S.; Omkaram, N.; Scheffer, J. R.; Trotter, J. Tetrahedron Lett. 1985, 2Jl, 5903. b) Gagosian, R. B.; Dalton, C.; Turro, N. L Am. Chem. Soc. 1970, 22., 4752. c) Turro, N. J.; Dalton, J.; Dawes, K.; Farrington, G.; Hautala, R.; Motron, D.; Niemezyk, M.; Schore, N. Acc. Chem. Res. 1972, 5., 92. d) Swenton, J. S. JL_ Chem. Ed. 1969, 4&, 217. e) Wagner, P. J. Acc. Chem. Res. 1971,1, 168. 0 Sugimura, T.; Paquette, L. A. J. Am. Chem. Soc. 1987,1Q£, 3107.
5. If no y proton is accessible, H-5 may be abstracted to yield a cyclopentanol (homo-Norrish photocyclization).
6 . Although a cyclobutanol would result, the previous preparation of cyclopropyl dodecahedranes suggested this compound would be stable, see: Paquette, L. A.; Kobayashi, T.; Gallucci, J. J. Am. Chem. Soc. 1988. 110. 1305.
7. J. Weber, unpublished results.
8. Shue, Y.-K.; Carrera, G. M. Jr.; Nadzan, A. M. Tetrahedron Lett. 1987, 2&, 3225.
9. 450 watt Hanovia lamp.
10. It was obvious that a dodecahedryl fragmentation pattern was present. 147
11. Dodecahedryl alcohol is one of the more insoluble dodecahedryl derivatives. If this material was the desired cyclobutanol, it might exhibit similar insolubility.
12. McLafferty rearrangement is the well-known mass spectral analog of Type II photoelimination, see: Reference 4e. The MS of 1 1 2 exhibits a large m /z 258 fragment.
13. a) Smidt, J.; Hafner, W.; Jira, R.; Sieber.R.; Sedlmeier, J.; Sabel, A. Angew. Chem.. Int. Ed. Eng. 1962,1, 80. b) Tsuji, J. Synthesis 1 9 8 4 , 369.
14. For leading references, see: Wagner, P. J.; Subrahmanyam, D.; Park, B.-S. J. Am. Chem. Soc. 1991, 113. 709. 148
Experimental Section
General Methods Melting points were measured using a Thomas Hoover (Uni- Melt) capillary melting point apparatus and are uncorrected. Infrared (IR) spectra were recorded with a Perkin-Elmer 1320 or Perkin-Elmer Series 1600 FT/IR spectrometer and are expressed in reciprocal centimeters (cm-1). The IR spectrum of 103 was measured on an Hewlet Packard GC/IRD/MSD. Proton nuclear magnetic resonance spectra (*H NMR) were measured at 300 MHz (Bruker WP 300 and Bruker AC 300 FT NMR spectrometers) or at 500 MHz (Bruker AM 500 NMR spectrometer) and the splitting patterns are designated as follows: s, singlet; br s, broad singlet; d, doublet; t, tripplet; q, quartet; and m, multiplet. Carbon-13 NMR were recorded at 75 MHz (Bruker WP 300 and Bruker AC 300 FT NMR spectrometers) or at 125 MHz (Bruker AM 500 NMR spectrometer). The chemical shifts are given in parts per million (6) and the coupling constants are expressed in hertz (Hz). Exact mass measurements were determined at the Ohio State University Chemical Instrument Center with a Kratos MS-30 mass spectrometer. Gas chromatography/mass spectrum (GC/MS) analyses were performed using a Hewlett Packard 5970 series mass selective detector GC/MS fitted with a 12 m x 0.2 mm methyl silicon gum column and a set flow rate of 0.2 ml/min. Capillary gas 149 chromatography (GC) analyses were performed using a Carlo Erba Strumentazione Fractovap 4130 GC fitted with a 30 x 0.2S mm Durabond 5 column set for a flow rate of 2 ml/min at 260° C and with a split ratio of 30:1. Preparative thin layer chromatography separations were carried out using pre-coated, 0 .S mm, 2 0 x 2 0 cm
Kieselgel 60 F254 plates. All solvents were reagent grade and pre-dried where appropriate during the preparation of dodecahedrane, but freshly distilled once the dodecahedryl nucleus was present. Reactions involving non-aqueous media were carried out under inert atmosphere.
Increased Scale Birch Reduction to 25.
PhOCH.
To ammonia (300 to 350 ml, freshly distilled from sodium) in a flame-dried, 500-ml round-bottomed flask equipped with a mechanical stirrer, dry ice condensor and Claisen adapter (with argon inlet and septa) was added lithium wire (600 to 640 mg) and the blue solution stirred at -78°C for 45 to 75 minutes, during which 150 time approximately 35 ml of dry THF was added. Approximately 85% of a solution of 20 (6.0 to 6.4 g in 60 ml dry THF) was added over 10 minutes. The dark blue solution was stirred at -78°C for an additional 30 minutes at which point the remaining 20 was added dropwise until the mixture became very light blue. The amount of remaining 2 0 was weighed and the exact amount added was thereby determined. Chloromethyl phenyl ether (1.0 equivalents, in 15 ml dry THF) was added in one continuous stream to the rapidly stirred mixture. After an additional 15 minutes of stirring, solid NH 4 CI (6 grams) was added and the ammonia was allowed to evaporate overnight under a stream of argon.
The resulting mixture was diluted with CH 2 CI2 (100 ml) and enough water was added to dissolve the inorganic salts. After separation, the aqueous layer was extracted (CH 2 CI2 , 3 x 60 ml), filtered and concentrated. Purification of the resulting mixture was accomplished with MPLC (silica gel, elution with 10:15:75
CH2C l2 :Et2 0 :PE) to deliver desired ketone 25 (2.6g) in 42% yield. 15 1
Scale-up for Preparation of 30
A solution of CH 2 CI2 (30 ml) and oxalyl chloride (0-1396 g, l.l mmol) was placed in a 100-ml three-necked round-bottomed flask equipped with a magnetic stirrer, and two pressure-equalizing dropping funnels containing DMSO (0.1719 g, 2.2 mmol) dissolved in
CH 2 CI2 (5 ml) and 29 (0.3886 g, 1 mmol in 10 ml CH 2 CI2 or a minimum amount of CH 2 CI2 -DMSO to dissolve the alcohol), respectively. The DMSO was added to the stirred oxalyl chloride solution at -50 to -60°C. The reaction mixture was stirred for 5 minutes and 29 was added within five minutes; stirring was
continued for an additional 15 minutes. Triethylamine ( 1-01 g, 10 mmol) was added and the reaction mixture was allowed to warm to 5°C prior to work-up. Water (100 ml) was added and th e aqueous
layer extracted with CH 2 CI2 (3 x 75 ml). The combined organic layers
were washed successively with dilute HC1 solution ( 1%), water, and
dilute N a 2 CC>3 solution (5%), then dried over anhydrous Na 2S0 4 . Solvent removal (in vacuo) left slightly colored crude aldehyde 30. Passage through a short silica gel column (elution whh 75:25 ether:PE) yielded the pure 30 (0.30 g, 78%). 152
Scale-up for Preparation of 34
CHO
A solution of CH 2 CI2 (30 ml) and oxalyl chloride (0.2792 g, 2.2 mmol) was placed in a 1 0 0 -ml three-necked round-bottom ed flask equipped with a magnetic stirrer, and two pressure-equalizing dropping funnels containing DMSO (0.3438 g, 4.4 mmol) dissolved in
CH 2CI2 (5 ml) and 33 (0.310 g, 1 mmol in 10 ml CH 2 CI2 or a minimum amount of CH 2 CI2-DMSO to dissolve the alcohol), respectively. The DMSO was added to the stirred oxalyl chloride solution at -50 to -60°C. The reaction mixture was stirred for 5 minutes and 29 was added within five minutes; stirring was continued for an additional 15 minutes. Triethylamine (2.02 g, 20 mmol) was added and the reaction mixture was allowed to warm to 5°C prior to work-up. Water (100 ml) was added and the aqueous layer extracted with
CH 2 CI2 (3 x 75 ml). The combined organic layers were washed successively with dilute HC1 solution (1%), water, and dilute Na 2CC>3
(5%) solution, then dried over anhydrous Na 2 S(>4 . Solvent removal
(in vacuo) left slightly colored crude keto-aldehyde 34. Passage through a short silica gel column (elution with 75:25 ether:PE) yielded pure 34 (0.23 g, 75%). 153
Improved Preparation of Seco Olefin 36
Seco alcohol 35 (40 mg, 0.14 mmol) dissolved in CH 2CI2 (10 ml) was treated with one drop of trifluoroacetic acid at room temperature. After 15 minutes, reaction was complete (by TLC) and the solvent was removed in vacuo (no heat). The crude olefin was passed through a short column (silica gel, elution with 1:1 hexane:benzene) to yield olefin 36 as a crystalline solid (31 mg, 8 6 %) after solvent removal.
Allyldodecahedrane (39)
To nitrogen blanketed bromododecahedrane (38, 4.1 mg, 0.012 mmol) in a flame-dried 1.0 ml Wheaton vial equipped with a magnetic stirrer and teflon-coated cap was added dry CH 2 CI2 (0.1 154 ml), allyltrimethylsilane (1 0 0 pi, excess) and anhydrous zinc iodide (50 mg). The reaction mixture was capped and stirred at room temperature for three days. After dilution with CH 2 CI2 (0.4 ml) and addition of enough water to dissolve the Znl 2 (0 .2 ml), the mixture was extracted with CH 2 CI2 (3x1 ml). The combined organic layers were passed through a short silica gel plug and dried (anhydrous
MgSC>4 and/or anhydrous Na 2 SC>4 ) prior to solvent removal (in vacuo) to yield 3 mg of 39 as a light tan powder (65%), mp 190°C (dec.); IR
(NaCl, cm-1) 3069, 2936, 1296, 990, 734; !H NMR (300 MHz, CDCI 3) 6 5.8-5.7 (m, 1 H), 5.1-4.95 (m, 2 H), 3.35 (br s, 16 H), 3.05 (br s, 3 H),
2.25 (d, 6.99 Hz, 2 H); NMR (75 MHz, CDCI 3) ppm 137.33, 116.40, 79.14, 71.79, 67.15, 66.61, 66.50, 65.31, 48.13; m/z calcd (M+) 300.1855, obsd 300.1866.
(Phenylseleno)dodecahedrane (47)
.SePh
To bromododecahedrane (38, 5.2 mg, 0.015 mmol) dissolved in dry dichloromethane (5 ml) was added anhydrous zinc iodide (15 mg, 0.047 mmol) and (trimethylsilyl)phenylselenide (500 mg, 2.18 mmol). The reaction mixture was stirred at room temperature for 155
1.5 days, at which point additional Znl 2 (5 mg, 0.016 mmol) and TMSSePh (250 mg, 1.09 mmol) was added and stirring was continued for a total of three days. Water (2 ml) was added and the mixture was extracted with chloroform (2x5 ml). After drying (anhydrous
M g S0 4 ), filtration, and solvent removal (in vacuo), air was blown over the phenylselenol-containing yellow solution for one day. The resulting solid diphenyldiselenide could be removed from the desired 47 either by sublimation (100°C, house vacuum, 3 to 5 days) or more conveniently by trituration with ether (3 x 0.3 ml). Product 47 remained as a slightly yellow, crystalline solid, which may be used directly or further purified by preparative TLC (silica gel,
5:10:85 Et2 0 : CH2CI2 : hexane; UV active product at Rf 0.5) to yield 47 as an off-white powder (4 mg, 63%), mp 226-227°C (sealed tube); IR (solid, cm*1) 3040, 2940, 2850, 1260, 1100, 1010, 820-720; 1H NMR
(300 MHz, C6 D 6) 8 7.68 (m, 2 H), 7.09-7.07 (m, 3 H), 3.74-3.65 (m, 3
H), 3.26-3.2 (br s, 6 H), 3.2-3.12 (s, 10 H); NMR (75 MHz, C 6 D 6) ppm 134.73, 133.63, 128.96, 127.38, 96.46, 82.46, 76.68, 67.06, 66.91, 66.59; m/z calcd (M+) 416.1043, obsd 416.1062. 156
Improved Synthesis of Dodecahedranol (53).
OH
Bromododecahedrane (38, 5 mg, 0.015 mmol) dissolved in moist CH2 CI2 (0.5 ml) in a 1.0 ml Wheaton vial containg Znl 2 (50 mg) was stirred in the dark for 3 days (or longer). After dilution with
CH 2CI2 (5 ml), a water wash removed the inorganic salts. The organic layer was dried (anhydrous MgSC> 4 ) and the solvent removed to yield pure alcohol 53 (4 mg, 98%).
l-Cyano-2-dodecahedryl ethane (57)
.CN
To nitrogen blanketed bromododecahedrane ( 3 8 , 6 .8 mg, 0.02 mmol) in a flame-dried 1.0 ml Wheaton vial equipped with a magnetic stirrer and teflon-coated cap was added dry benzene (0.15 157 ml), acrylonitrile ( 0.1ml, excess), and tri-n-butyltin hydride (80 pi, excess). AIBN (0.1 mg) was added and the mixture capped and heated at 75°C for 6 hours. After being cooled and diluted with benzene (0.5 ml) and ether (0.2 ml), the reaction was treated with a solution of KF(aq) (1 ml) and stirred vigorously. Extraction of the resulting mixture with CH 2 CI2 (2x3 ml), passage through a short silica gel plug (elution with CH 2 CI2 ), drying (anhydrous MgSC >4 and/or anhydrous Na 2 SC>4 ) followed by solvent removal (in vacuo) returned crude 57 readily purified by preparative TLC (silica gel; 5%
E t2 0 :Hexane; Rf 0.14, I 2 stain) to yield 4.1 mg pure 57 (65%), mp 174-175°C (dec., sealed tube); IR (NaCl, cm-l) 2942, 2620, 2242, 1730, 1444, 1422, 1344, 1294, 1260, 1160, 904; *H NMR (300 MHz,
CDCI3) 5 3.38 (br s, 16 H), 3.00 (br s, 3H), 2.31-2.26 (t, 2H), 1.82, 1.77
(t, 2H); 13c NMR (75 MHz, CDCI 3 ) ppm 120.50, 78.96, 71.26, 67.13,
66.98, 66.50, 66.44, 38.52, 14.02; m/z calcd (M+) 313.1831, obsd 313.1817. 158
3-(Dodecahedryl)cyclopentanone (64)
Dry benzene (0.4 ml) containing bromododecahedrane (38, 4.7 mg, 0.014 mmol), 2-cyclopentenone (40 pi, excess), NaBFfy (1 mg, excess), and hexabutylditin (2 pi) was irradiated at 10°C in a dry NMR tube for 18 hours. The solution was diluted with ether (0.3 ml) and benzene ( 0 .1 ml), then treated with HF(aq) solution (a few drops). The mixture was stirred until the NaBH 4 had been consumed and the organic layer was passed through a short silica gel column
(Pasteur pipette) containing some anhydrous MgSC >4 and Na 2 S 0 4 prior to solvent removal (in vacuo). Preparative TLC (silica gel; elution with 1:8 EtOAcrhexane, Rf 0.27) purification provides 1.7 mg 6 4 (35%), mp 205°C (dec.); IR (NaCl, cm*) 2938, 1737, 1296, 1158, 739;
1H NMR (300 MHz, CDCI3 ) 8 3.5-3.2 (br s, 16 H), 3.15-2.95 (br s, 3 H), 2.4-2.0 (m, 4H), 1.9-1.75 (m, 1H), 1.5-1.3 (m, 2H); NMR (75 MHz,
CDCI3 ) ppm 220.26, 82.64, 67.17, 67.14, 66.59, 66.51, 66.48, 47.09, 41.98, 39.34, 25.45; m/z calcd (M+) 342.1983, obsd 342.2022. 159
(Phenylthio)dodecahedrane (79)
,SPh
To bromododecahedrane (38, 4 mg, 0.012 mmol) dissolved in dry CH 2 CI 2 (5 ml) in a 25 ml recovery flask equipped with a condensor and argon inlet was added anhydrous zinc iodide (40 mg, 0.13 mmol) and trimethylsilylphenylsulfide (500 mg, 2.7 mmol). The mixture was stirred for 10 hours at 50°C. After cooling, water (3 ml) was added and the mixture was extracted with chloroform (2x5 ml).
Drying (anhydrous MgS 0 4 ), filtration, and solvent removal (in vacuo) gave an oily residue containing thiophenol which was removed under a stream of air to leave 79 as a light tan solid. After trituration with ether (3 x 0.3 ml), pure 79 remained as an off-white powder (4 mg,
0.01 mmol, 93%), mp 241-243°C (dec., sealed tube); IR (CHCI 3 , cm*1) 3040, 2940, 1580, 1475, 1435, 1295, 1260, 1120, 885, 660-690; 'H
NMR (300 MHz, C6 D 6) 8 7.55-7.45 ( m, 2 H), 7.70-7.35 (m, 3H), 3.75-
3.50 (br s, 3 H), 3.45-3.2 (br s, 6 H), 3.2-3.0 (s, 10 H); (75 MHz,
C 6 D6) ppm 138.88, 131.55, 128.87, 126.38, 87.57, 75.74, 67.09, 67.00,
66.75, 66.59; m/z calcd (M+) 368.1599, obsd 368.1588. 160
l,4-Di(dodecahedryl)benzene (92)
To a 3 ml Wheaton vial (equipped with a magnetic stir bar) containing phenyldodecahedrane (SO, 11 mg, 0.033 mmol) and bromododecahedrane (38, 11 mg, 0.032 mmol) dissolved in dry
CH 2CI2 (1.5 ml) was added one crystal of AICI 3 . The reaction mixture was capped under argon and stirred at room temperature for 5 days.
After the AICI3 was quenched with one drop of water, the light tan
solid was triturated with CH 2 CI2 and ether to yield the extremely insoluble 92 as an off-white solid (10 mg, 51%), mp > 250°C; IR (NaCl, cm-*) 3053, 2936, 2850, 1728, 1592, 1296, 906, 736; NMR
(500 MHz, CDCI3) 6 7.24 (s, 2 H), 3.7-3.6 (br s, 3 H), 3.6-3.5 (br s, 6 H),
3.45-3.4 (s, 10 H); 13C NMR (125 MHz, CDCI3 ) ppm 128.34, 124.61, 75.15, 67.16, 66.05, 66.67 (2 carbons not observed); m/z calcd (M+) 594.3224, obsd 594.3286. 161
Epoxydodecahedrane (103)
A dry 25-ml round-bottomed flask containing degassed, dry
benzene (1 0 ml), 4A molecular sieves (0.5 g), anhydrous sodium hydrogen phosphate (0.5 g) and m-chloroperbenzoic acid (150 mg,
0.87 mmol) was stirred for 6 hours under argon prior to the addition of solid 47 (11.2 mg, 0.027 mmol). Oxidation to phenyl-selenoxide 102 occurred rapidly (within 15 minutes by NMR). Stirring of the
buffered solution at room temperature for 12 hours resulted in trapping of the in situ generated olefin 94 as epoxide 103. Water
(1 0 ml) was added and the mixture was extracted with CH 2 CI2 ( 2 x 5 ml). The combined organic layers were washed successively with saturated sodium bisulfite solution (2 x 25 ml) and 3% sodium
hydroxide solution (25 ml). After drying (anhydrous Na 2 S 0 4 ), the
solvent was removed (in vacuo) and the remaining solids were triturated with diethyl ether (0.5 ml). This solid consisted of 53 and the desired 103. Separation by column chromatography (basic
alumina, activity III, 5 g, elution with 1:1:8 ether:CH 2 C l2 :h e x a n e \ containing 3 % triethylamine) yielded 3.4 mg (46%) of the less polar epoxide followed by 2 mg of alcohol 53 (27%). For 102: *H NMR 162
(300 MHz, C6 D6) 6 7.57 (m, 2 H), 7.05 (m, 3 H), 3.59, (br s, 3 H), 3.01 (br s, 16 H); 103: mp >240°C; IR (vapor phase, cm*1) 2958, 1440,
1409, 1291, 1208, 967; iH NMR (500 MHz, C 6 D6) 8 3.31-3.25 (m, 2 H),
3.17-3.01 (m, 12 H), 2.88 (dd, J=7.3, 10.6 Hz, 4 H); (125 MHz, CgDg) ppm 74.99, 72.34, 66.74, 66.58, 65.74, 61.53, 57.07; m/z calcd (M+) 274.1357, obsd 274.1351.
Dodecahedrylacetaldehyde (112)
CHO
To allyldodecahedrane (39, 11 mg, 0.037 mmol) dissolved in
THF/H 2 O (6 ml/1 m l), was added sodium periodate (40 mg, 0.186 mmol) and a catalytic amount of OsC^/pyridine (5 pi of 0.157 M ). After being stirred for 2 days at room temperature, the reaction mixture was diluted with CH 2 CI2 (5 ml), washed with brine and dried
(anhydrous Na 2S0 4 ). Solvent removal and trituration (Et 2 0 ) returned 7 mg of aldehyde 112 (63%). Further purification may be achieved by short colum chromatography (elution with 1 0 % Et2 0 /PE), mp 2 0 2 -
205°C; IR (NaCl, cm*1) 2938, 1690, 1430, 1300, 1250; >H NMR (300
MHz, C 6 D 6) 8 9.47 (t, J=1.98 Hz, 1 H), 3.35 (br s) and 3.21 (br s)(16 H 163 total), 2.08 (d, J=1.98 Hz, 2 H); 13c NMR (125 MHz, C6 D6) ppm 201.53,
73.25, 72.93, 67.21, 67.19, 6 8 .8 6 , 66.80, 66.72; m/z calcd (M+) 302.1670, obsd 302.1650.
l-Dodecahedryl-2-propanone (115)
Into a 5 ml, two-necked pear-shaped flask (prerinsed with 1%
HC1) was placed a magnetic stir bar, PdCl 2 (1 mg), CuCl2 (2 mg) and 2 ml of a DMF/H2 O (3:1) solution, Oxygen was introduced and the initially black solution slowly turned green as the O 2 was absorbed.
To the green solution was added 39 (8 mg, 0.026 mmol) in DMF (0.1 ml). Oxygen was bubbled through the mixture for 2 hours and the mixture stirred under O 2 for 1 day. The reaction mixture was diluted with 10% HC1 and extracted with CH 2 CI2 (3x5 ml). The combined organic layers were washed with water (1 0 ml) and dried
(anhydrous MgS 0 4 ). Solvent removal (in vacuo) left a DMF/115 mixture. The remaining DMF was removed by dissolving the crude mixture in ether, washing with water, drying and solvent evaporation to leave pure ketone 115 (7.8 mg, 95%). Further 164 purification could be achieved by preparative TLC (silica gel, 1:2:7
CH 2Cl2 :Et20:hexane, Rf 0.51), mp 159-161°C (sealed tube); IR (NaCl, cm-*) 2940, 1715, 1395, 1160; *H NMR (300 MHz, C 6 D 6) 8 3.48 (br s,
6 H), 3.27 (br s, 10 H), 2.97 ( br s, 3 H), 2.17 (s, 2 H), 1.63 (s, 3 H); 13c
NMR (75 MHz, C 6 D6) ppm 206.03, 76.79, 73.28, 67.23, 67.16, 67.00,
6 6 .8 8, 56.98, 30.32; m/z calcd (M+) 316.1827, obsd 316.1805. APPENDIX A Carbon Origins of Dodecahedrane
165 166
i
< ■fit 7
£> -S P h 2BF4 o
M e02C C = C C 0 2Me
Figure 3. Origins of Dodecahedryl Carbons
The von Baeyer/IUPAC name for 3 is undecacyclo- [9.9.9.02*9.03.7()4^O.O5.15.06.16.08,15.010,14.012,19 0l3.17]eicosane<
The official Chemical Abstracts nomenclature for 3 is perhydro-5,2,1,6,3,4-
[2,3]butanyl[l ,4]diylidenedipentaleno[2,1,6-cde:2\ 1 ’, 6 '-gha]pentalene APPENDIX B
Details of the X-Ray Crystal Structure Analysis for 17.
167 168 Table 3. Summary of Data Collection and Structure Refinement Parameters.
17
molecular formula C24H24O6 for. wt. 408.45 space group PI cell constants a, A 10.276(2) b, A 13.000(2) c, A 7.782(1) a,deg 98.06(1) M eg 97.63(1) 7 ,deg 84.43(1) cell vol, A3 1017 formula units/unit cell 2 Dpajc g cm-3 1.33 max crystal dimensions, mm 0.35 x 0.38 x 0.38 scan width 1.10 + 0.35 tane standard reflections 121; 101; 2 0 0 ; 611; 361; 554 reflections measured 4935 26 range, deg 4 < 20 ^ 55 range of h, k, 1 h£ 13 *16
169 170
HydroxylGroup I
These are two different positions for the same carbon atom.
Figure 4. Numbering Scheme for 21 171
Table 4. Summary of Data Collection and Structure Refinement Parameters.
21
molecular formula C19H22°5 for. wt. 330.38 space group P2,/c cell constants a, A 13.246(2) b, A 9.145(2) c, A 13.437(2) M eg 107.97(1) cell vol. A3 1548 formula units/unit cell 4 Dcaic* 8 cm ' 3 1.417 max crystal dimensions, mm 0.31 x 0.32x0.44 scan width 0.80+ 0.35 tane standard reflections 002; 002; 0 1 1 ; 011; 202; 200 reflections measured 3065 26 range, deg 4 < 29 < 52 range of h, k, 1 +h, +k, +1 reflections observed [F 0 > 5o(F0)] 2078 computer programs SHELX-76 structure solution MULTAN80 no. of parameters varied 216 R = IIIF0I-IFCII/IIF0I 0.069 Rw 0.072 largest feature final diff. map 0.14e/A3 Table 5. Bond Lengths for 21
atom atom distance atom atom distance
C(l) C(2) 1.552(5) C(9) C(13) 1.534(6) C(l) C(8) 1.538(4) C(10) C (ll) 1.510(7) C(l) C(17) 1.495(5) C(ll) C(12) 1.508(6) C(l) C(15) 1.627(5) C(12) C(13) 1.519(5) C(2) C(3) 1.524(5) C(12) 0(4) 1.469(5) C(2) C(6 ) 1.540(5) C(13) C(14) 1.506(5) C(3) C(4) 1.538(6) C(14) C(15) 1.544(4) C(4) C(5) 1.520(5) C(15) C(16) 1.486(5) C(4) 0(5) 1.388(5) C(16) 0(4) 1.326(5) C(5) C(6 ) 1.552(5) C(16) 0(3) 1.203(4) C(5) C(15) 1.545(5) C(17) 0(1) 1.187(4) C(6 ) C(7) 1.574(5) C(17) 0(2) 1.334(4) C(7) C( 8) 1.546(5) 0(2) C(18) 1.458(5) C(7) C(14) 1.524(5) C(18) C(19) 1.442(9) C(8) C(9) 1.529(5) C(18) C(19’) 1.436(4) C(9) C(10) 1.543(6) 0(3) 0(5) 2.770(5) 173
Table 6 . Intramolecular Bond Angles Involving the Nonhydrogen Atoms (degrees).
atom atom atom degrees atom atom atom degrees
C(2) C (l) C(8) 99.0(3) C(15) 0(16) 0(4) 117.4(3) C(2) C (l) C(15) 101.4(3) 0(3) 0(16) 0(4) 118.0(4) C(2) C (l) C(17) 111.9(3) C(16) 0(4) 0 (12) 125.2(3) C(8) C( 1) C(15) 101.6(3) C(l) 0(17) 0 (2 ) 111.3(3) C(8) C (l) C(17) 122.1(5) C(l) 0(17) 0 (1) 125.8(3) C(15) C( 1) C(17) 117.5(3) 0 (1) 0(17) 0 (2 ) 122.7(3) C(l) C(2) C(3) 113.4(3) C(2) 0 (6 ) 0(7) 103.3(3) C(l) C(2) C(6 ) 93.6(3) C(6 ) 0(7) 0 (8) 105.8(3) C(3) C(2) C(6 ) 107.2(3) C(6 ) 0(7) 0(14) 103.3(3) C(2) C(3) C(4) 103.1(3) C(8) 0(7) 0(14) 94.3(3) C(3) C(4) C(5) 104.1(3) 0(7) 0 (8) 0 (1) 92.8(3) C(3) C(4) 0(5) 115.6(3) 0(7) 0 (8) 0(9) 104.8(3) C(5) C( 4) 0(5) 117.7(4) 0 (1) 0 (8) 0(9) 116.2(3) C(4) C(5) C(6 ) 104.6(3) 0 (8) 0(9) 0 (10) 119.2(3) C(4) C(5) C(15) 117.5(3) 0 (8) 0(9) 0(13) 103.2(3) 0 (6 ) C(5) C(15) 92.6(3) C(10) 0(9) 0(13) 106.4(3) C(5) C(6 ) C(2) 93.7(3) 0(9) 0 (10) 0 (11) 106.2(4) C(5) C(6 ) C(7) 105.3(3) 0 (10) 0 (11) 0 (12) 107.7(4) C(13) C(14) C(7) 108.4(3) C(ll) 0 (12) 0(13) 105.1(4) C(13) C(14) C(15) 110.5(5) 0 (11) 0 (12) 0(4) 110.6(4) C(7) C(14) C(15) 93.8(3) 0(13) 0 (12) 0(4) 116.2(3) C(14) C(15) C(l) 101.7(3) 0 (12) 0(13) 0(9) 107.4(3) C(14) C(15) C(5) 98.2(3) 0 (12) 0(13) 0(14) 116.0(3) C(14) C(15) C(16) 112.1(3) 0(9) 0(13) 0(14) 103.8(3) C(l) C(15) C(5) 102.1(3) 0(17) 0 (2) 0(18) 116.9(3) C(l) C(15) C(16) 117.5(3) 0 (2 ) 0(18) 0(19) 111.1(5) C(5) C(15) C(16) 122.0(3) 0 (2 ) 0(18) C(19') 116.0(7) C(15) C(16) 0(3) 124.7(4) Appendix D. *H Nuclear Magnetic Resonance Spectra
174 175 PPM
o\ .SePh
(47)
11 * • • 11 * • • 111 ■< | ■ * i ■ 11111 j r | | ■ i ■ ■ ] i .h i.si i.ii i.si i.n i.ii lcs i.»i -J O' 177 178 S'» N44 » S‘« 1 • (w) - • *• • 1 - > » I SI It S'l -jsr n.tf tn uiiii ihu r t i u m t •fr ; * J f I ll'll 0t1 1100 I0*t-1 , •I •I 9 cs It ‘cost ll ‘ lltll ill -it, *4 •III !1 10 • •• W/IM 14/IN /N IS 01 4S *4 Ok Oil fll It 91 IN 11 04 10 40 •9 13 44 10 13 t4 14 r
i'i i i' m r r r i ' n i i i i i i i i i i | i n -r r n i i i v r i r j i i i i | i r r «.« 1.1 ).< J.l l.t 1.1 I .) .5 1.1 179 180
.s
£
.8 181 0*0 CHO
MSCV209.00I OMlTIMC l*U*U10:94 300.199 SF03 900.130 SY 310.0 01 5000.000 91 93760 TO 33760 9V 4000.000 HZ/PI .344 PH 9.6 TO 0.0 (112) AO 4.090 no 160 N9 97 Tt 303 8oe-R 150.9 12 K 135 FH 5000 03 60000.000 OP 7L PO 19 .300 66 .400 NC 4 OC 1.000 HZ/CM 79.779 PPM/CH .363 X9 2 69 3394.91
" " I 1"i 1'1 " I I..... I..... I...... J.80 8.00 1.80 t.00 S.80 1.00 . 80 0.0 00 to (115)
....w v. v. 183 BIBLIOGRAPHY
Benson, S. W. Thermochemical Kinetics. 2nd ed.; Wiley; New York, 1976.
Doering, W. v. W. In Theoretical Organic Chemistry - The Kekule Symposium: Butterworths: London: 1969 p 45.
Engler, E. M.; Schleyer, P. v. R. In MTP International Review of Science. Alicvclic Compounds: Parker, W., Ed.; Butterworths, London, 1973; Vol. 5.
Ermer, O. Aspekte von Kraftfeldrechnungen: Wolfgand Baur Verlag: Munich, W. Germany, 1981.
Fleming, I. in Comprehensive Organic Chemistry, ed. D. H. R. Barton and W. D. Ollis, Pergamon, Oxford, 1979, Vol. Ill, p. 541.
Fort, R. C. Adamantane-The Chemistry of Diamond Molecules: Marcel Dekker: New York, 1976.
Giese, B.: Radicals in Organic Synthesis: Formation of Carbon- Carbon Bonds: Pergamon Press: Oxford, 1986.
Huntress, E. H.; Lesslie, T. E.; Bomstein, J. Organic Synthesis Collective Volume 4, 329 (1963).
Kharasch, M. S.; Reinmuth, O. Grignard Reactions of Nonmetallic Substances: Prentice-Hall: New York, 1954; pp 1067 - 1125.
Kirby, S. P. Thermochemical Data of Organic Compounds. 2nd ed.; Chapman and Hall; London, 1986.
Matzner, E. A. Ph. D. Dissertation, Yale University, 1958.
184 185
Negishi, E. Organometallics in Organic Synthesis: Wiley: New York, 1980; p 38.
Rossi, R. A.; Pierini, A. B.; Palaaios, S. M. in Advances in Free Radical Chemistry. Vol 1. JAI Press Inc.; Greenwich, Conn., 1990; p 193 - 252.
Szeimies, G. In Reactive Intermediates: Abramovitch, R. A., Ed.; Plenum: New York, 1983; Vol 3, pp 299-366.
Weber, Jeff Ph. D. Dissertation, The Ohio State University, 1987.