. 83.3%. gli-Yfij... 1.... I 4.4»: . I... .313. . .. 3 . . .. . . mgfisvfi. _. . .... _ ... . . . «mm; 3.0 . .. . .. . .3 [JP 3!.- . . . . 3 . . 1 .c ”fir 3 s“ r mgr... L» .I- . 3 z I” . .fu 3-. e E .13... $.43... ., ”EMF...- fifi. 3m. . mu“... 1%»? n3.. E4133m . 5.» - 13W"... .3... $3...- . . 3.... .33 . ‘33.. :33 . 3.3. . . 3 t. . .- ... . 3 . .1 . .551"? 3 3.0; r 3.413. r... . 9433 . fi . . v . . . . , ._ 3. . .3 - r r .. .tuhn1. ....... . . . .. 3 . .. www.méfimam 3.... 3m, . ., . . 3 czfim...m .3. 3 . «I... .- erzvzmmmr . . : L 3 . V. 24 - u .3. : i .. _ .3 1mm...“ .nifl aaxbhwfir.nmv...3x. .- . i - ,. 33 m3 .. 3?».- u..-. 35%... .\ ' a”! ,1 IC . a 3 . .1..- .3; 3 13 .. fiqgidiJsfih r9. . g In”. 3% ”HANK: 4.3%? 3.» 3 3-. . . . .33 4. IVES“. 93.3.“... 3.3%.... 3.3.3390... .. . 3333.1? .23.». 3 . . 3215. ._mm.... a 3.333.»? . 3. 5......- .: «um-”43 L3“. c .3 . .. a... 3.5.33.3 .52 0.3.. www.3- an .. 4,. .. twp .. . 3m... 33...? 33... 3.33.3333-w3333..¢.£. r33.“ day; . . .. .. Jaw-WWW; 33 :33.- .aimfi. Euhfiwfi. 333 33.3 - . v . .‘ I c. 3. 3!? tr. .0 3: II . K. 3% . .33.}. :3. .3333».me 3.4mm; _ . - . . . 3 Div :3 .H 9.30, . 3% 3 %§.uh.%r - r \\l 133 0b: in 93¢". ‘1’9‘ 5'” f‘ .3 I“... 2L“ % . u. , mm” . . 3%.». Mafia-33.- $3.3“.3-33-Hflqw. . 3 .. ifi, .. mmw..3.wm.m»3-, 3", kiwi-3%.. 3W3. .3 33 53.....«333 h.- 33?»? ”I . cud-$1M 02...}... . ...... 4..» hi”, I . 1 t. i lrh. I .Otx1kmrd th.‘ 3 , . Saab“. 333 3a . .. . . . . . . - - «Mmfwa ferrfiauhmmwenwh! 3 .. . . .2. . . . . . . ..... ._ . 3. 5... Hagan 333 .3733: . - I . .. . . ., 3.3333... hr- .nfiuMWEW . . .. . .- . . - . . .. 3 3 .3 3 . 3 ..3. .3 .3 3 .33. . 3.1 3. PM.“ 33 3. . . .. . . . - Ihl. 3233334333313 . 3 3: . . 33. {L‘ . 3.. d.» 348.390“ 3 133 O..- o‘T' .3 3 3! .3 $943 . z... . I 1‘ (' 5‘. 3 3b: 8.31.3? mg... Jun-rm... 3.3.3333. v3 . umIJwJ... .s 33333 . . - . . 333%.... «Ekawmw. ...... .353. 3.3mm? .33 33... 1:43.. .33”... 33......uafifiqvvfi3t-h3m3u ,4me3 $6333.;- 3.mufiw« . 3.. I ‘ )3 I (Ola '3‘ .W ..3. 333W»? 333 .33 33 U 3... L 53.3.3 gums-“M $5 3.. mi- . ...3r3.~333nm1w3w3slu.33.3.. 4.83.... , 3 2.3.". . Ham-3...“.- $33331. . 3 «gm 3.3.3.3333?“ hfiwmfif .3 . . . . 43»de .3 .33Ufln-"uv3mmcli v 1. 3.. .meRrflnfl... an. .33. 3353 . .. 3 - s3“ .\.J...d.l.‘30“L3lV-. 1-..KIAlflllvv-mv I fiz.‘ igq‘f. 3 .\ 3 .. . . . . . . . .rv3333a. . 3. §3fiwwmmfia . 333w33¢332333$33 3. $53.34.: 333363 , . . . .. . .3 3 33.33 .3. L.- .-., mug-aw .33 3 .2»..- “3.....- .. . .. 3 , . .. . . : . - ! ‘V’ 3 v . u l .3 r1 , . - . . . .3 313333-3331 3 . n . .3 II II . i 3...“. a .3: . 33 3. 33:25.3 {at} urn“. 3...... 33 . 1 -V I. . flash? "3.3 3.3.39“. than": 31:33- nruu . . .33 v! 3 .- . 311;. $333 mp. lHW-tynvahvhd3 . 333-3qu 333 “$833.3.- O 333! 3353 3 . b. 3 33 . .3:- Libb 3. . dug-.0333 3 .....-.n $D3I333 3:13. 3333.33 “1433333 in.» 333333.33Mr 3... .333.II|33.unmn.lw3.3.umu.3 3333.0333333 .3033 . .i- 3... . .23".- ..3-3. . .. 3. -h - . . . .3331uu...n3r3-Wm3.3hmnm33 333333 354.393.33.3wfiaqu. .333..333€».333u....,mm..-mr-w31:33..w 333% -333 33hr 3.3».33mW3....33-33V3333.333-Ihm“3 {33335333. ”333“ . . E. 4 -. .3331 . . u ZQSAIIVLrJz‘r“ IIK 3.3.3133. I z . must-h. {F333- -333-3333n. 331' Hakim... 333333 .3.” . .33". 3833.333. 133$!!! - .. . 6...th . 33 .«VT‘atfl 1.3.332».-. .. !|I3 II . - 3 33! V 0 It 00“.. t 3 3 I .I'B .3! 3 IV. .l‘. 3 '0.qu LENA. 3 . 3. 33 3-. .3 . .. . 3 . - .3 . . .- . . 3 . - - .3 . 3 3 .- .3.3... 3033““ .333.- 33333 3 . . . ...Ivl-LHWII3DIV3DL3WfiW.V3" I x3 .3! 333 )3 L . -333 nu - at... 33.33.333.303‘333333313333333343333Q33 3, 3. 3.. L [3.1.3.3831 3. . 33.333 "‘33:. 3 Val-ND“ 10“.? - 1! . J. .. .. 4v] 40.... O 3. .l 1.... .3 1. . 5.3Hw33433hmt333 33333.9 5 3 .. 33? fikxfiddirlfihfi Mr vfiwflhfifitiir 3 in 33.133 '31.: . 3303.333 J. 03‘. 3.3. 133.. b." 3. 3.... 3.333 $3333.33... 3 31ml. :33! «Juli . . 2.33333 .13 ...e.%v..i31i-.r v 3... «3.33. . “$9.3. .wfily. .39 W3 . nhfimmfluwdnnf -. 3.3933333 3. .m3n..3w\3 .. :3 tum-3.3.3 1|. .3 0303].“: ”it: VI. 3, 3s- »I. I4 I” #firfiawsiuua33-J n. . 3-. avg-39 .3Fd333333éahdnnv3m-3w3. {I'lh‘itn‘t 33ta3du33. . cum-‘3‘. 3... ”vi-\YJI‘ “rahfiy \a p 3 IQIK . a 3, 6 31". 3 - 33 3 :V . . I- 3. 3-9033. 1:33-.th 33.3.. 3. , 3,33331L3..33um_3u3.3.3n33u “aim-‘33.}... . 33 nap-3 . . '95. .3x35313‘hun 3".01333. J3... .. 5335-3 3333 3. 33333-3333..“3333fl-33 . 3.3-33.3 . . .. , . 3. Simularnelrl- 33 Jun-.3». 3.330%. 3. Lam-33.1! .. . . 3 313 133.33. 3|333333|.3V-r3£3n4.3333.n3331n1.. 3 . . . .. . n . . 33.3 3M“.A3Tm3nmt.wh3-mlz .. . . .. . . 2.3.3? . 3 3 3k. . . . 33.3 33....333n3333..n-w3..$.3r 3-3..- 3 3352.. .. nub-r Q3 L3. :3 (-Hl3hI3I3-313Larfll33 .3 3-33. 3 o J3 3 3....) .kVu3Lanfln-ahnn‘w311n13353. .IIO .d 1.07 3A. 0‘3 .14 l 1’ 7" "5‘3. ”nu-338 33333330... 3. 3.! 33!.3\- 33.3 33933343“... $3333. 3 r 3.» . t. 333 331‘33333: fr; .333“. . 3.3333 sI‘ ,13 33 o alvv .annJ “333.3033. 3003-ur 33' ”fivflflh3331.03333F3..I.V333 ....l3'3\31|3.13» 3 t .33 333.33. r3933! .3! 3. .3330 313.333 3.3 .333.”- . o3 . .3 |.\H..333 t33h$.3h~uxflr13“r31. 3-:3 . .- . tr . 1U. (1. . an. nxallv. .3 .3. 33:3. 3333.331 ‘hmamzsnflvflum- . - - .l 1.3. 3.343 I _. 3.33.. 33,-. . .Hawna. cl! 3.1.33 3... 3 ---.3 .3 - 3.3333 13....- ltr‘. . . .L3.l .-u3\.u.u.3l Jun-3K :n 13343311.. .ffilklzkn. 3 . 33.333 .33 ‘13:? -3 ? 233-333;! . \3333333433. .flufifiz. 333 13333.13..I3¢333-P.n3.3.; . A ‘Ilo .. 3". II I '0'!‘ I...V.'I‘IIQ I'I‘.‘3Y.l 3 Dr?! :33 33133.3!!! 1th3! . 3 33...» 53-3.. 33.3313 11. . 3.1.3.3133! , - . . .. x 3 r: .333! .NO. 13 3 . ..333.33,..,-...»dl33a1. 3 3. 3.11.2.3 . 1.. .3 33.33“.- . “! «5.3.313 .;3 .3 - 3.3.33 3 a 33 3313-39 3'33bolnv33‘313i 3.3.3 3.335 . .3 33 .3 3- 33ll3~l1l33 3 L.» r 3-3uu3hllo3hl\3.hhwwn33.3dr...3r.3I|3..--|Yr.i ,3 If 3 J3 . 333333-3333 1....{hK-3Vrfw32u1u Vina-tr} xfl.v3-3-.h.3 .3.l- v. I - . , , - .I3 3153 33.. - - ----- .- i. 7.. 3.3. . .HH3_.3..333J..333333-...33-3%.3W3u3.333.I-133333 . n , u _ 3 .l..-33s3...-N3l3-3.33.,q3\-3.33-hr.-H.u.-333N|ru333 31 1.333” . 3\- l .3333!- 1|33l3|31 3|.l l ['3 ‘ Illnl ’ l 'l‘l IE 3" IIIIIIIIIIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIII IIIIIIIIIIIIIIII III I IIIIIIIIII‘ LIB Rig?“ Michigan State University — This is to certify that the dissertation entitled 000 - ELEC new r EMMA/6 f/v MEA/UM- BIA/6 8 form. [C 3 12/0 6 5 HEAD BAA/CALL 5 presented by UL} fir/UA CKAVCI'UA/ has been accepted towards fulfillment of the requirements for PHHD degreein CHEMISTRY Q4... Mjgflazw Date 40 JULY) I???’ " .‘ri'..;¢: .' b $53755" “4233 . PLACE Ii RETURN BOX to roman this checkout from your record. ' To AVOID FINES return on or before date duo. DATE DUE DATE DUE DATE DUE §I_T__ * _L_ I I I I_ MSU isAn Affirmative Action/EM Opportunity inflation WWI ODD-ELECTRON c BONDING IN MEDIUM-KIN G BICYCLIC BRIDGEHEAD RADICALS By Liliana Craciun A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1997 ABSTRACT ODD-ELECTRON 0' BONDING IN NIEDIUM-RING BICYCLIC BRIDGET-[BAD RADICALS By Liliana Craciun My research has combined multi-step organic synthesis with physical and computational studies to explore medium-ring bridgehead radicals. The results of MNDO and ab initio HF/6-31G* calculations, presented in Chapter 1, suggest that such species might show unusual stability and/or persistence, as well as interesting bridgehead- bridgehead interactions. Description of the synthesis, kinetics, spin trapping, EPR, ENDOR and computational studies, of bicyclo[3.3.3]undec-l-yl (l-manxyl) radical, a key reference species for medium-ring bridgehead radicals, now generated in solution from manxane by H—abstraction with tert-butoxyl radicals, is given in Chapter 2. The exceptional persistence of this sterically open radical is unique and is attributed to the high strain of all its decomposition products. [3.3.3]Propellane has been identified among the decay products of l-manxyl radical; its formation was rationalized by a novel a-disproportionation. This ' research is extended in Chapter 3, where efforts toward the bridgehead carbon-centered radical of l-azabicyclo[3.3.3]undecane (manxine) are described. My synthetic work centered on developing routes and efficient precursors to atrane-like bicyclics, whose corresponding bridgehead organic radicals could provide a potentially long series of compounds for the investigation of intrabridgehead through- ii space 0' interactions. A modified literature procedure for the preparation of 3-(2- hydroxyethyl)-1,S-pentanediol, along with the syntheses of the novel compounds, tris-2- aminoethyl-methane and tris(o-hydroxyphenyl)methane are depicted in Chapter 4. Chapters 5 and 6 describe additional computational work. Inspired by the hybridization change of the bridgehead carbons in manxane and manxine, associated with decreased one-bond C-H couplings, we explored the prediction, from standard quantum chemical models, of C-H couplings in a series of bi- and polycyclics. Lastly, the availability of the RHF/6-3IG*//RHF/6-31G* wavefunctions and energies obtained for the set of small- and medium-ring polycyclic compounds considered in the hybridization study, led us to reexamine the performance of the Wiberg and Ibrahim/Schleyer hydrocarbon group increments in calculating heats of formation from ab initio energies. The research described herein was motivated by the challenge of designing species that can be used to probe theories of structure and bonding. The unusual properties of the bicyclo[3.3.3] system are a consequence of the geometry and strain inherent in a bicyclic array made up entirely of eight-membered rings. Our foray in the field of medium-ring bicyclic radicals revealed unforeseen opportunities for firrther work in this area. iii ”It is not thy duty to compfete the work hut neither are thoufree to desist of it” ‘Ethics of the fathers (‘1’ he Tafmud) 2:21 DEDICATION Dedicated to those courageous Romanians who Est their lives in the fight against communism during the 1989 revofution, without their sacrifice I couhf not have reached to this dream, and to my mentors, Ioan ‘Voda, Darin Breazu, Son'n Myer, and flames 1:“. yachson, from whom not onfy I horned a great deafahout chemistry, hut ahso how to enjoy it. ACKNOWLEDGMENTS I [ike to believe that the whoh worht is myfamify, hut from those dosest to my heart; I want to thank Laura and Roda, whom I deprived of too many hours, my adoptive parents and Laura’s grandmothers, who provided us not onfy with free daycare, but also with [ots of love and support, and my wonderfrdadvrsor at Mil, Q’rof flames {E jackson, with the regret that I couhf not do more and hetter. vi TABLE OF CONTENTS LIST OF TABLES ..................................................................................................... x LIST OF FIGURES ................................................................................................... xiii LIST OF SCHEMES ................................................................................................... xvi CHAPTER 1. Odd-Electron o Bonding in Medium-Ring Bicyclic Bridgehead Radicals: A Theoretical Investigation .............................................................................. 1 1.1 An Overview of Research on Odd-Electron o Bonds ................................ 2 1.2 Intrabridgehead Interactions in Medium-Ring Bicyclics .............................. 10 1.2.1 Closed-Shell Interactions in Neutral Medium-Ring Bicyclics ........................................................................................ 11 1.2.2 Atranes ........................................................................................ 17 1.2.3 Radical Cations of Medium-Ring Bicyclic Diamines, Disulfides and Diphosphines ................................................................. 20 1.2.4 Medium-Ring Bicych Carbon-Centered Bridgehead Radicals ........................................................................................ 24 1.2.5 Bridgehead Phosphoranyl Radicals .......................................... 25 1.2.6 Intrabridgehead Indirect Interactions via Hydrogen ................... 25 1.3 Geometry, Strain and Odd-Electron o Bonding in Medium-Ring Bicyclic Bridgehead Radicals: A Semiempirical and Ab Initio HF/6-3 16* Analysis ....... 28 1.4 References ........................................................................................ 41 CHAPTER 2. l-Manxyl: A Persistent Tertiary Alkyl Radical that Disproportionates via e-Hydrogen Abstraction ........................................................................................ 50 2.1 Results and Discussion ............................................................................. 52 2.2 Spin Trapping Studies on l-Manxyl Radical .......................................... 72 vii 2.3 Kinetic Decay and Product Analysis ..................................................... 77 2.4 Conclusions ........................................................................................ 92 2.5 Experimental Section ............................................................................ 92 2.6 References ...................................................................................... 103 CHAPTER 3. 5-Manxinyl Radical: A Computational and Experimental Study ...... 107 3.1 Results and Discussion ........................................................................... 108 3.2 Experimental Methods ........................................................................... 121 3 .3 References ...................................................................................... 127 CHAPTER 4. Progress Toward the Synthesis of Atrane-Like Compounds ................. 130 4.1 3-(2-Hydroxyethyl)-1,5-Pentanediol and 3-(2-Aminoethyl)-1,5- Pentanediarnine ....................................................................................... 132 4.2 tris-(o-Hydroxyphenyl)-Methane ............................................................... 139 4.4 Experimental Methods ........................................................................... 143 4.5 References ...................................................................................... 152 CHAPTER 5. Correlation of Tertiary One-Bond l3C-lH Spin-Spin Coupling Constants with PM3 Calculated Structures in Some Bi- and Polycyclic Saturated Hydrocarbons ....................................................................................... 155 5.1 Introduction ....................................................................................... 156 5.2 Theoretical Model ........................................................................... 162 5.3 Results and Discussion ........................................................................... 164 5.4 Summary .................................................................................................. 183 5.5 Experimental Methods ........................................................................... 186 5.6 References ...................................................................................... 188 CHAPTER 6. Heats of Formation of Medium-Ring Strained Cyclo- and Polycycloalkanes: Comparison of Ab Initio Group Equivalent Schemes with viii the PM3 and MMX Methods ...................................................................................... 193 6.1 Results and Discussion ............................................................... 194 6.2 References .......................................................................... 209 APPENDD( ............................................................................................................. 212 Table 1.1 Table 1.2 Table 1.3 Table 1.4 Table 1.5 Table 1.6 Table 1.7 Table 1.8 Table 1.9 Table 1.10 Table 1.11 LIST OF TABLES CHAPTER 1 Experimental One- and Three-Electron Bond Energies ..................... 6 MM2 Steric Energies of Lowest Energy Conformations for Some Bicyclic Hydrocarbons ................................................................. 12 Heats of Some Formal Dehydrogenations .......................................... 16 Heats of Formation and Strain Energies of Some Bicyclic Hydrocarbons and Propellanes ..................................................... 16 HF/6-31G“ Total Energies (MNDO Heats of Formation), BDE’s, Intrabridgehead Distances and Radical Stabilization Energies Relative to the tert-Butyl Radical in 1, 44 and 45 .................. 31 HF/6-31G* Total Energies (MNDO Heats of F orrnation), BDE’s, Intrabridgehead Distances and Radical Stabilization Energies Relative to the tert-Butyl Radical in 8 and 46-50 .............................. 32 HF/6-3 16* Total Energies (MNDO Heats of Formation), BDE’s, Intrabridgehead Distances and Radical Stabilization Energies Relative to the tert-Butyl Radical in 51-55 .......................................... 33 I-IF/6-31G* Total Energies (MNDO Heats of Formation), BDE’s, Intrabridgehead Distances and Radical Stabilization Energies Relative to the tert-Butyl Radical in 57-62 .............................. 34 Spin Densities (p) and HF/6-31G* Intrabridgehead Distances (BB) in the Carbon-Centered Bridgehead Radicals of 1, 8 and 44-62 .................. 36 Rate Constants for Reactive Bridgehead Systems ............................. 40 HF/6-31G* Total Energies, Strain and Bond Dissociation Energies ....................................................................................... 40 Table 2.1 Table 2.2 Table 2.3 Table 2.4 Table 2.5 Table 2.6 Table 3.1 Table 5.1 Table 5.2 Table 5.3 Table 5.4 Table 5.5 CHAPTER 2 Calculated Heats of Formation, Strain Energies and Bond Dissociation Energies ................................................................. 58 INDO Predicted Hyperfine Coupling Constants (in G) for l-Manxyl Radical 2 ............................................................................ 62 UHF/6-31G“ (PM3) Geometrical Parameters for l-Manxyl Radical 2 ........................................................................................ 67 PM3 Atomic Cartesian Coordinates (in A) for l-Manxyl Radical 2 ........................................................................................ 68 UHF/6-31G“ Atomic Cartesian Coordinates (in A) for l-Manxyl Radical 2 ............................................................................ 69 Calculated Heats of Formation and Strain Energies .............................. 85 CHAPTER 3 Calculated Heats of Formation, Strain Energies and Bond Dissociation Energies ............................................................... 1 14 CHAPTER 5 13c NMR Chemical Shifts and Experimental ‘1 ,,C_,H Couplings ..... 166 Experimental One-Bond C-H Spin-Spin Coupling Constants (in Hz), and Calculated % sc Character of the C Hybrid F orrning the C-H Bonds in 1-39 ........................................................................... 168 Previously Reported Correlations of Experimental One-Bond C-H Coupling Constants with Hybridization, Bond Angles or Atomic Charges in Hydrocarbons ............................................................... 174 Semiempirical Relationships between Experimental One-Bond C-H Couplings and Hybridization, C-H distance, C-H Bond Order, Natural Atomic Charges on Carbon and Hydrogen, or Intemuclear Angles, Established by Least-Squares Analysis for the PM3 Optimized Geometries of Hydrocarbons 1-39 ................................................... 176 Semiempirical Relationships between Experimental One-Bond C-H Coupling Constants and Hybridization, Natural Atomic Charges on Carbon and Hydrogen, or Intemuclear Angles, Established by Table 6.1 Table 6.2 Table 1A Least-Squares Analysis for the HF/6-3 lG* Optimized Geometries of Hydrocarbons 1-39 ............................................................... 184 CHAPTER 6 Experimental and Calculated Heats of Formation ............................ 197 Comparison of the Wiberg and Ibrahim/Schleyer Group Equivalents with those Derived from the Ab Initio Energies of Table 6.1 ...................................................................................... 206 APPENDIX PM3 and HF/6-31G* Calculated Parameters for 1-39 ................. 213 xii Figure 1.1 Figure 2.1 Figure 2.2 Figure 2.3 Figure 2.4 Figure 2.5 Figure 2.6 Figure 2.7 Figure 2.8 LIST OF FIGURES CHAPTER 1 Plot of pyramidalization angle, AXC’C, vs. ABDE in the bridgehead radicals of bicyclics 1, 8 and 44-62. .............................. 38 CHAPTER 2 (a) EPR spectrum (9.1 GHz) of l-manxyl radical in cyclopropane at -55 °C (g = 2.0024). (b) Computer simulation. ....... 56 The ENDOR spectrum of l-manxyl radical 2 in toluene at -50 °C. Insert: the central part of the ENDOR spectrum of 2, which reveals small HFCs at 0.19 and 0.08 G. ..................................................... 61 Assignments of the hyperfine coupling constants in l-manxyl radical 2. ........................................................................................ 62 Selected UHF/6-31G“ calculated geometrical parameters and experimental hyperfine coupling constants (in G) of bridgehead radicals. Legend: ACC'CWS, refers to the average CC'C angles in degrees, obtained as 2(4CC’C)/3. ..................................................... 65 I-IF/6-31G* geometry optimized structures of manxane l, l-manxyl 2, 1-bicyclo[2.2.2]octy122, and l-adarnanty123 radicals. Legend (C 3 refers to the axis of symmetry): a = C3C'Cp angle, and 0 = C 3C‘CpHp torsion angle, in degrees. ....... 66 EPR spectrum (9.065 GHz) of the N-alkoxyanilino radical obtained by spin trapping of l-manxyl radical 2 with TBN (g = 2.003). ........................................................................................ 76 Kinetics of decay of l-manxyl radical in methylcyclopentane at 23 °C: (a) variation with time of the concentration of l-manxyl radical 2; (b) plot of the inverse concentration of 2 against time. ....... 79 Mass spectra showing the E1 fragmentation of: a) [3.3.3]propellane 31 (retention time 3.6 min), and of the peaks with 3.6 min. retention time xiii Figure 2.9 Figure 2.10 Figure 2.11 Figure 2.12 Figure 3.1 Figure 5.1 Figure 5.2 Figure 5.3 in the chromatograms from the analysis of decomposition products of 2 in b) neat di-tert-butyl peroxide, and in c) cyclopropane, which are assigned to 31. .......................................................... Mass spectra showing the E1 fragmentation of: a) manxane 1 (retention time 6.4 min), and of the peaks with 6.2 min. retention time in the GC-MS analysis of the decomposition products of 2 in b) neat di-tert-butyl peroxide, and in c) cyclopropane, which are assigned to manxene 30. .......................................................... Mass spectra showing the E1 fiagmentation of the peaks with a) 7.22 and b) 7.41 min. retention times (see Scheme 2.2 for tentative assignments) in the chromatograms from the analysis of the decomposition products of 2 in neat di-tert-butyl peroxide. Mass spectra showing the E1 fragmentation of the peaks with a) 10.4 and b) 10.6 min. retention times (see Scheme 2.2 for tentative assignments) in the chromatograms from the analysis of the decomposition products of 2 in neat di-tert-butyl peroxide. ..................................................................... Mass spectra showing the E1 fragmentation of the peak with 14.4 min. retention time (presumably l-benzylmanxane) in the chromatogram that resulted from the analysis of the decomposition products of 2 in neat di-tert-butyl peroxide. ........... CHAPTER 3 The EPR spectra (9.1 GHz) resulting from UV photolysis of manxine in a) di-tert-butyl peroxide/cyclopropane, and in b) AIBN/cyclopropane, at -90 °C. .............................................. CHAPTER 5 Experimental one-bond C-H spin-spin coupling constants vs. percent 8 character of the C hybrid in the OH bonding orbital obtained fiom NBO analysis of: a) PM3, and b) I-IF/6-31G* optimized geometries of 1-39. ....................... Plot of experimental vs. calculated (with semiempirical relationship 29, Table 5.4) one-bond C-H spin-spin coupling constants in 1-39. ..................................................................... Experimental one-bond C-H spin-spin coupling constants in 1-39 against: a) PM3 natural atomic charge on hydrogen, qH (61 data points, 38a is excluded fiom the correlation), and xiv ...... 82 ...... 83 ....... 89 ....... 90 ........ 91 ..... 120 ..... 177 ...... 181 Figure 6.1 b) PM3 atomic orbital coefficient on hydrogen (61 data points, 38a is excluded from the correlation). ............................ 182 CHAPTER 6 Plot of experimental heats of formation, AHf(exp), vs. calculated values, AHt(calcd), from the HF/6-31G* group equivalents evaluated in this work, for the compounds in Table 6.1. Slope 1.00 was taken for the correlation line. ................. 207 Scheme 2.1 Scheme 2.2 Scheme 3.1 Scheme 3.2 Scheme 4.1 Scheme 4.2 Scheme 4.3 Scheme 4.4 LIST OF SCHEMES CHAPTER 2 Manxane synthesis ............................................................................ 53 Analysis of products from decomposition of l-manxyl radical ....... 87 CHAPTER 3 Manxine synthesis ............................................................ I ............... 110 Newcome’s synthesis of 1-azoniatricyclo[3.3.3.0]undecane chloride ...................................................................................... 111 CHAPTER 4 Synthesis of 3-(2-hydroxyethyl)-1,S-pentanediol ............................. 133 Synthesis of 4-substituted tetrahydropyrans ........................................ 136 Synthesis of the tris acid chloride of methanetriacetic acid and of the tris(N-benzyl)methanetriacetamide ........................................ 138 Synthesis of tris(o-hydroxyphenyl)methane ........................................ 142 “Through doubting we come to questioning and through questioning we come to the truth” Peter A belard CHAPTER 1 ODD-ELECTRON 0' BONDING IN MEDIUM-RIN G BICYCLIC BRIDGEHEAD RADICALS: A THEORETICAL INVESTIGATION Abstract: The study of interactions and chemical reactions between two bridgehead atoms in medium-ling systems is reviewed. Intrabridgehead o-type bonding in bicyclic carbon-centered radicals with various donors and acceptor heteroatoms is examined by semiempirical (MNDO) and ab initio (HF/6-3 lG*) methods. It is found that the tertiary C-H bond dissociations that yield bridgehead radicals of the symmetrical [3.3.3] bicyclics investigated are considerably lower (by 5 to 26 kcal/mol at the HF/6-31G* level) than for tert-butyl radical, the prototype tertiary alkyl radical. Intrabridgehead o-bonding can amount to as much as 18 kcal/mol of the stabilization energy, with the highest values for the radicals where the opposite bridgehead site is occupied by aluminum. The computational results suggest that medium-ring bridgehead radicals might show unusual stability and/or persistence as well as interesting bridgehead-bridgehead interactions. 1.1 An Overview of Research on Odd-Electron o Bonds Two atoms’ o-type interactions can occur in four topological situations: ........ e e Intermolecular Intramolecular Transannular Intrabridgehead Furthermore, these interactions can be direct or via an intervening atom, e. g. as in hydrogen bonding, and may be classified by the number of electrons involved. One- and three-electron bonds play an important role in radical and electron transfer chemistry, and in many gas-phase processes involving radical ions. Experimentally, one-electronl and three—electron bonds2 are abundant and well-characterized. Three-electron bonding is a general concept that can be applied to many different bonding situations in both paramagnetic and diamagnetic molecules.3 Numerous examples have been reported and substantiated by experimental data and theoretical calculations; various (R2S SR2)+ radical cations,‘ (RS SR)‘ radical anions,5 and R2S.. SR neutral radicalsfl’fi have been identified, as well as N.-.N,7 P.-.P,8 As.-.As,8‘"9 Se.-.Se,10 I.-.I,ll and a wide variety of heteronuclear X.'.Y two-center three-electron bonds. ‘2 Having an odd electron, one- and three-electron bonded species are, in general, reactive and their generation and characterization, particularly in fluid solutions, is not necessarily straightforward. In rigid matrices, however, such species can be produced in situ by radiolysis or photolysis, and the observation of reactive species can be carried out at leisure. EPR spectroscopy has played a major role in the study of odd-electron bonded species,l3 along with pulse radiolysisl4 and mass spectrometry”. First described by Pauling in 1931,16 odd-electron bonds owe their stability to resonance between two limiting localized Lewis structures that are mutually related by charge transfer, as shown in (1) for one-electron bonds, and in (2) and (3) for typical three—electron bonds. A'B+ <—> NE (1) A°+B: e) A: B+ (2) A' B :' <—-> A:' B' (3) The bond strength depends on the energy difference between the two resonance structures (i.e. the difference in ionization potential between A and B), and stabilization will be significant only if the resonating structures are of almost equal energy. Thus, Meot-Ner et al. 17 observed that bonding energies in radical dimer cations of aromatic compounds are largest in symmetric associations and decrease as the difference in the ionization potentials of the neutral fragments increases. In (C6H6)2+ the dissociation enthalpy was measured as 17 kcal/mol, while in CeHeh €st it was about 11 kcal/mol; the difference, 6 kcal/mol, is ascribed to charge transfer resonance. ‘7 In MO theory the stability of one- and three-electron bonds is considered to arise from the fact that they possess one net bonding electron in the MO’s of the AB species. The formal AB bond order is 1/2 for both one- and three-electron bonds; for a single electron, the occupancy of the bonding MO yields a bond order of 1/2, whereas for the three-electron case the bonding effect of one of the two bonding electrons is canceled by that of the antibonding electron, leaving one net bonding electron. If overlap is included, the antibonding orbital is more destabilized than the bonding orbital is stabilized, originating a distinctive bond strength dependence on overlap for three-electron bonds. 18 The quantum-mechanical foundation of this unusual dependence on overlap for three- electron—bonded systems is illustrated by the orbital splitting diagram represented below, where the symbols (1, B and S represent the Hiickel coulomb, resonance, and overlap integrals, respectively. I \ I \ | I I l a 4+4 . \ \ \ l - .5 I ._H_, Since the advantage in magnitude of destabilization over stabilization increases with S, then under some circumstances the destabilization of a single-electron occupying the antibonding MO will outweigh the total stabilization of the two electrons occupying the bonding MO. Thus, if the maximum strength of a three-electron bond is half that of the corresponding two-electron bond, the strength of the bond falls off rapidly with increasing overlap integral. A simple mathematical evaluation fi'om the above expressions shows the interaction associated with a three-electron bond involving two initially degenerate levels to be net destabilizing if S exceeds 1/3; for nondegenerate levels the crossover from stabilization to destabilization occurs at even smaller overlaps. As the gap energy between initial levels is increased, the numerical advantage held by destabilization of the antibonding MO over stabilization of the bonding MO increases. In addition, electron repulsion in the three-electron case is a problem not explicitly acknowledged in the Huckel formalism; e. g. consider H2" which dissociates to H‘ and H'.” The bonding in many diatomic cations, including Hez'i, Nez'i, Krz'i, and Xez'i, involves two—center three-electron bonds and is well-characterized.20 A summary of some of the available data is shown in Table 1.1; however, the focus here is on odd-electron bonding through heteroatoms in organic molecules and despite the ample observations of such species, few experimental data exist even hinting at the strengths of their odd- electron bonds, and detailed thermodynamic data are remarkably sparse. Meot-Ner and F ield21 obtained thermodynamic parameters for the association reactions of CO'+ and N2'+ radical cations and of even-electron HCO+ and NzH+ ions with CO and N2, by equilibrium studies in pulsed high-pressure mass spectrometry. Alder et al.22 deduced bond energies for N.-.N three-electron bonds in radical cations of polycyclic diamines, fi'om ionization energies and proton afiinity measurements of the neutral amines, and fi'om kinetic decomposition studies of the radical ions. Illies et al.‘“*"’11b reported gas- phase measurements on the strength of iodine-iodine and sulfur-sulfur three-electron bonds, estimated from ion-molecule time-resolved equilibrium studies carried out in the high-pressure ion source of a mass spectrometer. Apart fi'om these experimental determinations of two-center three-electron binding energies, most of the information about odd-electron bonding energies comes from theoretical studies.23 Clark” has carried out systematic ab initio molecular orbital calculations on series of one- and three-electron bonded radical cation complexes of first- and second-row elements Li-Ar and their hydrides to address the question of whether significant 0 bonding Table 1.1 Experimental One- and Three-Electron Bond Energies Bond Energy' Reaction (kcal/mol) H; —)H++H 64.4“ Li; —)Li+ +Li‘ 29.4“ Naz'+ —> Nai + Na' 2271.: K; —> K”: + K' 18.31d Hez'+ —) Hei + He' 57426 N<32'+——>Ne+ +Ne' 31.12c Ar; _) Ar* + Ar' 28.82“ Xez'+ —-> Xe+ + Xe' 23 F2“ —+ F+ + F 29.72" €12" —+ C1' + Cl' 29.12“ Br; —) Br' + Br' 26.2 12" —) I’ + I 24.3 IBr" —> Br" + I 23.1 ' From NIST Standard Reference Database 25, Structures and Properties, version 2.02, 1994, by Lias, S. G.; Liebman, J. F.; Levin, R D. and Kafafi, S. A, unless otherwise noted. can occur in a given situation, and to identify the factors affecting odd-electron bond dissociation energies. He found that first row elements form stronger odd-electron bonds than their second row equivalents, while hydrogen and helium form the strongest odd- electron bonds, up to 65 kcanol. Within a given row of the periodic table the alkali metals and the noble gases form the weakest odd-electron bonds. Asymmetric three- electron bonds are of special interest since they generally possess a lower stability than their homonuclear analogs, as a consequence of the electronegativity difi‘erence between the two atoms. Clark24 proposed a general equation (4) to predict dissociation energies of both one- and three-electron bonds in unsymmetrical complexes: DAB = % (DAA + D133) exp (- AA AB Alp) (4) where DM, D33 and DAB are the binding energies of the symmetric and unsymmetrical dimers, Alp is the difi‘erence in the ionization potentials of A and B (the energy required to transfer an electron from one partner in the complex to the other), and IA and AB are adjustable preexponential factors characteristic of the elements involved. In preceding computational studies, Clark25 found an analogous exponential decrease of DAB in the three-electron bonded radical cation complexes of HCl, H28 and PH3, with increasing difference in Am. Also, since Alp can only be small for charged species, Clark24 concluded that neutral odd-electron bonded complexes should be weakly bound in the gas-phase because the charge transfer in (5) is strongly endothermic, but nevertheless, they may be stabilized in solution. AB —> NE (5) Gill and Radom26 performed similar calculations on the first- and second-row ion dimers Hez'i, (NH3)2'+, (1120);, (HF)2'+, Nez'i, (PH3)2'+, (H2S)2'+, (HC1)2'+, and Arz'i, to examine whether they exist as hydrogen-bonded ions or as hemibonded species, the latter involving binding through heavy atom-heavy atom three-electron bonds. The hydrogen- bonded systems are favored for all the first-row elements, while for the remaining second- row systems, the hemibonded isomers are preferred. Gill and Radomz‘5 calculate remarkably strong three-electron hemibonds with energies greater than 41 kcal/mol, concluding that if rearrangement to hydrogen-bonded species is precluded by appropriate substitution, the hemibonded species examined should be readily observable. The calculations performed by Clark24 and by Gill and Radom26 were carried out at both unrestricted Hartree-Fock (UHF) and Moller-Plesset perturbation (MP) levels, and exhibited important and sometimes intriguing features: (i) high levels of ab initio theory, including in particular electron correlation, are necessary to predict accurately odd- electron bonding energies; (ii) the MP2 level is satisfactory and provides geometries and bonding energies in good agreement with higher orders of perturbation theory; (iii) puzzlingly, UHF optimized geometries of odd-electron bonded species are similar with MP2 geometries, whereas bonding energies are exceedingly underestimated; (iv) the HF error is nonsystematic and always large for three-electron bonds, while the error is smaller in the case of one-electron bonds. A nonempirical remedy for the HF bias was proposed by Hiberty et al.23°; their established Uniform Mean-Field Hartree-Fock (UMHF) procedure involves orbital occupancy constraints and correction of the UHF resonance energies by nonempirical factors, and provides a routine inexpensive tool for obtaining odd-electron bond energies for large molecules. The UMHF approach was tested on one- and three-electron bonded systems, and was shown to yield bonding energies in satisfactory agreement with more sophisticated calculations (up to and beyond fourth order MP perturbation theory)?“ In contradiction with Clark’s24 theoretical results, J anssen et al.27 showed that three-electron bonded phosphoranyl radicals, R3P..SR 3‘1 (n = 0,1,2), are formed despite an unfavorable balance of the ionization potentials of the two fi'agments involved, implying that a large number of heteronuclear three—electron bonds between a variety of elements should be experimentally accessible in solution and in the solid-phase. Also, despite the fact that theory predicts first-row three-electron complexes to be more stable than their second-row analogues, most systems studied are formed from second-row or heavier atoms. Apart fi'om F2", first-row systems are rare. Recently, evidence for H3N.-.NH3'+ radicals has been presented,28 but these centers were very unstable, giving NH; at ca. 140 K, presumably via the reaction: H3N..NH3 —> NH3 + NH3°+ (6) In particular, cases where carbon participates in odd-electron bonding are relatively rare and poorly characterized.29 Regardless of disagreements between experiment and theory, computational chemistry remains a powerful tool that, when judiciously used, can help predict or confirm daring hypotheses. In this work, intrabridgehead o-type interactions of bicyclic carbon- Centered radicals with various donor and acceptor heteroatoms are examined by semiempirical and ab initio molecular orbital calculations. The results suggest that such 10 species might show unusual stability and/or persistence as well as interesting bridgehead- bridgehead interactions. 1.2 Intrabridgehead Interactions in Medium-Ring Bicyclics Our interest in through-space perturbation of unpaired electron centers has drawn us to the rich potential of intrabridgehead chemistry.30 The special structure/ strain relationship in medium ring bicyclic fi'ameworks has allowed the construction of many unusual chemical entities such as 1- and 3- electron bonds,31 symmetrical C-H-C hydride-bridged carbenium32 and N-H-N hydrogen-bonded ammonium cations,33 intrabridgehead donor- acceptor complexes,34 hyperstable olefins,3s near-planar aliphatic amines,36 stabilized bridgehead carbocations,” and rapidly autoxidizable arkanes”. In the following, the study of interactions and chemical reactions between two bridgehead atoms in medium-ring systems will be reviewed. Bicyclic compounds have essentially rigid molecular frameworks and well defined structures, and thus, allow control of orbitals and bonds toward a desired alignment. The optimum chain length/ring sizes for enforcing o-type interactions between the two bridgeheads are likely to be in the range of 3 to 5 atoms for each bridge to permit close approach of the bridgehead atoms without developing strain. Geometrical control of the intrabridgehead relationship provides an opportunity for the careful examination of fundamental questions of structure and bonding. 11 1.2.1 Closed-Shell Interactions in Neutral Medium-Ring Bicyclics The strain energy39 of medium-ring bicyclics is mainly due to nonbonded interactions between the bridges and torsional strain. To avoid intolerable H/H steric repulsions, bond angles are opened up, causing increased angle strain and framework rehybridization. In addition, bicyclic ring systems with large enough bridges to allow in/out isomerism40 are conforrnationally very complex. The prediction and understanding of possible conformations is a difficult matter; the borderlines where in,out- and in,in-isomers become possible are by no means obvious and depend strongly on the bridgehead atoms and their substituents. Whereas the out,out-, in,0ut- and in,in-isomers in compounds with carbon bridgeheads are separated by high barriers, the situation is quite different for bridgehead amines where nitrogen inversion allows equilibration of the isomers. The question of the relative thermodynamic stability of out,out-, in,out-, and in,in- isomers of bicyclic hydrocarbons is amenable to molecular mechanics (MM) calculations. Saunders41 used the stochastic (or Monte Carlo) search method for 32 bicyclic hydrocarbons ranging from bicyclo[3.2.2]nonane to bicyclo[6.6.6]eicosane, to locate all isomers and predict thermodynamic preferences (see Table 1.2). As expected, output- isomers are strongly preferred for systems built from small- and common-sized rings; their strain energy grows rapidly as the sizes of the constituent rings increase, reaching a maximum in the [4.4.4]system. Out,out-bicyclo[4.4.4]tetradecane, built entirely from ten- membered rings, has a strain energy which is more than three times that of cyclodecane‘z. Because of this high strain energy, input-isomers become preferred to output in medium- 12 ring systems. According to Saunders’ calculations, the input-isomers become the most stable for several bicyclotridecanes ([4.4.3], [5.3.3], and [542]), while the in,in-isomer is the most stable for bicyclo[5.5.5]heptadecane. Table 1.2 MM2' Steric Energies of Lowest Energy Conformations for Some Bicyclic Hydrocarbonsb Steric Energyc (in kcal/mol) Compound out,out in,out in,in Bicyclo[3 .2.2]nonane 24.3 81 .4 - Bicyclo[3.3.2]decane 29.9 66.8 130.2 Bicyclo[3.3.3]undecane 37.3 - 119.6 Bicyclo[4.3.3]dodecane 48.6 55.8 93.5 Bicyclo[4.4.3]t1idecane 58.4 54.8 82.4 Bicyclo[4.4.4]tetradecane 68.7 56.5 71 .9 Bicyclo[5.4.4]pentadecane 64.9 55.0 63.6 Bicyclo[5.5.4]hexadecane 63.8 54. 8 57. 1 Bicyclo[5.5.5]heptadecane 60.8 54.2 49.8 ' Allinger, N. L. J. Am. Chem. Soc. 1977, 99, 8127. b Reproduwd from ref. 41. ° The sum ofbond stretching, angle bending, torsion, and van der Waals terms, that form the force-field, is called the steric energy of a molecule; steric energy can be roughly interpreted as strain energy, and steric energy differences between stereoisomers can properly be understood as strain energy differences. In many respects, the most interesting cases are those where all the bridges are of the same length, and especially the symmetrical [3.3.3], [4.4.4], and [5.5.5] hydrocarbons. BicYclo[3.3.3]undecane 1 (manxane) was first prepared in 1970 as the prototype 13 compound which comprises together three eight-membered rings.43 The conformations of manxane and of all known derivatives indicate the output C31, symmetry conformation to be the energy minimum, but even this arrangement is strained in contrast to the flexibility observed for most eight-membered rings. Ab initio calculations carried out at the HF/6- 31G* level estimate a strain energy of 28.0 kcal/mol for manxane, in good agreement with the 27.2 chmol experimental value (Table 1.4). One structural manifestation of the strain is flattening of the bridgehead regions accompanied by widening of the angles in the bridges. X-ray structures of 1- azabicyclo[3.3.3]undecane (manxine) hydrochloride 2 and bicyclo[3.3.3]undecane-1,5-diol 3 show the expected structural features.44 The electron-difi‘raction data from manxane vapors also confirms both the bridgehead flattening and the C31. molecular symmetry.45 There is only limited experimental evidence concerning bicyclo[4.4.4]tetradecane and derivatives, while bicyclo[5.5.5]heptadecane is unknown. Saunders’ calculations41 (Table 1.2) predict the input-isomer as the most stable for the former and the in,in-isomer for the. latter, but clearly suggest that all isomers should be isolable. McMurry and Hodge32'4’ prepared in-bicyclo[4.4.4]tetradec-1-ene 5 in 30% yield by Ti-induced cyclization of 6-(4-oxobutyl)cyclodecanone 4 and were able to hydrogenate it slowly to input-bicyclo[4.4.4]tetradecane 6 (5 is a “hyperstable olefin”, where the alkene is less strained than the corresponding alkane). In addition to 6, a small amount of an isomeric 14 product (presumably the output-isomer, calculated to be 7.4 kcal/mol less stable than 5) was obtained in the cyclization reaction, but no further work on this material has been T1Cl3 Zn/Cu Pd/C H 6 The symmetric monoamines 1-azabicyclo[2.2.2]octane 7 (quinuclidine),46 manxine reported. 8,36‘I’43'4’ and out-6H-1-azabicyclo[4.4.4]tetradecane 9 (hiddenarnine),“7 form a series in which the nitrogen atom appears to be successively pyramidal out, essentially flat, and pyramidal in. Structural data on 8 and 9 have not been obtained, but the photoelectron spectrum of 8 is indicative of a flat amine,48 and the X-ray structure of the outside protonated ion of 1,6-diazabicyclo[4.4.4]tetradecane 10,49 a compound which should be structurally similar to 9, reveals an input conformation. 7 8 9 l-A2abicyclo[4.4.4]tetradec-5-ene 11 also appears to have an inwardly pyramidalized nitrogen since its photoelectron spectrum indicates a strong lone pair/n—bond interaction, and 11 reacts rapidly with acid to form the saturated azoniapropellane salt 12.50 15 10 11 12 ‘2 In the bicyclic bridgehead diamine series: 1,4-diazabicyclo[2.2.2]octane (DABCO) 13 is output,51 1,5-diazabicyclo[3.3.3]undecane 14 most likely has nearly flat nitrogens m KW F) V“ as We} 13 14 15 according to its photoelectron spectrum,52 and 1,6-diazabicyclo[4.4.4]tetradecane 15 adopts an in,in structure established by X—ray crystallography53 . The structure/strain situation in medium-ring bicyclic compounds forces inverting atoms like nitrogen to have inside lone pairs with interesting chemical consequences. Thus, any process that allows outside pyramidalized bridgehead atoms to planarize or pyramidalize inward brings considerable relief of strain. The most effective strain-relieving process, however, is intrabridgehead bond formation. Alder31 tried to estimate the thermodynamics for this process by calculating the energetics of the hypothetical dehydrogenation reaction which removes the bridgehead hydrogens from a bicyclic ring system and forms a propellane. His results (see Table 1.3) confirm once more that intrabridgehead bond formation brings relief primarily in medium-ring bicyclics. The chemistry of propellanes has been very well reviewed.S4 In small—ring propellanes the bridgehead carbons are severely distorted from the tetrahedral geometry 16 Table 1.3 Heats of Some Formal Dehydrogenationsa Dehydrogenation Heat of Dehydrogenationb Bicycloalkane Product (kcal/mol) Bicyclo[l . 1. 1]pentane [1 . 1 . 1]Propellane +39 Bicyclo[2.2.2]octane [2.2.2]Propellane +67 Bicyclo[3.3.3]undecane [3.3.3]Propellane -5 Bicyclo[4.4.4]tetradecane [4.4.4]Propellane -36 Bicyclo[5.5.5]heptadecane [5.5.5]Propellane +0.5 ‘ Only output isomers were considered. Reproduced from ref. 31b. b These results, presumably MMZ calculations, correlate reasonably well with those presented below in Table 1.4, with the exception of [2.2.2]prope11ane, where our calculated HF/6-3 16* structure is more strained, leading to a difference in the formal heat of dehydrogenation of 24 kcal/mol between M and HF/6-3 lG“ computations. Table 1.4 Heats of Formation and Strain Energies of Some Bicyclic Hydrocarbons and Propellanes Heat of Formation' Strain Energy” Compound Qccal/mcm (kcal/mol) Bicyclo[l . 1. 1]pentane 49.7 67.6 [1 . 1. 1]Propellane 84 97.9 Bicyclo[2.2.2]octane -23 9.7 [2.2.2]Propellane 68c (96.7) Bicyclo[3 .3 .3]undecane -21 .2d 27.2 [3.3.3]Propellane -28.7c (14.9) input-Bicyclo[4.4.4]tetradecane -10.5° 25.9 _[4.4.4]Propellane -43.6° (14.8) From NIST Standard Reference Database 25, Structures and Properties, version 2.02, 1994, by Lias, S. G. ,Liebman, J. F. ,Levin, R D. and Kafafi, S. A., unless otherwise noted. bStrain energy, from experimental (calculated) heats of formation and the Benson group equivalents (Benson, S. W. Thermochemical Kinetics; John Wiley: New York, 1976). cThis work; from HF/6-3 16* total energies ([2. 2. 2]propellane -309. 80932 H; [3.3. 3]propellane -427. 04326 H; in put-bicyclo[4. 4. 4]tetradecane -545.24313 H, [4. 4. 4]propellane -544. 14676 H) and the Wiberg group equivalents (Wiberg, K. B. J. Org. Chem. 1985, 50, 5285). Parker, W. Steele, W. V.’ ,Stirling, W. ,Watt, I. J. Chem. Thermodyn. 1975, 7, 795. 17 for central bond formation. The distortion is extreme in [1.1.1]propellane 16 where each bridgehead carbon is “inverted” with all four bonds to one side, while the hybridization at the bridgehead carbons in [4.4.4]propellane 18 is close to the normal sp3.55 16 17 18 The bond angle distortions in propellanes lead to both strain and unusual reactivity. The strain is lower in medium-ring propellanes, where only modest distortion of the bridgehead carbons is required to permit bonding. That intrabridgehead bond formation brings strain relief in medium-ring bicyclics is reflected in the lower strain energies in propellanes 17 and 18 than in the corresponding bicyclic hydrocarbons (Table 1.4). 1.2.2 Atranes Heterobicyclic esters of triethanolamine (TEA) are commonly known as “atranes”.34 This term was extended to define general structures of the type ZE(YCH2CH2)3N, where Y = CH2, 0, S or NR (e. g. when Y = NR the prefix aza is inserted) and E presently extends fi'om group 1 to group 15.56 Qualitatively, atranes can be viewed as donor-acceptor bonded propellanes and may be differentiated with respect to the strength of this transannular dative interaction. The intrabridgehead distance in atranes is quite variable, changing fiom the sum of the van der Waals radii of the atoms E and N or higher, as depicted in A, through intermediate distances, represented by B, to firll transannular 18 bonds, as shown in C. Z <:E\"' > Y’E\j HY) C: E"§> Pro-atrane Quasi-atrane Atrane A B C Main group element atranes (e. g. E = B, Al, Si, P) have been the most comprehensively studied. In particular, silatranes have aroused widespread interest not only among synthetic and structural chemists but also among pharmacologists and physiologists. The discovery of the high toxicity and specific biological activity of 1- arylsilatranes (e. g. l-phenylsilatrane is about twice as toxic as strychnine or hydrocyanic acid)57 originated an extensive search for new types of biologically active organosilicon compounds. Thus, many practically non-toxic or low toxicity silatranes display specific biological and pharmacological activity, having a broad spectrum of action with applications in health, agriculture, and industrial microbiology.58 The most intriguing aspect of atrane structure is the existence of the transannular dative bond, which leads to hypervalent bridgehead atoms and unique physical and chemical properties. The validity of this intrabridgehead interaction was initially demonstrated by Voronkov59 in silatranes, based on dipole moment measurements and infrared absorption spectra. Further overwhelming experimental data from X-ray crystallography,‘50 XPS, infrared spectroscopy, mass spectrometry, and NMR featuring several isotopes (III, 13C, 15N, 29Si, 27Al, 11B, 31P), confirmed this hypothesis.61 The 19 strength of the intrabridgehead bond is stereoelectronically controlled, depending on the electron-withdrawing power of Z and the steric properties of E and Y substituents. In silatranes the N-Si internuclear distance has been found to range from 2.89 A to 1.96 A60 These distances are considerably shorter than the 3.5 A sum of the van der Waals radii, yet they are longer than conventional N-Si covalent bonds of 1.7-1.8 A found in tetracoordinate silicon compounds. Structural correlations have been made between the N-Si bond length and a variety of parameters. The Si atom displacement (ASi) out of the plane of the three oxygens is linearly dependent on the N-Si distance.62 Tafi’s polar inductive parameter, 6*, of the substituent (R) attached to Si, as well as Si—R bond lengths, vary linearly with N-Si distance, and with 15N chemical shifts.63 The N—Si bond length decreases with increased electronegativity of R; considerable charge transfer from N to Si is observed by XPS when Si is bound to a very electronegative substituent.64 The anticipated increase of the binding energy of N1, and the decrease of that of Si;p in silatranes relative to TEA and triethoxysilane, was confirmed by X-ray photoelectron measurements. In addition, the correlation between N, and Sigp binding energies in silatranes with difi‘erent substituents on silicon, proves the existence of the intrabridgehead interaction.65 Voronkov et al.66 estimated the strength of N-Si bonds in a variety of silatranes from thermochemical parameters and ionization potential data; bond energies between 7 and 22 kcal/mol were obtained, reflecting a progression with increasing electron-withdrawing power of the silicon substituent. Azatrane chemistry is expanded considerably by the presence of the nitrogen substituents. The steric hindrance resulting from stepwise substitution of the NH 20 fimctionalities with bulky groups leads to a significant weakening of the N-Si bond in silazatranes, correlated with 29Si deshielding and increases in lJs;.c and 213”: for the Z substituents.“ Verkade‘58 demonstrated a gradation of hypervalent N-P interactions in phosphazatranes. The N-P distance varies from 1.9 A to 3.2 A depending on the apical substituent, Z, on phosphorus. Well-developed transannular N-P bonds emerge when the phosphorus lone-pair is strongly polarized by a positively charged Lewis acid and are associated with substantial upfield 31P chemical shifts.” Unusual phosphorus basicity is found in proazaphosphatranes of the type P(RNCH2CH2)3N (where R = H, CH3, CHzPh) producing the unexpectedly weak conjugate acids HP(RNCH2CH2)3N+ (pK. ~ 27 for 20 in DMSO).7° The flexibility of these versatile nonionic superbases with respect to transannulation gives rise to new and exciting chemistry that has valuable implications for synthesis and catalysis. Thus, the commercially available N-methyl derivative of P- CH3\ /CH3 }h\\N/ ”I" L j) __’ proazaphosphatrane 19 has found applications as a superior catalyst for aryl isocyanate trimerization;ll as well as for silylation of hindered tertiary alcohols and phenols”. 1.2.3 Radical Cations of Medium-Ring Bicyclic Diamines, Disulfides and Diphosphines Alder73 has repeatedly stressed the unique chemistry of medium-ring bicyclic compounds and demonstrated the potential of the intrabridgehead situation for studying weak o-type 21 bonding. Thus, the persistence of medium-ring bicyclic diamine radical cations in solution was interpreted on the basis of through-space intrabridgehead interactions presumed to generate three-electron o bonds.22b The first persistent radical cation discovered by Alder et al.?4 was that of naphtho[3.3.3]diamine, 21. Subsequently, oxidation of a wide range of medium-ring diamines in solution led to long-lived radical cationsm’75 Lifetimes of more than a second in CH3CN at 25 °C are obtained for the radical cations of [3.3.3], [4.3.3], [4.4.3], [4.4.4], [5.4.3], and [6.3.3] diamines. The perchlorate salt of the [4.4.4]diamine radical cation 22 is indefinitely stable as a crystalline solid. Vogel et al.76 prepared 23, a modified [3.3.3] structure whose perchlorate salt is also stable as a solid. (A1.+ —|.+ —‘|.+ IhC\N/CH3-—l+ N 1% KW 21 22 23 24 The first ionization energies of such medium-ring bicyclic diamines are exceptionally low, and their photoelectron spectra show two bands separated by ~ 1 eV. Alder et aim argued that this splitting is a measure of the through-space interaction of the nitrogen lone-pair orbitals. Thus, 1,6-diaza[4.4.4]tetraundecane is oxidized at a less positive potential than N,N,N ’,N ’-tetramethylphenylenediamine, the diamine that produces the well-known and indefinitely stable Wiirster’s blue radical cation, 24. Furthermore, the ESR spectra of these bicyclic diamine radical cations show hyperfine coupling to two nitrogens." DABCO 13 also forms an unusually persistent radical cation 25 in solution 22 (tm ~ 1 s in CH3CN at 25 °C) which shows two equivalent nitrogens in its ESR spectra, in contrast to the transient quinuclidin-4-yl radical 26 which does not show any spin T 1 .. 25 26 27 delocalization at nitrogen; this result was rationalized primarily by through-bond long- range electron delocalization, however, rather than a three-electron bond.78 Alder et 111.22" estimated the stabilization resonance induced by three-electron bonding in radical cations of medium-ring diamines as the difference in the N—H bond dissociation energies of the protonated diamine and the analogous monoamine. An energy of 11 kcal/mol was obtained for the three-electron bond in both the radical cations of [3.3.3] and [4.4.4] diamines, in agreement with a previous estimate of 14.5 kcal/mol for the three-electron bond in 27 .22‘ Further oxidation of bicyclic diamine radical cations with loss of a second electron produces stable propellane hydrazinium dications with central N-N two-electron o-bonds, also prepared quantitatively by alkylation of bicyclic bridgehead diamines.79 Their reductive cleavage affords a convenient route to medium- ring bicyclic diarnines.77' X-ray structural data for all three oxidation states of diamine 15 show progressive shortening of the N-N distance from the neutral amine to the dication.80 Crystals of the perchlorate salt of 22 were obtained in acetonitrile by a remarkably slow one-electron transfer reaction from 15 to the diperchlorate salt of 28. The three-electron bond in 22 is perhaps one of the few established bond lengths in a three-electron case. Q. Q “d [on Q N-Ndig. 2.8 A 2.3 A 1.5 A 15 22 28 Medium-ring disulfides undergo facile oxidation, too, where cation formation occurs concomitantly with coupling of the two sulfiir atoms. ”‘8' Even though most thioethers are easily oxidized, only the eight- (e. g. the radical cation of 1,5-DTCO 29) and nine-membered rings give long-lived radical cations. Subsequent oxidation gives dications 29 30 31 having Si-S+ bonds. Cycles with a thioether group transannular to other groups with lone pair electrons undergo oxidative coupling reactions to give stable cations; evidence of N.-.S bond formation was also obtained in several arninothioethers.81b Similarly, two- electron Pi-P+ bonds are found in medium-ring cyclic and bicyclic diphosphines; the X-ray structure of 30 shows a P-P distance significantly shorter than in neutral diphosphines despite the adjacent positive charges.82 A series of nucleophilic adducts of 30 have been described, with Y-P-P+ bonding. As with Verkade’s atrane-type superbases, the adduct with Y = H, 31, was very dificult to deprotonate.82"’83 24 1.2.4 Medium-Ring Bicyclic Carbon-Centered Bridgehead Radicals There are only two examples in the literature of intrabridgehead odd-electron bonded complexes with carbon participation: the radical cation of [3.3.3]propellane 32,29' and the l-azabicyclo[3.3.3]tetradec—4-yl radical 33.29b @ . E I 32 33 Ion 32 was generated as a transient species in CCl4 or CBr, matrices by y-radiolysis of [3.3.3]propellane at 77 K. The ESR spectrum of 32 shows strong coupling to 6 equivalent hydrogens, characteristic of C31. symmetry of the radical cation. The bridgehead radical 33 was obtained by y-irradiation of 1-azoniatricyclo[4.4.4.01’6]tetradecane tetrafluoroborate, either as the pure salt or in dilute methanol solution, at 77 K. In both media, the ESR spectrum of 33 shows a quartet of broad lines, which is assigned to hyperfine coupling to the three pseudo-equatorial equivalent hydrogens adjacent to the radical center. The spin density on nitrogen is not higher than 5%. Thus, despite the ideal structure of the bicyclo[4.4.4] system for intrabridgehead bond formation, the three-electron bonding in the neutral radical 33 is very weak, in agreement with Clark’s24 calculations and Harcourt’s3 theoretical predictions that three-electron bonding is destroyed by too much overlap. 25 1.2.5 Bridgehead Phosphoranyl Radicals Hamerlinck et al.84 reported that X-ray irradiation of 34-BF4 at 77 K produces phosphoranyl PV radicals. Their structure has been proposed based on single-crystal ESR measurements, which hint at initial formation of 35, followed by an irreversible transformation to 36 with temperature increase. Also, 36 could be obtained directly from 34-BF4 by UV laser irradiation. The evidence for the unprecedented structure of 36, where the unpaired electron is in apical position, has been disputed by Roberts.85 He interpreted the results to be consistent with structure 37, where the odd-electron is localized in a P-N 0* molecular orbital, generating a three-electron bond between phosphorus and nitrogen, however, there is still no definite answer to this problem. C;> Cm C33 C3 1.2.6 Intrabridgehead Indirect Interactions via Hydrogen All of the observed indirect interactions across medium-ring bicyclic compounds involve hydrogen, as they are normally too small to accommodate anything larger. In macrobicyclic compounds, interactions via other atoms (or ions) are possible, but they will not be discussed here. The [1.1.1]cryptand, for example, can hold two hydrogens or one 26 lithium cation, but it is certain that the interactions with the oxygen atoms are as important as the intrabridgehead bonding and no evidence of complexation other than for protons has been found in the analogous bicyclic diamines.31b Alder et al.33 converted l3 medium-ring bicyclic diamines to inside protonated monocations, all of which have intrabridgehead hydrogen bonds, by slow, conventional proton-transfer or by redox-promoted rearrangements. X-ray structures were obtained for seven inside-protonated ions and show N-H-N distances varying from 2.47 to 2.69 A, and N—H-N angles ranging from 180° (in 38) to 132°. An interesting question is whether these intrabridgehead hydrogen-bonded ions have single or double minimum potentials. On the basis of the 8A(1H,2H) test85 for equilibrating and resonance structural distinction, also known as isotopic perturbation of equilibrium, all inside protonated ions have double minima structures except for the in[4.4.4]H+ ion 38.87 Neutron difi‘raction studies show the inside hydrogen atom in 38 to be central even at 20 K; the N-H-N distance in 38 (2.53 A) is the shortest known for a linear hydrogen bond.87 A chemical indication of the strength of these N-H-N hydrogen bonds is their resistance to deprotonation. In fact, upon treatment with strong base, 38 slowly undergoes redox- mediated loss of proton (i.e. loss from one of the CH2 sites) rather than “simple” loss of the internal H'. 27 u-Hydrido-bridged carbocations of medium-ring systems give rise to transannular interaction by C-H-C three-center two-electron bonding.88 They are characterized by the high field 1H NMR chemical shift of the bridging hydrogen, anomalously low coupling constants involving this hydrogen, and very small isotope perturbation shifis. Whereas monocyclic ions, such as 40,89 are susceptible to loss of hydrogen and rearrangement at higher temperature, in the intrabridghead situation, e. g. 41 and 42, the inside hydrogen is well enclosed in the caged structure, making escape sterically impossible and, thus, inducing kinetic stability.88 T (113 40 41 42 Three-center two-electron bonds can exist in two distinct types, often referred to as “open” and “closed” geometries, although it is recognized that intermediate geometries are "open" "closed" possible. Sorensen and Whitworth90 prepared a series of ions based on a bridged bicyclo[3.3. 1]nonyl ring as in the general structure 41, to examine the efi‘ect of C-H-C bending on u—hydrido bridging. When n = 5 the formal C+ center and the potentially bridging remote H-C group are close enough to develop a fully u-hydrido—bridged structure; for larger sizes of the polymethylene-connecting link, one sees a gradation of 28 structures with progressive C-H-C bending, leading for n = 8 back to a normal tertiary ion. Similarly, McMurry and Lectka88 built a set of bicyclo[x.y.z] carbocations (x, y, z = 2 to 6) as in the general structure 42, all of which showed three-electron two-center bonding and progressive bending of the central C-H-C1+ bond with decreasing ring size. Particularly, in-bicyclo[4.4.4]-1-tetradecyl cation 43 is one of the most stable carbocations known; 43 was obtained by protonation of bridgehead alkene 5, as well as upon a remarkable protonolysis of 6 in glacial acetic acid at 40 °C.91 Q H ' H '11 'Hz 5 43 6 The present review demonstrates the potential of intrabridgehead chemistry for studying weak o-type interactions. When the interacting bridgeheads are of the same type, odd-electron bonding becomes a significant and easily observed phenomenon. 1.3 Geometry, Strain and Odd-Electron o Bonding in Medium-Ring Bicyclic Bridgehead Radicals: A Semiempirical and Ab Initio I-IF/6-31G* Analysis The unique properties of medium-ring bicyclic compounds are intimately connected with special structure/strain relationships. For example, the experimental rates of solvolysis of bridgehead derivatives correlate well with the calculated strain (steric) energy differences between substrate and the intermediate carbenium ion.92 Solvolysis reactions occurring at 29 bridgehead positions are mechanistically simple and homogeneous, since most of the potentially competing pathways are forbidden for structural reasons. The relative rates of bridgehead derivatives are dominated by steric effects and essentially independent of leaving groups or solvent. On these grounds a unified reactivity scale for solvolysis of bridgehead derivatives was proposed,92c while the experimental data for solvolytic bridgehead reactivities were used to develop a revised force-field for tertiary carbenium ions.92c Radical reactivities also parallel those of the corresponding bridgehead carbenium ions.92"b Bridgehead compounds provide a calibrated series of widely varying reactivities, spanning 22 orders of magnitude, which permits a general, reasonable reactivity prediction for similar substrates. Rate enhancements of larger magnitude than in typical acyclic analogs have been reported for bridgehead systems.93 1-Chlorobicyclo[3.3.3]undecane (1- manxyl chloride) is ca. 104 times more reactive than tert-butyl chloride in solvolysis reactions?” Consistent with the results of solvolysis studies and the experimental observation that manxane reacts rapidly with atmospheric oxygen to produce a mixture of bridgehead peroxides and hydroperoxides, empirical force-field calculations suggest that enhanced reactivities at these sites are due to 6.8 kcal/mol relief of strain when the bridgehead converts to a trigonal center (sp3 —> sp2 rehybridization in the transition state)?“ All the CCC bond angles in manxane are considerably larger than the ideal tetrahedral value and a sp2 hybridized carbon is more readily accommodated at the bridgehead, also reducing the repulsive nonbonded interactions between the bridges.” The structure/reactivity relationship in manxane suggests the symmetrical [3.3.3] system as the archetypal medium-ring bicycle for studying intrabridgehead o-type interactions. 3O Semiempirical (MNDO)9‘ and ab initio (HF/6-31m)” calculations were performed on various symmetrical [3.3.3] bicyclics and their bridgehead carbon-centered radicals, to evaluate the bridgehead C-H bond dissociation energy and the strength of potential intrabridgehead odd-electron o-bonds in the radicals. The bond dissociation energies of the bridgehead C-H bonds in l, 8 and 44-62 were estimated relative to the tert—butyl radical (see Tables 1.5 to 1.8).96 In all cases, the calculated BDE’s (Tables 1.5 to 1.8) show strikingly low values for the tertiary C-H sites, consistent with the strain relief upon bridgehead flattening discussed above. Aliphatic carbon—centered radicals are considerably stabilized by lone pair donors or acceptors which can delocalize the unpaired electron through n-resonance as shown below.” In the radicals considered here the semioccupied orbital is collinear with the opposite bridgehead and their interaction occurs by o-delocalization. The substantial shortening of the BB distance (see Tables 1.5 to 1.8) in the bridgehead radicals of 1, 8 and 44-62, may suficiently augment contact of the two bridgeheads to form odd-electron 0 bonds. n-Delocalization o-Delocalization \N ....... .. \N \N / ‘ r"'7 “"7 The C-H bond dissociation energy differences (ABDE’s; see Tables 1.5 to 1.8) relative to tert-butyl radical reflect both the strain energy relief due to bridgehead flattening and the electronic stabilization by intrabridgehead a bonding. However, the ABDE’s of the radicals with carbon atoms in the opposite bridgehead represent 31 Table 1.5 HF/6-31G“ Total Energies (MNDO Heats of Formation), BDE’s, Intrabridgehead Distances and Radical Stabilization Energies Relative to the tert-Butyl Radical‘ in 1, 44 and 45 TE (R-H) TE (R') R-H (AHf) BB” (R-H) (7311.) 1313b (R‘) ABDE‘ BDE“ ABB" -428.l7907 3.401 -427.56808 3.097 8.1 87.9 0.304 (-217) (3.357) (-971) (3.024) (4.7) (91.3) (0.333) 1 42635559 3.012 -426.76778 2.172 22.7 73.3 0.840 (181.6) (3.006) (188.4) (2.369) (9.9) (86.1) (0.637) 44 T 427.50194 3.306 426.89565 2.491 11.9 84.1 0.815 C3 (-14.8) (3.035) (-4.8) (2.623) (6.7) (89.3) (0.412) 45‘1 ‘ HF/6-3 16" total energies (TE) are given in hartrees, 1 H = 627.5 kcal/mol; MNDO heats of formation (AH) and bond dissociation energies (BDE) are given in kcal/mol; distances are given in angstroms. b BB is intrabridgehead distance. ° Stabilization energy relative to tert-butyl radical; from isodesmic reactions vs. isobutane/tert-butyl radical. " Based on relative stabilities vs. tert-butyl radical and BDE (tert-Bu-H) = 96.0 kcal/mol (ref. 96). ° ABB is the difference between the intrabridghead distance of R-H and that of the corresponding bridgehead radical. d The earbanion calculations are done at 6-31+G* level, since a proper description of anions requires basis sets which incorporate diffuse functions. Total energies at 6-31+G* level: isobutane -157.31456 H; terr- butyl radical -156.68935 H 32 Table 1.6 I-IF/6-31G* Total Energies (MNDO Heats of Formation), BDE’s, Intrabridgehead Distances and Radical Stabilization Energies Relative to the tert-Butyl Radical in 8 and 46-50‘ TE(R-H) TE(R’) b b R-H (AHf) BB (R-H) (AHf) BB (K) ABDE BDE ABB° < 2 41442449 3.032 -413.82346 2.520 14.4 81.6 0.512 (-345) (3.023) (-252) (2.645) (7.4) (88.6) (0.378) 46 < Q) -63l.61403 3.171 -631.03176 2.474 26.2 69.8 0.697 A1, (-300) (3.128) (-26.1) (2.610) (12.8) (83.2) (0.518) 47 .3 -444.55189 3.119 -443.55189 2.801 8.6 87.4 0.318 N (.40) (3.124) (8.1) (2.784) (4.6) (91.4) (0.340) 8 C3) -679.24119 3.531 -678.64296 3.012 16.2 79.8 0.519 s1 (-48.0) (3.462) (-410) (3.060) (9.7) (86.3) (0.402) H 48 < )7 -730.84ll4 3.477 -730.24028 2.930 14.5 81.5 0.547 111 (-1301) (3.386) (137.737) (2.978) (9.4) (86.6) (0.408) 49 -730.44256 3.739 -729.83717 3.341 11.7 84.3 0.398 (-48.7) (3.530) (-399) (3.146) (7.9) (88.1) (0.384) 50 ' HF/6-3 lG“ 'I'Es are given in hartrees,; MNDO heats of formation (AH;) and bond dissociation energies (BDEs) are given in kcal/mol; distances are given in angstrdms. BB is intrabridgehead distance. b Stabilization energy relative to tert-butyl radical. BDE based on relative stabilities vs. tert-butyl radical and BDE (tert-Bu-H) = 96.0 kcal/mol (ref. 96). ° ABB is the difference between the intrabridgehead distance of R-H and that of the corresponding bridgehead radical. 33 Table 1.7 HF/6-3 16* Total Energies (MNDO Heats of F orrnation), BDE’s, Intrabridgehead Distances and Radical Stabilization Energies Relative to the tert-Butyl Radical in 51-56' TE(R-H) TE(R') b 1. R-H (AHf) BB (R-H) (AHf) BB (K) ABDE BDE ABB° O’V‘O ) -535.65528 3.238 -535.03929 3.011 5.0 91.0 0.227 C ‘0} (-127.6) (3.263) (-1124) (2.967) (1.5) (94.5) (0.296) 51 o— "0 B”~o ) -522.03574 2.827 -521.42545 2.548 8.6 87.4 0.279 C J (-l76.4) (2.840) (-l63.92) (2.519) (4.2) (91.8) (0.321) 52 ‘ o—-A1:‘0 0) -739.23863 2.996 -738.64586 2.413 19.6 76.4 0.583 C 7 (-182.3) (3.072) (-176.2) (2.633) (10.6) (85.4) (0.439) 53 o’Siz-O O -786.85057 3.373 -786.24312 3 .033 10.4 85.6 0.340 (-2217) (3.440) (-2121) (3.423) (7.1) (88.9) (0.170) 54 H lei. ‘ -838.36592 3.313 -837.75614 2.957 14.2 81.8 0.356 (30.1) (3.326) (41.2) (2.956) (5.6) (90.4) (0.370) 55 o’P‘x'ao C D -837.98495 3.568 -837.37264 3.297 7.3 88.7 0.271 56 (-196.4) (3.505) (-184.3) (3.173) (4.6) (91.4) (0.332) ' HF/6-3lG‘ 'I'Es are given in hartrees; MNDO heats of formation (Am) and BDEs are given in kcal/mol; distances are given in angstrbms. BB is intrabridgehead distance. b Stabilization energy relative to tert- butyl radical; fiom isodesmic reactions vs. isobutane/terr-butyl radical. Based on relative stabilities vs. tert-butyl radical and BDE (tert-Bu-H) = 96.0 kcal/mol (ref. 96). ° ABB is the difference between the intrabridgehead distance of R-H and that of the corresponding bridgehead radical. 34 Table 1.8 HF/6-31G“ Total Energies (MNDO Heats of Formation), BDE’s, Intrabridgehead Distances and Radical Stabilization Energies Relative to the tert-Butyl Radical in 57-62‘ TE(R-H) TE(R') b R-H (am) BB (R-H) (AHf) BB (K) ABDE" BDE ABB° HN 476.13947 3.360 47552549 3.114 6.3 89.7 0.246 (15.6) (3.323) (29.0) (3.006) (3.3) (92.7) (0.317) 57 NH HN— 5“) 46249187 2.935 -461.88426 2.638 10.3 85.7 0.297 (-31.0) (2.912) (-202) (2.572) (5.9) (90.1) (0.340) 58 HN—Al’m ‘ -679.67332 3.129 -679.07967 2.560 19.0 77.0 0.569 C i (-25.8) (3.120) (-205) (2.676) (11.4) (84.6) (0.444) 59 H Her—Cit!“ -728.28875 3.482 -727.68399 3.109 12.1 83.9 0.373 (-58.0) (3.455) (-500) (3.068) (8.7) (87.3) (0.387) 60 mi -778.85268 3.406 -778.24526 3.024 10.4 85.6 0.382 (161.9) (3.336) (171.271) (2.941) (7.3) (88.7) (0.395) 61 HN’P‘NH 50) -778.45020 3.610 -777.84163 3.279 9.7 86.3 0.331 (46.6) (3.525) (-36.0) (3.177) (6.1) (89.9) (0.348) 62 ' HF/6-31G" TEs are given in hartrees; MNDO heats of formation (AH) and BDEs are given in keal/mol; distances are given in angstrdrns. BB is intrabridgehead distance. b Stabilization energy relative to tert- butyl radial; fi'om isodesmic reactions vs. isobutane/tert-butyl radical. Based on relative stabilities vs. tert-butyl radical and BDE (tert-Bu-H) = 96.0 kcal/mol (ref. 96). ° ABB is the difference between the intrabridgehead distance of R-H and that of the corresponding bridgehead radical. 35 exclusively the strain energy changes upon radical formation. Accordingly, if the bridgehead radicals of 1, 51 and 57 are taken as references for their set of compounds, then the difl‘erence in the relative BDE’s for the other radicals can be approximated as a measure of stabilization by o—delocalization over the opposed bridgehead. Based on Clark’ s” findings, the cation radical 44 and anion radical 45, where the bridgeheads are of the same type, would give best (upper limits for the one- and three-electron BDEs) charge delocalized one- and three-electron bonds (see Table 1.5). The bridgehead C-H bond dissociation energy in manxine 8 is similar to the BDE of the bridgehead C-H bonds in 1 (Table 1.6), suggesting that there is no significant stabilization by o-delocalization in this case. Analogously, the EPR study of the quinuclidin—4-yl radical 63 revealed very little delocalization of the unpaired spin to the nitrogen,98 in contrast to the radical cation of 13,99 whose EPR spectrum shows two equivalent N’s. This lack of stabilization was considered to originate in the nondegeneracy @6 The bridgehead atoms are in close contact in all radicals considered in Tables 1.5- of the interacting orbitals. 1.8; in each one, the BB distance is shorter than the sum of the van der Waals radii of the bridgehead atoms (van der Waals radii: C 1.65 A, B 1.7 A, N 1.55 A, A1 2.15 A, Si 2.10 A, P 1.85 A)100 and may allow intrabridgehead o—type interactions (Table 1.9). The calculated geometries of the radicals presented below show inward pyramidalization of the radical center, having the semioccupied orbital directed toward the opposite bridgehead in 36 Table 1.9 Spin Densities' (p) and HF/6-31G“ Intrabridgehead Distances (BB) in the Carbon-Centered Bridgehead Radicals of 1, 8 and 44-62 R-H R’ Compound X BB BB ABB" p 23m,c CH (1) 3.401 3.097 0.304 0.00213 ' 2.30 "'“" B (46) 3.032 2.520 0.512 0.05324 2.35 > A1 (47) 3.171 2.474 0.697 0.04505 3.80 N (8) 3.119 2.801 0.318 0.01227 2.20 $111 (48) 3.531 3.012 0.519 0.00115 3.75 PH (49) 3.477 2.930 0.547 0.00156 3.50 p (50) 3.739 3.341 0.398 0.00167 3.50 OJ.” CH (51) 3.238 3.011 0.227 0.00030 2.30 ‘o B (52) 2.827 2.548 0.279 0.01836 2.35 C l) A] (53) 2.996 2.413 0.583 0.03489 3.80 SiH (54) 3.373 3.033 0.340 0.00455 3.75 P‘H (55) 3.313 2.957 0.356 0.00577 3.50 p (56) 3.568 3.297 0.271 0.00003 3.50 HNJ....Nn CH (57) 3.360 3.114 0.246 0.00140 2.30 ‘ B (58) 2.935 2.638 0.297 0.01834 2.35 C ”)1 Al (59) 3.129 2.560 0.569 0.02774 3.80 SiH (60) 3.482 3.109 0.373 0.00456 3.75 PH (61) 3.406 3.024 0.382 0.00636 3.50 P (62) 3.610 3.279 0.331 0.00042 3.50 44 3.012 2.172 0.840 0.49576 2.30 45 3.306 2.491 0.815 0.55237 2.30 ' Spin density at the bridgehead opposite to the radical center, calculated by NBC analysis of the HF/6-3 16* wave-functions. b ABB is the difference between the intrabridghead distance of R-H and that of the corresponding bridgehead radical. ° Sum of the van der Waals radii of the bridgehead atoms. 37 a favorable arrangement for o bonding. The pyramidalization is greatest for aluminum compounds (47, 53 and 59), whose carbon-centered radicals exhibit considerably high stabilization energies relative to the tert-butyl radical. The calculated bridgehead spin densities (p) on aluminum in 47, 53 and 59 are 0.04505, 0.03489 and 0.02774 atomic units, respectively (Table 1.9). Increased spin densities on the bridgehead opposite to the radical center are calculated also for the radicals of the boron-containing compounds 46, 52 and 58, of 0.05324, 0.01836 and 0.01834 atomic units, respectively (Table 1.9). 3.3 [BCC = 91.8° X = CH2: AAIC‘C = 97. 1° ASiC'C = 92.7° APC‘C = 92.8° X = O: éAlC’C = 93.8° X = N: AAJC'C = 94.1° A good linear correlation of the relative ABDE’s with the pyramidalization angle of the radical center, AXC'C, is obtained for all compounds included in Tables 1.5 to 1.8, A XC’C """" ’ x (correlation coeflicient 0.96; Figure 1.1). It is difficult, however, to separate the effects of strain energy relief from stabilization by intrabridgehead o bonding. As mentioned previously, the difl‘erence in the relative BDE’s of the bridgehead radicals vs. the reference radicals of bicyclics l, 51 and 57, can be viewed as an upper limit ABDE (kcal/mol) Figure 1.1 28 24 20 16 12 38 82 84 86 88 90 92 94 96 98 Pyramidalization Angle (degrees) Plot of pyramidalization angle, AXC‘C, vs. ABDE in the bridgehead radicals of bicyclics 1, 8 and 44-62 (the best fit was taken for the correlation line). 39 for stabilization by o-delocalization over the opposed bridgehead. Examination of ABDE’s from Tables 1.6 to 1.8 reveals that intrabridgehead o—bonding can amount to as much as 18.1 kcal/mol for the bridgehead radical of 47, which also exhibits considerable shrinkage of the BB distance and substantial inward pyramidalization of the radical center. The strain relief is large when Si or P is placed at the bridgehead, since due to longer Si-C and P-C bonds the methine bridgehead is more strained then in the parent hydrocarbon and one needs to “push” harder to flatten the bridgehead regions. The ABDE’s are smaller for the compounds fiom Tables 1.7 and 1.8 relative to those in Table 1.6. In the bridgehead radicals of 52 and 58 it is conceivable that boron is less available for o delocalization because of n-resonance with the lone pairs of the adjacent oxygen or nitrogen atoms, but it sure looks like aluminum (compounds 47, 53 and 59) offers good opportunities. Parker et al.93' used empirical force field calculations to predict bridgehead reactivities, in a quest to find systems significantly more reactive than tert-butyl chloride. Their data (Table 1.9) suggested 1-chlorobicyclo[4.4.4]tetradecane to be even more reactive than l-manxyl chloride. Our computational results (HF/6-31G“) show 21.7 kcal/mol strain energy relief when 6 is converted to the corresponding bridgehead radical 64 (Table 1.10). Radical 64, expected to be persistent by analogy with the corresponding bridgehead cation 43, appears to be an excellent objective for experimental investigation. UHF/6816* parameters: (11 = den = 1.07 A 0. d2=dHC.=1.93A ACCH = 104.7° ACC'H = 943° 64 40 Table 1.10 Rate Constants for Reactive Bridgehead Systems“ Predicted rate constants" Bingharn force field Engler force field Compound lr(exp)b ASE“ k(calcd) ASE“ k(calcd) 1-Bicyclo[3.3.3]undecyl Chloride 17.4 -6.77 2.9 -8.36 2.5 1-Bicyclo[4.4.4]tetradecyl Chloride - -1494 2.0x103 -20.8 1.2x105 ‘ Reproduced from ref. 93a. b Experimental rate of solvolysis in 80% ethanol at 70°C; in s". ° Calculated from the semiempirical correlations of experimental solvolysis rates in bridgehead chlorides with strain energy differences between substrate and the intermediate carbenium ron, estimated with various force fields, in s dStrain energy difference between carbenium ron and corresponding hydrocarbon. Table 1.11 HF/6-31G* Total Energies, Strain, and Bond Dissociation Energiesa Compound Total Energyb AH? SE" BDE° Bicyclo[4.4.4]tetradecane 6 -545.24313 -10.5 52.7 - 1-Bicyclo[4.4.4]tetradecyl Radical 64 -S44.65377 11.8 31.0 74.3 ' In kcal/mol; structures were fully optimized at the HF/6-3 16* level, using Spartan 4.0 (Wavefunction Inc., Irvine, CA). b Total energies are given in hartrees, 1 H = 627 .5 kcal/mol. ° Calculated fi'om Wiberg’s group equivalents (Wiberg, K. B. J. Org. Chem. 1985, 50, 5285). The BDE estimates were used to calculate values for radieal 64. d Strain energy; from calculated AH: and Benson’s group equivalents (Benson, S. W. Thermochemical Kinetics", John Wiley: New York, 1976) for 6, and from isodesmic reactions vs. isobutane/tert-butyl radical for 64. ° Based on BDE (tert-Bu-H) = 96.0 kcal/mol (ref. 96). 41 In a letter addressed to Professor James E. Jackson, McMurry101 wrote that cyclic voltammetry studies on the cation 43 showed a one-electron reduction to generate a persistent radical, but no ESR work was pursued filrther. The bridgehead bicyclo[4.4.4]- tetradec—l-ene is “hyperstable” and we might see rapid loss of 64 by conventional B- disproportionation. Nevertheless, radical 64 provides a unique opportunity for examination of a caged H abstraction where one can address the question of single- or double-well potential for 1,6-H° migration. Given the expected stability of the bridgehead radicals, significant hydrogen abstraction at other sites in these molecules seems unlikely. Hence, hydrogen abstraction from the parent compounds by tert-butoxyl radicals generated photolytically from di-tert- butyl peroxide should selectively produce the bridgehead radicals. Such species promise to become new examples with unforeseen properties in the already unique chemistry of medium-ring bicyclics. 1.4 References 1 For hydrogen and alkali radical cations see: (a) Bates, D. R.; Ledsham, K.; Stewart, A. L. Philos. Trans. R. Soc. London 1953, A246, 215. (b) Mathur, B. P.; Rothe, E. W.; Reck, G. P.; Lightrnan, A. J. Chem. Phys. Lett. 1978, 56, 336. (c) Carlson, N. W.; Taylor, A J.; Jones, K. M.; Schawlow, A. L. Phys. Rev. 1981, A24, 822. (d) Leytwyler, S.; Herrman, A.; Woeste, L.; Schumacher, E. Chem. Phys. 1980, 48, 253. For other one- electron bonded radical cations: (e) Gilbert, T. L.; Wahl, A. C. J. Chem. Phys. 1971, 55, 5247 and references therein. (1) Shida, T.; Kubodera, H.; Egawa, Y. Chem. Phys. Lett. 1981, 79, 179. (g) Wang, J. T.; Williams, F. J. Chem. Soc., Chem. Commun. 1981, 666. 2 For leading references, see: (a) Asmus, K.-D. In Sulfur-Centered Reactive Intermediates in Chemistry and Biology; Chatgilialoglu, C., Asmus, K.-D. Eds; Plenum Press: New York and London, 1990; p 155. (b) Gilbert, B. C. In SulfiIr-Centered Reactive Intermediates in Chemistry and Biology; Chatgilialoglu, C., Asmus, K.-D. Eds; Plenum 42 Press: New York and London, 1990; p 135. 3 Harcourt, R D. Lect. Notes Chem. 1982, 30, 80. ‘ (a) Gilbert, B. C.; Hodgeman, D. K. C.; Norman, R. D. C. J. Chem. Soc., Perkin Trans. 2 1973, 1748. (b) Illies, A J .; Livant, P.; McKee, M. L. J. Am. Chem. Soc. 1988, 110, 7980. (c) Ekern, S.; Illies, A.; McKee, M. L.; Peschke, M. J. Am. Chem. Soc. 1993, 115, 12510. (d) Deng, Y.; Illies, A. J .; McKee, M. L.; Peschke, M. J. Am. Chem. Soc. 1995, 117, 420. (e) James, M. A; McKee, M. L.; Illies, A. J. J. Am. Chem. Soc. 1996, 118, 7836. (i) Chaudhri, S. A.; Mohan, H.; Anklarn, E.; Asmus, K.-D. J. Chem. Soc., Perkin Trans. 2 1996, 383. 5 (a) Symons, M. C. R J. Chem. Soc., Perkin Trans. 2 1974, 1618. (b) Nelson, D. J.; Peterson, R L.; Symons, M. C. R. J. Chem. Soc., Perkin Trans. 2 1977, 2005. (c) Marignier, J. L.; Belloni, J. Chem. Phys. Lett. 1980, 73, 461. (d) Marignier, J. L.; Belloni, J. J. Phys. Chem. 1981, 85, 3100. (e) Surdhar, P. S.; Armstrong, D. A J. Phys. Chem. 1987, 91, 6532. 6 (a) Gilbert, B. 0; Marriott, P. R. J. Chem. Soc., Perkin Trans. 2 1979, 1425. (b) Giles, J. R M.; Roberts, B. P. J. Chem. Soc., Perkin Trans. 2 1980, 1497. (c) Franzi, R; Geom’oy, M.; Reddy, M. V. V. S. J. Phys. Chem. 1987, 91, 3187. 7 (a) Ganghi, N.; Wyatt, J. L.; Symons, M. C. R. J. Chem. Soc., Chem. Commun. 1986, 1424. (b) Alder, R W. Acc. Chem. Res. 1983, 16, 321 and references therein. (c) Alder, R W.; Bonifacic, M.; Asmus, K.-D. J. Chem. Soc., Perkin Trans. 2 1986, 277. (d) Dinnocenzo, J. P.; Banach, T. E. J. Am. Chem. Soc. 1988, 110, 971. (e) Gerson, F.; Knébel, J .; Buser, U.; Vogel, E.; Zehuder, M. J. Am. Chem. Soc. 1986, 108, 3781. 8 (a) Lyons, A a; Symons, M. C. R. J. Chem. Soc., Faraday Trans. 2 1972, 68, 1589. (b) Symons, M. C. R.; McConnachie, G. D. G. J. Chem. Soc., Chem. Commun. 1982, 851. 9 Hudson, R. L.; Williams, F. J. Phys. Chem. 1980, 84, 3483. ‘° lehikada, K.; Williams, F. Chem. Phys. Lett. 1975, 34, 302. 11 (a) Mohan, H.; Asmus, K.-D. J. Phys. Chem. 1988, 92, 118. (b) Livant, P.; Illies, A. J. Am. Chem. Soc. 1991, 113, 1510. ‘2 (a) Musker, w. K.; Hirschon, A s.; Doi, J. T. J. Am. Chem. Soc. 1978, 100, 7754. (b) Asmus, K.-D.; Bahemenn, D.; Bonifacic, M.; Gillis, H. A. Discuss. Faraday Soc. 1977, 63, 1748. (c) Symons, M. C. R.; Petersen, R. L. J. Chem. Soc., Faraday Trans. 1978, 210. (d) Bonifacic, M.; Asmus, K.-D. J. Chem. Soc., Perkin Trans. 2 1980, 758. (e) Asmus, K.-D.; Goebl, M.; Hiller, K. 0.; Mahling, S.; Moning, J. J. Chem. Soc., Perkin Trans. 2 1985, 641. (f) Raynor, J. B.; Rowland, 1. J.; Symons, M. C. R. J. Chem. Soc., 43 Dalton Trans. 1987, 421. (g) Symons, M. C. R.; Chanfra, H.; Alder, R. W. J. Chem. Soc., Chem. Commun. 1988, 844. (h) Breslow, R.; Brandl, M.; Hunger, J .; Turro, N.; Cassidy, K.; Krogh-Jespersen, K.; Westbrook, J. D. J. Am. Chem. Soc. 1987, 109, 7204. (i) Bobrowski, K.; Holcman, J. J. Phys. Chem. 1989, 93, 6381. (j) Abu—Raqabah, A.; Symons, M. C. R J. Am. Chem. Soc. 1990, 112, 8614. (k) Abu-Raqabah, A.; Symons, M. C. R J. Chem. Soc., Faraday Trans. 1990, 86, 3293. (1)Kishore, K.; Asmus, K.-D. J. Phys. Chem. 1991, 95, 7233. ‘3 (a) Williams, E; Sprague, E. D. Acc. Chem. Res. 1982, 12, 408. (b) Shida, T.; Haselback, E.; Bally, T. Acc. Chem. Res. 1984, 17, 180-186. 1‘ (a) Asmus, K.-D. Acc. Chem. Res. 1979, 12, 436. (b) Asmus, K.-D.; Janata, E. In The Study of Fast Processes and Transient Species by Electron Pulse Radiolysis; Baxendale, J. H., Busi, F. Eds; Reidel, Dordrecht, 1982. 1’ Meot-Ner, M. Acc. Chem. Res. 1984, 5, 186. 1‘ (a) Pauling, L. J. Am. Chem. Soc. 1931, 53, 1367. (b) Pauling, L. J. Chem. Phys. 1933, 1, 56. ‘7 Meot-Ner, M.; Hamlet, P.; Hunter, E. P.; Field, F. H. J. Am. Chem. Soc. 1978, 100, 5466. ‘8 Baird, N. C. J. Chem. Educ. 1977, 54(5), 291. ‘9 Harcourt, R D. J. Phys. B 1987, 20, 617. 2° Gilbert, T. L.; Wahl, A. C. J. Chem. Phys. 1971, 55, 5247 and references therein. For an X-ray structure of a salt of Xef, see: Seppelt, K.; Drews, T. Angew. Chem. Int. Ed Engl. 1997, 36,273. 2‘ Meot-Ner, M.; Field, F. H. J. Chem. Phys. 1974, 61(9), 3742. 22 (a) Nelsen, S. F.; Alder, R. W.; Sessions, R. B.; Asmus, K.-D.; Hiller, K.-O.; Gobl, M. J. Am. Chem. Soc. 1980, 102, 1429. (b) Alder, R. W.; Arrowsmith, R. J.; Casson, A.; Sessions, R B.; Heilbronner, E.; Kovaé, R; Huber, H.; Taagepera, M. J. Am. Chem. Soc. l981,103,6137. 23 For recent theoretical papers see: (a) McKee, M. L. J. Phys. Chem. 1992, 96, 1675 and references therein. (b) Ekern, S.; Illies, A.; McKee, M. L. J. Am. Chem. Soc. 1993, 115, 12510. (c) Hiberty, P. C.; Humbel, S.; Danovich, D.; Shaik, S. J. Am. Chem. Soc. 1995, 117, 9003. 2“ Clark, T. J. Am. Chem. Soc. 1988, 110, 1672. 44 25 Clark, T. J. Comput. Chem. 1983, 4, 404. 26Gill, P. M. W.; Radom, L. J. Am. Chem. Soc. 1988, 110, 4931. 27 Janssen, R A J.; Aagaard, o. M.; van der Woerd, M. J.; Buck, H. M. Chem. Phys. Lea. 1990, 171, 127. 28 Ganghi, N.; Wyatt, J. L.; Symons, M. C. R. J. Chem. Soc., Chem. Commun. 1986, 1424. 29 (a) Alder, R W.; Sessions, R. B.; Symons, M. C. R. J. Chem. Res, Synop. 1981, 82. (b) Symons, M. C. R; Chandra, H.; Alder, R. W. J. Chem. Soc., Chem. Commun. 1988, 844. 3° (8) Jang, S.-H.; Bertsch, R. A.; Jackson, J. E.; Kahr, B. Mol. Cryst. Liq. Cryst. 1992, 211, 289. (b) Jang, S.-H.; Lee, H.-I.; McCracken, J .; Jackson, J. E. J. Am. Chem. Soc. 1993, 115, 12623. (c) Dostal, S.; Stoudt, S. J.; Fanwick, P.; Sereatan, W. F .; Kahr, B.; Jackson, J. E. Organometallics 1993, 12. 31 For reviews on intrabridgehead chemistry, see (a) Alder, R. W. Acc. Chem. Res. 1983, 16, 321. (b) Alder, R W. Tetrahedron 1990, 46, 683. 32 (a) McMurry, J. E.; Hodge, C. N. J. Am. Chem. Soc. 1984, 106, 6450. (b) McMurry, J. E.; Lectka, T.; Hodge, C. N. J. Am. Chem. Soc. 1989, 111, 8867. (c) Sorensen, T. S.; Whitworth, S. M. J. Am. Chem. Soc. 1990, 112, 8135. 33 Alder, R W.; Moss, R E.; Sessions, R. B. J. Chem. Soc., Chem. Commun. 1983, 997. 3‘ (a) Voronkov, M. G.; Dyakov, V. M.; Kirpichenko, S. V. J. Organomet. Chem. 1982, 233, 1. (b) Miiller, E. Ph.D. Thesis, ETH, Ziirich, 1982. (c) Verkade, J. Acc. Chem. Res. 1993, 26, 483. (d) Verkade, J. Coord Chem. Rev. 1994, 13 7, 233. 35 (a) Maier, W. F.; Schleyer, P. v. R. J. Am. Chem. Soc. 1981, 103, 1891. (b) Schleyer, P. v. R; McEwen, A B. J. Am. Chem. Soc. 1986, 108, 3951. 36 (a) Leonard, N. J.; Coll, J. C.; Wang, A. H. J.; Missavage, R. J.; Paul, I. C. J. Am. Chem. Soc. 1971, 93, 4628. (b) Wang, A. H.; Missavage, R. J.; Bym, S. R.; Paul, I. C. J. Am. Chem. Soc. 1972, 94, 7100. ' 3’ Saunders, M.; Jiménez-Vazquez, H. A. Chem. Rev. 1991, 91, 375. 3‘ Korcek, s.; Chenier, J. H. 8.; Howard, J. A.; Ingold, K. U. Can. J. Chem. 1972, 50, 2285. 39 The strain energy is the difference between the experimental (calculated) heat of 45 formation of a compound and that of a hypothetically “strainless” model. ‘° Alder, R W.; East, 8. P. Chem. Rev. 1996, 96, 2097. 4‘ Saunders, M. J. Comput. Chem. 1989, 10, 203. ‘2 Pedley, J. B.; Naylor, R D.; Kirby, S. B. In Thermochemical Data of Organic Compounds, 2nd ed., Chapman and Hall: London and New York, 1986. ‘3 (a) Doyle, M.; Parker, W. Tetrahedron Lett. 1970, 42, 3619. (b) Leonard, N. J.; Coll, J. C. J. Am. Chem. Soc. 1970, 92, 6685. (c) Coll, J. C.; Crist, D. R.; Barrio, M. d. C. G.; Leonard, N. J. J. Am. Chem. Soc. 1972, 94, 7092. ‘4 For the crystal structure of manxine hydrochloride see refs. 35 a-b. The crystal structure of bicyclo[3.3.3]undecane-1,5-diol is described in: Murray-Rust, P.; Murray-Rust, J .; Watt, C. I. F. Tetrahedron 1980, 36, 2799. ‘5 Gundersen, G.; Murray-Rust, P.; Rankin, D. w. H.; Seip, R.; Watt, C. I. F. Acta Chem. Scand. 1983, A37, 823. ‘6 Wada, T.; Kishida, E.; Tomiie, Y.; Suga, H.; Seki, S.; Nitta, 1. Bull. Chem. Soc. Jpn. 1960, 33, 1317. ‘7 Alder, R W.; Arrowsmith, R. J. J. Chem. Res, Synop. 1980, 163; J. Chem. Res, Miniprint 1980, 2301. 4" Aue, D. H.; Webb, H. M.; Bowers, M. T. J Am. Chem. Soc. 1975, 97, 4136. ‘9 Alder, R W.; Orpen, A. G.; White, J. M.; Saunders, M unpublished results. 5° Alder, R W.; Arrowsmith, R. J.; Boothby, C. St.; Heilbronner, E.; Zhong-Zhi, Y. J. Chem. Soc., Chem. Commun. 1982, 940. 51 (a) Nitta, 1. Acta Crystallogr 1960, 13, 1960. For a complete X-ray crystal analysis, see ref. 45 and: (b) Weiss, G. S.; Parkes, A. S.; Nixon, E. R.; Hughes, R. E. J. Chem. Phys. 1964,41,3759. ’2 For proton affinities and photoelectron spectra of bicyclic amines and diamines, including 14, see: Alder, R. W.; Arrowsmith, R. J .; Casson, A.; Sessions, R. B.; Heilbronner, E.; Kovaé, B.; Huber, H.; Taagepera, M. J. Am. Chem. Soc. 1981, 103, 6137. 53 For the X-ray Structure of 15, see: Alder, R. W.; Orpen, A. G.; Sessions, R. B. .1. Chem. Soc., Chem. Commun. 1983, 999. 46 5‘ Wiberg, K. B. Acc. Chem. Res. 1984, 17, 379. 55 A single crystal X-ray analysis of [1.1.1]propellane is described in: Seiler, P. Helv. Chim. Acta 1990, 73, 1574. For the structure of [4.4.4]propellane by X-ray analysis, see: Ermer, 0.; Gerdil, R; Dunitz, J. D. Helv. Chim. Acta 1971, 54, 2476. ’6 For reviews on main group atranes, see ref. 33. For a review on metallatranes, see: Voronkov, M. G.; Baryshok, V. P. J. Organomet. Chem. 1982, 239, 199. For recent papers on transition metal atranes, see: (a) Scheer, M.; Mt'iller, J .; Hfiser, M. Angew. Chem, Int. Ed Engl. 1996, 35, 2492. (b) Schrock, R. R; Cummins, C. C.; Wilhelm, T.; Lin, S.; Reid, S. M.; Kol, M.; Davis, W. M. Organometallics 1996, 15, 1470. (c) Duan, Z.; Verkade, J. G. Inorg. Chem. 1995, 34, 1576. (d) Zanetti, N.; Schrock, R R; Davis, W. M. Angew. Chem, Int. Ed Engl. 1995, 34, 2044. (e) Scheer, M.; Schuster, K.; Budzichowski, T. A; Chisholm, M. H.; Streib, W. E. J. Chem. Soc., Chem. Commun. 1995, 1671. (t) Laplaza, C. E.; Davis, W. M.; Cummins, C. C. Angew. Chem, Int. Ed Engl. 1995, 34, 2042. (g) Kol, M.; Schrock, R R; Kempe, R; Davis, W. M. J. Am. Chem. Soc. 1994, 116, 4382. (h) Shih, K.-Y.; Schrock, R. R.; Kempe, R. J. Am. Chem. Soc. 1994, 116, 8804. (i) Cummins, C. C.; Schrock, R. R; Davis, W. M. Organometallics 1992, 11, 1452. (j) Nugent, W. A. J. Am. Chem. Soc. 1992, 114, 2768. ’7 Baltkais, J. J.; Voronkov, M. (3.; Zelchan, G. I. Latv. Kim. Z. 1964, N2, 102. ’8 Voronkov, M. G. Top. Curr. Chem. 1979,84, 77. 5° Voronkov, M. G. Pure Appl. Chem. 1966, 13,35. 6° Structural parameters for 46 crystallographically determined silatranes are provided by: Hencsei, P.; Parkanyi, L. In Reviews on Silicon, Germanium, Tin and Lead Compounds; Gielen, M. Ed; Scient. Pub. Div. Freund Pub. House: Tel-Aviv; Vol. VIII, No. 2 and 3; p 191. Also, see: Hencsei, P. Struct. Chem. 1991, 2, 21. 6‘ For a comprehensive review on silatranes, see ref. 33a and references therein. Also, see: (a) Tandura, S. N.; Voronkov, M. G.; Alekseev, N. V. Top. Curr. Chem. 1986, 131, 99. (b) Voronkov, M. G.; Baryshok, V. P.; Petukhov, L. P.; Rakhlin, V. 1.; Mirskov, R. G.; Pestunovich, V. A J. Organomet. Chem. 1988, 358, 39. (c) Lukevits, E.; Pudova, 0.; Sturkovich, R In Molecular Structure of Organosilicon Compounds; Ellis Horwood: Chichester, 1989; p 262. (d) Cervean, G.; Chuit, C.; Corrin, R. J. P.; Nayyar, N. K.; Reye, C. J. Organomet. Chem. 1990, 389, 159. (e) Chuit, C.; Corrin, R. J. P.; Reye, C.; Young, J. C. Chem. Rev. 1993, 93, 1371. 62 (a) Sidorkin, S. F.; Pestunovich, V. A.; Voronkov, M. G. Russ. Chem. Rev. 1980, 49, 414. (b) Greenberg, A.; Wu, G. Struct. Chem. 1990, 1, 79. 63 (a) Zhu, J.; Sun, X.; Wu, H.; Jing, L.; Chen, P.; Wu, G. Acta Chim. Sin. 1985, 43, 47 1151. (b) Iwamiya, J. H.; Maciel, G. E. J. Am. Chem. Soc. 1993, 115, 6835. 64 Gray, R. C.; Hercules, D. M. Inorg. Chem. 1977, 16, 1426. 65 (a) Gradock, S.; Ebsworth, E. A.; Muiry, J. B. J. Chem. Soc., Dalton Trans. 1975, 25. (b) Wang, D.; Zhang, D.; Lu, K.; Wu, G. Sci. Sin., Ser. B 1982, 10, 875. (c) Wang, D.; Lu, K.; Wu, Y.; Wu, G. Chem. Bull. (Beijing) 1982, 10, 12. ((1) Wang, D.-X.; Zhang, D.; Lu, K.-J.; Wu, Y.-X.; Wu, G.-L. Sci. Sin, Ser. B 1983, 26, 9. 66 (a) Brodskaya, L. I.; Voronkov, M. G. Bull. Acad Sci. USSR, Div. Chem. 1986, 7, 1694. (b) Voronkov, M. G.; Klyuchnikov, V. A.; Korchagina, A. N.; Danilova, T. F.; Shvets, G. N.; Baryshok, V. P.; D’yakov, V. M. Bull. Acad Sci. USSR, Div. Chem. 1986, 9,1976. 67 (a) Gudat, D.; Verkade, J. G. Organometallics 1989, 8, 2772. (b) Gudat, D.; Daniels, L. M.; Verkade, J. G. J. Am. Chem. Soc. 1989, 111, 8520. (c) Gudat, D.; Daniels, L. M.; Verkade, J. G. Organometallics 1990, 9, 1464. ‘8 (a) Lensink, C.; Xi, s. K.; Daniels, L. M.; Verkade, J. G. J. Am. Chem. Soc. 1989, 111, 3478. (b) Xi, S. K.; Schmidt, H.; Lensink, C.; Kim, S.; Wintergrass, D.; Daniels, L. M.; Jacobson, R A; Verkade, J. G. Inorg. Chem. 1990, 29, 2214. (c) Tang, J .-S.; Laramay, M. A H.; Young, V.; Ringrose, S.; Jacobson, R. A.; Verkade, J. G. J. Am. Chem. Soc. 1992, 114, 3129. ‘9 (a) Millbrath, D. s; Verkade, J. G. J. Am. Chem. Soc. 1977, 99, 6607. (b) Miller, E.; Biirgi, H.-B. Helv. Chim. Acta 1987, 70, 1061. (c) Laramay, J. G.; Verkade, J. G. J. Am. Chem. Soc. 1990, 112, 9421. 7° (8) Laramay, M. A. H.; Verkade, J. G. Z. Anorg. Allg. Chem. 1991, 605, 163. (b) Tang, J.-S.; Verkade, J. G. Tetrahedron Lett. 1993, 34, 2903. (c) Tang, J .-S.; Verkade, J. G. J. Am. Chem. Soc. 1993, 115, 1660. 71 (a) Tang, J .-S.; Verkade, J. G. Angew. Chem, Int. Ed Engl. 1993, 32, 896. (b) Tang, J.-S.; Mohan, T.; Verkade, J. G. J. Org. Chem. 1994, 59, 4931. ’2 D’Sa, B. A.; Verkade, J. G. J. Am. Chem. Soc. 1996, 118, 12832. 73 See refs. 30 and 39. Also, see: (a) Alder, R. W. Chem. Rev. 1989, 89, 1215. (b) Alder, R W.; Heilbronner, E.; Honegger, E.; McEwen, A. B.; Moss, R. E.; Olefirowicz, E.; Petillo, P. A.; Sessions, R. R; Weisman, G. R.; White, J. M.; Yang, Z.-Z. J. Am. Chem. Soc. 1993, 115, 6580. 74 (a) Alder, R. W.; Goode, N. C.; King, T. J .; Mellor, J. M.; Miller, B. W. J. Chem. Soc., Chem. Commun. 1976, 173. (b) Alder, R. W.; Gill, R; Goode, N. C. J. Chem. Soc., 48 Chem. Commun. 1976, 973. 7’ Alder, R W.; Sessions, R B. J. Am. Chem. Soc. 1979, 101, 3651. 76 Gerson, F .; Knobel, J .; Buser, U.; Vogel, E.; Zehnder, M. J. Am. Chem. Soc. 1986, 108, 3781. 77 (a) Alder, R W.; Sessions, R. B.; Mellor, J. M.; Rawlins, M. F. J. Chem. Soc., Chem. Commun. 1977, 747. (b) Symons, M. C. R; Smith, I. G. J. Chem. Res, Synop. 1979, 382. (c) Kirste, B.; Alder, R W.; Sessions, R. B.; Bock, M.; Kurreck, H.; Nelsen, S. F. J. Am. Chem. Soc. 1985, 107, 2635. 7‘ McKinney, T. M.; Geske, D. H. J. Am. Chem. Soc. 1965, 87, 3013. 79 Alder, R W.; Sessions, R. B.; Bennet, A. J.; Moss, R. E. J. Chem. Soc., Perkin Trans. 1 1982, 603. 8° (a) Alder, R W.; Sessions, R. B. J. Am. Chem. Soc. 1979, 101, 3651. (b) Alder, R. W.; Orpen, A G.; White, J. M. J. Chem. Soc., Chem. Commun. 1985, 949. 81Muslrer, W. K. Acc. Chem. Res. 1980, 13, 200. 82(a) Alder, R W.; Ganter, C.; Harris, C. J.; Orpen, A. G. J. Chem. Soc., Chem. Commun. 1992, 1170. (b) Alder, R W.; Ganter, C.; Harris, C. J.; Orpen, A G. J. Chem. Soc., Chem. Commun. 1992, 1172. 83 Alder, R W.; Ganter, C.; Harris, C. J .; Orpen, A. G. Phosphorus, Sulfilr, Silicon 1993, 77, 234. 8‘ Hamerlinck, J. H. H.; Schipper, P.; Buck, H. M. J. Am. Chem. Soc. 1983, 105, 385. 3‘ Roberts, B. P. Tetrahedron Lett. 1983, 24, 3377. 8‘ (a) Saunders, M.; Telkowski, L.; Kates, M. R. J. Am. Chem. Soc. 1977, 99, 8070. (b) Altman, L. J .; Laungani, D.; Gunnarsson, G.; Wennerstrom, H.; F orsen, S. J. Am. Chem. Sbc.l978,100,8264. 87 Alder, R W.; Moss, R. E.; Sessions, R. B. J. Chem. Soc., Chem. Commun. 1983, 1000. 88 (a) McMurry, J. E.; Lectka, T. Acc. Chem. Res. 1992, 47. (b) McMurry, J. E.; Lectka, T. J. Am. Chem. Soc. 1993, 115, 10167. 89 Kirchen, R P.; Sorensen, T. S. J. Chem. Soc., Chem. Commun. 1978, 769. 49 9° Sorensen, T. s.; Whitworth, S. M. J Am. Chem. Soc. 1990, 112, 8135. 9‘ (a) McMurry, J. E.; Lectka, T.; Hodge, C. N. J. Am. Chem. Soc. 1989, 111, 8867. (b) McMurry, J. E.; Lectka, T. J. Am. Chem. Soc. 1990, 112, 869. 92 (a) Gleicher, G. J.; Schleyer, P. v. R. J. Am. Chem. Soc. 1967,89, 583. (b) Bingham, R. C.; Schleyer, P. v. R. J. Am. Chem. Soc. 1971, 93, 3189. (c) Bentley, T. W.; Roberts, K. J. J. Org. Chem. 1985, 50, 5852. (d) Miiller, P.; Blanc, J .; Mareda, J. Helv. Chim. Acta 1986, 69, 635. (e) Miiller, P.; Mareda, J. Helv. Chim. Acta 1987, 70, 1017. 93 (a) Parker, W.; Tranter, R. L.; Watt, C. I. F.; Chang, L. W. K.; Schleyer, P. v. R. J. Am. Chem. Soc. 1974, 96, 7121. (b) Lomas, J. S.; D’Souza, M. J.; Kevill, D. N. J. J. Am. Chem. Soc. 1995, 117, 5891. 9“ Dewar, M. J. s.; Theil, W. J. Am. Chem. Soc. 1977, 99, 4899. 9’ Hariharan, P. C.;P0p1e, J. A. Chem. Phys. Lett. 1972, 66, 217. 96 All structures were firlly optimized employing the computer program Spartan; see: Hehre, W. J .; Huang, W. W.; Burke, L. D.; Shusterrnan, A. J. A SPARTAN Tutorial, version 4.0, Wavefimction Inc: Irvine, CA, 1995. The BDE estimates are based upon: BDEWMLH = 96.0 kcal/mol, see Gutman, D. Acc. Chem. Res. 1990, 23, 375; AHWDO (isobutane) = -26.8 kcal/mol; AHWDQ (tert-butyl radical) = -10.1 kcal/mol; and the I-IF/6— 31 G“ total energies for isobutane -157.29898 H and tert-butyl radical -156.67501 1-1, respectively. 97 (a) Grotewald, J.; Lissi, E. A.; Scaiano, J. C. J. Chem. Soc. B 1971, 1187. (b) Burkey, T. J.; Castelhano, A.; Griller, D.; Lossing, F. P. J. Am. Chem. Soc. 1983, 105, 4701. (c) Crans, D.; Clark, T.; Schleyer, P. v. R. Tetrahedron Lett. 1980, 21, 3681. 9‘ Bank, 5.; Cleveland, w. K. S.; Griller, D.; Ingold, K. U. J. Am. Chem. Soc. 1979, 101, 3409. 99 (a) McKinney, T. M.; Geske, D. H. J. Am. Chem. Soc. 1965, 87, 3013. (b) Eastland, G. W.; Symons, M. C. R Chem. Phys. Lett. 1977, 45, 422. 10° Bondi, A J. Phys. Chem. 1964, 68, 441. 101 McMurry, J. Private Communication 1988. “When you have eliminated the impossible, whatever remains, however improbable, is the truth” A. Conan Doyle CHAPTER 2 l-MANXYL: A PERSISTENT TERTIARY ALKYL RADICAL THAT DISPROPORTIONATES VIA a-HYDROGEN ABSTRACTION Abstract: Bicyclo[3.3.3]undecane (manxane) shows high bridgehead reactivity. With atmospheric oxygen it autoxidizes to form a mixture of bridgehead peroxides and hydroperoxides. 1-Manxyl chloride undergoes solvolysis ca. 104 times faster than tert- butyl chloride. The enhanced reactivity at these sites is due to relief of strain when the bridgehead converts to a trigonal center, as indicated by earlier molecular mechanics and new ab initio results. The l-manxyl radical 2 has now been generated in solution from manxane 1 by hydrogen abstraction with tert-butoxyl radicals. The EPR spectrum of 2, which shows anomalously low B hyperfine coupling constants, is reported here for the first time. Continuous-wave ENDOR experiments have helped to confirm the values of the hyperfine splittings. The decay of the radical is birnolecular with a rate constant of 0.5 M'ls'1 in methylcyclopentane at 23 °C; one of the decay products of 2 has been identified as the [3.3.3]propellane 31, formed presumably by an unusual e-disproportionation. 1-Manxyl is the first example of a persistent alkyl radical whose exceptionally long lifetime arises not from steric protection, but from the high strain of all its decomposition products. 50 51 Bicyclo[3.3.3]undecane (manxane)l 1 was first synthesized in 1970 independently by Leonard et al.2 and by Doyle et 81.3 as the prototype compound which comprises together three eight-membered rings. The conformations of manxane and some of its derivatives have been studied by dynamic NMR3 and molecular mechanics“. Calculations point to the C31. boat-chair conformation as the energy minimum, but even this arrangement is strained in contrast to the flexibility observed for most monocyclic eight-membered rings. Confirmation of the high ground strain of manxane has been provided by experimental measurements of its enthalpy of formation, AHr(CuH20, g) = -21.2 kcal/mol.5 One structural manifestation of the strain is a flattening of the bridgehead regions, accompanied by widening of the angles in the bridges. Bridgehead flattening in 1 has been related to increased p character in the methine C-H bond, and this hybridization change is reflected in the low value of the corresponding ‘Jcsr (120.0 Hz). X-ray structures of 1- azabicyclo[3.3.3]undecane hydrochloride“ and bicyclo[3.3.3]undecane-1,5-diol7 show the expected structural features. The electron-diffraction data fiom manxane vapors confirmed the C31, molecular symmetry.8 At room temperature manxane is in rapid conformational equilibrium between two degenerate forms. In a temperature dependence study of the 1H H la 1b 52 NMR spectrum of 1, Doyle et al.3 obtained the “frozen” spectrum, corresponding to the slow exchange between 1a and 1b, at -80 °C with CDC13/CD2C12 (1:1) as solvent, and calculated a free energy of activation for the inversion process of 11i2 kcal/mol. Our interest in through-space perturbation of unpaired electron centers9 has drawn us to the rich potential of interbridgehead chemistry, for which the bicyclo[3.3 .3]undec-1- yl, or l-manxyl, radical 2 is a key reference species. With its 27.2 kcal/mol strain energy (SE) (Table 2.1) and high bridgehead reactivity,lo manxane 1 readily undergoes hydrogen abstraction by tert-butoxyl radicals to yield radical 2. Herein we present EPR ENDOR, spin trapping, product studies, and ab initio results for the l-manxyl radical 2. This sterically open radical shows remarkable persistence and unexpectedly small [3- hydrogen hyperfine couplings. 2.1 Results and Discussion Manxane 1 was prepared in a multistep synthesis involving double-ring expansion of the short bridge of bicyclo[3.3. 1]nonan-9-one 8, following Leonard et al.2, with modifications to obtain an overall optimized yield (Scheme 2.1) of 2.2%. Bicyclo[3.3.1]nonan-9-one 8 was made from cyclohexanone 3 in four steps according to the method of Foote and 53 Scheme 2.1 it 0 (\o . C E.) v p-CH3-C6H4-SO3H + CH2=CHCHO “ o ..W .. , 85% 65% 3 4 1. H202 30% H2, 3atm. Diazald CH3OH, I’d/g CH3OH KOH 2, A, 15 torr Pd/C 10°o CH3OH/H20> 25% 95% 70% (CH3)3COK MCPBA $21513 (C6H5)3CH3PBI’ NaHC NaN DIWF C6H5, reflux CHCl3/HzO 75% 66% 1. H2, 3 atm. a0 2NH3CI 11. 2. HCl, Eth ’ . CH3COOH ’ u. ' :2; 63% from 11 98% 13 14 15 NH2NH2.H20 NH2NH2-2HC1\ I) KOH/TEG 77% l 54 Woodward“. The morpholine enamine of cyclohexanone 4 was condensed with acrolein in THF to give 2-N-morpholinyl-bicyclo[3.3.1]nonan-9-one 5. The mechanism of this remarkable condensation is somewhat obscure; at some stage in the reaction the nitrogen and oxygen firnctions must exchange positions.11b Conversion of the aminoketone to the N-oxide 6 by oxidation with hydrogen peroxide in methanol followed by pyrolysis at 120 °C (Cope elimination) yielded bicyclo[3.3.1]non-2-en-9-one 7, which was hydrogenated over MIC 10% to give 8. Ring expansion of bicyclo[3.3. 1]nonan-9-one 8 with methanolic diazomethane afl‘orded bicyclo[3.3.2]decan-9-one 9. The original experimental procedures of Leonard et al.2 for conversion of 9 to 9-methylenebicyclo- [3.3.2]decane 10 and its subsequent epoxidation to 11 were replaced by a revised “frttig reaction for methylenation of sterically hindered ketones with tert-BuOK and (CsHs)3CH3PBr in refluxing benzene,12 and respectively, by epoxidation with m- chloroperbenzoic acid in an alkaline biphasic system (N aHCOg, H20/CHC13)13. The resulting 9-epoxymethylenebicyclo[3.3 .2]decane 11 was cleaved by sodium azide in DMF to the hydroxyazide 12, and reduction in ethanol with hydrogen over Adams’ catalyst, followed by Demjanov-Tifieneau ring expansion of the hydrochloride salt 13 yielded a 3 :1 mixture of bicyclo[3.3.3]undecan—9- and 10-ones 14 and 15. Wolff-Kishner reduction of the ketone mixture afl‘orded 1. Manxane 1 is autoxidized by air to a mixture of bridgehead peroxides and hydroperoxides, and l-manxyl chloride undergoes solvolysis ca. 104 times faster than tert- butyl chloride, consistent with a molecular mechanics estimate of 6.8 kcal/mol strain relief for bridgehead conversion to a trigonal center. 1° Given the enhanced reactivity of the 55 bridgehead sites, hydrogen atom abstraction from manxane 1 by photochemically generated tert-butoxyl radicals provides a convenient technique for generating the bridgehead radical 2.14 h tert-BuO-O-tert-Bu —U> 2 tert-BuO' tert-BuO° + Manxane (l) —> l-Manxyl' (2) + tert-BuOH Reaction of l-manxyl chloride with triethylsilyl (Et3Si’) or tri-n-butyl-tin (n-Bu3Sn') radicals provides in principle a direct route to 2;15 the bridgehead chloride, however, is troublesome to synthesize, has never been isolated pure, and solvolyzes completely to the alcohol on exposure to air. 1° This route was therefore not attempted. Cyclopropane, with C-H bond dissociation energies (BDE’s) of 106.3 kcal/mol,16 is a convenient solvent for the hydrogen abstraction procedure. ’4 Figure 2.1 shows the EPR spectrum obtained from photolysis of a cyclopropane solution of manxane 1 and di- tert-butyl peroxide at -55 °C. Identical EPR spectra arise in toluene or methylcyclopentane solutions, and in neat di-tert-butyl peroxide. On shuttering the photolysis beam, the spectrum of 2 decays extremely slowly, i.e., the radical lifetimes are, depending on temperature and solvent, on the order of days or even weeks. The photolysis temperature can be widely varied; in cyclopropane the best EPR spectra are obtained between --60 and -40 °C, but in toluene and neat di-tert-butyl peroxide, room temperature gives the optimum experimental conditions. Remarkably, the EPR spectrum of 2 in frozen toluene, obtained after gradual cooling of a toluene solution of l-manxyl radicals, displays all the features of the spectrum recorded in liquid phase. 56 (a) 5G 0)) Figure 2.1 (a) EPR spectrum (9.1 GHz) of l-manxyl radical in cyclopropane at -55 °C (g = 2.0024). (b) Computer simulation. 57 We assign this EPR spectrum to l-manxyl radical 2 on the following grounds: (1) the radical is tertiary, showing neither an or C-H hyperfine coupling constant, nor a corresponding splitting in the 2,4,6-tri-tert-butyl-nitrosobenzene spin trapping product (see section 2.3); (2) simulation of the spectrum (Figure 2.1) requires five difi‘erent sets of three equivalent protons; (3) the radical decays via an extraordinarily slow birnolecular process, and trapping by addition of n-Bu3SnH immediately after photolysis turns off production of its disproportionation products, of which one is [3.3.3]propellane (see section 2.2); (4) the known autoxidation of 1 is Specific for the bridgehead site. With 18 secondary and only 2 tertiary C-H bonds in 1, significant secondary hydrogen abstraction might be expected on statistical grounds, but no evidence for the secondary 2- and 3-manxyl radicals, 16 and 17, is seen in the EPR spectra under any 16 17 conditions. Generally, in compounds where more than one type of hydrogen atoms are present, the EPR spectrum observed belongs to that radical produced by hydrogen abstraction fiom the weakest bond.17 The BDEs of the OH bonds in manxane were estimated at HF/6-31G* level from isodesmic reactions vs. isobutane/tert-butyl radical for 2, and propane/isopropyl radical for 16 and 17.18 Besides being the unique tertiary sites in manxane, the bridgeheads also afford the greatest strain relief upon hydrogen abstraction, resulting in BDE difi‘erences of 6.9 and 7.9 kcal/mol vs. 16 and 17, respectively (Table 58 Table 2.1 Calculated Heats of Formation, Strain Energies and Bond Dissociation Energies'I Compound Total Energyb AH; SBd ASE’ BDE' Manxane 1 -428.17907 -204 (.212)8 28.0 (27.2) l-Manxyl Radical 2 427.56808 14.6 19.9 -73 87.9 2-Manxyl Radical 16 42755704 21.5 24.6 -2.6 94.8 3-Manxyl Radical 17 42755538 22.5 25.6 -1.6 95.8 'In kcal/mol; structures were fully optimized at 1-1F/6-3 16* level, using Spartan 4.0 (Wavefunction Inc., Irvine, CA). t’Total energies are given in hartrees, 1 H = 627.5 kcal/mol. ° Calculated (experimental) from Wiberg’s group equivalents (Wiberg, K. B. J. Org. Chem. 1985, 50, 5285) for manxane. The BDE estimates were used to calculate heats of formation for the product radicals. d Strain energy, from ealculated (experimental) AH; and Benson’s group equivalents (Benson, S. W. Thermochemical Kinetics, John Wiley: New York, 1976) for manxane, and from isodesmic rections vs. isobutane/tert-butyl radical for l-manxyl radical, and propane/isopropyl radical for 2- and 3-manxyl radieals. ° Defined vs. SE of manxane. ‘Based on BDE (t-Bu-H) = 96.0 kcal/mol (Gutman, D. Acc. Chem. Res. 1990, 23, 375), and BDE (iso-Pr- H) = 98.2 keal/mol (Russell, J. J.; Seetula, J. A.; Gutman, D. J. Am. Chem. Soc. 1988, 110, 3092). 3 Ref. 5. 59 2.1). A recent model relating activation energies to reaction exothermicities suggests that for tert-butoxyl abstracting H from alkanes, barrier heights change by roughly 30-40% of reaction energy differences. ’9 Thus, even a fraction of the difference between H- abstraction transition states would easily outweigh the 9:1 statistical factor between secondary and tertiary sites in 2. The experimental EPR spectrum of 2, essentially independent of temperature, can be simulated with the following hyperfine constants: an = 5.3 G (3H), an = 2.4 G (3H), an = 0.99 G (3H), an = 0.88 G (3H) (see Figure 2.1). The EPR simulation program employed in this work was written at MSU by Dr. Andrew S. Ichimura, for use with the non-linear least squares fitting program KINFIT.20 The resonance fields were calculated to first order, and the hyperfine splitting constants and the line widths were varied until a minimum in the rms error was found between the observed and calculated spectra. The EPR spectrum of 2 was also analyzed using the computer program MATCH, kindly provided to us by Professor R. A. Jackson fi'om University of Sussex, UK.21 MATCH was designed to determine accurate coupling constants and line width data for EPR spectra, based on correlation methods. The analysis is efficient even for low intensity or complex spectra; in our case MATCH produced coupling constants identical with the values determined from simulation. The procedure involves comparison of the experimental EPR spectrum with a matching “test spectrum”, using a product firnction produced by cross- correlation of the test spectrum with the experimental spectrum, as the optimization criterion for improvement of fit. 60 1H ENDOR (Electron Nuclear DOuble Resonance)22 resonance measurements were performed on samples containing 2 in toluene solution, in order to confirm the values of the hyperfine couplings obtained by simulation of its experimental EPR spectrum. In the ENDOR experiment nuclear spin transitions in paramagnetic molecules are induced by means of a suitable radio frequency (RF) field and are detected by a change in the EPR signal intensity. The ENDOR spectrum consists of pairs of lines that correspond to the types of protons in the molecule, each symmetrically split fiom the fi'ee proton nuclear magnetic resonance frequency of 14.44 MHz by the appropriate electron nuclear hyperfine interaction. The principal advantages and improved resolving power of ENDOR over ESR are those of simplifying complex spectra and giving precise values of the hyperfine coupling constants (I-H" C), which can be extracted without difficulty and usually unambigously without need for computer simulations. The ENDOR studies on 2 confirmed the previously determined HF Cs and revealed two more couplings at 0.19 and 0.08 G (see Figure 2.2). The ground state conformation of 2 has C3 symmetry, and accordingly, the maximum number of different hyperfine couplings is 7 (6 sets of 3 equivalent HS each, and one H in the opposed bridgehead). INDO (Intermediate Neglect of Differential Overlap)23 calculations performed on PM3 and UHF/6-31G* geometries of 2 (see Table 2.2) reproduce the magnitude of the smaller couplings well, but predict a B-hydrogen hyperfine of ~ 20 G, well above the largest HF C to hydrogen, an, observed (5.3 G). The 2.4 G coupling is assigned to one set of y-hydrogens related to the semioccupied orbital via a W arrangement that commonly leads to a strong interaction with the unpaired electron. 61 13.8 14.2 14.6 15 4 8 12 16 20 24 v [MHz] Figure 2.2 The ENDOR spectrum of l-manxyl radical 2 in toluene at -50 °C. Insert: the central part of the ENDOR spectrum of 2, which reveals small HFCS at 0.19 and 0.08 G. '62 Table 2.2 INDO Predicted Hyperfine Coupling Constants (in G) for l-Manxyl Radical 2' 0g AIL-b Method 1 2 3 4 5 6 7 UHF/6.316“ 20.9 0.8 2.0 -1.0 0.9 0.2 1.2 14.6 PM3 235 1.1 2.7 -12 0.6 0.3 1.6 -143 ABBc (A) +0.3 18.6 0.5 3.3 -13 1.3 0.3 0.6 -91 +0.2 20.3 0.7 3.2 -12 1.1 0.3 0.8 -120 +0.1 22.0 0.9 3.0 -12 0.8 0.3 1.1 -137 .01 24.8 1.4 2.2 -1.1 0.3 0.3 2.5 -137 .0.2 25.6 1.7 1.7 -09 0.1 0.2 3.8 -122 -05 25.5 2.1 1.1 -0.8 -02 0.2 5.9 -97 0(C2C1C5C4)‘ 15° 27.9 2.0 2.3 -10 0.7 0.2 1.4 30° 29.4 3.1 1.7 -1.0 0.5 0.0 1.4 60° 26.7 3.0 3.0 -0.6 -02 0.2 3.0 Exp. 5.3 0.88 2.4 0.99 0.08 0.19 ' Structures were fully optimized using Spartan 4.0. H’s are labeled as below, in Figure 2.3. b Heats of formation in kcal/mol. ° ABB is defined as an inward (-)/outward (+) displacement of the spin-bearing bridgehead carbon along the symmetry axis (C3) from the BB (bridgehead-bridgehead) distance in the PM3 geometry Optimized structure (3.0127 A). A constraint is defined as the new BB distance (elongated or contracted by ABB), and the new structure is geometry optimized at the PM3 level. dDihedral angle in degrees (1.9(C2C1C5C4) = 0°; 22 9(C2C1C5C4) = 0.430); equal to 9(C3C1C5C5) and e(C9C1C5Cll)- 5.3 G (H1) 0.88 G (Hz) 2.4 G (H3) 0.99 G (H4) 0.08 G (1'15) H7 H5 0.19 (H7) Figure 2.3 Assignments of the hyperfine coupling constants in l-manxyl radical 2. 63 Tentative an assignments, based on INDO results, are: 5.3 G and 0.88 G for B-H, 2.4 G and 0.99 G for y-H, 0.08 G for 5-H, and 0.19 G for the e-H (Figure 2.3). The B—hydrogen splitting has been rationalized in terms of a hyperconjugative mechanism, described by the familiar McConnell relationship, aHB = A + B cos20, where A is small and usually neglected, B is assumed to be 2 x aHB of the tert-butyl radical (z 50 G), and 0 is the angle between the H-C-C plane and the axis of the spin bearing orbital.24 Under conditions where rotation about the Ca(2p)-Cp bond is rapid, the average value of 00829 is 0.5 and aHB z 27 G.25 The hyperfine interactions are expected to be small for B- protons, provided that the Cp-Hg bond lies in the nodal plane of the Ca(2p) orbital. The angular dependence of the B-proton coupling, along with the variation with temperature of the EPR HF Cs and line shapes, have been commonly employed to distinguish preferred conformations and to determine rotation and ring inversion barriers of alkyl and cycloalkyl radicals?” According to the McConnell relation, the 5.3 G B-H splitting in 2 is unexpectedly low. The analogous delocalized D31. radical cations of [3.3.3]propellane, 18,27 and 1,5-diazabicyclo[3.3.3]undecane, 19,28 Show B-H couplings of 17 G and 22 G, :9 U U respectively, interpreted as reflecting nearly planar bridgeheads with 0 angles of approximately 30°. The calculated structures (UHF/6-31G“) of 2 and of radical cations 18 64 and 19 show similar torsion angles (0) of the B-hydrogens with the half-occupied orbital (333°, 319° and 326°); however, the radical center in 2 is pyramidalized syn to the Cp- Hn bonds, which should make hyperconjugation less effective.29 The EPR spectra of bicyclo[2.2.0]hex-1-yl30 20 and l-cubyl31 21 radicals also show exceptionally low [3- hydrogen HFCs (12.4 G and 6.2 G, respectively; see Figure 2.4) considering that 0 is formally zero and thus optimum for overlap. The more comparable aHB values of 6.64 and 6.58 G for the localized bridgehead radicals 1-bicyclo[2.2.2]octyl 22 and l-adamanty123 are attributed to pyramidal geometries at the radical sites (Figure 2.5).32 Bridgehead radicals are strongly pyramidal with B-carbons tied back by the cage structure leaving the radical center sterically uncongested. In bridgehead radicals the orientation of the SOMO with respect to the orbitals of the B-C-H bonds is usually less favorable for overlap, and the rigid structure prevents rotation to improve it. In addition, hyperconjugative structures will contain strained “anti-Bredt” bridgehead alkene units. Thus, most bridgehead radicals have aHB values lower in magnitude than predicted by the McConnell relation, while they Show large long-range HFCS.33 For l-manxyl radical 2, however, the UHF/6-31G“ (or PM3) structure shows only modest pyramidalization and B-hydrogens that are more nearly eclipsed than those in 22 and 23 (see Figure 2.5, and Tables 2.3 to 2.5), leaving the low “H5 value somewhat puzzling. The l-norbomyl radical 24 gives B-H HF Cs of 9.81 G for HMO, 0.49 G for Hpfido, and 2.35 G for the two B-Hs fiom the one-carbon bridge; this set fits linearly with cos20 but with a B coeficient of about a quarter of the corresponding constant for planar 65 .QA000VVN mm 8:350 .mooemoc E 833 0.00 ommeofi 2: 8 3&8 9.0.00V ”teamed 28:08 Umonowctn («o A0 :5 $5388 wezasoo 85093 358588 98 386888 32.568on 33328 cm: $28.5 vogue—om Ya charm 3 3 an 5 an .03 u 0.84 some n 0.00m. .32 u 0004 age u 0.84 .442 u 9ro 3.3: 8.0: 38m $.on 3.8 GEE . meme £80 . 8.3 . . 8.0m m . 7 . e8 68 Am. C AN6Vm 66 Cb d) or. = 104.0° (exp. 105. 1°) on = 94.8° on = 106.6° on = 106.2° 9 = 32.5°; 80.9° 0 = 333°; 829° 9 = 59.4° 9 = 59.9° l 2 22 23 Figure 2.5 HF/6-31G“ geometry optimized structures of manxane l, l-manxyl 2, 1-bicyclo[2.2.2]octy122, and l-adamanty123 radicals. Legend (C3 refers to the axis of symmetry): a = C 3C'Cp angle, and 6 = C 3C'CBI-Ig torsion angle, in degrees. 67 Table 2.3 UHF/6-31G* (PM3) Geometrical Parameters for l-Manxyl Radical 2' Selected distances (r), bond angles (4), and torsion angles (6) UHF/6-31G“ (PM3) LUCs ' Distances in A, angles in degrees; C3 refers to the three-fold axis of symmetry. r(C1-C2) r(C2-C3) r(C1-C5) 1(C2C1Cs) 4(C1C2C3) 4(C2C3C4) 1(C5C1C2) 9(C3C1C2H1) 9(C3C1C2H2) 0(czclcsc4) 1.5052 (1.4799) 1.5362 (1.5234) 3.0970 (3.0127) 119.3 (199.7) 113.7 (112.2) 117.0 (113.9) 85.2 (86.8) 33.3 (35.5) 97.1 (99.9) 0.43 (7.8) 68 Table 2.4 PM3 Atomic Cartesian Coordinates (in A) for l-Manxyl Radical 2 Atom x y z H 1 -0.5784940 -2.0251974 -1.0658355 C 2 -1.0681364 -1.0356499 -1.2087390 H 3 -1.7478700 -1.1612237 -2.0752316 C 4 0.0000000 0.0000000 -1.5951783 C 5 -1.9242512 -O.7247960 0.0089550 H 6 -2.3332975 0.3067356 -0.0844577 C 7 -1.1896046 -0.8763683 1.3349257 H 8 -2.8044188 -1.3977577 0.0116290 C 9 0.0000000 0.0000000 1.4177657 H 10 -0.9012690 -1.9355879 1.4913718 H 11 -1.8711708 -0.6202584 2.1715529 C 12 -0.1641549 1.4684120 1.3349257 C 13 1.3537595 -0.5920437 1.3349257 H 14 -l.4646254 1.5135892 -1.0658355 H 15 -0.1317142 2.0943117 -2.0752316 H 16 0.0000000 0.0000000 -2.7151618 C 17 1.4309674 -0.4072083 -1.2087390 C 18 1.5898174 -1.3040524 0.0089550 H19 2.0431194 0.5116082 -1.0658355 H 20 1.8795842 -0.9330880 -2.0752316 H 21 0.9010079 -2. 1740627 -0.0844577 C 22 -0.3628309 1.4428582 -1.2087390 H 23 1.4727449 -1.3103522 2.1715529 H 24 2.1269028 0.1872721 1.4913718 H 25 0.3984259 1.9306106 2.1715529 H 26 -1.2256338 1.7483158 1.4913718 C 27 0.3344338 2.0288484 0.0089550 H 28 1.4322896 1.8673271 -0.0844577 H 29 0.1917158 3.1275767 0.0116290 H 30 2.6127030 -1.7298191 0.0116290 69 Table 2.5 UHF/6-31G* Atomic Cartesian Coordinates (in A) for l-Manxyl Radical 2 Atom x y 2 H 1 0.0000000 0.0000000 2.7009321 C 2 0.0000000 0.0000000 1.6142657 C 3 0.0593796 -1.5134558 1.2630577 C 4 1.2810014 0.8081521 1.2630577 C 5 -1.3403809 0.7053037 1.2630577 H 6 2.1405044 0.1443425 1.3266034 H 7 1.4208356 1.5340192 2.0608630 C 8 1.3544058 1.6073025 -0.0462537 H 9 -2.0389174 0.4634702 2.0608630 C 10 -2.0691677 0.3692985 -0.0462537 H 11 -1.1952565 1.7815600 1.3266034 H 12 -0.9452480 -1.9259025 1.3266034 H 13 0.6180818 -1.9974893 2.0608630 C 14 0.7147619 -1.9766011 -0.0462537 H 15 -3.0129518 0.9105534 -0.0425296 H 16 -2.3383691 -0.6826451 -0.0421622 C 17 -1.3324475 0.6885334 -1.3559473 H 18 2.2950383 2.1540161 -0.0425296 H 19 0.5779966 2.3664096 -0.0421622 C 20 1.2625112 0.8096666 -1.3559473 C 21 0.0699363 -1.4982001 -1.3559473 H 22 0.7179135 -3.0645695 -0.0425296 H 23 1.7603 726 -1.6837645 -0.0421622 C 24 0.0000000 0.0000000 -1.4827090 H 25 -1 .2073288 1.7641697 -1.4505304 H 26 -1.9787037 0.3868696 -2. 1808085 H 27 -0.9241514 -1.9276623 -1 .4505304 H 28 1.3243907 1.5201729 -2.1808085 H 29 2.1314802 0.1634926 -1.4505304 H 30 0.6543130 -1.9070424 -2.1808085 70 radicals.34 In bicyclo[1.1.l]pent-1-yl radical 25, where the Cp-Hg bonds are basically orthogonal to the axis of the Ca(2p) orbital, an!3 is 1.2 G.35 Such strained small-ring bicycloalkyl radicals have been studied by EPR mostly for the assessment of through-bond (TB) and through-space (TS) interactions. Thus, the bridgehead hydrogen HF C increases steeply fi'om 24 (2.5 G), with the odd-electron delocalized onto the bridgehead H atom through a TS mechanism, to bicyclo[2. 1. 1]hex-1-yl radical 25 (22.5 G),36 where both TS and TB mechanisms operate, and to 26 (69.6 G), where the TB interaction is prevalent. Bicyclo[3.3.3]undec-1-yl cation 27, prepared by Olah et al.37 from 1-chloro- or 1- hydroxybicyclo[3.3.3]undecane with SbF5/S02C12 at -78 °C, shows the same temperature- independent behavior as 2. As observed by 1H and 13 C NMR the solution of 27 does not change between -135 and -30 °C, and it slowly decomposes at high temperatures. This behavior is surprising in comparison with manxane l or the bridgehead manxyl dication 28, where the intriguing bridge flipping process (see below) is fiozen at low temperature. + + + 27 28 Olah et a1.37 suggested either a rapid ring flipping in 27, faster than can be detected on the NMR time scale, or a ve1y slow inversion of conformation due to additional strain in 27 introduced by the sp2 hybridized carbon at Ca(2p). However, bridgehead planarization in 27, if anything, brings relief of strain when compared to l, which leaves the first alternative as more probable. By analogy with 27, 2 might also undergo rapid vibrational 71 averaging with the net effect of reducing aHB' Furthermore, it is of interest to mention that the methine protons in manxane 1 (6 2.38 ppm; width ~ 24 Hz) and the methine proton in 1-azabicyclo[3.3.3]undecane (manxine) (8 2.57 ppm; width ~ 181-12) are broad multiplets with no discemable couplings.2 The broad signal becomes a well-resolved septet upon addition of dipivaloylrnethanatoeuropium(III) complex to manxine, as well as in manxine hydrochloride (J = 5 Hz).2 The dihedral angle ((0) dependence of vicinal spin-spin couplings, 31“.“, is described by the Karplus relation:38 3J1“; = A + B coso + C cos2

(ST-R') spin trap spin adduct afl'mity for radicals, is added to the reactive radical to give a particularly persistent new 73 fi'ee radical (the “spin adduct”), whose concentration will build to readily detectable levels (> 10'7 - 1045 M). The success and value of the spin-trapping experiment depend upon how fast and selective is the trapping reaction, how persistent is the resulting radical, and if the identity of R‘ can be readily discerned fiom the EPR spectrum of the spin adduct. Although many different unsaturated groups have been used to trap various radicals, the vast majority of investigations or applications of the spin-trapping technique depend on the use of C-nitroso compounds or nitrones, to yield relatively stable aminoxyl (or nitroxide) flee radicals, which are readily detected by EPR spectroscopy. The preeminent advantage of C-nitroso-compounds over nitrones as spin traps is that in the spin adduct the scavenged radical is directly attached to the nitroxide nitrogen. As a result, the ESR R- + R'—N=O ——> / \ . C-nitroso compound R R / k. nitrone \ ,0 \ 0' R- + C=N+ ———> R—C—N: RI spectrum of the spin adduct is likely to reveal splittings from magnetic nuclei in the trapped radical, which facilitate its identification. Spin trapping of the l-manxyl radical 2 by the nitroxide method was attempted with 2,4,6-tri-tert-butyl-nitrosobenzene (TBN) as spin scavenger.40 The main benefits of TBN over other spin traps are that it functions as an ambident (“bifunctional”) spin trap, and that it is stable to light both in solution and in the solid state, which makes it useful for application to photoradical reactions. Thus, TBN reacts at either the N or the O atoms of 74 the nitroso-group, depending on the steric hindrance of the attacking radicals, to give as spin adducts the corresponding nitroxide or N-alkoxyanilino radicals. Primary alkyl radicals react at nitrogen, secondary alkyl radicals react at both trapping sites, while tertiary alkyl radicals react exclusively at the oxygen atom. It is therefore possible to distinguish between attacking primary, secondary and tertiary alkyl radicals fi'om the EPR spectra of the spin adducts, since nitroxides have significantly difi‘erent an and an splittings than N-alkoxyanilino radicals. The alkoxyaminyl radicals have a lower g—value than the nitroxides (ca. 2.004 vs. 2.006) and their spectra are therefore centered at higher field positions than those of nitroxides. Splitting patterns are also significantly difl‘erent; the spectra of the alkoxyaminyls show much larger splittings from the meta-protons of the aryl rings than do the nitroxides, but aN is smaller. In addition, TBN is monomeric and does not dimerize. - O R N=O °N—OR ‘N’ t-Bu r-Bu t-Bu t-Bu t-Bu t-Bu O —-+”" O + O t-Bu t-Bu t-Bu TBN N-alkoxyanilino radical nitroxide radical g = 2.003 - 2.004 g = 2,006 aN=9.6-12.3G aN=ll.7-l3.7 aH=l.7-ZG aH=O.6-1G TBN reacts with l-manxyl radical 2 to produce a persistent N-alkoxyanilino radical, the EPR spectrum of which (Figure 2.6) shows the following g value and coupling constants: g = 2.003, (In = 9.0 G (1N), an = 1.8 G (2H). The spin trapping experiments were performed either by adding a toluene solution of TBN to an irradiated sample of 75 manxane in di-tert-butyl-peroxide, which contained 2 in concentrations of ~ 10'3-104 M, or by UV irradiation of a solution containing manxane, TBN and di-tert-butyl-peroxide, directly in the cavity of a Varian E4 spectrometer. Identical EPR spectra were obtained in both cases, in agreement with the experimental observations that TBN is not a good trap for radicals other than alkyl, and it can be used successfully in situ when the alkyl radicals are generated by H atom abstraction from substrates with tert-butoxyl radicals. Irradiation of TBN itself, in solid state or in di-tert-butyl peroxide solution, gave no detectable EPR signal. The 1.8 G meta-H hyperfine and the absence of B-hydrogen splittings in the EPR spectrum of the TBN spin adduct of 2 indicate exclusive addition at the O atom of TBN by an unreactive tertiary radical such as 2, consistent with the observed multi-minute trapping time. The rate constant for the reaction of TBN with tert-butyl radical has been experimentally determined as 2.3x 105 M'ls‘l at 24 °C in benzene." Ifwe extrapolate this value for the reaction of TBN with 2, under the pseudo first-order conditions of the spin trapping experiment the above rate constant gives reaction times on the order of miliseconds; however, as expected, the trapping rate of l-manxyl radical 2 by TBN is considerably slower, since it takes a few minutes for the addition of l to TBN to be complete. Thus, the choice of TBN as spin trap to elucidate the nature of the radical obtained on H atom abstraction from 1 is validated: the long lived radical obtained from photolysis of manxane and di-tert-butyl peroxide reacts slowly, at the 0 position of TBN, to yield a tert-alkyl alkoxyaminyl radical, and the significant steric efi‘ects revealed in the trapping reaction, all strongly support the assignment of the initial EPR spectrum to an inflexible, bulky alkyl radical such as 2. 76 J n " Figure 2.6 EPR spectrum (9.065 GHz) of the N-alkoxyanilino radical obtained by spin trapping of l-manxyl radical 2 with TBN (g = 2.003). .1310 77 2.3 Kinetic Decay and Product Analysis The kinetics of radical disappearance for l-manxyl radical 2 were readily obtained fi'om spin resonance experiments due to its remarkable persistence. Photolysis of l and di-tert- butyl peroxide in methylcyclopentane at room temperature (23 °C) generated l-manxyl radical 2, whose decay was monitored by EPR. The number of electron spins present in the cavity during the EPR measurement was obtained by comparison of the area under the absorption curve with that of a reference radical. DPPH (diphenylpicrylhydrazyl) solutions ofknown concentration (4x10‘4 M, 2x104 M, 1x10‘4 M, 8x105 M, 6x105 M, and 4x10" M) in benzene were employed as standards for spin concentration determinations.42 OzN Q... .. @ 3C} OZN DPPH The EPR spectra of both 2 and the reference samples were recorded at 23 °C with identical microwave power levels. No saturation was observed for any of the radicals under the conditions of the experiment. However, in computing the absolute number of spins, a correction had to be applied because of different modulation amplitude and gain settings. From a consideration of the various errors involved in such a determination, a deviation of :tSO% is usually assigned to concentration, which, nevertheless, does not change the order of magnitude of the rate constant for radical disappearance.42 The areas resulting fi'om double integration of the EPR derivative signals of DPPH solutions were plotted against DPPH concentration for calibration. The calibration curves 78 were validated by UV measurements of the DPPH absorption (at 327 and 520 nm), which established, as expected, a linear variation of DPPH concentration with UV absorption. The number of spins corresponding to l-manxyl radical 2 was computed from the area of the EPR absorption curve, obtained by double integration of the derivative signal, relative to that of the standard. The plot of the inverse concentration (l/c) of l-manxyl radical against time (1:) is linear at longer times and indicates second order kinetics in 2 (see Figure 2.7).43 The rate constant for the radical disappearance, k, is calculated from the slope of the line (best linear fit of No against 1) whiCh equals 2k, and the half-life rug, is determined as 1/(2k[l-manxyl]o), where [l-manxyl]o is the initial radical concentration equal to the intercept of the line. Thus, the decay of 2 in methylcyclopentane, monitored by EPR, is second order (n = 2) with a rate constant of 0.5 Mls'l at 23 °C and a half-life (1: 1/2) of 6 hours for a 5x10‘5 M initial radical concentration (ci). Such exceptional persistence is unique considering the lack of steric protection around the radical center.44 A few representative examples of persistent secondary and tertiary alkyl radicals are given below, where the long lifetime of the radicals is a consequence of steric factors.44 (MesC)2CH° (M62CH)3C' (M63030 bis(tert-butyl)- 2,2,4,6,6-pentamethyl- 2,2,4,4,6,6,-hexamethyl- tris(isopropyl)- tris(tert-butyD- methyl radical cyclohexyl radical cyclohexyl radical methyl radical methyl radical 1713: 1.1 min 113:4.211'1111 113:16.7m111 113:4.2111111 110:8.3111111 (n=1) (n=1) (n=1) (n= 2; c.= 10'5M) (n=1) Many tertiary alkyl radicals decay with first-order kinetics presumably via intramolecular hydrogen transfer or B-scissions.45 In general, B-scission occurs readily if 79 6.00E-05 5.00E-05 4.00E-05 3.00E-05 - 2.00E-05 - 7 Concentration [M] Am“ 1.00E—05~ “A AA A all- p 1- ul- 0.00E+00 o 500 1000 1500 2000 Time [min] 120000 (b) 8 1000004 80000- ’5 40000 4- l/Concentration [M"] 20000 0 500 1000 1500 2000 Time [min] Figure 2.7 Kinetics of decay of l-manxyl radical in methylcyclopentane at 23 °C: (a) variation with time of the concentration of l-manxyl radical 2; (b) plot of the inverse concentration of 2 against time. 80 the semioccupied molecular orbital can assume an eclipsed conformation with respect to the bond about to break, or if it brings considerable relief of strain, as in 3- or 4-membered ring cycloalkyl or cycloalkylrnethyl radicals. In bridgehead radicals, both internal strain and the degree of steric exposure of the radical center control their reactivity. The EPR spectra of bridgehead radicals showed that they have lifetimes in solution of the same order of magnitude as other transient alkyl radicals. It is remarkable that even radicals with as much strain as 21 or 26 could be directly observed. The orientation of the semi-occupied molecular orbital (SOMO) particularly influences the rates of unimolecular reactions such as decomposition and rearrangements. Bridgehead radicals are reluctant to rearrange due to unfavorable stereoelectronic efi‘ects. Even radicals with potentially strongly exothermic ring opening processes, such as 20, 21, or 26, require harsh conditions for B-scissions to occur. Generally, in bridgehead radicals the SOMO and the orbitals of the bond to break, Cp-C.,, are poorly aligned for overlap and considerable structural reorganization must take place during rearrangement, which kinetically is inhibited. Instead, bridgehead radicals abstract hydrogen or halogen, add to unsaturated molecules, and take part in combination reactions. Thus, facile rearrangements are not expected for l-manxyl radical. In agreement with the finding that 2 decays by second-order processes, the combination of two 1- manxyl radicals can lead to either 1,1-bimanxyl 29 by dimerization, or l-manxene 30 and 1 by conventional B-disproportionation. The reaction mixtures resulting from the decay of 2 were examined by GC-MS. The samples utilized for product analysis were prepared by generating 2 in high concentration in cyclopropane, toluene or neat di-tert-butyl peroxide, and allowing it to 81 Dirnerization D 29 2 Disproportionation ’ + G 30 l decompose at room temperature. The radical disappearance was monitored by EPR to ensure total consumption of 2. When no more EPR signals were detected, the EPR tubes were opened to air and subjected immediately to GC-MS analysis. The chromatograms obtained from the decay of 2 in cyclopropane or in neat di- tert-butyl peroxide were essentially identical, and besides unreacted 1, showed major peaks at 150, 220, 222, 235, and 237 amu. No 1,1-bimanxyl 29 is detected in either case, but the calculated F -strain in this compound is large, ca. 21 kcal/mol,46 making dimerization less exothermic (AHdim = -18.1 kcal/mol) than for ordinary alkyl radicals. One of the 150 amu peaks in the product mixture was identified as [3.3.3]propellane 31 by independent synthesis47 and GC-MS analysis (Figure 2.8). A second 150 amu product seen by GC-MS is tentatively assigned to l-manxene 30, the Bredt alkene fi'om conventional B- hydrogen disproportionation of 1 (Figure 2.9). That both these products are derived from 2 is confirmed by their absence in samples where 2 has been quenched after short photolysis times by the addition of n-Bu3SnH. The presence of 31 among the decomposition products of 2 was rationalized by a novel e—disproportionation. This process is reminiscent of the case ofhalobicyclo[1.1.1]- 100 4 1P7 1 R i a 1 ’ e i ) l . I so « a J 1 . t . l v 601 t e 4 A 1 150 ’ 4 t b 40 J 79 i _ u 4 j i . n l ' d . I , i a 20 -4 67 i L 3 1 5} i 77:: Y 9.4 H 122 I 1 1 " 1 H i i C . 1 1 fillLi Y 111? Y iiliL Y Y11.Y Y 11 ”1 Y 41% Y Y Li Y 60 80 I00 120 140 160 M/Z I00 4 ”)7 Y R i b C ‘ ) ’ l 80 ~ . a < l l V 60 e 1 . A 1 b 401 . U 4 15Y0 ‘ n i 84 d i 79 l a 20 4 1 ! n 1 57 71 i l' " 122 C 7 . 1' . e 1 53.! 6; I 7.711: ii 9194 1191 141 1 fi 14111 r fil L 1 11 7‘ L1: +liv 1 11iir Y 1; 1 Y 1 Y1 j E1 i Y 60 80 100 120 140 160 M/Z '00 814 R . 1 * 1‘ °) " 1 80 4 H r a 1 57 Ii . 1 < '! 1 J i . V 60 '1 1 r C 4 l A 1 71 (j '07 i b 40 ~ 1 . t: 1 l 1 : d . ' 6,91 l a 20 — 5 '3 n . j g I; 150 c 1 53 67'i 7'9 1‘ 122127 '4' L e 1 2 - Y Y LAY Y #LiliLfi 1"1 iii 1 9) lY r lly‘J—lw Tl Yi Y ij . Y fi 60 80 I00 120 140 160 M/Z Figure 2.8 Mass spectra showing the El fragmentation of a) [3.3.3]propellane 31 (retention time 3.6 min), and of the peaks with 3.6 min. retention time in the chromatograms from the analysis of decomposition products of 2 in b) neat di-tert-butyl peroxide, and in c) cyclopropane, which are assigned to 31. Vttlbt-ttu.r.ith-s-b.r1_1i-1-.r.r11P-i11 ml viiiiliiiFI-P b r P 1i? 1 1 1 u 1 1 -07 i i it bli 1 _ bill-Iii - M .. M f 6.. -1 5 iiLllllli. L. 0 1| 0 I1 III C 5 ilrr 51"1 1M .- 11m 5 - 5 .1 21111111 3111- .u. l l v 4' 11'. 1 Lr D 11 1 2 1 1 1 2 111 1 211 - 1.11.11 1 Z-ll 11111111-1-1...11-.-.H1 (K I 7 t l a. a. r 1 9 - 1-11-11 - . 0-111- 11 - .11.... 7 w-W11111-111 7 1 11.114191 1 01111 111111-1- A 01 1- 1111 1111,- l tlr I IL A L 2 w .m \. 1| 1 m,- 1 - - 11111111111-11-1h-W-.1+ [- - - 11.-. 5 - t I 1 l] l --1 ) i lu 3 Ill-1|- iI-ltt 1].!1 1111- i 5). 1| ‘ 1 l O, 11- 1-1. 9 I 9 9 9 -- - f 1w - 1 - 11111 .1-1- h on A on I- 1 11- .9 .1 1 n on I Wl-H-H 11 “111180 on 1--- ....-1 . .. 7 7.. -1 R W111 -| 11 1 1- 111 11 1.7- 11 11W1 8 W 11- 1- 1 7- -.1. 7 2 7 7 9 - 1- 1 . 1 7 111%.. -11111- -1 - 7111- 1111-1111w11-1- 7 - - - 1 -1..-.-- .1 6 11' 6 .3 1 1 6 5 111..” ,hu 1 [nu .1-.1 1 - 6 a .n n ( IL ll .11 71.1 .3-111111.1111- n.1hr .311111111m. 321-I11- NV» 5 1-1 11111. A 5 311..“ 5 3. 11111-21. ) ) 5 1. ) 5 a 2 b c T r 1411-11141 1Jt1..41-41Ji1i4lt1—lldll4lldlll«ll]lillur JJlJlllflll-‘wr-Jlfl I4llJvi4l-1lfiJlllellq-J11411411mu v.11— 114'11141 .4 «1414 l1-14114-1«114i4114»l<1|14|43 U f.\ ) 0 O O O D O 0 0 0 . 0 U U 0 m 8 6 4 2 ( 8 6 4 2 w 8 6 4 _ Relative Abundance Relative Abundance Relative Abundance M/Z I 60 I40 ‘r l()() 811 ()0 MS analysis of the decomposition products of 2 in b) neat di-tert-butyl time 6.4 min.). and of the peaks with 6.2 min. retention time iii the GC- peroxide, and in c) cyclopropane, which are assigned to manxene 30. Mass spectra showing the El fragmentation of: a) manxane 1 (retention ‘l’ Figure 2.9 84 —> +C©> + +t—Bu0' 1 31 30 —’ — 1-——’ l 2 +n--Bu3SnH a l pent-l-yl radicals, where evidence was found for a new y-disproportionation process in which the y-fluorine (or chlorine) atom was transferred from the 3-fluoro (or 3-chloro) radical to a triethylsilyl or to a second bicyclo[l.1.1]pent-1-yl radical to yield, in both cases, [1.1.1]propellane.48 [3.3.3]Propellane was also detected in reaction mixtures after longer photolysis times, followed by quenching of 2 with n-Bu3SnH; conceivably, 31 may also be formed by bridgehead H abstraction from 2 with tert-butoxyl radicals. The ab initio results in Table 2.6 indicate that s-disproportionation is thermochemically favored over classical B-disproportionation (AHeaimp, = -7 8.3 kcal/mol; Athi,M_ = -3 6.6 kcal/mol) by more than 40 chmol. 2 2 31 l The olefinic strain (OS) of manxene 30 calculated at the HF/6-31G“ level is 7 chmol (see Table 2.6), higher than a previous MM] estimate of3.9 kcal/mol.” os is used to interpret and predict the stability and the reactivity of bridgehead olefins.so 85 Table 2.6 Calculated Heats of Formation and Strain Energies‘ Compound Total Energyb AH;c SE‘I Manxane (l) -428. 17907 -20.4 (-21.2)c 28.0 (27.2) l-Manxyl Radical (2) 42756808 14.6 19.9 1,1-Bimanxyl (29) -855. 16181 11.1 80.8 l-Manxene (30) 42697830 13.0 35.0 [3.3.3]Propellane (31) 42704326 -28.7 14.9 ' In keel/mo]; structures were fully optimized at HF/6-3 16* level, using Spartan 4.0 (Wavefunction Inc., Irvine, CA). " Total energies are given in hartrees, l H = 627.5 kcal/mol. ° Calculated (experimental) from Wiberg’s group equivalents (Wiberg, K. B. J. Org. Chem. 1985, 50, 5285) for l, 29, 30 and 31. The heat of formation for 2 was estimated from the isodesmic reaction vs. isobutane/tert—butyl radical. d Strain energy; from calculated (experimental) AHf and Benson’s group equivalents (Benson, S. W. Thermochemical Kinetics", John Wiley: New York, 1976) for l, 29, 30 and 31, and from the isodesmic motion vs. isobutaneltert-butyl radical for 2. ° Ref. 5. 86 According to empirical rules deduced from comparison of OS values with experimental behavior,49 30 should be an isolable olefin (OS .<_ 17 kcal/mol), kinetically stable at room temperature, at least long enough to allow reactions and spectroscopic measurements to be carried out. A compound can not be unambiguously identified solely on the basis of its mass spectrum and further studies to confirm the assignment of 30 are necessary. Nevertheless, the analysis of the mass spectrum attributed to 30, hints at a compound with the manxane skeleton but with higher unsaturation. Under electron impact manxene could fiagrnent by breaking one of the C3-C7 bonds fiom the fiilly saturated bridges to give the 122 and 135 amu cations by loss of either methyl or ethyl radicals, which is exactly what is observed experimentally (see Figure 2.9). The ratio of 30 to 31 in all runs analyzed by GC-MS is relatively constant, at about 3:1, which suggests that 30 and 31 must be formed by kinetically parallel reaction pathways. Thus, while 31 is thermodynamically favored, 30 is the kinetic product. This is not surprising since 2 is sterically uncongested and does not hinder the approach of a second tert-butoxyl or manxyl radical to give 30, while the bridgehead diradical-like TS en route to 31 needs more internal motion to collapse to [3.3.3]propellane. Under continuous photolysis and thus, high concentration of radicals, 30 may undergo a second H abstraction to form the allylic rt-type radical 32, which then adds intramolecularly to the double bond to form the less strained [3 .3 3]propellane skeleton via 33. Further, combination with another tert-butoxyl radical gives the 222 amu product (Scheme 2.2). We believe abstraction of an allylic H from 30 to be less probable because the resulting radical 34 is severely twisted, hindering allylic conjugation. Addition of tert-butoxyl 87 1‘“ N N N 6:928 +H mm .5928 1 5.8.. / 31:2 NM El .m+ O N 51:8 8.28.0. M? a ! N.N «Begum 1!. =m-t2-O EO=m1t3 1 Cam-:8 + .Osm128 + 1 88 radicals to the double bond of 30 to form 35 is also conceivable (Scheme 2.2); however, in the competition between addition to n-systems versus H atom abstraction, tert-butoxyl has shown almost exclusively H-abstraction,51 which makes the addition pathway less likely. Nevertheless, both 222 and 220 amu products display in their mass spectra intense peaks at 57 amu (tert-Bu“), which confirms tert-butoxyl incorporation (Figure 2.10). Further H abstraction in the already substituted bridge of these two compounds and combination with tert-butoxyl radicals gives rise to minor products, presumably also tert-butoxyl ethers, which do not exhibit the molecular ions in their mass spectra,52 but whose fragmentation parallels that of the above compounds (highest peaks at 235 and 237 amu, respectively; see Figure 2.11). The GC-MS analysis of the reaction mixture obtained from the decay of 2 in toluene is consistent with the above interpretation and all the compounds discussed above can be easily identified in the GC chromatogram. The most intense peak, however, in this case is dibenzyl, confirming production of benzyl radicals fi'om toluene under H atom abstraction conditions. The benzyl radical could not be observed by EPR because of the remarkable persistence of the concomitantly produced l-manxyl radical 2, but photolysis of di-tert-butyl peroxide and toluene alone yields the spectrum of benzyl radical. '4 A new peak, however, appears at 14.4 min. retention time, with a molecular ion of 242 amu, 89 100 107 11l9 a) R .1 e i 57 67 145 1 l 80- 31 1 1‘ ‘ 1 i 1 1 95 163 1 V 601 1 1 [ e -< i } I A . ‘ | 1 b 40« 1 1 1 1 U 1 1 i 11 1 1 1 1 1 1 g 1 1 11 1 111 l 1 135 ’ n 20 -1 I 1 11 11 I 1.11" 1 1 1 ’ c 1 I 1 h 11 '1 11 1111‘ 11 '1 1’ 220 1 e G . ,1111111‘ 111111- f 11114 . 11111'f.11111111.11111111111 .1311; . 1+ . 1 11 1 .1 . 1 . fiff . 1 50 100 150 200 M/Z '00 1 57 11109 [ R . 1 b) C 1 ' 1 l 80 . 1 + a * 1 i . 166 v 60 « e j 617 149 A d 11 119 , B 40: 8'1 91 1 1 , n . 11 1 I d 1 11 1 1 1 a 20‘ 11 111 1' 1 2 ; .l ,- ,- 11 c _, ' 1 1‘ :5 '1'! (‘fl‘iT [ii 11 11111? 711:1; 1. 112T 1‘1 1' Y 1111111? '111' f 11 T V ‘1‘1‘1‘7 fl Y Y 237 , i’fi—I 1 ‘0 100 150 200 M/Z Figure 2.10 Mass spectra showing the El fragmentation of the peaks with a) 7.22 min. and b) 7.41 min. retention times (see Scheme 2.2 for tentative assignments) in the chromatograms from the analysis of the decomposition products of 2 in neat di-tert-butyl peroxide. 9O U1 80. 16] n< --~m -rb;u 1 179 anamascc> 50 100 150 200 on V 1)) rb<-~n>—-rbx 57 221 40‘ IQ \O [J 79 1 119 135 1 67 191 1 111 ' ‘ Y 1 L 11.111.111.111 , 1 50 100 150 anamaacc> 'J O LLJLJJA CD 1. Figure 2.11 Mass spectra showing the El fragmentation of the peaks with a) 10.4 min. and b) 10.6 min. retention times (see Scheme 2.2 for tentative assignments) in the chromatograms from the analysis of the decomposition products of 2 in neat di-lert-butyl peroxide. 9l 100 1 (y 15] 1 R 1 3 1 C 1 . l 30 « . a " 1 1 j 161 : V 60 ‘ r e 1 109 1 A 1 . b 40 — 81 1 u . 242 n 1 1 1 f f fil fl d . r 10.0 1 a 20 « _ 1C" ‘ 53 69 91 1 C 1 G 11 , 1.11. .11 1 11 .1.. 1.1.3.18. 11.1511 .-1 - T . f - , , , ,. . so 100 150 200 250 Ml Figure 2.12 Mass spectra showing the E1 fragmentation patterns of the peak with 14.4 min. retention time (presumably l-benzylmanxane) in the chromatogram that resulted from the analysis of the decomposition products of 2 in di- tert-butyl peroxide/toluene. 92 which is believed, based on appropriate fragmentation, to be l-benzyl-manxane (Figure 2.12). 2.4 Conclusions Surprisingly, unlike their small-ring cousins, simple bridgehead radicals of medium-ring bicycloalkanes have not been reported, although computational results suggest that such species might show unusual stability and/or persistence. Furthermore, to date, persistent alkyl radicals have depended on steric protection by bulky groups around the radical center. The l-manxyl radical 2 is the first example of a persistent simple medium-ring alkyl radical whose exceptionally long lifetime arises not from steric protection, but from the high strain of all its decomposition products. The remarkable persistence and puzzlingly low hyperfine splittings for the B-hydrogens in 2 suggest that even such simple entities as bridgehead alkyl radicals have not yet given up all their secrets. 2.5 Experimental Section General Methods. Melting points were determined on a Thomas Hoover capillary melting point apparatus and are uncorrected. F ourier-transform infrared (IR) spectra were recorded on Manson-Galaxy F T-IR 3000 or Nicolet IR/42 spectrometers; samples were measured either as thin layers on a NaCl plate (liquids) or as KBr pellets (solids). Electron impact (EI) mass spectra were run on a F isons VG Trio-1 MS spectrometer which 93 operates in line with a Hewlett Packard 5890 gas chromatograph for GC-MS measurements. High-resolution mass spectra for analysis of the decay products from 2 were carried out on a JEOL AX-SOSH double-focusing mass spectrometer coupled to a Hewlett-Packard 5890] gas chromatograph via a heated interface. GC separation employed a DBSMS fused-silica capillary column (30 m length x 0.25 mm ID. with a 0.25 um film coating). Direct (splitless) injection was used. Helium gas flow was approximately 1 ml/min. The GC temperature program was initiated at 100 °C with an increase of 10°/min. MS conditions were as follows: interface temperature 280 °C, ion source temperature ca. 250 °C, electron energy was 100 eV, scan rate of the mass spectrometer was 1 s/scan over the m/z range 45-500. Routine 1H and 13 C NMR spectra were obtained at 300 MHz, on Varian GEMINI 300 or VXR-3 00 spectrometers. All spectra were recorded at ambient temperature and are referenced to solvent signals. Peak multiplicities are abbreviated: s singlet, d doublet, t triplet, q quartet, and m multiplet. Coupling constants (J) are reported in Hertz. Two- dimensional HMQC (‘H-detected heteronuclear Multiple Quantum Coherence) and 2D Heteronuclear J-Resolved experiments were performed on a Varian VXR-S 00 spectrometer at 25 °C. EPR spectra were recorded with a Varian E4 X-band spectrometer equipped with a quartz Dewar insert for variable temperature operation. The temperature was controlled by passing N2 gas through cooling coils immersed in liquid nitrogen and was measured by a thermocouple inserted into the flow Dewar immediately below the cavity. Samples were 94 prepared in 3 mm id. quartz EPR tubes (Wilmad), modified with quartz —) Pyrex graded seals so they could be attached to a Kontes Right Angle Hi-Vac valve with a PTFE plug. TheEPR tubes were connected to a Schlenk line through the side arm of the Kontes valve, degassed, and photolyzed directly in the cavity of the spectrometer with the unfiltered light of a 500 W Oriel high-pressure Hg lamp. Absolute values of the g factor were obtained directly from measurements of the microwave fi'equency with a Microwave Inc. EIP Model 25B fi'equency counter and of the magnetic field with a Bruker ER 035M gaussmeter. ENDOR spectra were recorded on a Bruker ESP 300E spectrometer. The di- tert-butyl peroxide used in the EPR experiments was purified by passing it over activated alumina to remove traces of tert-butyl hydroperoxide, followed by distillation at reduced pressure (hp. 50 °C at 90 torr). All air-sensitive reactions were performed in oven-dried glassware using regular syringe/cannula techniques. Gravity and flash column chromatography were performed on E. Merck silica gel (230-400 mesh). Starting materials and solvents were used as supplied from commercial sources or purified according to standard procedures. Cyclohexanone Morpholine Enamine (4). Cyclohexanone (250 ml; 2.4 mol), morpholine (294 ml; 2.4 mol) and a few crystals of p—toluenesulfonic acid were refluxed in benzene (~ 1 1) until no more water was collected in the Dean-Stark trap, and GC analysis of aliquots from the reaction mixture showed total consumption of the cyclohexanone. Usually it takes about 1 day until all the water is azeotropically distilled and separated in the Dean-Stark trap, and the reaction stops. The solvent was removed by vacuum distillation and the enamine was distilled at reduced pressure to give 340.7 g (2.04 mol) of 95 4(bp10m 116-118 °C;1it.53 bp10mm 117-120 °C; yield 85%). IR 1647, 1450 cm"; 1H NMR (300 MHz, CDC13) 5 4.58 (t, 1H), 3.63 (t, 4H), 2.68 (t, 4H), 2.05-1.91 (m, 4H), 1.68-1.42 (m, 4H); 13C NMR (300 MHz, CDC13) 6 145.5, 100.5, 67.0, 48.4, 26,8, 24.4, 23.2, 22.8. 2-N-MorpholinyI-bicyclo[3.3.1]nonan-9-one (5). Cyclohexanone morpholine enamine 4 (340.7 g; 2.04 mol) was dissolved in THF (750 ml fi'eshly distilled fi'om Na/benzophenone) and cooled to 0 °C with stirring. Acrolein (136 ml; 2.04 mol) was added dropwise at such a rate that the temperature remained below 10 °C. The homogeneous solution was allowed to warm to room temperature and was stirred overnight. The THF was removed on the rotary evaporator, and the residue distilled at reduced pressure to give 5 (296.6 g; 1.33 mol) as a viscous pale yellow oil (pr m 142- 147 °C; lit.“ bp1mm 141-147 °C ; yield 65%). IR 1713 cm"; 1H NMR (300 MHz, CDC13) 5 3.66 (t, 2H), 3.61 (t, 2H), 2.49-2.1 (m, 2H), 2.31-2.49 (m, 7H), 1.30-2.19 (m, 8H). N-(2-Bicyclo[3.3.l]nonan-9-one) Morpholine N-Oxide (6). To 2-N- Morpholinyl-bicyclo[3.3.1]nonan-9-one 5 (296.6 g; 1.33 mol) were added an equal volume of methanol (600 ml) and hydrogen peroxide, H202 (30% in water; 218.9 g; 1.93 mol). The solution was refluxed for two hours and allowed to cool to room temperature. As the solution was still slightly basic, additional hydrogen peroxide was added (200.6 g H202 30%; 1.77 mol) and the solution was again refluxed for two hours and then cooled to room temperature. To the homogenous reaction mixture Pd/C 10% was added slowly in batches and with vigorous stirring to destroy the excess peroxide, and the resulting suspension was stirred for several days. The palladium was filtered off and the solvent 96 removed on a rotary evaporator at 50 °C and water pump pressure to afiord crude N- oxide 6 (310.7 g; 1.3 mol; yield 98%) as a glassy oil, which was not characterized and was used in the next step without firrther purification. Bicyclo[3.3.1]non-2-en-9-one (7). The crude N-oxide 6 (310.7 g; 1.3 mol), in a flask fitted with a short path distillation head followed by an ice-cooled trap (reversed to avoid plugging) connected in series with a dry-ice trap, was further dried at l torr pressure for one hour. The temperature was then slowly raised to 110-120 °C (in the oil bath) with stirring of the amine N-oxide with a teflon-covered magnetic bar, at which time pyrolysis began. In one large-scale pyrolysis the temperature was raised too rapidly, causing dangerously fast decomposition and pressurization of the system, forcing the distillation head from the flask and spewing resinous material. Proper safety precautions should be taken. After about two hours, the reaction was complete, leaving a large amount of hard resin in the pyrolysis flask. The product which was collected in the traps was poured into 6 N HCl (294 ml) and extracted with ether (4x 140 ml). The ether extracts were washed once with 6 N HCl, 10% Na2C03 aqueous solution, and water, then dried over anhydrous MgSO4 and filtered. The ether was removed on a rotary evaporator, leaving a semicrystalline sweet-smelling ketone. Sublimation at 80 °C and 12 torr yielded colorless crystals of7 (45.15 g; 0.332 mol; yield 255%) with mp 95-96 °C (lit.2b mp 98-99 °C). IR 1730 cm'l; 1H NMR (300 MHz, CDCl3) 5 5.87 (dt, 1H), 5.53 (m, 1H), 2.80-2.32 (m, 4H), 1.98-1.40 (m, 6H); 13C NMR (300 MHz, CDC13) a 216.5, 129.8, 126.9, 47.56, 45.4, 45.3, 36.6, 33.1, 16.8; MS (EI) m/z (relative intensity): 136 M, 77), 108 (12), 94 (10), 95 (64), 91 (23), 80 (53), 79 (100), 78 (14), 77 (33), 68 (24), 67 (53). 97 Bicyclo[3.3.l]nonan-9-one (8). Bicyclo[3.3.1]non-2-en-9-one 7 (45.15 g; 0.332 mol) was hydrogenated with 10% palladium on charcoal (700 mg) in methanol (200 ml), in a Parr hydrogenator at room temperature and 3 atm. The suspension was filtered, the methanol distilled on a rotary evaporator, and the residue sublimed at 80 °C and water aspirator pressure to give colorless waxy crystals of the ketone 8 (43.5 g; 0.315 mol; yield 95%) with a distinct camphor-like odor and mp 155-158 °C (lit.2b 153-155 °C). IR 1725 cm"; 1H NMR (300 MHz, CDC13) 5 2.40 (m, 2H), 2.08-1.96 (m, 12H); 13C NMR (300 MHz, CDCl3) 5 219.3, 46.6, 34.3, 20.6; MS (EI) m/z (relative intensity): 138 M, 36), 122 (30), 93 (22), 82 (40), 81 (75), 80 (35), 79 (38), 68 (25), 67 (100), 55 (27), 41 (45). Bicyclo [3.3.2]decan-9-one (9). A solution of N-methyl-N-nitroso-p- toluenesulfonamide (Diazald; 134.8 g; 0.630 mol) in methanol (1350 ml) was added dropwise to a stirred solution containing bicyclo[3.3.1]nonan-9-one 8 (43.5 g; 0.315 mol), potassium hydroxide (22.4 g; 0.40 mol), water (740 ml) and methanol (130 ml) at 0 °C over a period of 6 hours. The mixture was allowed to warm gradually to 20 °C and was stirred overnight. The suspension was filtered and the filtrate was concentrated in vacuo. The filtered salt was washed with ether, the ether washes were combined with the concentrate, more ether was added, the whole organic phase was washed with water and dried over MgSO4, filtered and the ether was removed in vacuo. The residue in ether- hexane (1:19) was placed on a silica column in the same solvent and eluted to give, in first recovery, bicyclo[3.3.2]decan-9-one 9, which afier vacuum sublimation at 60 °C (10 torr) had mp 177-179 °C, lit.2b mp 177-179 °C (28.73 g; 0.189 mol; yield 70%). IR 1689 cm'l; lHNMR (300 MHz, CDC13) 8 2.84 (m, 1H), 2.48 (d, 2H, J= 6 Hz), 2.24 (m, 1H), 1.92- 98 1.41 (m, 12H); 13C NMR (300 MHz, CDC13) 6 222.1, 46.6, 34.3, 31.7, 24.4, 21.5, 20.6; MS (EI) m/z (relative intensity): 152 M, 50), 110 (24), 109 (3 8), 108 (45), 97 (44), 96 (91), 95 (65), 82 (63), 81 (100), 68 (40), 67 (71). 9—Methylenebicyclol3.3.2]decane (10). To a stirred suspension of potassium tert-butoxide (21.88 g; 0.195 mol) in dry benzene (380 ml; freshly distilled over Na) under nitrogen was added an equirnolar amount of methyltriphenylphosphonium bromide (67.52 g; 0.189 mol), and the mixture was heated to reflux (the oil bath was preheated to 80 °C). After 15 min. most of the benzene was distilled ofl‘ until the temperature of the remaining slurry reached 90 °C. Ketone 9 (28.73 g; 0.189 mol) was added at once as a saturated solution in benzene via a syringe, causing a vigorous exothermic effect and a significant rise in temperature (10—20 °C). Heating was continued for two more hours at 90-100 °C. Pentane (280 ml) and water (140 ml) were added to the cooled reaction mixture with vigorous stirring, the organic layer was decanted, the heterogeneous residue was extracted again with pentane, and the combined organic layers were washed with water and dried (MgSO4). The solvent was removed on a rotary evaporator and the residue was distilled at reduced pressure to afl‘ord pure 9-methylenebicyclo[3.3.2]decane 10 (bpz 60 °C; lit.2b bpzs 67-69 °C; 18.75 g; 0.125 mol; yield 66%). IR 1610 cm"; 1H NMR (300 MHz, CDC13) 5 4.68 (dd, 1H, JAB = 2.7 Hz, JAX = 2 Hz), 4.57 (dd, 1H, JBA = 2.7 Hz, JBx = 2 Hz), 2.85 (m, 1H), 2.54 (m, 2H), 2. 15 (m, 1H), 1.78-1.42 (m, 12H). 9-Epoxymethylenebicyclo[3.3.2]decane (11). Solid m-chloroperbenzoic acid (25.7 g 85%; 0.126 mol) was slowly added in small portions to a mechanically stirred mixture of 9-methylenebicyclo[3.3.2]decane 10 (18.75 g; 0.125 mol) in CHC13 (1250 ml) 99 and aqueous sodium (or potassium) bicarbonate (15.95 g NaHCO3; 0.19 mol in 380 ml H20). The mixture was stirred at room temperature for 2 hours following the addition of the peracid (the consumption of peracid was tested with starch-12 paper) and the two phases were separated. The organic phase was washed successively with 1 N aqueous sodium hydroxide and water, dried (N a2SO4) and filtered. The solvent (CHzClz can be used, too, instead of CHC13) was removed under reduced pressure to yield crude 11 as a mixture of diastereomers, which was firrther purified by sublimation (70 °C; 10 torr) to give waxy colorless crystals (15.6 g; 0.094 mol; yield 75%) with mp 96-97 °C (lit.2b mp 97-98 °C). IR 2915, 2861, 1452 cm"; 1H NMR (300 MHz, CDC13) 6 2.60 (dd, 2H), 2.12 (m, 1H), 2.01 (dd, 1H), 1.83 (dd, 1H), 1.78-1.46 (m, 13H); MS (EI) m/z (relative intensity): 166 (20), 148 (18), 135 (47), 123 (65), 122 (37), 109 (41), 95 (76), 93 (40), 81 (100), 67 (74), and 166(8), 148 (22), 123 (60), 122 (39), 109 (45), 95 (84), 93 (29), 81 (100), 67 (63). 9-Azidomethylbicyclo[3.3.2]decan-9-ol (12). The epoxide 11 (15.6 g; 0.094 mol) in DMF (520 ml) was treated with sodium azide (20.2 g; 0.31 mol) and boric acid (20.2 g; 0.32 mol) at reflux for 3 hours. The solvent was removed in vacuo and the residue was partitioned between ether and water. The ether extracts were washed with water, dried (N a2SO4) and filtered. The solvent was removed carefully on a rotary evaporator leaving 9-azidoethylbicyclo[3.3.2]decan-9-ol 12 as an oily residue (13.8 g; 0.066 mol; yield 70%). IR 3441, 2102 cm'1 (lit.2b IR 3420, 2100 cm"); 1H NMR (300 MHz, CDC13) 6 3.96 (s, 1H), 3.44 (d, 1H, JAB = 13 Hz), 3.31 (d, 1H, JAB = 13 Hz), 2.25- 2.02 (m, 3H), 1.92-1.37 (m, 13H). 100 The Hydrochloride Salt of 9-Aminoethylbicyclo[3.3.2]decan-9-ol (13). The hydroxyazide 12 (13.8 g; 0.066 mol) in ethanol (100 ml) was shaken with hydrogen over Adams catalyst (PtOz x H20; 750 mg) at 3 atm for 2 hours at room temperature in a Parr hydrogenator. The catalyst was removed by filtration and the ethanol was distilled on a rotary evaporator. Dried ether (freshly distilled over Na/benzophenone) was added to the residue and the resulting solution was saturated with gaseous hydrogen chloride (obtained by adding dropwise concentrated H2804 to NaCl) until no more precipitate was formed. The filtered solid was recrystallized from ethanol to afl‘ord white crystals of the hydrochloride salt of 9-aminoethylbicyclo[3.3.2]decan-9-ol 13 (13.1 g; 0.059 mol; yield 90%) with mp 240-242 °C (lit.2b mp 241-242 °C). IR 3225, 3195 cm"; 1H NMR (300 MHz, D20) 6 3.24 (d, 1H, JAB = 13 Hz), 3.04 (d, 1H, JAB =13 Hz), 2.27 (m, 1H), 2.10- 1.44 (m, 15H). Bicyclo[3.3.3]undecan-9- and -10-ones (l4 and 15). The amine hydrochloride 13 (300 mg; 1.37 mmol) in water (6 ml) containing acetic acid glacial (0.3 ml; 5.25 mmol) was treated with sodium nitrite (0.3 g; 4.35 mmol) in water (3.3 ml) dropwise at 0 °C and warmed on a steam bath for 1 hour after the addition. The suspension was cooled and extracted with ether. The ether extracts were washed with water, sodium bicarbonate solution (10%), again water, dried (N a2SO4) and filtered. The solvent was removed in vacuo to yield a semicrystalline white solid, which, based on its'H NMR spectrum, was a 2.7:1 mixture of bicyclo[3.3.3]undecan-9-one 14 to bicyclo[3.3.3]undecan-10-one 15 (227.4 mg; 1.37 mmol; yield 100%). IR 1690 cm1 (lit.2b IR 1690 cm"); 1H NMR (300 101 MHz, CDC13) 5 2.84 (q, 1H, CHC=0 in 14), 2.56 (m, 2Hx0.73 cmc=o in 14 and 4Hx0.23 CH2COCH2 in 15). Bicyclo[3.3.3]undecane (l). The mixture of ketones 14 and 15 (227.4 mg; 1.37 mrnol) in triethylene glycol (30 ml) was heated with hydrazine hydrate (4.51 g; 0.09 mol) and hydrazine dihydrochloride (1.15 g; 0.011 mol) at 130 °C for 2.5 hours. Potassium hydroxide pellets (1.70 g; 0.03 mol) were added cautiously and the temperature was raised slowly to 210 °C with distillation of hydrazine-water. The mixture was heated for a firrther 2.5 hours and the product, bicyclo[3.3.3]undecane I, collected on the cool part of the condenser where it had steam distilled or sublimed. Purification by sublimation (50 °C, 10 torr) afl'orded white crystals of 1 (160.3 mg; 1.05 mmol; yield 77%) with mp 191 °C (sealed tube; lit.” mp 192 °C; lit.3 mp 191-193 °C); 1H NMR (300 MHz, CDC13) 5 2.38 (m, 2H), 1.45-1.55 (m, 18H), in accord with previous literature”; 13 C NMR (300 MHz, CDC13) 6 30.74 (2xCH, J13C-H = 120 Hz), 28.96 (6xCH2, J13¢_H =124.2 Hz), 20.1 (3 xCHz, J13C-H =125 Hz); MS (EI) m/z (relative intensity): 152 (M, 31), 124 (27), 109 (47), 96 (100), 81 (91), 67 (85), 55 (60). [3.3.3]Propellane (30). Our thanks go to Professor Roger Alder, who kindly provided us with [3.3.3]propellanedione, converted to [3.3.3]propellane by Kishner-Wolff reduction according to the literature procedure.“ [3.3 .3]Propellane-3,7-dione (0.15 g; 0.84 mrnol) was added to a mixture of hydrazine (0.8 ml 95%), potassium hydroxide (0.7 g), and triethyleneglycol (3 ml). The slurry was refluxed at 136 °C for 2.5 hours after which the water was distilled from the reaction until the pot temperature reached 220 °C. During the distillation [3.3.3]propellane crystallized on the condenser. The product was 102 removed fi'om the condenser and the distillate by washing with ether. The combined extracts were dried (N a2SO4) and the ether removed by distillation at room temperature to provide a white solid which was further purified by slow sublimation in vacuum to give 53 mg of the highly volatile [3.3.3]propellane 30 (0.35 mmol; yield 42%) with mp 129 °C (lit.48 mp 130 °C); 1H NMR (300 MHz, 0130,) 5 1.53 (s); 13c NMR (300 MHz, CDC13) 5 60.3, 40.3, 24.6; MS (EI) m/z (relative intensity): 150 M, 48), 122 (11), 109 (19), 108 (21), 107 (100), 94 (18), 91 (20), 79 (50). EPR Spectra. Manxane 2 (5 mg) was dissolved in di-tert-butyl peroxide (30 1.11). This solution was placed in a quartz EPR tube and degassed on a vacuum line by 3 freeze- pump-thaw cycles. The solvent, e. g. cyclopropane (ca. 260 11.1), was distilled in and the tube was sealed. Experiments were also carried out with toluene or methylcyclopentane in place of cyclopropane, in which case the fieshly distilled solvent was added to the EPR tube prior to the freeze-pump-thaw cycles. ENDOR Spectra. lH ENDOR resonance measurements were performed on samples containing l-manxyl radicals 2 in toluene. Spin Trapping. Spin trapping experiments were performed by adding a solution of 3 mg 2,4,6-tri-tert-butyl-nitrosobenzene (TBN) in 250 111 toluene, to an irradiated sample of 4 mg manxane l in 25 11.1 di-tert-butyl-peroxide, which contains 2 in concentrations of ca. 10'3-10‘4 M. Identical EPR spectra were obtained by irradiation of a solution of 4 mg manxane, 2 mg TBN and 25 1.11 di-tert-butyl-peroxide directly in the cavity of a Varian E4 spectrometer, with light from a 500 W Oriel high-pressure Hg lamp. 103 2.6 References l The name suggested for this compound was inspired by the similarity between the hydrocarbon structure and the official coat of arms of the Isle of Man, a tiny, independent country surrounded by Ireland, Scotland and England. Most of the inhabitants of the isle, as well as the dialect spoken, are Manx. The emblem, known as triskelion, consists of three armored legs, which seems to be “kicking at Scotland, ignoring Ireland, and kneeling to Englan ”. Nickon, A.; Silversmith, E. F. In Organic Chemistry: The Name Game; Pergarnon Press: New York, 1987; p 122. 2 (a) Leonard, N. J.; Coll, J. c. .1. Am. Chem. Soc. 1970, 92, 6685. (b) Coll, J. c; Crist, D. R; Barrio, M. d. C. G.; Leonard, N. J. J. Am. Chem. Soc. 1972, 94, 7092. 3 Doyle, M.; Parker, W. Tetrahedron Lett. 1970, 42, 3619. l (a) Engler, E. M.; Andose, J. D.; Schleyer, P. v. R J. Am. Chem. Soc. 1973, 95, 8005. (b) A study of the conformational flexibility of manxane by adiabatic mapping revealed two energy minima, corresponding to the C31. and C. conformations. The C. conformation is higher in energy by 5.7 kcal/mol, mostly due to valence angle strain. Sessions, R B.; Osguthorpe, D. J .; Dauber-Osguthorpe, P. J. Phys. Chem. 1995, 99, 9034. 5 Parker, W.; Steele, W. V.; Stirling, W.; Watt, 1. J. Chem. Thermodyn 1977, 7, 795. 6 (a) Leonard, N. J.; Coll, J. C.; Wang, A. H. J.; Missavage, R J.; Paul, I. C. J. Am. Chem. Soc. 1971, 93, 4628. (b) Wang, A. H.; Missavage, R J.; Bym, S. R; Paul, I. C. J. Am. Chem. Soc. 1972, 94, 7100. 7 Murray-Rust, P.; Murray-Rust, J.; Watt, c. I. F. Tetrahedron 1980, 36, 2799. 8 Gundersen, G.; Murray-Rust, P.; Rankin, D. W. H.; Seip, R; Watt, C. I. F. Acta Chem. Scand. 1983, A37, 823. 9 (a) Jang, S.-H.; Bertsch, R. A.; Jackson, J. E.; Kahr, B. Mol. Cryst. Liq. Cryst. 1992, 211, 289. (b) Jang, S.-H.; Lee, H.-I.; McCracken, J.; Jackson, J. E. J. Am. Chem. Soc. 1993, 115, 12623. (c) Dostal, S.; Stoudt, S. J.; Fanwick, P.; Sereatan, W. F.; Kahr, R; Jackson, J. E. Organometallics 1993, 12. 1° Parker, W.; Tranter, R. L.; Parker, W.; Watt, C. I. F.; Chang, L. W. K.; Schleyer, P. v. R J. Am. Chem. Soc. 1974, 96, 7121. “ (a) Foote, c. S.; Woodward, R. B. Tetrahedron 1964, 20, 687. (b) ) Foote, c. s. Ph.D. Thesis, Harvard University, 1962. 104 ‘2 Fitjer, L.; Quabeck, U. Synth. Commun. 1985, 15, 855. ‘3 Anderson, W. K.; Veysoglu, T. J. Org. Chem. 1973, 38, 2267. 1‘ Radicals (R') can be generated by ultraviolet irradiation of a static solution of di-tert- butyl peroxide in the presence of a hydrogen donor (R-H). Krusic, P. J .; Kochi, J. K. J. Am. Chem. Soc. 1968, 90, 7155. 15 Hudson, A; Jackson, R. A J. Chem. Soc., Chem. Commun. 1969, 1323. 1‘ McMillen, D F.; Golden, D. M. Annu. Rev. Phys. Chem. 1982, 33, 493. 17 Hudson, A; Hussain, H. A J. Chem. Soc. B 1969, 793. ‘8 HF/6-3 16* total energies for: propane -118.26365 H; isopropyl radical -117.63614 H; isobutane -157.29898 H; tert-butyl radical -156.67501 H. AH;(propane) = -25.0 kcal/mol see Lias, S. G.; Bartmess, J. E.; Liebman, J. F .; Holmes, J. L.; Levin, R D.; Mallard, W. G. J. Phys. Chem. Ref Data 1988, 17, supplement 1; AHAisopropyl radical) = 21.1 kcal/mol (BDEan = 98.2 kcal/mol) see Russell, J. J .; Seetula, J. A; Gutman, D. J. Am. Chem. Soc. 1988, 110, 3092; AH1(isobutane) = -32.3 kcal/mol and AHAtert-butyl radical) = 11.6 kcal/mol (BDE,.,,M..H = 96.0 kcalmol) see Gutman, D. Acc. Chem. Res. 1990,23,375. ‘9 Roberts, B. P.; Steel, A. J. J. Chem. Soc., Perkin Trans 2 1994,2155. 2° Dye, J. L.; Nicely, V. A. J. Chem. Educ. 1971, 48,443. 2‘ (a) Jackson, R A J. Chem. Soc, Perkin Trans 2 1983, 523. (b) Jackson, R A J. Magn. Reson. 1987, 75, 174. (c) Jackson, R. A. J. Chem. Soc., Perkin Trans 2 1993, 1991. 22 (a) Kurreck, H.; Kirste, B.; Lubitz, W. Angew. Chem, Int. Ed Engl. 1984, 23, 173. (b) Kurreck, H. In Electron Nuclear Double Resonance Spectroscopy of Radicals in Solution: Application to Organic and Biological Chemistry; VCH: New York, 1988. 23 The semiempirical MO method known as INDO has been developed by Pople, Beveridge and Dobosh and represents the lowest level of approximation on which unpaired electron distributions in free radicals can be accommodated, since the one-center exchange integrals are necessary to introduce spin exchange polarization effects. Pople, J. A; Beveridge, D. L.; Dobosh, P. A. J. Chem. Phys. 1967, 47, 2026. 2‘ Heller, c.; McConnell, H. M. J. Phys. Chem. 1960, 32, 1535. 25 (a) Griller, D.; Ingold, K. U. J. Am. Chem. Soc. 1974, 96, 6715. (b) Kochi, J. K. Adv. Free Radical Chem. 1975, 5, 189. 105 26 (a) Geske, D. H. Prog. Phys. Org. Chem. 1967, 4, 125. (b) Lloyd, R V.; Wood, D. E. J. Am. Chem. Soc. 1977, 99, 8269. (c) Kemball, M. L.; Walton, J. C.; Ingold, K. U. J. Chem. Soc., Perla'n Trans 2 1982, 1017 . (d) MacCorquodale, F.; Walton, J. C. J. Chem. Soc., Faraday Trans 1 1988, 84, 3233. (e) Ingold, K. U.; Walton, J. C. Acc. Chem. Res. 1989, 22, 8. 27 Alder, R W.; Sessions, R. B.; Symons, M. C. R. J. Chem. Res, Synop. 1981, 82. 28 Symons, M. C. R; Chandra, H.; Alder, R. W. J. Chem. Soc., Chem. Commun. 1988, 844. ‘9 (a) Paddon-Row, M. N.; Houk, K. N. J. Am. Chem. Soc. 1981, 103, 5046. (b) Paddon- Row, M. N.; Houk, K. N. J. Phys. Chem. 1985, 89, 3771. 3° Walton, J. C. J. Chem. Soc., Perkin Trans 2 1988, 1371. 3‘ (a) Della, E. W.; Elsey, G. M.; Head, N. J.; Walton, J. C. J. Chem. Soc., Chem. Commun. 1990, 1589. (b) Della, E. W.; Head, N. J .; Mallon, P.; Walton, J. C. J. Am. Chem. Soc. 1992,114, 10730. 32 Krusic, P. J.; Rettig, T. A.; Schleyer, P. v. R. J. Am. Chem. Soc. 1972,94, 995. 33 Walton, J. C. Chem. Soc. Rev. 1992, 105. 3‘ Kawamura, T.; Matsunaga, M.; Yonezawa, T. J. Am. Chem. Soc. 1975,97, 3234. 3’ Maillard, B.; Walton, J. C. J. Chem. Soc., Chem. Commun. 1983, 900. 36 Kawarnura, T.; Yonezawa, T. J. Chem. Soc., Chem. Commun. 1976, 948. 37 Olah; G. A; Liang, G.; Schleyer, P. v. R; Parker, W.; Watt, C. I. F. J. Am. Chem. Soc. l977,99,966. 3“ Karplus, M. J. Am. Chem. Soc. 1963, 85, 2870. 39 (a) Lagercrantz, C. J. Phys. Chem. 1971, 75, 3466. (b) Janzen, E. G. Acc. Chem. Res. 1971, 4, 31. (c) Perkins, M. J. Adv. Phys. Org. Chem. 1980, 17, l. (d) Janzen, E. G.; Haire, D. L. Adv. Free Rad Chem. (Greenwich) 1990, 1, 253. 4° (a) Terabe, S.; Konaka, R. J. Am. Chem. Soc. 1971, 93, 4306. (b) Terabe, S.; Konaka, R J. Chem. Soc., Perkin Trans 2 1973, 369. 4‘ Doba, r; Ichikawa, T.; Yoshida, H. Bull. Chem. Soc. Jpn. 1977,50, 3158. ‘2 Weil, J. A; Bolton, J. R; Wertz, J. E. In Electron Paramagnetic Resonance: 106 Elementary Theory and Practical Applications; John Wiley: New York, 1994, p 497. ‘3 Steinfeld, J. I.; Francisco, J. S.; Hase, W. L. In Chemical Kinetics and Dynamics; Prentice Hall, Inc.: Englewood Cliffs, 1989, p 8. ‘4 Griller, D.; Ingold, K. U. Acc. Chem. Res. 1976, 9, 13. ‘5 Beckwith, A L. J .; Ingold, K. U. In Rearrangements in Ground and Excited States; ed. P. de Mayo, Academic Press: New York, 1980, vol. 1, p 161. ‘6 The fi'ont strain (F -strain) in 29 is given by six CH3-CH3 gauche-gauche interactions, ~ 6x 1.0 kcal/mol, along with twice the Strain caused by bridgehead pyramidalization, 2x7 .3 kcal/mol (see Table 2.5), which adds to approximately 21 kcal/mol. ‘7 Weber, R W.; Cook, J. M. Can. J. Chem. 1978, 56, 189. ‘8 Adcock, W.; Binmore, G. T.; Krstic, A. R; Walton, J. C.; Wilkie, J. J. Am. Chem. Soc. 1995, 117, 2758. ‘9 Olefin Strain (OS) is defined as the difference between the strain energy of the olefin and that of its parent saturated hydrocarbon. Maier, W. F.; Schleyer, P. v. R J. Am. Chem. Soc. 1981, 103, 189]. 5° McEwen, A B.; Schleyer, P. v. R. J. Am. Chem. Soc. 1986, 108, 3951. ’1 (a) Walling, C.; Thaler, W. J. Am. Chem. Soc. 1961, 83, 3877. (b) Erben-Russ, M.; Michel, C.; Bors, W.; Saran, M. J. Phys. Chem. 1987, 91, 2362. ’2 In a single run, the compound with the highest mass fi'agment at 237 amu exhibited a very weak peak at 292 amu, which most probably is the molecular ion. Such a compound would result fi'om the 222 amu product by substitution of a H for a tert-butoxyl group. ’3 Hiinig, S.; Benzing, E.; Liicke, E. Chem. Ber. 1957, 90, 2833. CHAPTER 3 S-MANXINYL RADICAL: A COMPUTATIONAL AND EXPERIMENTAL STUDY Abstract: A modified literature procedure for the preparation of l-azabicyclo[3.3.3]- undecane (manxine) is described. Our attempts to produce 1-azabicyclo[3.3.3]undec-5-yl radical by bridgehead H-abstraction from the amine with tert-butoxyl radicals, or by y- irradiation either of manxine in adamantane matrix, or of 1-azoniatricyclo[3.3.3.0]- undecane bromide or tetrafluoroborate salts, are presented. 107 108 In view of the exceptional persistence of l-manxyl radicals, a logical subsequent target of our study appears to be the bridgehead radical of 1-azabicyclo[3.3.3]undecane (manxine), where the efl‘ect of through-space o interactions with the opposite nitrogen atom are to be probed. The synthesis of manxine 1 described herein represents a modified but efficient route to this compound, based on the original published procedure of Leonard et all; <1”) <11” > 1 2 however, the preparation and characterization of the corresponding bridgehead radical, 5- manxinyl 2, remain an unachieved goal. EPR investigations aiming to produce 2 by H- abstraction fi'om l, or by y—irradiation either of manxine l in an adamantane matrix, or of 1-azoniatricyclo[3.3.3.0]undecane bromide or tetrafluoroborate salts, failed to reveal evidence for the S-manxinyl radical. 3.1 Results and Discussion l-Azabicyclo[3.3.3]undecane 1 (manxine) was prepared following the procedure of Leonard et al.1 fi'om l-azoniatricyclo[3.3.3.0]undecane bromide 3 by reduction with sodium and liquid ammonia (Scheme 5.1). The l-azoniapropellane salt 3 was readily accessible employing the convenient synthesis of Sorrn and Beranekz. Several modifications were introduced, however, in the synthesis of tris(2-carboethoxyethyl)- 109 nitromethane 10 and its reduction to 5,5-bis(2-carboethoxyethyl)-2-pyrrolidone 9. The triethyl ester 10 was obtained by an alternative route which involves a one pot threefold Michael addition of nitromethane to ethyl acrylate in high yield,3 instead of going through the sequential synthesis of tris(2-cyanoethyl)nitromethane, hydrolysis of the trinitrile and esterification of the triacid, as in the method of Sorm and Beranekz. Subsequently, reduction of 10 to the pyrrolidone 9 was successfirlly achieved under moderate pressures (60 psi) at 80 °C with 30% Pd/C as catalyst, whereas initially, drastic reaction conditions (1500 psi and 110 °C) were employed for this chemical transformation. The activated T-l Raney nickel catalyst,4 commonly used in hydrogenations carried out at low pressures (2 60 psi) and temperatures (2 60 oC), failed in our hands to reduce 10 to 9. A similar six—step route to 1-azoniatricyclo[3.3.3.0]undecane chloride was developed by Newcome et al.5 (Scheme 5.2). In an attempt to reduce the number of steps for preparation of 3, we converted 10 to tris(3 -hydroxypropyl)aminomethane in one step by lithium aluminum hydride reduction; the experimental yield, however, was moderate (3 5%) and we made no efl‘orts to improve it filrther. NMR analysis of manxine in CH2C12:CHC13 (1:1) revealed a “fiozen spectrum” near -80 °C, a temperature in close agreement with that found for manxane.l Both 13C and 1H NMR spectra of 1 indicate the unusual nature of the methine carbon and proton. The one-bond C-H coupling constant was estimated as 121i5 Hz for the bridgehead C-H bond in manxine hydrochloride.l Overlap of signals in the off-resonance decoupled spectra of l precluded accurate measurement of the C-H direct couplings at the time of its first synthesis. We obtained the values of the C-H coupling constants in 1 from its 2D 110 .xeolllllall. hmH.~.~.HZ\J+1/ZmE\/' {can cm: Allllml. _ cmaaemuamo on? ._ measure measure e a .58 xlllmllla mz\J+1/z_\J t: $8 ‘llllnlll :80 mo mONBNaNB s2 cmm moafiafiafi moaaaaea moooafiamu a o a ea .58 .5“ cc 2 £8 E 1weamw mz Axlllleeemwan SmoouamoamBQZNO Ammumbmfi .moovmonamo + aozamo 0 moooamcsc .u moooscea fin 08055 111 .53. C 5&2 d M: C m 5.8.2 .2052 +2 2 a 0 mo :0 e5; :0 N .55» mo .53 A55 2 468 $12 m A a N5 29m $12.0 .25. him no 22 68m E. 20 208 .53 20 .532 308 x: 20 05385 \ Cum - 1 - 26. JWMWIS. 2.0 A 48223..“ 20 20196 + ~02 £0 208 20 mm 28:5 112 Heteronuclear J-Resolved6 spectrum taken in CDC13 at ambient temperature, finding an even lower bridgehead C-H coupling, 120 Hz, for the fi'ee amine then for its hydrochloride. The flattening of the bridgehead regions of the bicyclo[3 .3.3] system is confirmed by X-ray crystallographic studies on manxine hydrochloride.7 In the crystal, the manxinium cation possesses C3 symmetry with each of the three constituent eight-membered rings in boat-chair conformation. The internal strain is obvious from the angles obtained by X-ray analysis: 117-120° for the CCC angles in the methylene bridges and 114-116° for the bridgehead CCC and CNC angles. More evidence of the planar nitrogen configuration in 1 comes from basicity measurements,"8 UV1 and photoelectron spectroscopy studies on 1,8 and fi'om linear sweep voltammetrf. The intrinsic basicity of the lone pair p electrons in manxine was measured by equilibrium ion cyclotron resonance techniques in the gas-phase, relative to tris-n-propylarnine.8 Competition between hybridization and strain energy efl‘ects, which oppose each other in 1, results in a proton aflinity 3 chmol lower than that for tris-n- propylamine. Solution-phase basicities show a similar outcome; manxine-HG] is a stronger conjugate acid than quinuclidine‘HCl, i.e. pK. 8.8. vs. 10.05 in 66% aqueous DMF, and 9.9 vs. 10.9 in water, respectively.1 The photoelectron spectrum of manxine, with a remarkable difi'erent appearance from that of an ordinary tertiary amine, displays a sharp and narrow band shifted to lower energies (7.05 eV), which is interpreted as vertical ionization fi'om a preferred planar geometry in 1 to a planar radical cation.8 The exceptional shift to longer wavelength (240 nm; a = 2935 in ether) for the n—rp transition 113 in the UV spectrum of 1, reflects a reduction in the energy difference between the ground state and the excited state, where nitrogen is expected to approach coplanar bonding.l Analogously, the ease of oxidation of 1 (the oxidation peak potential appears at 0.38 V in aqueous alkaline solution compared to 0.73 V for triethylamine) arises fiom relief of angular strain that accompanies formation of the sp2 hybridized aminium radical.9 Other spectroscopic and photophysical studies on manxine 1 include reports of its fluorescence spectrum and adiabatic ionization potential,“10 of the two-photon resonance- enhanced multiphoton ionization (REMPI) spectrum for the lowest excited electronic state of l,11 as well as flash photolysis studies of 1 in acetonitrile solution at 248 nm, where the resultant transient spectra were assigned as the absorption of the solvated aminium radical cation of 112. Aliphatic carbon-centered radicals are significantly stabilized by lone pair donors or acceptors which can delocalize the unpaired electron. ‘3 Despite such additional stabilization by n-delocalization over the N atom in 11, the BDE estimates (HF/6-31G“) of the methine C-H and methylenic C-H bonds next to nitrogen in 1 (Table 3.1), point to the tertiary site in l as the one which affords the greatest strain relief upon H-abstraction. N ..... ll 114 Table 3.1 Calculated Heats of F orrnation, Strain Energies and Bond Dissociation Energies' Compound Total Energyb AH.c SE“ ASE" BDE‘ Manxine 1 44416223 1.5 28.5 (-192) (7.8) S-Manxinleadicalz 443.93979 36.8 19.9 -8.6 87.4 (-27) (7.1) (-0.7) (95.3) 2-Manxinyl Radical 11 443.54295 42.4 23.3 -5.2 93.0 (4.0) (5.3) (-25) (95.7) l-Manxinium-S-yl Radical 12. 443.93979 -6.0 90.0 (178.2) (7.2) (88.8) l-Manxinium Radical Cation 13 443.95605 -18.8 (154.3) (-8.7) 'In kcal/mol; structures were fully optimized at HF/6-3 16* (MNDO) level, using Spartan 4.0 (Wavefunction Inc., Irvine, CA). bTotal energies are given in hartrees, 1 H = 627.5 kcal/mol. ° Heat of formation; calculated fi'om isodesmic reactions vs. trimethylamine, pentane, isobutane and ethane. The BDE estimates were used to calculate heats of formation for the product radicals. d Strain energy; from calculated AH; and Benson’s group equivalents (Benson, S. W. Jhermochemical Kinetics; John Wiley: New York, 1976) for manxine l, and from isodesmic rections vs. isobutane/tert- butyl radical for 2 and 12, vs. propane/isopropyl radical for 11, and vs. trimethylamineltrimethyl- ammonium radical cation for 13. ° Defined vs. SE of manxine 1. fBased on BDE (t-Bu-H) = 96.0 kcal/mol (Gutman, D. Acc. Chem. Res. 1990, 23, 375), and BDE (iso-Pr- H) = 98.2 kcal/mol (Russell, J. J.; Seetula, J. A.; Gutman, D. J. Am. Chem. Soc. 1988, 110, 3092). ‘ HF/6-3 16* total energy for protonated manxine: 44455425 H; MNDO heat of formation for protonated manxine: 165.5 kcal/mol. 115 Bridgehead H-abstraction in the protonated manxine would yield radical 12, where delocalization of the unpaired electron over the opposite bridgehead is precluded by protonation. The strain energy relief calculated for this process is slightly lower (6.0 kcal/mol; Table 3.1) than upon formation of 2 (8.6 kcal/mol; Table 3.1). In view of the puzzlingly low B-hyperfines in l-manxyl radical, it seems of interest to examine the manxinium radical cation 13, too. Flash photolysis of 1 in CH3CN with a KrF excirner laser at 248 nm produced transient spectra with first-order decay, assigned to the aminium radical 13.12 The lifetirne reported for the radical cation 12, of 4.6 us, is lower than for the radical cations of DABCO, 12 us, triethylamine, 14 us, or quinuclidine, 6.3 us. 513 £13 Quinuclidine DABCO The reversibility of the electrochemical oxidation of amines is also a measure of the radical cation lifetime and it has been used as a test to recommend which aminium radicals might be good candidates for EPR studies. This strategy led to the discovery of the exceptionally persistent 9-tert-butylazabicyclo[3.3.1]nonane radical cation 14, whose stability is based on stereoelectronic grounds. 1‘ Rapid loss of a Ca-H proton from tertiary amine radical cations, leading to an easily oxidized aminoalkyl radical and hence very rapid destruction, is usually responsible for their decay, 1‘ whereas in 14 the a-H is constrained to lie in the nodal plane of the formal charge-bearing p-orbital at nitrogen, which results in a dramatic increase in the radical cation lifetime. However, the cyclic voltammetry oxidation wave of 116 l is irreversible,9 which does not leave much hope for the observation of 13 by EPR By analogy with the EPR studies on the radical cations of quinuclidine 1515 and 1,3,6,8- ”3‘\ .. N —‘.. 1' A ‘1 AN Q~—© QED tetraazatricyclo[4.4.].13'8]dodecane 16,16 we attempted to produce 13 by one-electron chemical oxidation of l with tris(p-bromophenyl)aminium hexachloroantimonate in butyronitrile at -100 °C, but no EPR spectra were obtained. It is almost certain that 13 is formed under these conditions, but most likely it reacts so fast that it can not reach detectable concentrations. Aminium radicals are frequently formed in high-energy ionizing irradiation of appropriate amine precursors, however, this is an “over ' ” method and generally not a clean source of radicals. ‘7 We irradiated manxine in chloroform with 6°Co y-rays at 77 K to obtain a strong but unresolved EPR signal, circa 80 G wide. As mentioned previously, aminium radicals’ lifetimes are principally controlled by their rates of deprotonation, although in several instances they appear to decompose by C- C bond cleavage. In highly acidic media the deprotonation rate is decreased and the lifetime of the aminium radicals increases appreciably to allow detection by EPR spectroscopy. Thermal or photolytic decomposition of N-haloamines in highly acidic media has successfillly generated aminium radicals. ‘7 UV photolysis of the appropriate amine'Clz adducts in CF3S03H at 0 to -50 °C produced bridgehead aminium radicals l7, 117 18 and 19, and readily allowed their characterization by EPR18 This alternate method appears as a conceivable route to produce 13 in high enough concentration that would allow detection by EPR and remains to be tested in future work. T, T .. T 513 5 l7 l8 19 We tried several methods to prepare the bridgehead carbon-centered radical 2, however, all attempts to generate 2 in solution or in matrix have been unsuccessfill. It has been shown that impurities added to a solid adamantane matrix undergo selective radiation damage to give trapped fiee radicals which exhibit solution-like, isotropic EPR spectra at room temperature. ’9 X-rays irradiation of aliphatic amines in adamantane matrix cleanly afl’ord a-aminoalkyl radicals. The size of the amine is limited to that which can replace an adamantane molecule in the adamantane crystal lattice without crowding for isotropic spectra to be obtained. In the case of tertiary amines, triethylarnine gives a good spectrum but upon addition of only one more carbon atom (e. g. diethyl-n-propylamine) an isotropic spectrum is not obtained. The radicals difl‘use only very slowly through the adamantane matrix and typically exhibit half-lives of 10 h at room temperature. Incorporation of the amine was accomplished in this study by dissolving adamantane in the desired amine followed by evaporation or by precipitation and filtration. Manxine 2 is slightly larger than adamantane, by ca. 10%, but since both molecules are globular, very close in shape and size, we assumed that 1 might be sumciently flexible to fold into the volume of an 118 adamantane molecule and attempted to generate it in the matrix. We irradiated solid samples of 10% manxine (by weight) in adamantane at 77 K in a 60C0 y-ray source with doses of ca. 1 Mrad to get, however, unresolved weak EPR signals. The analogous 1-azabicyclo[4.4.4]tetradec—6-yl radical 20 was formed by y- irradiation of 1-azoniatricyclo[4.4.4.0]tetradecane tetrafluoroborate, either as the pure salt or in dilute frozen CD30D solution.20 The EPR spectrum of 20 showed a broad quartet of lines (am3 = 24 G) with no significant coupling to nitrogen. Our similar experiments on 1- azoniatricyclo[3.3.3.0]undecane tetrafluoroborate 21 or bromide 3, resulted in strong but featureless EPR spectra from the pure salts. y-Irradiation of 21 or 3 in dilute frozen CD30D solutions produced a strong septet of broad lines (6.5 G), while the matrix developed an intense purple color which disappeared above 150 K. However, the control probe of pure CD3OD yielded upon y-irradiation identical EPR signals, which we believe are due to trapped electrons in the y—irradiated methanol-d4.21 N N+-X' @ 8 20 21 X = BF4' 3 X=Br We also attempted to produce 2 by H-abstraction fi'om 1 with tert-butoxyl radicals. UV photolysis of cyclopropane (250 ul) solutions of 2 (5 mg) and di-tert-butyl Peroxide (25 ul) yielded unresolved, featureless EPR spectra, with widths of circa 30 G (Figure 3.1a). The reactions of amines with photolytically produced tert-butoxyl radicals 119 have been shown previously to occur by H-abstraction at the carbon adjacent to nitrogen and are several orders of magnitude faster than in typical hydrocarbon substrates. This technique, which is perfectly satisfactory for radical generation,22 was equally unsuccessful for Griller et al.”, who meant to characterize by EPR the a-aminoalkyls resulted by H- abstraction from a variety of amines. They concluded that the large number of hyperfine interactions coupled with the general absence of sharp spectral lines preclude easy detection of these radicals. In other cases, however, a-aminoalkyl radicals generated by H- abstraction were unambiguously characterized by EPR.24 Azobisisobutyronitrile (AIBN), which decomposes thermally or photolytically to generate 2-cyano—2-propyl radicals,25 was also employed in reaction with 1. UV photolysis of a solution of manxine l (5 mg) and AIBN (20 mg) in cyclopropane (250 ul) directly in the cavity of a Varian E4 spectrometer afforded a weak transient EPR spectrum (Figure 3.1b) which could be resolved into a triplet (1:1:1; aN = 11.5 G) and a doublet (1:1; an = 3 G). The assignment of this spectrum is by no means obvious, since 2 should exhibit a much larger hyperfine to the geminal hydrogen, while 13 should display at least a quartet due to coupling with the B-hydrogens. In the control experiment, photolysis of AIBN alone in cyclopropane, generated, as expected, the EPR spectrum of the persistent 2- cyanoisopropyl radical, which is not observed when manxine is present. Generation of the bridgehead carbon-centered radical 2 proved to be much more dificult than we initially expected. Our efforts to produce it under a variety of conditions were fi'uitless, but by no means did we use up all the methods developed for making aminoalkyl radicals. Besides adamantane, matrices such as SFh,26 GeCL1,27 camphane,28 120 20 G a) . b) ‘ 20 G J Figure 3.1 The EPR spectra (9.1 GHz) resulting from UV photolysis of manxine in a) di—tert-butyl peroxide/cyclopropane, and in b) AIBN/cyclopropane, at -90 °C. 121 urea inclusion compounds,29 silica gel30 and others, have been used to trap rapidly reorienting free radicals. Radiolytic generation of radicals,30 as well as other chemical and photochemical means, have been successful in particular cases to allow explicit EPR studies. Ultimately, we can at least hope that the knowledge acquired will help us in future endeavors. 3.2 Experimental Methods Melting points were determined on a Thomas Hoover capillary melting point apparatus and are uncorrected. Fourier-transform infrared (IR) spectra were recorded on a Nicolet IR/42 spectrometer. Electron impact (EI) mass spectra were run on a Fisons VG Trio-1 MS Spectrometer which operates in line with a Hewlett Packard 5890 gas chromatograph for GC-MS measurements. Routine 1H and 13C NMR spectra were obtained at 300 MHz, on Varian GEMINI 300 or VXR-300 spectrometers. All spectra were recorded at ambient temperature and are referenced to solvent signals. 2D Heteronuclear J-Resolved experiments were performed on a Varian VXR-SOO spectrometer at 25 °C. 'y-Irradiation experiments were performed on a US Nuclear Corporation variable flux y—irradiator, model E—0117-M-1, by exposing samples inserted in a Dewar flask filled with liquid N; and placed in the cavity areas, to doses of circa 1 Mrad of y—rays. tris(2-Carboethoxyethyl)nitromethane (10). To a stirred solution of nitromethane (15.3 g; 0.25 mol), Triton B (40% benzyltrimethylammonium hydroxide in water; 1 ml) and dirnethoxyethane (40 ml) was added dropwise ethyl acrylate (75 g; 0.75 122 mol) over 30 min at a rate such that a temperature of 72-7 8 °C was maintained. Additional Triton B was added twice when the temperature started to decrease; then stirring was continued for an additional 45 min. Afier concentration in vacuo, the residue was dissolved in CHC13 (250 ml), washed with 0.5 N HCl (100 ml), and then brine (3x80 ml), dried over anhydrous MgSOr, filtered and concentrated in vacuo to afford the crude triester, which was column chromatographed on silica gel eluting with EtOAc/hexane (1 :5) to give triester 10 as a light yellow oil (72.2 g; 0.2 mol; yield 80%). IR 2984, 1734, 1541, 1188 cm'1 (lit.3 IR 1738, 1542 cm"); 1H NMR (300 MHz, CDC13)5 4.08 (q, 6H), 2.24 (m, 12H), 1.20 (t, 9H), in accord with previous reports3; 13C NMR (300 MHz, CDCl;;) 6 171.58, 91.81, 60.83, 30.09, 28.55, 13.99. 5,5-bis(2-Carboethoxyethyl)-2-pyrrolidone (9). The experimental procedure used is based on the reduction of the analogous tris(2-carboethoxymethyl)nitromethane.31 The triester 10 (72.2 g; 0.2 mol) in methanol (120 ml) was hydrogenated over 30% Pd/C (1.3 g) in a stainless steel autoclave at 60 psi and 80 °C for 12 hours. The reaction product was filtered to remove the catalyst, diluted with ethanol, shaken with activated charcoal and filtered. The filtrate was taken to dryness under reduced pressure to give crude 5,5- bis(2-carboethoxyethyl)-2-pyrrolidone 9 (55.9 g; 0.196 mol; yield 98 %) as an oil which crystallized upon standing in the refiigerator (mp 45 °C; lit.2 mp 46 °C), and was used in the next step without firrther purification. IR 2980, 1732, 1691, 1305, 1186 cm"; 1H NMR (300 MHz, CDC13) 5 6.34 (s, 1H), 4.25 (q, 4H), 2.48-2.32 (m, 6H), 2.00-1.94 (m, 123 6H), 1.26 (t, 6H); 13C NMR (300 MHz, CDC13) 6 177.33, 173.05, 60.68, 60.61, 34.53, 30.19, 30.14, 28.93, 14.05. H2—Carboethoxyethyl)—3,5-dioxopyrrolizidine (8). Pyrrolidone 9 (55.9 g; 0.196 mol) was heated on an oil bath at 205-210 °C and 12 torr for 5 hours. The solidified reaction product was triturated with ether, the undissolved portion filtered ofi‘ and recrystallized from ethanol/diethyl ether to afford pure 8-(2-carboethoxyethyl)—3,5- dioxopyrrolizidine 8 (37.5 g; 0.157 mol; yield 80%). mp 95-96 °C (lit.2 103 °C); 1H NMR (300 MHz, CDC13) 6 4.13 (q, 2H), 2.90-1.90 (m, 12H), 1.25 (t, 3H). 8-(3-Hydroxypropyl)—pyrrolizidine (7) and 2,2-bis-(3-Hydroxypropyl)- pyrrolidine (6). A solution of 8 (37 .5 g; 0.157 mol) in THF (200 ml; freshly distilled over Na/benzophenone) was added in the course of 7 hours to a suspension of LiAlIL (19 g; 0.5 mol) in THF (600 ml) maintaining the reaction temperature at 70-80 °C. The reaction mixture was then refluxed for 2 more hours and allowed to stand overnight. The excess LiAlIL was carefully decomposed with water (24 ml) under stirring and cooling. Aqueous sodium hydroxide (36 g NaOH in 180 ml H20) were than added dropwise at 30-40 °C to decompose the reaction complex. The resultant white precipitate was filtered and the THF solution was taken to dryness under reduced pressure and subjected to fractional distillation. 8-(3-Hydroxypropyl)-pyrrolizidine 7 was collected at 140-160 °C and 15 torr (lit.2 bp.2mm 130-150 °C; 13.26 g; 78.5 mmol; 50%). 1H NMR (300 MHz, CDC13) 5 3.53 (t, 2H), 3.04-2.86 (m, 2H), 2.60-2.49 (m, 2H), 1.82-1.54 (m, 13H); 13C NMR (300 MHz, CDC13) 6 72.57, 63.37, 55.28, 40.56, 38.23, 27.27, 24.62. A higher boiling fi’action was 124 collected at 170-175 °C and 1.5 torr (lit.2 bpo. m 165-180 °C) which yielded 2,2-bis-(3- hydroxypropyl)pyrrolidine 6 as a viscous colorless oil (8.5 g; 45.5 mmol; yield 29%).‘H NMR (300 MHz, CDC13) 6 4.20 (s, broad, 1H), 3.51 (m, 4H), 2.92 (t, 2H), 1.82-1.38 (in, 14H); 13C NMR (300 MHz, CDC13) 6 63.71, 62.89, 45.41, 36.98, 35.94, 27.69, 25.81. 8-(3-Bromopropyl)-pyrrolizidine Hydrobromide (5). 8-(3-Hydroxypropyl)- pyrrolizidine 7 (13.26 g; 78.5 mmol) and a solution of HBr 31% in glacial acetic acid (40 ml) were placed in a high-pressure reactor and heated to 100 °C in an oven for 11 hours. The reaction mixture was worked up by driving off the acid under reduced pressure, diluting the residue with water, extracting with ether and taking the aqueous solution to dryness to yield 5 (mp 123 °C; lit.2 mp 123 °C; 23.3 g; 74.6 mmol; yield 95%). 2,2-bis-(3-BromopropyI)-pyrrolidine Hydrobromide (4) The pyrrolidine derivative 6 (8.5 g; 45.5 mrnol) and a solution of 3 1% HBr in glacial acetic acid (80 ml) were placed in a high-pressure reactor and heated to 100°C in an oven for 14 hours. The reaction mixture was worked up by driving off the acid under reduced pressure, diluting the residue with water, extracted with ether and taking the water layer to dryness to afford to 4 (mp 95°C; lit.2 mp 95-96 °C; 16.4 g; 43.2 mmol; yield 95%). l-Azoniatricyclol3.3.3.0]undecane Bromide (3). (a) A solution of 5 (23.3 g; 74.6 mrnol) in water (600 ml) was poured under vigorous stirring over freshly precipitated silver oxide prepared fi'om Silver nitrate (37 g; 0.218 mmol) and sodium hydroxide (8.9 g). The reaction mixture was stirred for 30 rrrinutes, allowed to stand overnight, filtered, the filtrate taken to the boil, again filtered and treated with a calculated amount of picric acid 125 (17.1 g; 74.6 mrnol). The picrate precipitated as yellow needles, was recrystallized fiom 70% ethanol and triturated with water and 48% aqueous I-IBr. The liberated picric acid was extracted with ether, the aqueous solution was filtered with active charcoal and taken to dryness under reduced pressure or precipitated with TI-IF to yield 1-azoniatricyclo- [3.3.3.0]undecane bromide 3 as white prisms (16.6 g; 71.6 mmol; yield 96%). (b) A solution of the hydrobrorrride 4 (16.4 g; 43.2 mrnol) in water (450 ml) was poured over fieshly prepared silver oxide from AgNO3 (33 g; 194 mmol) and NaOH (7.9 g). The resulting mixture was stirred for one hour and allowed to stand overnight. The precipitated AgBr was filtered ofl', the filtrate was briefly taken to the boil and filtered again. The aqueous solution of 1-azoniatricyclo[3.3.3.0]undecane hydroxide was either converted to the picrate as described previously, or transformed directly into the bromide by neutralization with aqueous HBr. The aqueous solution was shaken with ether, filtered over charcoal and taken to dryness under reduced pressure or precipitated with THF to afford 3 (9.6 g; 41.5 mmol; yield 96%). l-Azoniatricyclo[3.3.3.0]undecane picrate: mp 319°C (lit.2 mp 318 °C); 1H NMR (300 MHz, CDC13) 6 9.03 (S, 2H). 3.76 (t, 6H), 2.14 (m, 6H), 2.05 (q, 6H), 1.22 (s, 1H). 1-Azoniatricyclo[3.3.3.0]undecane bromide 3: mp > 275 °C (lit.2 mp > 350 °C); 1H NMR (300 MHz, D20) 5 3.26 (t, 6H), 2.58 (nr, 6H), 2.03 (q, 6H); 13C NMR (300 MHz, D20, TMSP Sodium) 6 93.64, 64.56 (t), 3735, 23.29. l-Azoniatricyclol3.3.3.0]undecane Tetrafluoroborate (20). l-Azoniatricyclo- [3.3.3.0]undecane tetrafluoroborate was prepared either from 1-azoniatricyclo[3.3.3.0]- undecane bronride 3 reacted in aqueous solution with a stoichiometric amount of NaBFa, 126 or from l-azoniatricyclo[3.3.3.0]undecane picrate reacted with a calculated amount of 48% aqueous I-IBFa. The tetrafluoroborate salt 18 was isolated from the reaction mixture by concentration in vacuo and coprecipitation with diethyl ether. Recrystallization from ethanol arrorded pure 20, mp > 275 °C. 1H NMR (300 MHz, CDC13) 5 3.57 (t, 6H), 1.86 (nr, 121-I); 13C NMR (300 MHz, D20, TMSP Sodium) 6 96.89, 67.14 (t), 39.51, 25.74. Manxine (1). 1-Azoniatricyclo[3.3.3.0]undecane bromide 3 (1 g; 4.3 mrnol) was added to liquid ammonia (circa 50 ml) in a well-stirred, cooled flask, and small pieces of fi'eshly cut sodium metal were added. Fresh sodium was added as the blue color disappeared, and the addition was continued until the blue color persisted. The reaction vessel was allowed to warm to room temperature and the ammonia evaporated slowly. Water and ether were carefully added and the ether layer was washed, dried over Na2S04, filtered and evaporated to dryness. The crystalline residue was sublimed at 30 °C and 15 mm to give manxine 1 as a white volatile solid (260 mg; 1.7 mmol; yield 40%). mp 150- 152 °C (lit.“’ mp 150-152 °C); 1H NMR (300 Nfiiz, CD3CN) 5 2.72 (t, 6H), 2.53 (septet, 1H), 1.59-1.43 (m, 1211); 13C NMR (300 MHz, CD3CN) 6 49.54 (3 xNQHz, J13C-H = 134.4 Hz), 31.99 (CH, J13C-H =120.8 Hz), 28.21 (3 xNCHzCHz, J13C-H = 123.9 Hz), 24.28 (3 xNCH2CH2QH2, J13C-H = 123.9 Hz); MS (EI) m/z (relative intensity): 153 (1W, 39), 138 (17), 124 (50), 110 (26), 97 (34), 96 (58), 84 (17), 83 (31), 82 (100), 69 (25), 58 (17), 55 (22), 43 (25), 42 (50), 41 (62). Manxinium Tetrafluoroborate (21). Manxine l (15 mg; 0.1 mrnol) was dissolved in 48% aqueous I-IBF. (20 pl) diluted with water (0.5 nrl). Diethyl ether was 127 added and the resulting white precipitate was filtered to afford pure manxinium tetrafluoroborate 21, mp > 225 °C with decomposition (15.7 mg; 0.065 mmol; yield 65%). 1H NMR (300 MHz, CDC13) 6 8.72 (s, broad, 1H), 3.28 (m, 6H), 2.58 (septet, 1H), 1.87 (m, 6H), 1.65 (m, 6H); 13C NMR (300 MHz, CDC13) 5 49.93 (J13C.H = 140 Hz), 28.09 (J13C.H = 123 Hz), 26.48 (Jl3c_H = 125 Hz), 18.54 (J13c_H = 126 Hz). 3 .3 References ‘ (a) Leonard, N. J.; Coll, J. C. J. Am. Chem. Soc. 1970, 92, 6685. (b) Coll, J. C.; Crist, D. R; Barrio, M. d. C. G.; Leonard, N. J. J. Am. Chem. Soc. 1972, 94, 7092. 2 Sorrn, F .; Beranek, J. Collect. Czech. Chem. Commun. 1954, 19, 298. 3 Weis, C. D; Newkome, G. R. J. Org. Chem. 1990,55, 5801. 4 Dominguez, x. A; Lopez, 1. C; Franco, R. J. Org. Chem. 1961, 26, 1625. 5 Newkome, G. R; Moorefield, C. N.; Theriot, K. J. J. Org. Chem. 1988, 53, 5552. 6 Bax, A T wo-Dimensional Nuclear Magnetic Resonance in Liquids; Delfi University Press: Delft, Holland, 1982; p 99. 7 (8) Leonard, N. J.; Coll, J. C.; Wang, A. H.-J.; Missavage, R J.; Paul, I. C. J. Am. Chem. Soc. 1971, 93, 4628. (b) Wang, A. H.-J.; Missavage, R. J.; Bym, S. R; Paul, I. C. J. Am. Chem. Soc. 1972, 94, 7100. 8 Aue, D. H; Webb, H. M.; Bowers, M. T. J. Am. Chem. Soc. 1975, 97, 4136. 9 Smith, J. R. L.; Masheder, D. J. Chem. Soc., Perkin Trans 2 1976, 47. ‘0 Halperrr, A. M. J. Am. Chem. Soc. 1974, 96, 7655. “ Weber, A. M.; Acharya, A.; Parker, D. H. J. Phys. Chem. 1984,88, 6087. 128 ‘2 Halperrr, A; Forsyth, D. A.; Nosowitz, M. J. Phys. Chem. 1986,90, 2677. 13 (a) Crans, D.; Clark, T.; Schleyer, P. v. R. Tetrahedron Lett. 1980, 21, 3681. (b) Griller, D.; Lossing, F. P. J. Am. Chem. Soc. 1981, 103, 1586. (c) Burkey, T. J.; Castelhano, A.; Griller, D.; Lossing, F. P. J. Am. Chem. Soc. 1983, 105, 4701. 1‘ Nelsen, S. F.; Kessel, C. R. J. Chem. Soc., Chem. Commun. 1977, 490. 1’ Dinnocenzo, J. P; Banach, T. E. J. Am. Chem. Soc. 1988, 110, 971. ‘6 Nelsen, s. F.; Buschek, J. M. J. Am. Chem. Soc. 1974, 96, 6424. n (a) Danen, W. C.; Neugebauer, F. A. Angew. Chem, Int. Ed Engl. 1975, 14, 783. (b) Chow, Y. L.; Danen, W. C.; Nelsen, S. F.; Rosenblatt, D. H. Chem. Rev. 1978, 78, 243. ‘3 Danen, W. C; Rickard, R. C. J. Am. Chem. Soc. 1975, 97, 2303. 19 (a) Wood, D. E.; Lloyd, R. V. J. Chem. Phys. 1970, 52, 3840. (b) Wood, D. E.; Lloyd, R V. J. Chem. Phys. 1970, 53, 3832. 2° Symons, M. C. R; Chandra, H.; Alder, R. W. J. Chem. Soc., Chem. Commun. 1988, 844. 2‘ Bonin, M. A.; Takeda, R; Williams, F. J. Chem. Phys. 1969, 50, 5423. 22 Griller, D. Magn. Reson. Rev. 1979, 5, 1. 23 Griller, D; Howard, J. A.; Marriott, P. R; Scaiano, J. C. J. Am. Chem. Soc. 1981, 103, 619. 2‘ Danen, w. C.; West, C. T. J. Am. Chem. Soc. 1974, 96, 2447. 2‘ Engel, P. S. Chem. Rev. 1980,80, 99. 2‘ Fessenden, R W.; Schuler, R. H. J. Chem. Phys. 1966, 45, 1845. 2’ Roncin, J.; Debuyst, R. J. Phys. Chem. 1969, 51, 577. 28 Bennett, J. E. Mol. Spectr. Proc. Conf 1968, 313. 2° Grinith, o. H. J. Chem. Phys. 1965, 42, 2651. 3° Wardmarr, P; Smith, D. R. Can. J. Chem. 1971, 49, 1869. 129 3‘ Butler, D. E. Eur. Pat. App]. EP 95,278 (Cl. C07D207/267); CA 100: P138947. CHAPTER 4 PROGRESS TOWARD THE SYNTHESIS OF ATRANE-LIKE COMPOUNDS Abstract: A modified literature procedure for the preparation of 3-(2-hydroxyethyl)-1,5- pentanediol (1), along with the syntheses of the novel compounds, 3-(2-aminoethyl)-1,5- diaminopentane (2) and tris(o-hydroxyphenyl)methane (3), as potential precursors to atrane-like bicylics, are described. Our preliminary attempts to cyclize l were unsuccessful so far, resulting in polymeric materials or 4-substituted tetrahydropyrans (6 and 7). Derivatization of 1 led to the novel tris(N-benzyl)methanetriacetamide (12). 130 131 In view of the predicted properties for the heterocyclic medium-ring bridgehead radicals studied computationally in section 1.3, atrane-like radicals emerge as good candidates for examination of organic radical o—type interactions with heteroelements and promise the requisite geometrical stability and intrabridgehead distances appropriate for this work. Atrane Atrane-like Radicals Atranesl are typically derived from the condensation reaction of N(CH2CH2Y)3, where Y is usually OH or NH;, with an appropriate heteroelement halide, orthoester or triarnine. Despite the ease of synthesis of atranes, atrane-like compounds with carbon at the bridgehead have not been reported. Herein, synthetic efl'orts centered on developing routes and eflicient precursors to carbatranes, whose corresponding carbon-centered bridgehead radicals could provide a potentially interesting series of compounds to probe the effects of positioning heteroatoms at the bridgehead opposite to the radical center, are presented. For conceptual Simplicity, these compounds are named as trisubstituted methanes - i.e. 1: tris(2-hydroxyethyl)methane (THEM); 2: tris(2-aminoethyl)methane (TAEM), and 3: tris(o-hydroxyphenyl)methane (THPM). Use of the known THEM l to H H H / 1 1 O ~O OH HOHO NHz ILIzNH2N OH H031; l 2 3 132 make atrane-like bicyclics was unsuccessful so far, resulting in polymeric materials or tetrahydropyrans. The syntheses of the novel TAEM 2 and THPM 3, aimed to favor the desired atrane-like bicyclics with respect to the polymers, were accomplished. Thus, as observed previously for similar compounds, usage of the N-alkylated derivatives of 2 should sterically hinder polymerization, while the rigidity of 3 prorrrises to prefer entropically the monomeric compounds as regards to the polymers. Future syntheses of atrane-like compounds will undoubtedly take advantage of these potential precursors toward novel medium-ring bicyclics. 4.1 3-(2-Hydroxyethyl)-l,5-pentanediol and 3-(2-Aminoethyl)-1,5-pentanediamine 3-(2-Hydroxyethyl)-1,5-pentanediol l (THEM) was synthesized by reduction of triethyl methanetriacetate 4 (3-ethoxycarbonylmethyl-glutaric acid diethyl ester) with lithium aluminum hydride in THF (Scheme 4.1). Paul and Tchelitcherr2 and Nasielski et al.3 report yields of 70% and 50%, respectively, for this reaction. Wetzel and Kenyon4 reduced 4 to THEM 1 with LiBH. in 47% yield, while Lukas et al.’ obtained 87% yield for the LiAlH. reduction of trimethyl methanetriacetate. In order to improve the above experimental yields, the alurrrinum alkoxide obtained by hydride transfer from LiAlH4 to the ester STOUps in 4, was hydrolyzed at 0 °C followed by addition of aqueous NaOH at 30-40 °C to tI'ansforrn the aluminum hydroxide into an easily filterable granular precipitate.6 This mOdification of the published procedure gave almost quantitative yields in THEM from 4. 133 ll) u v fl08 m .523 5% 29m N m ..S £2. aw. Doom E m . . o mfiommommovom tail Smooonmovom 16:5": .2 A :05 829m mnw + A mooo ammo Doom £08 £08 m 08 m Doom m Doom 2158 m + A 5 £08 29m .8205 m __ A4982 29m UN on 80 :4 625.2 2 A Doom a «.58 ms .52 A m How—M noon :9... ..sz 620 + .5. HQ vac—Em 134 The Na metal reduction of triethyl methanetriacetate 4 in ethanol, reported by Walton? to produce 1 in 80% yield, was unsuccessful in our hands. Other reducing systems such as NaBHJLiCl8 and NaBHJAlCh" employed with 4, and BH3-THF‘° used with methanetriacetic acid, yielded only partially reduced products. Triethyl methanetriacetate 4 was obtained based on previously published methods, by Michael addition of the diethylrnalonate anion to diethylglutaconate 5.3“ The diethyl glutaconate employed in the synthesis of 4 was prepared either fi'om Na diethylrnalonate and chloroform by the method of Kohler and Reid”, with moderate yields (50%), or fi'om Na diethylrnalonate and diethyl ethoxymethylenemalonate,12 available commercially, with experimental yields which consistently exceeded the literature values (80%). Previous reports showed that bicyclic ortho esters can be synthesized by directly reacting a trio] with strong organic acids such as trifluoroacetic acid, di- and trichloroacetic acid, or 3,5-dinitrobenzoic acid. 13 The difference in the experimental behavior of acids having electron attracting groups from other weaker acids, which give only mixtures of ordinary esters, was rationalized on the basis of dissirnilarities in equilibria or in rates. 13’” As illustrated below for the general reaction of 2-hydroxymethyl- 2-methyl-1,3-propanediol with RCOOH, which renders stable bicyclic[2.2.2] ortho esters, when R is a strong electron-attracting group, the concentration of intermediate A would be increased.13 The relative rate of reaction of B with water to regenerate A as compared to the rate of cyclization to product is critical. An electron-attracting R group could inductively increase the positive charge on the carbon atom of B and in this way facilitate intramolecular attack by the third hydroxyl to yield the orthoester. By analogy, the 135 0‘13 012011 G13 0'1on 0'13 + 11+ 84' - H20 + H O r \H o o 2 no X R 113-(IV R R H 2-Hydroxymethyl-2- Bicyclic -methyl-l,3-propanediol A B ortho ester condensation of THEM l with trifluoroacetic acid was attempted under similar experimental conditions. The product of the reaction, however, was identified as the novel 4-substituted tetrahydropyran 6 instead of the desired [3 .3.3] bicyclic. Thus, as noticed also in other reactions employing 1, cyclization to the strainless tetrahydropyranic ring can become a major impediment to the desired derivatization of 1. Formation of 6 can be easily understood considering the leaving group aptitudes of trifluoroacetate anion and the favorable six-membered ring closure in which -OH displaces CF 3COO' instead of adding to the OCOCF3 (Scheme 4.2). The resulting 4-(2-hydroxyethyl)-tetrahydropyran is esterified to form ultimately trifluoroacetate 6. Analogously, reaction of l with the dimethylforrnarnide dimethyl ketal, HC(OCH3)2N(CH3)2, afl‘orded 4-(2-hydroxyethyl)- tetrahydropyrane 7, a compound whose synthesis was previously reported,”15 but which, to our knowledge, has not been characterized prior to our work. In a selective spin decoupling experiment16 the lH-IH splitting patterns in 6 were established (see experimental part). In addition, THEM l reacted with trimethylborate, tris(dimethylamino)borane, tris(dimethylamino)phosphine or tris(dirnethylamino)silane under a variety of experimental 136 b 20 02 no mo © 432822520902 2 2 m6 oz 74. e. o 4 cease. - p \j _ 2 mo \7; 08.6 20 on 1 me A 1 mo e mooonmo o o Mmuooo mo 41 1:33:38 .. m m m m 2 a... .335 137 conditions (longer or shorter reaction times, slower or concomitant additions of reactants, high dilutions) to form insoluble oligomeric/polymeric materials. Such polymers might be able to undergo pyrolytic breakdown to yield the bicyclic orthoesters,l6 but we found that heating (up to 220 °C) under high vacuum (10'5 torr) did not produce any sublimable monomeric compounds. Apparently, the bicyclic carbaboratrane forms when reaction of l with B(OCH3)2 or B(N(CH3)2)3 is carried out in pyridine, as shown by the decay of the 1H NMR signals of 1 and the growth of new triplet signals, presumably due to the monomer. However, we failed to isolate this compound fi'om the reaction mixture. THEM l was also converted into the novel TAEM 2 via the reaction sequence presented below, which follows a modified procedure for synthesis of alicych primary H H ) H2, 3 atm. ) H210 98% T50) 50% EtOH NHZHZN OH OTs TsO N3 HzN 75°/ 1 8 ° 2 polyarnines fi'om the corresponding alcohols. 17 Reaction of 1 with p-toluenesulfonyl chloride in pyridine at 0 °C for 30 minutes afi‘ords the p-toluenesulfonic triester 8 as a white solid. Any increase in the reaction temperature or contact time allows monocyclization to compete with simple substitution, leading, once again, to tetrahydropyrans as major products. Subsequently, the p-toluenesulfonic triester 8 reacts with sodium azide in DMSO to produce the triazide 9, which is used without firrther 138 «— MAAmconomzoonmuvum .58 20.5 1 «Eamonmeo : 808N282 .03 .N mOmZONm A .53. 3 1.5151 mEooommovum 4 4.50m 5.0mm: .305 .m CEONE A m... 080:5 v momooommuvom £08 @08 £08 139 purification for catalytic reduction to 3-(2-aminoethyl)-1,5-pentanediamine 2. N-Substituted tris(arninoethyl)methanes are also accessible via the acid trichloride of methanetriacetic acid 1 l18 (3 -chlorocarbonyhnethyl-pentanedioyl chloride) obtained fi'om 10 by reaction with thionyl chloride (Scheme 4.3). Methanetriacetic acid 10 was either isolated during the synthesis of 4 as the product resulted the hydrolysis and decarboxylation of the tetraester intermediate or obtained by saponification of 4. Conversion of the acid chloride 11 into tris(N-benzyl)methane-triacetamide 12 was accomplished by reacting 11 with benzylanrine in acetonitrile at -15 °C, once again, to avoid nucleophilic substitution with subsequent condensation to piperidin-2,6-dione which is likely to occur at higher temperature. Further reduction of the triarrride was not attempted, but it should easily render the corresponding trianrines. In principle, TAEM 2 and its N-alkyl derivatives could provide access to medium- ring bicyclics just as tris(2-aminoethyl)amine, “tren”, can be used to make azatranes.""l° It is expected that the bulkiness of the N-substituent will reduce the nucleophilicity of the amine firnctionalities and hinder polymerization. 4.2 tris(o-Hydroxyphenyl)methane In our quest for a better ligand to form carbatranes we sought to synthesize the more rigid tris(o-hydroxyphenyl)methane (THMP) 3 by analogy with the tetradentate tripod ligand tris(o-hydroxyphenyl)amine 15,19 which does not easily form transannulated structures, probably owing to reduced flexibility of the bridges imposed by the benzo rings.20 140 Phosphite 16 shows a bicycloundecane framework; no significant N- . -P interaction is present, as illustrated by a N-P distance of 3.14 A in the crystal.21 The phosphate 17 O 1. O N 1. O N 1 I 0 15 l6 17 has probably a structure very similar to 16, with no or very little N- - -P interaction, as judged fi'om the chemical shift of the protons ortho to the N-atom of the ligand.21 The boron complex 18, however, shows an transannular N—rB dative bond of 1.68 A in a strained tricyclo-[3.3.3.0]undecane chelating system.21 The complex reacts with nitrogen bases such as pyridine, quinuclidine and others, to form adducts in which the intramolecular N—rB bond is replaced by one between B and the external nucleophile (see below the adduct with Py, 19). In solution, this nucleophilic displacement, studied by \JA °6i© temperature-dependent 1H NMR spectroscopy, is reversible.22 Analogous complexes with I .rlo 2‘0 O\T 20 A1 show a central N—rAl dative bond, where Al is 5-coordinate in an approximately 141 trigonal-bipyranridal environment, in which the 3 donor O-atoms of the ligand occupy the equatorial and the N-atom one of the axial position; the remaining apical position is occupied by an external nucleophile (OH', pyridine or an O-atom of a second unit; see the dimer 20 obtained by high vacuum sublimation, at 400 °C and 0.05 torr, of the corresponding pyridine adduct).23 tris(o-Hydroxyphenyl)methane24 3 was prepared from tris(o-methoxyphenyl)- methane 14 by ether cleavage with trimethylsilyl iodide (Scheme 4.4).25 Addition of the Grignard reagent of o-bromoarrisole to methyl o-methoxybenzoate produced carbinol 13,26 which was firrther reduced by treatment with refluxing ethanol/HCl26b to afl‘ord tris(o- methoxyphenyl)methane 14. Verkadel" has suggested, based on NMR monitoring and molecular modeling of the possible intermediates, that generation of atranes occurs by transannular bond formation at an initial stage of the reaction, followed by successive stepwise substitution and ring closure. The precursors proposed here toward atrane-fike compounds lack this stabilization by dative-bond formation, which, along with facile polymerization due to their high firnctionality, might be the reason why 3-(2-hydroxyethyl)-l,5-pentanediol l, for example, did not succed to make carbatranes. The less flexible THPM 3, however, pronrises to overcome this insufliciency and appears to be a reasonable candidate for assembly of novel atrane-like bicyclics. m0 142 ..W m Axeon Roma momma .n A1x=u2 MAO—.5 A—mflnmov A A0: 19m 8 m$00 8 EU a 382808896 88.: $2. 1232 .Osm N a m 0 Am 32 5.0. e. .e. um»: A50 0 v... magnum 143 4.3 Experimental Methods Diethyl Glutaconate (5). (a) From diethyl malonate and chloroform: diethyl malonate (32 g; 0.2 mol) was added slowly from a dropping firnnel to a solution of sodium (1 1.5 g; 0.5 mol) in absolute ethanol (400 ml), followed, while the mixture was still hot, by rapid addition of chloroform (16.3 g; 0.13 mol) without losing control of the reaction. The solution boils so vigorously that it was usually necessary to use a double—jacketed coiled condensor, whereas cooling the liquid or adding the chloroform more slowly greatly diminished the yield. The liquid was allowed to stand overnight, when a mixture of the sodium derivative of the ester of dicarboxyglutaconic acid and sodium chloride separated. Water (500 ml) was added under stirring, followed by removal of ethanol on a rotary evaporator. Hydrochloric acid 5% (110 ml) was added and the reaction mixture was extracted with diethyl ether (4x100 ml). The ether was vacuum distilled and the flee ester was hydrolyzed and cleaved by boiling it with aqueous alcoholic hydrochloric acid (30 ml EtOH 95%; 30 ml H20; 30 ml HCl conc.) until solubilization was complete. The glutaconic acid, isolated by evaporating this solution under diminished pressure, was dried by azeotropic removal of water with toluene and esterified by refluxing it for 5 hours with absolute ethanol (50 ml) and concentrated H2804 (0.6 ml). The ethanol was vacuum distilled, cold water was added (200 ml) and the reaction mixture was extracted with ether (4x50 ml). The ether extracts were washed with cold water, dried over Mg2804 and filtered. The solvent was removed on a rotary evaporator and the oily residue subjected to vacuum distillation. The fi'action collected at 90-93 °C and 2.3 torr contained pure diethyl 144 glutaconate 5 (Lit. ‘2 bpos 84-87 °C; 9.3 g; 0.05 mol; yield 50%). (b) From diethyl malonate and diethyl ethoxyrnethylenemalonate: diethyl malonate (32 g; 0.2 mol) was added dropwise to a solution of sodium (4.6 g; 0.2 mol) in absolute ethanol (160 ml), followed by dropwise addition of diethyl ethoxymethylenemalonate (43.2 g; 0.2 mol). After the mildly exothermic reaction was complete, the reaction mixture was allowed to stand at room temperature for 24 hours, during which time the solution solidified. A mixture of glacial acetic acid (30 ml), concentrated hydrochloric acid (20 ml), and water (200 ml) was added, and the solution was extracted with ether. The ether was removed from the extract in vacuo, and the liquid residue was refluxed with dilute hydrochloric acid (60 ml HCl 18%) for 24 hours. The water and the other volatile materials were removed in vacuo, the residue was dissolved in absolute ethanol, dried with MgSO4, filtered, and again concentrated in vacuo. Absolute ethanol (60 ml) and concentrated sulfiiric acid (1 ml) were added and the solution was refluxed overnight. The reaction mixture was processed as in part a, to afi‘ord after vacuum distillation pure 5 (14.9 g; 0.08 mol; yield 80%). 1H NMR (300 MHz, cock) 5 6.95 (dt, 1H, JAB = 15.7 Hz, JAC = 7.2 Hz), 5.87 (dt, 1H, JAB = 15.7 Hz, JAC =1.5 Hz), 4.16 (quintet, 4H, JAB = 7.2 Hz), 3.20 (dd, 2H, JAB = 7.2, JAG = 1.5 Hz), 1.27 (t, 3H, JAB = 7.2 Hz), 1.25 (t, 3H, JAB = 7.2 Hz), in accord with previous literature reports”. Triethyl Methanetriacetate (4). To absolute ethanol (80 ml) was added with cooling sodium (1.86 g; 0.081 mol). When reaction of sodium was complete, diethylrnalonate (14.3 g; 0.088 mol) was added dropwise, followed by fieshly distilled diethyl glutaconate (14.9 g; 0.08 mol). The reaction mixture was heated at reflux for 6 145 hours, cooled and the solvent was distilled. Cold water was added (45 ml), followed by concentrated HCl (6 ml). The solution was ether extracted, the solvent was removed in vacuo and an aliquot of the oily residue was subjected to NMR analysis, confirming the identity of the tetraester intermediate (C2H5C02CH2)2CH-CH(C02C2H5)2I 1H NMR (300 MHz, CDC13) 6 4.15 (q, 4H), 4.08 (q, 4H), 3.70 (d, 1H), 2.98 (sextet, 1H), 2.62-2.38 (m, 4H), 1.22 (t, 6H), 1.18 (t, 6H), in accord with previous literature reports3. Ethanol (20 ml), water (20 ml) and concentrated HCl (20 ml) were added, and the biphasic mixture was refluxed for 2 days when solubilization was complete. The volatile materials were removed on a rotary evaporator, the resulting oil was throughly dried by heating it at 40 °C under vacuum, and esterifed by refluxing it with ethanol (55 ml) and concentrated H2S04 (0.9 ml) for 6 hours. Most of the ethanol was vacuum distilled, the ester was extracted with ether, the ether extracts were washed with aqueous KHCO3 (10%) and cold water, and dried over MgSOz. The ether was removed in vacuo and the resulting oil was vacuum distilled. The fraction collected at 148 °C and 3 torr contained pure triethyl methanetriacetate 4 (Lit.29 bp14 172-173 °C, bplg 200-205 °C; 17.7 g; 0.056 mol; yield 70%). IR 2984, 1734, 1377, 1159, 1030 cm"; 1H NMR (300 MHz, CDCla) a 4.11 (q, 6H, J= 7.1 Hz), 2.73 (heptet, 1H, J: 6.6 Hz), 2.43 (d, 6H, J= 6.6 Hz), 1.22 (t, 9H, J= 7.1 Hz), in accord with previous literature reports“; 13C NMR (300 MHz, CDCl;) 5 171.76, 60.28, 37.60, 28.62, 14.01; MS (EI) m/z (relative intensity): 230 (M, 13), 229 (100), 201 (23), 200 (43), 187 (14), 173 (13), 154 (40), 141 (60), 126 (14), 113 (65), 85 (13). 3-(2-Hydroxyethyl)-1,5-pentanediol (1). Triethyl methanetriacetate 4 (17.7 g; 0.056 mol) in anhydrous THF (184 ml; freshly distilled over Na/benzophenone) was added 146 slowly to a suspension of lithium aluminum hydride (13.3 g; 0.35 mol) in THF (3 70 ml; freshly distilled over Na/benzophenone) under nitrogen, at 0 °C. The reaction mixture was stirred for 15 hours at 40 °C and then cooled with ice to 0 °C. Water (11 ml) was added dropwise with cooling to destroy the excess of LiAlH4, followed by gentle heating to 40 °C and addition of aqueous sodium hydroxide (22 ml NaOH“I 15%). The white suspension was stirred for 3 more hours and filtered. The filtrate was concentrated in vacuo and subjected to vacuum distillation to give quantitatively pure 3-(2-hydroxyethyl)- 1,5-pentanediol 1 (15p2 190-192 °C, lit.2 bp2 189-190 °C; 8.1 g; 0.055 mol; 98% yield). IR 3338, 2931, 1433, 1376, 1055, 1011, 668 cm"; 1H NMR (300 MHz, CDC13)6 5.54 (t, 3H, J= 6.5 Hz), 3.70 (q, 6H, J= 7 Hz), 2.15 (heptet, 1H, J= 7 Hz), 1.41 (q, 6H, J= 7 Hz), in accord with ref. 7; 1H NMR (300 MHz, D20) 5 3.45 (t, 6H), 1.44 (heptet, 1H), 1.36 (q, 6H), in accord with ref. 4; 1H NMR (300 MHz, CD30D) 6 3.49 (t, 6H), 1.58 (heptet, 1H), 1.43 (q, 6H); 1H NMR (300 MHz, Py-d5) 5 6.12 (s, 3H), 4.18 (t, 6H), 2.51 (heptet, 1H), 2.12 (q, 6H); 13c NMR (300 MHz, CDCl3) 5 57.66, 33.68, 21.78; 13C NMR (300 MHz, CD30D) 6 60.84, 37.82, 29.43; MS (CI) m/z (relative intensity): 149 ([M+1]+, 60), 133(8), 131 (20), 129 (12), 125 910), 123 (13), 121 (17), 119 (15), 113 (13), 111 (17), 109 (42), 107 (33), 105 (25). 4-(2-Trifluoroacetoxyethylytetrahydropyran (6). A mixture of triol 1 (0.5 g; 3.4 mmol) and trifluoroacetic acid (0.39 g; 3 .4 mmol) in benzene was refluxed for 2 days. The solvent was removed on a rotary evaporator and the remaining oily residue was distilled at room temperature to afford pure 6 (0.65 g; 2.9 mmol; yield 75%) as a colorless liquid. IR 2931, 2762, 1786, 1220, 1166, 1094 cm"; 1H NMR (300 MHz, CDClg) 5 4.50 147 (t, 2H, CF3COOCH , J = 6.4 Hz), 3.98 (dd, 2H, H2,“, J2...” = 12 Hz, J2,3,, = 4 Hz, J2“. = 2 Hz), 3.39 (td, 2H, H2,“, .1222. = 12 Hz, J2.-3. = 12 Hz, J2..3e = 2 Hz), 1.69 (q, 2H, CF3COOCH2CHz, J = 6 Hz), 1.7-1.6 (m, 1H, H“), 1.62 (apparent dd, 2H, 113.58, .1303. = 12 Hz, J3..2. = 2 Hz, $9.2. = 4 Hz), 1.33 (qd, 2H, H3,5., J3”. = 12 Hz, J3..2. = 12 Hz, J3”,1 = 12 Hz, J3”. = 2 Hz); 13C NMR (300 MHz, CDC13) 5 157.41 (q, chz = 42 Hz, CF3QOO), 114.42 (q, J01: = 285 Hz, QF3), 67.64 (OQHZ), 65.64 (CF3COOQH2), 34.78 (CF3COOCHLQH2), 32.43 (OCHngz), 31.53 (OCHzCHZQ); MS (EI) m/z (relative intensity): 226 (NY, 16), 83 (99), 82 (18), 81 (20), 79 (36), 70 (56), 69 (99), 68 (27), 67 (88), 55 (100), 54 (97), 53 (20), 45 (43), 43 (23), 41 (68), 39 (34). 4—(2-Hydmxyethylytetrahydropyran (7). A mixture of triol l (0.5 g; 3.4 mrnol) and dimethylformamide dimethyl ketal (0.41 g; 3 .4 mmol) in benzene was refluxed for 2 days. The solvent was removed on a rotary evaporator and the remaining oily residue was distilled at room temperature to afford pure 7 (hp; 106 °C, 1it.’ bpz 104-106 °C; 0.3 g; 2.3 mmol; yield 68%) as a colorless liquid. IR 3394, 2925, 2849, 1442, 1092, 1055, 1016 cm' 1; 1H NMR (300 MHz, CDC13) 5 3.91 (dd, 2H, Hm), 3.63 (t, 2H, HOCHZ), 3.34 (td, 2H, H2,“), 1.95 (s, 1H, OH), 172-1.6 (m, 1H, 113.), 1.56 (apparent dd, 2H, £13,“), 1.47 (q, 2H, HOCHZCL-IJ), 1.25 (qd, 2H, 11359930 NMR (300 MHz, CDC13)6 67.96 (OQHz), 59.84 (HOQHZ), 39.56 (HOCHggHz), 32.98 (OCHZQHZ), 31.44 (OCHzCHng); MS (EI) m/z (relative intensity): 130 M, 10), 112(11), 100 (15), 83 (100), 67 (68), 55 (80). tris(TosylhydroxyethyI)methane (8). 3-(2-Hydroxyethyl)-1,5-pentanediol l (120 mg; 0.8 mrnol) was dissolved in pyridine (0.5 ml) and cooled to 0 °C. Tosyl chloride (560 mg; 3 mrnol) was added in small portions and the reaction mixture was stirred at 148 0 °C for halfan hour and filtered. The precipitate was dissolved in chloroform, washed with water, aqueous HCl 10%, again water, dried over Na2SO4 and filtered. Evaporation of the solvent gave tris(tosylhydroxyethyl)methane 8 as a white solid (474 mg; 0.78 mmol; yield 98%). 1H NMR (300 MHz, CDClg) 5 7.75 (d, 6H), 7.34 (d, 6H), 3.90 (t, 6H), 2.42 (s, 9H), 1.63 (heptet, 1H), 1.50 (q, 6H). 3-(2-Azidoethyl)-1,5-pentanediazide (9). tris(Tosylhydroxyethmeethane 8 (474 mg; 0.78 mrnol) and sodium azide ( mg; mmol) were dissolved in dimethyl sulfoxide and stirred under argon at 135 °C for 16 hours. After cooling, the mixture was poured into water and extracted with ether. The ethereal solution was dried over Na2SO4, filtered, treated with activated charcoal and filtered again. Upon evaporation of the solvent the 3- (2-azidoethyl)«l,5-pentanediazide 9 was obtained as a colorless oil (87 mg; 0.39 mmol; yield 50%) which was used immediately in the reduction step to obtain the triarnine. IR 2097 cm“; 1H NMR (300 MHz, CDClg) 5 3.32 (t, 6H, J= 6.9 Hz), 1.78 (heptet, 1H, J= 6.9 Hz), 1.57 (q, 6H, J= 6.9 Hz). 3-(2-Aminoethyl)-1,5-pentanediamine (2). 3-(2-Azidoethyl)-l,S-pentanediazide 9 (87 mg; 0.39 mmol) was reduced in a Parr hydrogenator with H; and PtOz (5 mg) in ethanol (2 ml) at 3 atm and room temperature for 6 hours. The catalyst was filtered and the solvent was removed by vacuum distillation to give 3-(2-aminoethyl)-1,5- pentanediamine 2 as a colorless oil (42 mg; 0.29 mmol; yield 75%). 1H NMR (300 MHz, CDCI3) 5 2.69 (t, 6H, J = 7.3 Hz), 1.56 (s, broad, 7H), 1.41 (q, 6H, J = 7.3 Hz). Methanetriacetic Acid (10). Triethyl methanetriacetate 4 (5.5 g; 20 mmol) was hydrolyzed by refluxing it with water (5 ml), concentrated HCl (5 ml) and ethanol 95% (5 149 ml) for 14 hours. The volatile materials were distilled in vacuo to afford an oil which was firrther dried by azeotropic removal of water with benzene. The resulting solid was recrystallized from ether to give methanetriacetic acid 10 (mp 112.5-113 °C, lit?“ 113.5- 114.5 °C; 3.4 g; 18 mmol; 90% yield). 1H NMR (300 MHz, DMSO-616)?) 12.17 (s, 3H), 2.44 (heptet, 1H, J = 6 Hz), 2.31 (d, 6H, J = 6.1 Hz), in accord with ref. 3. Acid Trichloride of Methanetriacetic Acid (1 1). Methanetriacetic acid 10 (3.4 g; 18 mmol) was heated with excess thionyl chloride (10.7 g; 90 mmol) for 30 minutes at 60 °C. The unreacted SOC12 is distilled and the remaining residue was recrystallized from cyclohexane to afi‘ord the acid trichloride of methanetriacetic acid 11 (mp 58-60 °C, litm 55-60 °C; 1.8 g; 7.4 mmol; 41%).‘H NMR (300 MHz, CDC13) 5 3.14 (d, 6H, J= 6.3 Hz), 3.70 (heptet, 1H, J= 6.3 Hz); 13C NMR (300 MHz, CDC13) 5 57.66, 33.68, 21.78. I tris(N-Benzyl)methanetriacetamide (12). The acid trichloride of methanetriacetic acid 11 (100 mg; 0.4 mmol) was dissolved in anhydrous acetonitrile (25 ml; freshly distilled over CaHz) and cooled to 0 °C with ice. Benzylamine (340 mg; 3.17 mmol) was added under stirring and the resulting precipitate was recrystallized from methanol to afl‘ord tris(N-benzyl)methanetriacetamide 12 (mp > 265 °C; 146 mg; 0.32 mmol; yield 80%). IR 3282, 3069, 1641, 1549, 1454, 744, 695. 1H NMR (300 MHz, CD3OD) 5 7.22-7.43 (m, 15H), 4.44 (d, 6H), 3.98 (s, 6H), 2.25 (heptet, 1H); MS (EI) m/z (relative intensity): 457 (1W, 14), 176 (25), 149 (15), 107 (11), 106 (57), 105 (10), 92 (13), 91 (100). tris(o-Methoxyphenyl)methanol (l3). Magnesium tumings (4.2 g; 173 mrnol) and a crystal of iodine were placed in a thoroughly dried flask under argon. Diethyl ether 150 (50 ml; fieshly distilled over Na/benzophenone) and a small quantity of bromoanisole (2.8 g; 15 mmol) were added into the flask. The flask was gently warmed to initiate the reaction and a crystal of iodine was added if necessary. The onset of the reaction was accompanied by the disappearance of the iodine color, the development of cloudiness and bubbles being released from the metal surface. When the reaction was progressing well, sufficient ether (250 ml) was added to cover the magnesium and the stirrer was set in motion. The remainder of the bromoanisole (25.7 g; 0.137 mmol) was added dropwise at such a rate that the reaction proceeds smoothly. When the solution commenced to cool and only a small amount of metal remains, methyl o-methoxybenzenoate ester (12.6 g; 76 mmol) was added to the well-stirred solution at such a rate that the mixture refluxed gently. The flask was cooled in a pan of cold water during the addition. After the addition was complete, the mixture was refluxed on a steam-bath for one hour, cooled in an ice-salt bath and then poured slowly with constant stirring into a mixture of cracked ice (~ 20 G) and sulfilric acid 2M (15 ml). The resulting white precipitate was filtered, washed with water, dried and recrystallized from benzenezhexane (1:1) to give tris(o-methoxyphenyl)- methanol 13 (mp 180 °C, lit?“ 181 °C; 20 g; 57 mmol; yield 75%). IR 3530, 2936, 1596, 1487, 1460, 1438, 1246, 1027, 755 cm"; 1H NMR (300 MHz, CDC13) 5 7.25-7.12 and 6.90-6.82 (m, 1211), 5.42 (s, 1H), 3.43 (s, 9H), in accord with ref. 26d; 13C NMR (300 MHz, CDC13) 8 157.42, 133.63, 129.71, 128.17, 120.11, 112.38, 80.28, 55.59; MS(EI) m/z (relative intensity): 350 (M, 13), 243 (44), 215 (11), 136 (14), 135 (100), 121 (19), 77 (23). 151 tris(o-Methoxyphenyl)methane (l4). tris(o-Methoxyphenyl)methanol 13 (20 g; 57 mmol) was dissolved in boiling ethanol (400 ml). Concentrated HCl (60 ml) was added and the solution was refluxed until the violet color disappears. Upon cooling the solution, the tris(o-methoxyphenyl)methane l4 crystallized out as white fine crystals. (mp 136 °C, lit.” 136-137 °C; 18.7 g; 56 mmol; 98%). IR 3068, 3009, 2933, 2835, 1587, 1489, 1460, 1437, 1288, 1220, 1163, 1107, 1030, 754 cm"; 1H NMR (300 MHz, CDC13) 5 7.12-7.22 (m, 3H), 6.86-6.69 (m, 9H), 6.50 (s, 1H), 3.66 (s, 9H), in accord with previous literature reports”; 13C NMR (300 MHz, CDClg) 5 157.30, 132.54, 129.62, 126.98, 119.91, 110.72, 55.75, 36.93; MS(EI) m/z (relative intensity): 335 (25), 334 (1W , 97), 319 (16), 303 (39), 227 (15), 226 (16), 195 (17), 181 (20), 165 (19), 152 (15), 121 (100), 107 (39), 91 (52). tris(o-Hydroxyphenyl)methane (3). To a stirred solution of tris(o-methoxy- phenyl)methane 14 (18.7 g; 56 mmol) in chloroform (3 60 ml; freshly distilled over P205) under argon was added neat trimethylsilyl iodide (74 g; 370 mmol; fieshly distilled) via a dry syringe. The reaction was heated at 60 °C on an oil bath for 24 hours. At the completion of the reaction the excess trimethylsilyl iodide was destroyed and the intermediate trimethylsilyl ethers formed during the reaction were hydrolyzed to the alcohols by pouring the reaction mixture into methanol (90 ml). The volatile components were removed at reduced pressure and the residue was filrther purified by column chromatography on silica gel (etherzhexane 2: 1) and recrystallized from benzene to give tris(o-hydroxyphenyl)methane 3. (mp 193-194 °C; 8.2 g; 28 mmol; yield 50%). IR 3344, 3060, 3009, 2883, 2746, 2623, 1612, 1500, 1454, 1394, 1327, 1269, 1180, 1089, 831, 761 cm"; 1H NMR (300 MHz, CDC13) 5 7.35-7.02 and 6.82-6.68 (m, 1211), 5.93 (s, 1H), 152 4.74 (s, 3H); 13C NMR (300 MHz, CDC13)8 153.45, 129.76, 128.52, 127.24, 121.22, 116.29; MS (EI) m/z (relative intensity): 292 (M*, 25), 199 (25), 197 (40), 181 (100), 152 (15), 115 (14). 4.4 References 1 (a) Voronkov, M. G.; Dyakov, V. M.; Kirpichenko, S. V. J. Organomet. Chem. 1982, 233, 1 (b) Verkade, J. Acc. Chem. Res. 1993, 26, 483. (c) Verkade, J. Coord Chem. Rev. 1994, 137, 233. 2 Paul, R; Tchelitchefi', S. Comptes Rendus Acad Sc. Paris. 1951, 323, 1939. 3 Nasielski, J .; Chao, S.-H.; Nasielski-Hinkens, R. Bull. Soc. Chem. Belg. 1989, 98, 375. ‘ Wetzel, R B; Kenyon, G. L. J. Am. Chem. Soc. 1974, 96, 5189. ’ Lukas, R; Strouf, o; Ferles, M. Collect. Czech. Chem. Commun. 1957,22, 1173. 6 Nenitescu, C. D. In General Chemistry; Ed. Didactica 8i Pedagogica: Bucharest, 1978, p 888. 7 Walton, J. C. J. Chem. Soc., Perkin Trans 2 1983, 1043. 8 Hamada, Y.; Shibata, M.; Sugiura, T.; Kato, S.; Shioiri, T. J. Org. Chem. 1987, 52, 1252. 9 Brown, H. C.; Subba Rao, B. C. J. Am. Chem. Soc. 1956, 75, 2582. 1° Yoon, N. M.; Pak, C. S.; Brown, H. C.; Krishnamurthy, S.; Stocky, T. P. J. Org. Chem. 1973, 38, 2786. “ Kohler, E. P.; Reid, G. H. J. Am. Chem. Soc. 1925, 47, 2803. ‘2 Schaefl‘er, H. J.; Baker, B. R. J. Org. Chem. 1958, 23, 626. ‘3 Barnes, R A; Doyle, G.; Hoffman, J. A. J. Org. Chem. 1962,27, 9o. ‘4 (a) Pittman, c. U. Jr.; McManus, s. P.; Larsen, J. w. Chem. Rev. 1972, 72, 357. (b) Guthrie, J. P. Can J. Chem. 1976, 54, 202. (c) Pindur, U.; Miiller, J .; Flo, C.; Witzel, H. 153 Chem. Soc. Rev. 1987, 16, 75. ‘5 Prelog, V.; Kohlbach, D.; Cerkovnikov, E.; Rezek, A.; Piantanida, M. Justus Liebigs Ann. Chem. 1937, 532, 69. 16 Silversteirl, R M.; Bassler, G. C .; Mom'll, T. C. In Spectrometric Identification of Organic Compounds; John Wiley & Sons: New York, 1991, p 198. ‘7 Fleischer, E. B; Gebala, A. E.; Levey, A.; Tasker, P. A. J. Org. Chem. 1971, 36, 3042. 18 (a) Stetter, H.; Stark, H. Chem. Ber. 1959, 92, 732. (b) Font, J .; Lopez, F.; Serratosa, F. Tetrahedron Lett. 1972, 25, 2589. ‘9 Frye, C. L.; Vincent, G. A.; Hauschildt, G. L. J. Am. Chem. Soc. 1966, 88, 2727. 2° Correspondence with professor Edgar Miiller is gratefully acknowledged. Miiller, E. Ph.D. Thesis, ETH, Ziirich, 1982. 2‘ Miiller, E.; Biirgi, H.-B. Helv. Chim. Acta 1987, 70, 1063. 22 Miiller, E.; Biirgi, H.-B. Helv. Chim. Acta 1987, 70, 499. 23 Miiller, E.; Biirgi, H.-B. Helv. Chim. Acta 1987, 70, 520. 2‘ In Chemical Abstracts, tris(o—hydroxyphenyl)methane 3 is mentioned twice, as being identified among the hydrolysis products of phenol-formaldehyde novolac resins (ref. 24b), and produced from formaldehyde and phenol in methanolic sodium methoxide (ref. 24a), respectively. Neither of the two citations explicitly refers to isolation or synthesis of 3. (a) Ulsperger, E.; Richter, L.; Mainas, F. Ger. (East) 46,455, CA 71 :8820d. (b) Jaroslav, R Sb. Prednasek, ‘M4KROYEST 1973” 1973, 1, 191 (CA 80:121401s). In Beilstein, however, there is no reference regarding compound 3. 2’ Jung, M. E.; Lyster, M. A. J. Org. Chem. 1977, 42, 3761. 26 (a) Baeyer, A.; Villiger, V. Chem. Ber. 1902, 25, 3013. (b) Lund, H. J. Am. Chem. Soc. 1927, 49, 1346. (c) Bachmann, W. E.; Hetzner, H. P. Org. Synth. Coll. Vol. 3 1955, 839. (d) Rufanov, K. A.; Kazennova, N. B.; Churakov, A. V.; Lemenovskii, D. A.; Kuz’mina, L. G. J. Organomet. Chem. 1995, 485, 173. 2’ Ml’iller, E.; Biirgi, H.-B. Acta Crystallogr. 1989, C45, 1403. 28 Doyle, M. P.; Dorow, R. L.; Tamblyn, W. H. J. Org. Chem. 1982, 47, 4059. 29 (a) Ingold, C. K; Thorpe, J. F. J. Chem. Soc. 1921, 119, 501. (b) Dreifuss, M. H.; 154 Ingold, C. K. J. Chem. Soc. 1923, 123, 2964. 3° Huszthy, P.; Lempert, K.; Simig, G.; Tamas, J .; Hegedus-Vajda, J. J. Chem. Soc., Miniprint 1985, 7, 2524. 155 CHAPTER 5 CORRELATION OF l3C-‘H COUPLING CONSTANTS WITH ELECTRONIC STRUCTURE 1N BI- AND POLYCYCLOALKANES: A PM3 AND HF/6-3 lG* ANALYSIS Abstract: Ml'iller-Pritchard type (1113C-1H = a x % sc) and related expressions are explored for the prediction, from standard quantum chemical models, of one-bond C-H Spin-spin coupling constants, in a series of bi- and polycyclics. Correlations of experimental lJ with quantities computed from NBC analyses of PM3 and HF/6- ”OJH 31G"I wavefilnctionsflgeometries are critically examined for 39 aliphatic hydrocarbons (>150 C-H sites; J range >100 Hz). Experimental vs. calculated coupling constants are best fit when the model includes contributions from atomic charges (qr; and qc) along with s-character at carbon (% Sc). The proposed semiempirical formula (equation 29) estimates lJ ‘30-‘11 with a 3.8 Hz average deviation from experimental values (62 data points, s.d. = 4.8 Hz). Previously used geometrical measures of hybridization are also discussed. The relationships obtained can be employed to easily predict one-bond C-H coupling constants at tertiary sites in polycyclic saturated hydrocarbons with experimentally useful accuracy. By using common computational chemistry methods for a large data set, we ofi‘er both a predictive tool for the practicing chemist, and insights into the validity of hybridization- based interpretations of coupling. 156 5.1 Introduction Our interest in bicyclo[3.3.3]undecane (manxane), which exhibits unusually high bridgehead reactivity and whose bridgehead radical we have investigated by EPR and ab initio computations,l turned our attention to the use of one-bond C-H spin-spin coupling constants, lJ as a physical property characteristic of hybridization efi‘ects on ‘30-'11 ’ carbon. The bridgehead flattening seen in the bicyclo[3.3.3]undecane system has been related to increased p character in the bridgehead C-H bond, and this hybridization change is reflected in the low value of the corresponding lJ,.,C_1H (120.0 Hz for the methine C in manxane). Historically, experimental lJ ,,C_,H values have been interpreted in terms of the hybridization of the carbon orbitals in C-H bonds. Modern quantum chemical tools now allow easy access to self-consistent geometrical and structural data, even for fairly large molecules. This work describes a search for a simple expression relating experimental tertiary 1J0}; values over a wide range of compounds to the hybridizations obtained fiom routine semiempirical and ab initio calculations. The results present both a broader test of the simple notion that hybridization determines C-H coupling, and a predictive tool that may help confirm structural assignments for unknown compounds. Much of the early interest in one-bond C-H spin-spin coupling constants has centered around theoretical models relating observed lJ '30—‘11 values to hybridization or, more specifically, to the fi'actional 8 character of the carbon hybrid orbital. The interpretation of the mechanism of spin-spin coupling is based on three types of electron- mediated interactions: a) a Fermi contact interaction between the electron and nuclear 157 spins; b) a magnetic dipolar interaction between the electron and nuclear spins, and c) an orbital interaction between the magnetic field produced by the orbital motion of the electrons and the nuclear magnetic dipole.2 It is generally accepted that couplings involving H are dominated by the Fermi contact interaction,3 a quantity that depends on the close approach of an electron to the nucleus and accordingly, is a measure of the density of the bonding electrons at the nuclei. Since only s-orbitals have non-zero values at the nucleus and can therefore contribute to the contact interaction, the magnitude of the Fermi term is a measure of the 8 character of the bond at the two nuclei. Based on the idea that the contact term is predominantly responsible for the C-H interactions, Miiller and Pritchard4 proposed a linear relationship (1) between '1 l3c-‘n and the fraction of 8 character, Sc, in the carbon hybrid orbital bonding to hydrogen. This equation has been used in its original form or in modified versions to make quantitative predictions for nuclear spin couplings and to test theoretical models of molecular systems. lJ.;,CJH = 500 Sc (HZ) (1) Hybridization arguments are based largely upon valence-bond (VB) or molecular orbital (MO) developments from Ramsey’s second-order perturbation formula2 for the Fermi contact term, using the average excitation energy (AE) approximation, ABE.5 Though such empirical assumptions have been criticized,6 the procedure is justified by its success in describing qualitative features of spin-spin coupling constants. Mathematical dificulties associated with the choice of a suitable algorithm for computing the ground state VB wavefunction in large molecules renders the VB method less satisfactory than the MO approach.7 For these reasons, recent calculations of spin-spin couplings have been 158 mainly carried out on LCAO-MO wavefimctions using SOS (sum-over—states),6b’8 FPT (finite perturbation)9 and SCP (self-consistent perturbation) methods. 1° Theory has become indeed very successful in reproducing the experimental nuclear spin-spin coupling constants between directly bonded nuclei in simple molecules. Spin-spin coupling is a subtle phenomenon, however, and considerable computational effort is required to achieve quantitative agreement with experiments even for small systems. 11 For larger molecules of viaa interest it is therefore more convenient to approach prediction of 1J '3C—‘H semiempirical strategy. Equation 2 shows one of the several equivalent forms which results fi'om a SOS MO treatment of the contact interaction in which the average AB is invoked.8 In this expression h is the Planck constant, 113 is the Bohr magneton, y C and y H are the nuclear magnetogyric ratios, sf: (0) is the orbital density of a carbon 2s orbital at the C nucleus, Si. (0) is the orbital density of a hydrogen 1s orbital at the H nucleus, and PSCSH is the carbon 28-hydrogen 1s element of the bond-order matrix. '..C_.,,J = (4/3)2h as 'YC YH (AE)“ s50) 401133.... (2) Interpretation of 1J '3c—ln in terms of hybridization, or carbon s character, is based on the evaluation of the bond-order component Pics“ , and efl‘ectively assumes the factor (AB)1 8% (0) 8%, (0) to be constant. If valence molecular orbitals (MO) are constructed from atomic orbitals Isa, 2sc and 2pc, and overlap integrals are neglected, the PSCSH term is directly proportional to a x b, where a and b represent atomic orbital coeficients for 18a and 28c in the C-H bonding MO, ‘1’}, (3). ‘P8 = a (18H) + b (28c) + C (2pc) (3) 159 According to Miiller and Pritchard, if all other contributing terms are neglected, under the assumptions of perfect pairing and AEE, the coupling constant can be written as: 'J,,C_,H = J. a2 52 (4) where J. is a constant to be determined empirically. In addition, normalization of the MO (again, ignoring overlap) requires a2+b2+c2 = 1, and spu hybridization at carbon implies that b2 = czln. Using the symbol % Sc for the percent 8 character of the carbon atomic orbital in the C-H bond (% Sc = 100 SC), it follows that: % sC = 100 b2 / (b2+c2) = 100 b2/(1-a2)= 100 'J,,C_1H /J., a2 (la?) (5) The well known relationship of Miiller and Pritchard (1) is derived fi'om this semiempirical equation for a2(l-a2) = 0.25, the value for a pure covalent bond, and Jo = 2000, as determined from the formal sp3 hybridization and the observed value of 125 Hz for 11130—111 in methane.“’12 Despite the drastic approximations involved, the linear correlation of 11130—111 with % Sc provides an example of good agreement between experiment and theory, especially for small data sets where hybridization has been crudely estimated as spn (n = 1, 2, 3), based on Simple coordination numbers. 13 The interpretation of this relation has been the subject of much controversy, since substitution may cause large changes in the couplings, in which case the exact correlations can not be foreseen. It is commonly thought that difficulties concerning the linear dependence of 1J on hybridization are only encountered when dealing with the ”OJH effects of heteroatoms. Karabatsos and Orzechl4 pointed out that the contact term is not adequate to explain their observations on the coupling constants for compounds having 160 heteroatoms; in the case of hydrocarbons, however, the other contributing terms (spin- dipolar and orbital) are small or relatively constant and the criterion might still be applicable. Factors of possible importance in determining spin-spin coupling constants other then changes in hybridization have been extensively discussed in the literature: orbital electronegativities,15 efl‘ective nuclear charge,16 bond polarity," and excitation energy;18 in hydrocarbons, where these factors are not expected to vary sharply fiom molecule to molecule, the simple model of Miiller and Pritchard (MP) is generally regarded as valid. Other correlations dealing with hybridization have been proposed in the literature. Maksic’ et al.19 introduced a modified relationship of the lJ,_,C_1H dependence on the % sc character by including the C-H bond overlap, as calculated by the maximum overlap method (MOM). Similar studies have been published by Newton et al.20 and F igeys et al.”, which estimate a linear dependence between the directly bonded C-H spin-spin coupling constant and the percents character in the C-H bonding hybrid, calculated from INDO molecular orbitals via a localized molecular orbital procedure (LMO). In a recent study of bridgehead C-H bonds in a series of polycyclic hydrocarbons, Kovaéek et al.22 found an analogous linear dependence between 1J and % so, calculated from AMI [BC-1H optimized geometries by using the LMO method of Trindle and Sinanoglu. Gil23 argued that 1J couplings should be proportional to (% sc)3’2 as a result of orbital [SC-1H delocalization effects, and that this proportionality should replace the previous linear correlations which involve large additive constants; however, despite the reduction of the additivity constant, his suggestion showed no improvement over previously established 161 empirical correlations of 'J ‘3c—1n with % Sc.22 Hu and Zhan24 used the maximum bond order hybrid orbital (MBOHO) procedure to examine the basic relations proposed by Miiller and Pritchard,4 by Maksié et 211.19 and by GilB, and concluded that better agreement with experiment is obtained for hydrocarbons vs. heterosubstituted hydrocarbons, while best results in both cases are attained when using the relationship derived by Maksié et al. ‘9, in which bond overlap is replaced by bond order. Subsequently, starting fi'om a further theoretical analysis of the Fermi contact coupling interaction with inclusion of ionic terms to the C-H bond, Zhan and Huzs proposed a novel generalized relationship for calculation of 1J ”C-‘H , which includes contributions fi'om hybrid orbitals and net atomic charges, and is suitable for both hydrocarbons and heterosubstituted hydrocarbons. Nevertheless, the optimal form of the relationship between 11130-111 and hybridization at carbon depends upon the compounds investigated, the particular definition of percent s character and the method of calculation (localization of ab initio or semiempirical molecular orbitals into hybrid atomic orbitals,“ or construction of bonding orbitals from hybrid atomic orbitals”). One-bond coupling constants serve as probes of steric strain and angle distortions, since bond angles and hybridization are closely related. Accordingly, correlations of lJ 13c—1lr coupling constants have been explored with geometrical surrogates for hybridization, such as internuclear CCC bond angles, age = (22CCC°)/3,28 and the sum of internuclear bond angle distortions, ZAOCCC = 2(109.5°-ACCC°).29 Miller and Pritchard12 also suggested a dependence of 'J13C_1H values on interorbital rather than internuclear angles, since bent bonds are frequently found in organic compounds”. 162 Mislow31 used this approach to express the relationship between one-bond l3C-lH coupling constants and interorbital bond angles. Tokita et al.32 correlated lJ ”OJ“ with the strain energy calculated by the Allinger force-field method; in this case, however, the data comprise only rings fi'om cyclopropane to cyclohexane, and no correlation was found for other systems33 . Generally, the correlations described above employ parameters derived fi'om experimental geometries, when available. In some cases, geometries assuming standard bond lengths19 or optimized by molecular mechanics or semiempirical methods (INDOZO’ZI, AMI”) were considered. The need to restrict the correlations to a given fiagment type, and to be consistent with regard to geometries for the compounds under study, led us to reevaluate the MP type relationships for strained aliphatic hydrocarbons, where previous methods gave less satisfactory results. With the ready availability of wavefunctions for geometry optimized structures from which hybridization information can be directly drawn, it seems appropriate to seek a correlation by which C-H coupling constants at tertiary sites can be predicted from easily obtained computational results for compounds of nontrivial size. 5.2 Theoretical Model Optimized geometries of compounds 1-39 were obtained by using the semiempirical PM334 and the ab initio I-IF‘/6-31G"‘35 methods. All calculations were carried out employing the computer program SPARTAN .3 6 163 Reported average errors in PM3 calculated molecular geometries are 0.036 A for bond lengths (average errors of 0.009 A and 0.017 A for C-H and C-C bonds, respectively), 39° for bond angles and 149° for torsion angles.37 In general, the PM3 method is an improvement over previous semiempirical methods (MNDO38, AMI”). Errors in bond angles and torsion angles are slightly higher than for the AMI method (average errors 33° for bond angles and 125° for torsion angles), but bond lengths are significantly better reproduced by PM3 calculations (AMl average error in bond lengths is 0.050 A, with average errors of 0.014 A and 0.017 A for C-H and C-C bonds, respectively)” Since optimized geometries are used to compute carbon atom hybridizations, upon which C-H bond distances depend, the PM3 method was chosen for study. The good agreement between HF/6-3 lG* calculated and experimental geometries of systems incorporating small strained rings suggests the application of this moderately large polarized basis set40 as a comparison model for the performance and reliability of the essentially minimal basis set-based semiempirical PM3 method. Hybridizations of carbon atoms and atomic charges in 1-39 were computed fi'om PM3 and HF/6-31G* wavefunctions using the Natural Bond Orbital (NBO) analysis41 as implemented in the SPARTAN package. This method makes use of the first-order reduced density matrix of the wavefunction, which is converted into a localized form corresponding to a conventional valence structure description of the molecule, dubbed the “natural Lewis structure”.41 With the density matrix transformed in a basis of atomic orbitals, the program forms for each pair of atoms the two-center density matrix and the associated matrix depleted of any lone-pair eigenvectors, searching for bond vectors 164 whose occupancy exceeds a preset pair threshold. Ifthere is a simple bond between two atoms, the depleted matrix is expected to have a unique eigenvector with double occupancy, which is decomposed into normalized hybrid contributions from each atom. Hybrids fi'om each center participating in different bonds are symmetrically orthogonalized to remove intraatomic overlap. The set of localized electron pairs found in this way constitutes the “natural Lewis structure” to describe the system. The resulting natural hybrids agree well with hybrids determined by other methods and with known trends such as those summarized in Bent’s rule.26a The natural atomic charges and hybridizations are calculated based on occupancies (natural populations) of the natural atomic orbitals (NAO) on each atom. The NAO’s are the orthonormal atomic orbitals of maximum occupancy for the given wavefunction and are obtained as eigenfimctions of the first-order density matrix. In a study on compounds spanning a wide range of ionic character, Reed et al.42 found computed natural charges to be in good agreement with empirical measures of charge and ionic character. The NBO analysis is applicable at any level of ab initio or semiempirical theory and is computationally efficient, the effort required being modest as compared to that for calculation of the wavefilnction. 5.3 Results and Discussion The 13C NMR chemical shifts and one-bond carbon-hydrogen coupling constants measured experimentally in this work for bicyclo[3.3.1]nonane 36, bicyclo[3.3.2]decane 37 and bicyclo[3.3.3]undecane 39 are presented in Table 5.1. The data Show, as expected, 165 a decrease in the coupling between the bridgehead C and its attached H with successive lengthening of the variable bridge and accordingly, flattening of the bridgehead region. The series of compounds considered in this study, which provides experimental 1J ”0,“ values ranging from 120 Hz to 215 Hz, was obtained by a systematic literature search for small and medium ring saturated bicyclics with reported one-bond C-H coupling constants, and substantially augmented with other polycych saturated hydrocarbons. In addition, this work includes all similar compounds referenced in previous studies. Table 5.2 lists the experimental ‘J13C_1H values for compounds 1-39, together with the percent 8 character % Sc in the C-H bonding hybrids computed by NBC analysis for PM3 and HF/6-31G* optimized geometries. The expected increase in C-H bond % So with decreasing ring size is well reproduced and is particularly evident if closely related compounds are compared. Also, enhanced C-H bond p character accompanied by wide CCC angles is associated with reduced experimental ‘J '30—! couplings. Selected PM3 H and HF/6-31G* geometrical parameters and atomic charges for the bridgehead sites in 1- 39 are included in the Appendix (Table 1A). The changes in the PM3 geometries of 1-39 vs. the corresponding ab initio HF/6-31G* geometries are significant only regarding C-H bond lengths, which are shorter at the ab initio level (without d-type fimctions, included in the 6-31G“ basis set, bonds to heavy elements are consistently too long)43 and correlate surprisingly poorly with the semiempirical values (the correlation coefficient, R, for a linear fit of PM3 vs. HF/6-31G“ C-H bond lengths is 0.8). Correlation of hybridization with C-H bond length is better for the PM3 method (R = 0.97) than for the ab initio I-IF/6- 31G* method (R = 0.85). The atomic orbital coefficients on C and H are more polarized 166 Table 5.1 ‘3 C NMR Chemical Shifis and Experimental lJ 13C—‘H Coupling Constants Compound Carbon 6 (ppm) 1J ‘30-'11 (Hz) Bicyclo[3.3.1]nonane 36" 9 1 27.9 129.4 '/ 2 31.6 127.4 1 3 22.5 125.6 2 3 9 35.0 128.3 Bicyclo[3 .3 .2]decane 37b 9 3 1 33.7 125.2 2 32.9 123.4 1 2 3 22.8 124.3 9 30.4 125.3 Bicyclo[3.3.3]undecane 39c 3 1 30.7 120.0 2 28.9 124.2 1 2 3 20.1 125.0 ' The 13C NMR spectrum of 36 is in agreement with previous literature reports (see ref. 62). b The ”C NMR signals of 37 are attributed to the corresponding carbons based on proton assignments and PVC correlations from the 2D HMQC spectrum of 37. ° The individual assignments of the 13C peaks of 39 are based on the relative intensities of the signals and their multiplicity in the off-resonance proton decoupled spectrumof39. 167 at the HF/6-31G* level of calculation, most likely due to inclusion of d-type functions in the 6-31G* basis set. Regardless of bond length differences, the PM3 and I-IF/6-31G* hybridizations of the carbon hybrids in the C-H bonding orbitals in this work correlate extremely well (% scpm = 1.09 x % Sam/6-316 + 3.22, R = 0.996). Bond angles, 0:“: and 0:00 , change only slightly from PM3 geometries to HF ab initio optimized geometries (the slopes of plots of PM3 bond angles vs. HF/6-31G* bond angles are 0.98 for 92230 and 1.03 for 9:00 , with R values of 0.99, respectively). In previous studies of empirical relationships between 'J 13c-1rr and hybridization or bond angles (summarized in Table 5.3), the choice of compounds was arbitrary and those with large deviations of calculated vs. experimental lJ couplings, such as '3C—‘H strained polycyclics, were generally excluded, obviating meaningfiil comparisons between different correlations. Most studies used both experimental and calculated geometries (employing INDOW'”, CNDO/224'25, AMl22 or 1141/1298” methods) based on standard bond lengths and bond angles, which could be a source of systematic deviations, too. Thus, the “problem” cases (bicyclobutane, cyclopropane) encountered by Szalontaiz” when studying the relation of 1J 1,0,” with ZAGCCC , the sum of internuclear angle distortions (equation 14, Table 5.3), were also problematic for the molecular mechanics based calculations of the 1980’s.44 Conformational averaging was also ignored in most cases. Hybridization parameters were extracted with different methods (MOM‘9, LMOZO’ZI'B, thOHOZ4’25); most gave the same general picture,26’27 but some (e. g. the MOM procedure) gave unsatisfactory results for highly strained cyclopropane ring compounds. Among previously reported MP type relationships (equations 6-10, Table Table 5.2 Experimental One-Bond C-H Spin-Spin Coupling Constants (in Hz), and Calculated % Sc Character of the C Hybrid Forming the C-H Bonds in 1-39 No. Compound 'JBC_1H % sc PM3 % sc HF/6-316“ Symmetryb la h 215° 41.2 34.9 C. 2a éb 212d 41.2 34.9 C2. 3 A 210° 42.5 36.4 C2. 1b w 209° 41.1 34.9 C. 4 178.1h 37.0 31.0 C. 11 169 12a 13a 14a 1c 13b 15a 16 14b 8b 17a 1d 18* 19a 20a 854 9 8575:5- rheaeea 175m 1 74'1 171° 171c 171“ 169p 167.8q 166' 1661‘ 1661 166c 165.7r 165' 1641 36.3 36.1 36.1 35.5 35.9 35.5 36.8 35.3 35.7 33.8 36.6 34.6 33.9 33.8 30.3 30.3 30.1 29.3 29.8 29.2 30.3 29.2 29.6 27.6 30.4 28.7 27.6 27.6 C3v sz sz D3h sz sz sz 170 2b 17b 21“ 7c 22a 23a 24 5b 15b 23b 25 22b 12b 26 @O@ba&fi m@@0%5 163d 161‘ 160c 1588“ 157.9t 157k 154.5u 154.2c 152n 152‘° 151.8m 148.8q 148i 148" 36.4 33.8 32.9 35.9 36.7 36.5 36.5 34.0 34.6 35.2 34.5 34.1 33.0 34.5 30.3 27.7 26.95 29.8 30.7 30.8 30.8 28.2 28.5 29.3 28.5 28.3 27.3 28.8 sz C3v sz D411 C2v sz C3v C3v Dsh 171 27 g 147.9q 32.5 27.8 on 23c % 146“ 33.4 27.5 C2. 6b K) 145f 31.9 26.4 C2 28 gm 144.9q 32.3 26.5 C. 13c @ 144‘ 32.7 26.9 C2. 29 db 141.0w 31.8 26.3 C2. 30 fl 137.01) 30.1 24.5 C31. 14c 6% 137‘ 30.2 24.5 C2. 17c @V 136.2‘ 29.8 24.3 C. 31* % 136f 31.0 25 C2 32* (ll 135x 29.6 24.1 C1 33 Lb 134.3° 29.1 24.0 D3h 19b % 134.2p 29.5 24.2 C. 34.. pm 133.7y 29.9 24.5 C. 172 35 g 1334‘ 29.2 23.8 Td 20b % 133‘ 29.8 24.2 C2. 36* Ah 129.42 28.1 22.8 C2. 7 37* fib 125.2° 26.8 21.5 C. 38a* 1224“ 25.7 21.9 C3 39* £7 120.02 26.5 21.0 C3,. 38b* 111.2" 25.2 20.1 C3 ' For several compounds considered here, various literature reports present different values for the one- bond carbon-hydrogen coupling constants; in such cases the most recent literature reference was considered 5 Symmetry of lowest energy geometry. ° Christl, M Chem. Ber. 1975, 108, 2781. “ Christl, M.; Brilntrup, G. Chem. Ber. 1974, 107, 3908. ° Andrews, G. D.; Baldwin, J. E. J. Am. Chem. Soc. 1977, 99, 4851. ‘Withrich, K.; Meiboom, S.; Snyder, L. C. J. Chem. Phys. 1970, 52, 230. 3 Star * means it may need conformational averaging, even if they are nondegenerate. ‘fDella, E. w; Hine, P. T.; Patney, H. K. J. Org. Chem. 1977, 42, 17. Christ], M.; Hemog, C. Chem. Ber. 1986, 119, 3067. ’ Christl, M.; Leininger, H.; Mattauch, B. Spectros. Int. J. 1983, 2, 184. “ Figeys, H P.; Geerlings, P.; Raeymaekers, P.; Van Lommen, G.; Defay, N. Tetrahedron 1975, 31, 1731. ‘Katz, T. J.; Acton, N. J. Am. Chem. Soc. 1973, 95, 2738. m Olah, G. A.; White, A M. J. Am. Chem. Soc. 1969, 91, 3954. n Hamlin, J. E.; Toyne, K. J. J. Chem. Soc., Perkin Trans 1 1981, 2731. ° Gunther, H.; Herrig, W.; Seel, H.; Tobias, S. J. Org. Chem. 1980, 45, 4329. P Christl, M; Herbert, R. Org. Magn. Reson. 1979, 12, 150. q Lazzaretti, P.; Malagoli, M.; Zanasi, R; Della, E. W.; Lochert, I. J.; Giribet, C. G.; Ruiz de Azna, M. C.; Contreras, R. H. J. Chem. Soc., Faraday Trans 1995, 91 , 4031. ' Shustov, G. V.; Denisenko, S. N.; Chervin, I. 1.; Asfandiarov, N. L.; Kostyanovsky, R G. Tetrahedron 1985, 41, 5719. ' De Meijere, A.; Schallner, O.; Weitemeyer, C.; Spielmann, W. Chem. Ber. 1979, 112, 908. ‘ Della, E. W.; Cotsaris, E.; Hine, P. T.; Pigou, P. E. Aust. J. Chem. 1981, 34, 913. 173 “ Axenrod, T.; Liang, B.; Bashir-Hasheuri, A.; Dave, P. R; Reddy, D. S. Magn. Reson. Chem. 1991, 29, 88. ' Eaton, P. E.; 01', Y. S.; Branca, S. J. J. Am. Chem. Soc. 1981, 103, 2134. ' Schneider, H J.; Heiske, D.; Hoppen, W.; Thomas, F. Tetrahedron 1977, 33, I769. " Maruyama, K.; Muraoka, M.; Naruta, Y. J. Org. Chem. 1981, 46, 983. 3' Kovaéek, D.’, Maksié, Z. 8.; Elbel, S.; Kudnig, J. J. Mol. Struct. 1994, 304, 247. z This work. "' McMurry, J. E.; Lectka, T.; Hodge, C. N. J. Am. Chem. Soc. 1989, 111, 8867. 5.3) the reduced slope and positive intercept of equation 10 are strikingly different. This correlation, based on 7 data points,22 appears to be the exception rather than the rule, since the methods of calculation used are no different than those utilized in other studies (AMI optimized geometries; hybrids estimated from AMI wavefiinctions by the LMO procedure of Trindle and Sinanoglu)”. Thus, some of the correlations presented in Table 5.3 are based on too few compounds to be of general use. Furthermore, in light of Gil’s23 finding that residual delocalization makes excitation energy dependent on carbon coordination number, it is arguably inappropriate to directly include primary, secondary and tertiary C-H sites in the same correlation, which all previous studies have done. Instead, we have focused this initial effort on prediction of tertiary C-H coupling constants for the widest possible range of hydrocarbons. The highly strained small-ring compounds included in this series provide a supplementary test for the adequacy of methods of calculation used, and extend the established relationships to more general use. The basic MP type relationships are reexamined for the hydrocarbons listed in Table 5.2. The PM3 correlations established by least-squares analysis“ are presented in Table 5.4. In comparison with the original relationship of Miiller and Pritchard3 (l) we 174 Table 5.4 Semiempirical Relationships between Experimental One-Bond C-H Couplings and Hybridization, C-H Distance, C-H Bond Order, Natural Atomic Charges on Carbon and Hydrogen, or Intemuclear Angles, Established by Least-Squares Analysis for the PM3 Optimized Geometries of Hydrocarbons 1-39‘ Semiempirical relationships Eq. No. s.d.b 'J,,C_1H = 4.66 (% sc) (19) 8.3 ‘J,,C_,H = 5.60 (% sc) - 32.70 (20) 7.4 ‘J,,C_1H = 0.18 (% sc)2 - 6.594 (% so) + 170.98 (21) 6.5 'J,,C_,H = 0.65 (% sc)” + 29.14 (22) 7.0 ‘J,,C_1H = 15.5 (% sc) / (2 + 0.2%,, + 0.6 d§_,, ) - 20 (23)° 7.6 ., 'J,,C_,H = 5.99 (% sc) / (0.2 +P§_H) - 21.39 (24)“ 6.9 ‘J,,C_,H = -360.11qc + 120.27 (25) 14.4 ‘J,,C_,H = 869.73 qH + 75.63 (26) 8.3 ‘J,,C_1H = -2666.15 qH qC + 128. 56 (27) 10.5 ‘J,,C__1H = 3.24 (% sc) - 2.87 (% sc) qc + 187.01 q1., + 19.05 (28) 5.4 ‘J,,C_,H = 3.79 (% so) - 2239.27 qH qC + 137.17 qC — 83.80 qH + 26.51 (29) 4.8 ‘J,,C_11H = 2.64 93cm 157.20 (30) 7.8 ‘J,,C_1H = -195 0220+ 344.65 (31) 6.3 'J,,C_,H = 131.25 + 0.66 zneccc - 6.35 x 10'5 (2710CCC )2 (32) 6.3 'J,,C_,H = 893.01 qH + 74.27 (33)° 4.7 ‘J,,C_,H = -2284.54 alsH + 1694.31 (34)° 4.7 ' Correlations 19-32 include all 62 independent data points from Table 5.2. b One-bond C-H coupling constants and standard deviations (s.d.) are given in Hz. ° dc_H is C-H bond distance in A. Based on equation 14 (Table 5.3) and the reported linear dependence of San on den (ref. 44), the denominator in 175 found, as have others before us, that better concordance between experimental and calculated ‘J ”0,“ values is obtained when a constant term, usually negative, is added to (1), (see equations 19 and 20, Table 5.4). This constant term is generally considered to originate in the deficiencies of AEE approach and the assumption of Fermi contact term predominance.46 Maksic et al.19 suggest that the constant term results fi'om the ionic character of C-H bonds, a point examined (and discarded) by Mfiller and Pritchard‘ themselves. The plot of ‘J13C_1H vs. percent s character shows a slight curvature (Figure 5.1), and accordingly, the correlation is improved if a second order expression is considered (equation 21, Table 5.4). There is no justification for such an empirical fit, but with so many approximations already inherent in the method, the enhanced predictive power of a better fit, however nonphysical, is worthy of exploration. Since the semiempirical PM3 method may introduce errors, we have also used the ab initio HF/6- 31G“ model to see whether the agreement between ’J 13c and % Sc can be refined by a —‘H higher level calculation. No improvement was found in the correlation of 1J‘3C—‘H with percent 8 character determined from the HF/6-31G* wavefilnctions of 1-39 (equations 20 and 36, Figure 5.1), which suggests that the deviations fi'om linearity seen in subsequent correlations are not an artifact of the PM3 method.47 The difference between experimental and calculated 1J13c—‘11 is especially high when the carbon atom at the tertiary site is contained in at least two 3- or 4-membered rings. It is very probable that these deviations occur as a result of breakdown of the AEE approximation in strained rings. The relationship proposed by Gil23 was also investigated (equation 22, Table 5.4), 176 Table 5.4 Semiempirical Relationships between Experimental One-Bond C-H Couplings and Hybridization, C-H Distance, C-H Bond Order, Natural Atomic Charges on Carbon and Hydrogen, or Intemuclear Angles, Established by Least-Squares Analysis for the PM3 Optimized Geometries of Hydrocarbons 1-39" Semiempirical relationships Eq. No. s.d.b 'J,,C_1H = 4.66 (% sc) (19) 8.3 ‘J,,C_,H = 5.60 (% sc) - 32.70 (20) 7.4 'J,,C_,H = 0.18 (% sc)2 - 6.594 (% sc) + 170.98 (21) 6.5 ‘J,,C_,H = 0.65 (% sc)“2 + 29.14 (22) 7.0 ‘J,,C_,H = 15.5 (% sc) / (2 + 0.2dC_H + 0.6 d3“ ) - 20 (23)° 7.6 ‘J,,C_1H = 5.99 (% sc) / (0.2 +Pg_,,) - 21.39 (24)“ 6.9 ‘J,,C_,H = -360.11qc +120.27 (25) 14.4 ‘J,,C_,H = 869.73 qH + 75.63 (26) 8.3 ‘J,,C_1H = -2666.15 qH qC + 128. 56 (27) 10.5 ‘J,,C_11H = 3.24 (% sc) - 2.87 (% sc) qC + 187.01 qH + 19.05 (28) 5.4 'J,,C_,H = 3.79 (% so) — 2239.27 qH qC + 137.17 qc - 83.80 qH + 26.51 (29) 4.8 'J,,C_IH = 2.64 exec-157.20 (30) 7.8 ‘J,,C_1H = -1.95 0300+ 344.65 (31) 6.3 ‘J,,C_,H = 131.25 + 0.66 ZAOCCC - 6.35 x 10'5 (2210CCC )2 (32) 6.3 'J,,C_1H = 893.01 q... + 74.27 (33)° 4.7 ‘J,,C_1H = -2284.54 a15H + 1694.31 (34)° 4.7 ' Correlations 1932 include all 62 independent data points from Table 5.2. " One-bond C-H coupling constants and standard deviations (s.d.) are given in Hz. ° dc_fl is C-H bond distance in A. Based on equation 14 (Table 5.3) and the reported linear dependence of Scar on dc-H (ref. 44), the denominator in equation 23 was approximated as a second-order polynomial regression in dc.H. d P“, is Mulliken C-H bond order. ° Equations 33-34 use only 61 independent data points; 38a is excluded. 177 2,0 _ a) PM3 _ ‘JC.H = 5.60 (% sc) - 32.70 (20) .3 . 200 _ s.d. = 7.4 HZ : 180 -4 lJC-H d (H2) 160 _ 140 ~— 120 — .1 100 I . T 20 ,2, _ b) HF/6-31G* - ‘JC.H = 6.12 (% so) - 14.47 (36) E 200 ._ s.d. ‘2 7.8 H2 0 l JC-H (HZ) 160- 20 25 30 35 40 % SC Figure 5.1 Experimental one-bond C-H spin-spin coupling constants vs. percent 8 character of the C hybrid in the C-H bonding orbital obtained fi'om NBO analysis of: a) PM3, and b) HF/6-31G“ wavefunctions for optimized geometries of 1-39. 178 and even though it gives a smaller s.d. (7.0 Hz) than the linear correlation of lJ l3c-‘n with % Sc (s.d. = 7.4 Hz), the result in the present case can not be explained on the basis of the variation of AEE with carbon coordination number, which is constant in 1-39. The Spartan sofiware does not explicitly report overlap integrals, so we examined the basic relationship of Maksic et al.19 (equation 14, Table 5.3) for 1-39 by replacing bond overlap with either C-H distance (overlap is a nearly linear firnction of distance in the range of interest“) or C-H bond order (NBO-derived), as proposed by Zhan and Hum (equation 15, Table 5.3). The relations obtained (equations 23 and 24, Table 5.4) do not show significant improvement over the simple linear dependence of 'J with % Sc. ‘30-‘H The best correlations are obtained by including the atomic charges qc and qH calculated by natural population analysis for carbon and hydrogen atoms (equations 28-29, Table 5.4). The calculated charges agree well with Bent’s rule,49 which states that atomic 8 character concentrates in orbitals directed toward electropositive substituents. Thus, small-ling compounds, where the distorted geometries cause the ring C atoms to rehybridize in such a manner as to augment the 8 character in the C-H bond, show increased C-H bond ionicity. Previously, Guillen and Gasteiger50 used the iterative partial equalization of orbital electronegativity method (PEOE) for calculating atomic partial charges in hydrocarbons with 3- and 4-membered rings and established a linear correlation between 11130—111 and the product of carbon and hydrogen charges (equation 17, Table 5.3). The PEOB procedure reproduces surprisingly well small trends in the coupling constants, even though hybridization states, calculated fiom substitution patterns, are taken to be artificially equal 179 for distinct compounds, as for example 4, 5a and 11. Zhan and Hu25 introduced a generalized relationship suitable for hydrocarbons and molecules with -I" and -1‘ substituents, in which the 5 character of the hybrids and the net atomic charges on C and H are involved for calculation of lJ ,,C_,H (equation 18, Table 5.3). Such a correlation applied to compounds 1-39, gives a much lower s.d. as compared to equations 19-24, which indicates that while hybridization is important in the study of one-bond C-H spin- spin coupling constants, the ionic contribution to bridgehead C-H bonds can not be neglected. Various forms of a possible semiempirical relationship of lJ ‘30-‘H vs. hybridization, qn, and qc, have been tested, among which equation 29 gave the lowest standard deviation. In their treatment of the F erlni contact contribution to spin coupling between directly bonded atoms, using electron pair theory, Karpluss1 and Grant and Litchman16 showed that besides hybridization, 'J 1 values depend also on the efi‘ective 3c-‘H nuclear charge, which is a filnction of the C-H bond covalency. Our results Show that ionic contributions to bridgehead C-H bonds significantly refine the classical MP relationships between 1I ‘30—‘H and % Sc, in which case bond ionicity can not be ignored. Interestingly, the best single-parameter correlations are the PM3 qr; or the PM3 atomic orbital coeflicient on H, 31s,, (or a, see equation 3), and experimental 1J 1,0,“ (equations 33-34, Table 5.4).52 If38a (the in-C-H bond of bicyclo[4.4.4]tetradecane) is excluded fi'om the correlations,53 linear relationships are obtained via least-squares analysis with standard deviations of only 4.7 Hz (Figure 5.3). The polarization of C-H bonds was also considered, as the atomic orbital coefficients for C and H (a, b and c) are given by NBO analysis. The correlation of 180 1le0IH with % Sc and 31er as given in equation (5), did not bring any improvement over previously discussed relationships. A possible semiempirical relationship of 1J ‘30—‘11 with CH bond order was also explored, but no improvement over those involving only hybridization and atomic charges was obtained. The relationship between one-bond carbon-hydrogen spin—spin coupling constants and calculated bond angles has been investigated for the compounds under study, too (equations 30-32, Table 5.4). Average CCC and HCC angles, 0:20 and 030‘: (03%: (ZAHCC°)/3 ), were considered for the general case of three substituents attached to a methine carbon; again, conformational averaging was included where necessary. The PM3 empirical relationships established via least-squares analysis are recorded in Table 5.4, and show similar standard deviations for plots of 'J,,C_1H vs. the average CCC angles, 02:20 , or the sum of internuclear angle distortions, ZAGCCC . It is recognized that bent bonds30 are frequently found in organic compounds and internuclear bonds do not always correspond to bond paths,54 defined as the path of maximum charge density between the bonded atoms. Hybridization is more closely related to interorbital rather than internuclear angles. A simple analysis of the correlation of lJ 1,0,“ with bond path angles vs. internuclear angles in methine systems with C3. symmetry, supports this idea and allows for a qualitative estimate of the amount of bond bending. Thus, we converted the corresponding hybridization, Sp“, at carbons with local C3” symmetry into interorbital angles, 93cc 7 11$ng Coulson’s relation:55 0 °CCC 2 arccos (- l) (9) n The results show improved correlation of 1J 130, H with 900cc (s.d. is 4.4 Hz for PM3 181 220 m o u .. . 200 — o _ . . o 180 - O o. O l d EXP. Jon '0 160 ‘ 0 .0 :0 (HZ) ° 00 o ' ~ 0 140 — . o . o 120 "1 .. o 100 l I I l r I I l 1 I’ 100 150 140 160 180 200 220 1 C310 JQH (Hz) Figure 5.2 Plot of experimental vs. calculated (with semiempirical relationship 29, Table 5.4) one-bond C-H spin-spin coupling constants in 1-39. 182 Agata—oboe 05 80b cows—88 mm can .353 8% Ev commie»: no 826808 33.5 2an.» gm 3 EB Acowflotg 2: 8on 32:88 mm «an .358 8% Ev xv dump—PE co amaze 0383 358: gm Q panama an; E 3:383 manage 592-5% 3-0 989-28 Ecogonxm $9 am 5. u .3 2.32 + as $.82- u no: 3 oo— o2 03 oo— ow— com cum ova QB 3-0—.2 a: S. n .3 r GO 2.: + as 8.93 n so: ? com on— ov— cm: 9: com can ova «rm 2:5 QB v2.0: 183 geometries and 4.2 Hz for HF/6-3 lG* geometries) compared with GCCC (s.d. is 5.1 Hz for PM3 geometries and 4.8 Hz for HF/6-31G* geometries)?4 A similar analysis was performed for the HF/6-31G* optimized geometries of l- 39. The relationships obtained are presented in Table 5.5 and, analogously with the PM3 results, show that inclusion of C and H atomic charges improve considerably the simple correlation of 1‘7”an with hybridization. Nevertheless, the 6-31G* results are less correlated with experiment than those fiom the PM3 method, in accord with the conclusion of Edison et al.6‘, that better agreement with experimental values is obtained for calculated nuclear spin-spin coupling constants when using modest levels of MO theory. More disturbing are the HF/6-31G* natural atomic charges on hydrogen and carbon in 1- 39, whose oscillating behavior and poor correlation with PM3 charges is surprising. The discrepancy of the H and C atomic charges in 38a vs. other bridgehead sites with similar hybridization at carbon, however, is reduced at the HF/6-31G* level of calculation. 5.4 Summary (1) The experimental values of 13 C NMR chemical shifts and one-bond carbon-hydrogen coupling constants in bicyclo[3.3. 1]nonane 36, bicyclo[3.3.2]decane 37, and bicyclo[3.3.3]undecane 39 are reported. (2) Semiempirical relationships of experimental l - 0 av av - J 13c-1 With A: sc, qu and qc, alsH, BCCC , OHCC, and ZAGCCC , are examined for H compounds 1-39, and show reasonable agreement of calculated vs. experimental lJ l3C-lH values (Tables 5.4 and 5.5). The PM3 model shows real promise; the computations 184 Table 5.5 Semiempirical Relationships between Experimental One-Bond C-H Coupling Constants and Hybridization, Natural Atomic Charges on Carbon and Hydrogen, or Intemuclear Angles, Established by Least-Squares Analysis for the HF/6-31G“ Optimized Geometries of Hydrocarbons 1-39‘l Semiempirical relationships Eq. No. s.d.” ‘J,,C_1H = 5.62 (% sc) (35) 8.0 ‘J,,C_1H = 6.12 (% sc) — 14.47 (36) 7.8 'J,,C_,H = 0.17 (% sc)2 - 3.58 (% 5C) + 120.12 (37) 7.3 'J,.,C_,H = 0.77 (% so)” + 41.73 (38) 7.4 ‘J,,C_,H = -2403 (% sc) qC - 4.28 (39) 6.0 'J,,C_IH = 1.47 (% sc) - 19.35 (% sc) qC - 144.70 qH + 19.29 (40) 5.5 ‘J,,C_1H = 6.19 (% sc) - 1296.56 qn qc - 287.87 qC - 480.79 qH - 46.16 (41) 5.6 'J,,C_,H = -192 0300+ 341.88 (42) 6.4 'J,,C_,H = 131.98 + 0.65 ZABCCC- 1.59 x 10‘4 (meme)2 (43) 6.4 ' Generations 3543 include all 62 independent data points from Table 5.2. b One-bond C-H coupling constants and standard deviations (s.d.) are given in Hz. 185 required for PM3 geometry optimization of 1-39 and NBC analysis are modest and can be carried out with readily available electronic structure packages. Correlation of experimental 11130-111 with PM3 hybridization is considerably improved by inclusion of natural atomic charges on C and H (equations 28-29) to give best fits of experimental vs. calculated 'J,3C_1H coupling constants (s.d. = 4.8 Hz for equation 29; 62 data points). Such an empirical relation is useful for predicting 1J H for hypothetical compounds for l3c_l comparison to experiment, but offers little physical insight into the coupling mechanisms. However, surprisingly good single-parameter linear correlations of ‘J with PM3 qH '30-‘H (equation 33; 61 data points, s.d. = 4.7 Hz), or alsH (equation 34; 61 data points, s.d. = 4.7 Hz), are found for 1-39, when the distant outliner 38a is removed. (3) That '1 ‘30—‘11 depends on carbon orbital hybridization is part of the canon of organic chemistry. Numerous equations have been previously proposed based on modest data sets and various measures of hybridization. However, in most cases the choice of compounds was arbitrary and their geometries inconsistent, while the correlations established gave less satisfactory results for strained polycyclics. On the basis of the comparison between various MP type relationships and the critical evaluation of their performance for our wide range of compounds, we conclude that ionic contributions to C-H bonds are important, at least in bridgehead C-H sites, for a suitable correlation of experimental C-H couplings with carbon orbital hybridization. The relationships obtained, particularly equation 29, which includes natural atomic charges along with hybridization at carbon, can be used to easily predict one-bond C-H coupling constants at tertiary sites in polycyclic saturated hydrocarbons with experimentally USCfiJi accuracy. Equations 33-34 ofl‘er simplified, more 186 physically understandable alternatives for predictions of 11130—111 values from modest computational data; however, their use is limited by the poor performance of the PM3 model in situations like 3811 and similar cases should be treated with caution. (4) The overall agreement of calculated with experimental data confirms that the Fermi contact interaction, as modulated by hybridization, is the dominant factor in determining the magnitude of the coupling between directly bonded carbon and hydrogen atoms. The polarity of C-H bonds, however, can not be ignored even in hydrocarbons. 5.5 Experimental Methods Bicyclo[3.3.1]nonane 36 was synthesized from bicyclo[3.3.1]nonan-9-one57 by Clemmensen reduction with amalgamated zinc and hydrochloric acid.58 Ring expansion of bicyclo[3.3. 1]nonan-9-one with methanolic diazomethane gave bicyclo[3.3.2]decan-9- one”, which was reduced under Wolff-Kishner conditions to afl‘ord bicyclo[3.3.2]decane 376°. Bicyclo[3.3.3]undecane 39 was prepared from bicyclo[3.3.1]nona-9-one by a modified synthesis following Leonard et a1.61 Physical and spectroscopic data of 36, 37 and 39 were in agreement with those reported in the literature. Bicyclo[3.3.l]nonane (36). Zinc metal (7.5 g 20—30 mesh) was added to mercuric chloride (15 ml Hng 10%) and the resulting suspension was stirred for an hour, decanted and washed with water. Bicyclononanone (500 mg; 3 .6 mmol) and concentrated hydrochloric acid (10 ml) were added to the fieshly prepared amalgamated zinc, and the reaction mixture was refluxed for half an hour, cooled, extracted with pentane and dried 187 over Na2SO4. The pentane was removed on a rotary evaporator, and the solid residue was sublimed to afford pure bicyclo[3.3.1]nonane 36 (17 8 mg; 1.44 mmol; 40 °/o yield); mp 144-146 °C (lit.58 145-146 °C); 1H NMR (300 MHz, CDC13) 6 1.78-1.94 (m, 4H), 1.6- 1.68 (m, 8H), 1.45-1.55 (m, 4H), in accord with previous reports“; 13C NMR (300 MHz, CDC13) 5 35.01, 31.59, 27.89, 22.52, in accord with previous reports”; MS(EI) m/z CeHla 124 (M), 109,96, 81 (base), 67, 55,41. Bicyclo[3.3.2]decane (37). Bicyclononane (200 mg; 1.3 mmol) was added to a solution obtained from sodium (70 mg), diethylene glycol (3.2 ml) and hydrazine hydrate (84 mg 100% NHZNH; x H20), and the reaction mixture was refluxed for one hour, cooled, diluted with water (5 ml), and extracted with pentane. The pentane extracts were dried over Na2S04, filtered, the solvent was vacuum distilled and the solid residue was purified by sublimation to give bicyclo[3.3.2]decane 37 (90 mg; 0.65 mmol; yield 50%); mp 177-179 °C (lit.59 177-178 °C); 1H NMR (300 MHz, CDC13)6 2.25-2.35 (1n, 2H), 1.41-1.75 (m, 16H), in agreement with refs. 59-60; 13C NMR (300 MHz, CDCl3) 5 33.67, 32.87, 30.36, 22.78; MS(EI) m/z Clon 138 (1W), 123, 110, 95, 81, 67 (base), 55, 41, 39. Bicyclo[3.3.3]undecane (39). See chapter 2 for experimental details: mp 191 °C (lit.61 192 °C); 1H NMR (300 MHz, CDCl3) 8 2.38 (m, 2H), 1.41-1.55 (m, 18H), in agreement with ref. 61; 13c NMR (300 MHz, CDCl3) 5 30.74, 28.96, 20.1; MS(EI) m/z C11H20152 (NF), 124, 109, 96, 81, 67, 55. Melting points were measured with a Thomas Hoover capillary melting point apparatus and were uncorrected. 1H and 13C NMR spectra were recorded on a Varian F T- 188 NMR 300 MHz at ambient temperature, and were referenced to solvent signals. Mass spectra were obtained using a VG Trio-1 GC-MS spectrometer. The 13 C NMR spectrum of bicyclo[3.3.2]decane 37 is reported here for the first time, and assignments to the corresponding carbons are made based on the HMQC (lH-detected heteronuclear Multiple Quantum Coherence)64 spectrum of 37 , which reveals all the crosspeaks from the secondary and tertiary carbons to the respective protons. Overlap of signals in the off- resonance decoupled spectra of 36 and 37 did not allow accurate measurement of the C-H direct couplings and thus, they were obtained fiom the corresponding 2D Heteronuclear J- Resolved spectra“, which showed contour peaks at each carbon in accordance with the number of protons directly connected. The l3C-IH spin-spin coupling constants in bicyclo[3.3.3]undecane 39 were determined from the ofilresonance proton decoupled spectrum of 39. All 2D NMR spectra were recorded on a Varian VXR 500 MHz spectrometer at 25 °C. 5 .6 References ' Craciun, L.; Jackson, J. E. J. Am. Chem. Soc. 1996, 118, 12232. 2 Ramsey, N. F. Phys. Rev. 1953, 91, 303. 3 Fermi, E. z Phys. 1930, 60, 320. 4 Ml’iller, N.; Pritchard, D. E. J. Chem. Phys. 1959, 31, 768. 5 McConnell, H. M. J. Chem. Phys. 1956, 24,460. 6 (a) Gil, V. M. S. J. Magn. Reson. 1973, 11, 268. (b) Edison, A. S.; Markley, J. L.; 189 Weinhold, F. J. Phys. Chem. 1993, 97, 11657. 7 Ando, 1.; Webb, G. A. Iheory of NMR Parameters; Academic Press: London, 1983; p 88. 8 Pople, J. A; Sentry, D. P. M01. Phys. 1963, 8, 1. 9 Pople, J. A.; McIver, J. w; Ostlund, N. 8. J. Chem. Phys. 1968, 49,2960. ‘° Blizzard, A C.; Santry, D. P. J Chem. Phys. 1971,55, 950. " Kowaleski, J .; Laaksonen, A Theoretical Models of Chemical Bonding; Springer- Verlag: Berlin, 1991; p 487. ‘2 Mt'iller, N.; Pritchard, D. E. J. Chem. Phys. 1959, 31, 1471. ‘3 In fact, rigorous sp3 hybridization at carbon is not required even for methane, as the C 28 and 2p orbitals differ in their energies and may participate to difierent extents in the bonding orbital combinations with symmetry adapted H 1s orbitals. ‘4 Karabatsos, G. J.; Orzech, c. E. J. Am. Chem. Soc. 1964,86, 3574. 1’ Considine, w. J. J. Chem. Phys. 1965, 42, 1130. ‘6 Grant, D. M.; Litchman, w. M. J. Am. Chem. Soc. 1965,87, 3994. ‘7 Brown, T. L.; Puckett, J. C. J Chem. Phys. 1966, 44, 2238. ‘8 (a) Saika, A. J. Chem. Phys. 1966, 45, 2715. (b) Cyr, N.; Cyr, T. J. R. J. Chem. Phys. 1967, 47, 3082. ‘9 Maksic, Z. B.; Eckert-Maksic', M.; Randic', M. Theoret. Chim. Acta 1971, 22, 70. 2° Newton, M. D.; Schulman, J. M.; Manus, M. M. J. Am. Chem. Soc. 1974, 96, 17. 21 (a) Figeys, H. P.; Geerlings, P.; Raeymaekers, P.; Van Lommen, G.; Defay, N. Tetrahedron 1975, 31, 1731. (b) Van Alsenoy, C.; Figeys, H. P.; Geerlings, P. Yheoret. Chim. Acta 1980, 55, 87. 22 Kovaéek, D.; Maksic, z. B.; Elbel, s; Kudnig, J. J. Molec. Struct. 1994, 304, 247. 23 Gil, V. M. s. Theor. Chim. Acta 1989, 76, 291. 2‘ Hu, Z.-M.; Zhan, C.-G 777e0r. Chim. Acta 1993, 84, 521. 190 25 Zhan, C.-G.; Hu, Z.-M Magn. Reson. Chem. 1994, 32,465. 2‘ (a) Foster, J. P; Weinhold, F. J. Am. Chem. Soc. 1980, 102, 7211. (b) Zhan, C.-G. and Hu, Z.-M. Theor. Chim. Acta 1993, 84, 511. 2" Randic', M.; Maksic, z. B. Chem. Rev. 1972, 72,43. 28 (a) Foote, S. C. Tetrahedron Lett. 1963, 9, 579. (b) Laszlo, P.; Schleyer, P. v. R J. Am. Chem. Soc. 1964, 86, 1171. (c) Baum, M. W.; Guenzi, A.; Johnson, C. A.; Mislow, K. Tetrahedron Lett. 1982, 23, 31. 29 (a) Szalontai, G. Tetrahedron 1983, 29, 1783. (b) Zhao, C. Y.; Xiao, T.; Xu, X. F.; You, X. Z. Chem. J. Chin. Univ. 1996, 17, 769. 3° Wiberg, K. B. Acc. Chem. Res. 1996, 29, 229. 3‘ Mislow, K. Tetrahedron Lett. 1964, 22, 1415. 32 Tokita, K.; Kondo, S.; Takeda, M. J. Chem. Res. (S) 1980, 63. 33 Werstiuk, K. B.; Taillefer, R.; Bell, R. A.; Sayer, B. Can. J. Chem. 1973,51, 3011. 3“ Stewart, J. J. P. J. Comput. Chem. 1989, 10, 209. 3’ Hariharan, P. C.; Pople, J. A. Chem. Phys. Lett. 1972, 66,217. 3‘ Hehre, w. J.; Huang, w. W.; Burke, L. D.; Shusterman, A. J. A SPARTAN Tutorial, version 4.0, Wavefirnction Inc.: Irvine, CA, 1995. 37 Stewart, J. J. P. J. Comput. Chem. 1989, 10, 221. 3* Dewar, M. J. S; Thiel, w. J. Am. Chem. Soc. 1977, 99,4899. 39 Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P. J. Am. Chem. Soc. 1985, 107, 9. 4° Franc], M. M.; Pietro, W. J .; Hehre, W. J .; Binkley, J. S.; Gordon, M. S.; DeFrees, D. J.; Pople, J. A J. Chem. Phys. 1982, 77, 3654. 4‘ Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. Rev. 1988,88, 899. ‘2 Reed, A E.; Weinstock, R. B; Weinhold, F. J. Chem. Phys. 1985, 83, 735. ‘3 Hehre, W. J. Practical Strategies for Electronic Structure Calculations; Wavefilnction 191 3‘ Dewar, M. J. s; Thiel, w. J. Am. Chem. Soc. 1977, 99,4899. 3‘9 Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P. J. Am. Chem. Soc. 1985, 107, 9. 4° Francl, M. M.; Pietro, w. J.; Hehre, w. J.; Binkley, J. s; Gordon, M. S; DeFrees, D. J.; Pople, J. A. J. Chem. Phys. 1982, 77, 3654. 4‘ Reed, A. E.; Curtiss, L. A; Weinhold, F. Chem. Rev. 1988,88, 899. ‘2 Reed, A E.; Weinstock, R. B; Weinhold, F. J. Chem. Phys. 1985,83, 735. ‘3 Hehre, W. J. Practical Strategies for Electronic Structure Calculations; Wavefirnction Inc.: Irvine, CA, 1995; p 23. ‘4 Engler, E. M.; Andose, J. D; Schleyer, P. v. R. J. Am. Chem. Soc. 1973,95, 8005. ‘5 Program used for least-squares analysis: Curve Fitter, version 4.0, written at MSU by David C. Young for use with MS-DOS computers. ‘6 Duijneveldt, F. B. v.; Gil, V. M. S.; Murrell, J. N. Yheoret. Chim. Acta 1966, 4, 85. ‘7 Edison et al.“b also found better performance for a less complete basis set (3-216) than for 6-31G* in their more rigorous analysis. 4‘ Maksic, z. B; Randié, M. J. Am. Chem. Soc. 1970, 92, 424. ‘9 Bent, H. A Chem. Rev. 1961, 61,275. 5° Guillen, M. A.; Gasteiger, J. Tetrahedron 1983, 39, 1331. 51Karplus,M.; Grant, D. M. Proc. Natl. Acad USA 1959, 45, 1269. 52 The linear correlation of C-H couplings (61 data points; 38a is excluded) with the PM3 atomic orbital coefficient on carbon, bZSC (or b, see equation 3) gives a Standard deviation of 4.9 Hz. ’3 The PM3 qc and qg values for the in-C-H bond in bicyclo[4.4.4]tetradecane 38a are calculated much higher than for other bridgehead sites with similar hybridization at carbon, while at the HF/6-31G* level of calculation this discrepancy is considerably reduced. Similarly, in-[34’1°][7]metacyc10phane (see Pascal, R. A. Jr.; Grossman, R. B.; Van Engen, D. J. Am. Chem. Soc. 1987, 109, 6878), a compound not formally included in our series 192 average CCC bond path angle. For 1-39 the second-order correlation of 1J 13c-‘1r with average CCC interorbital angles derived from PM3 hybridizations by using Coulson’s relation, gives a s.d. of 5.9 Hz. 5’ Foote, C. s; Woodward, R. B. Tetrahedron 1964,20, 687. ’8 Cope, A. C; Synerholm, M. E. J. Am. Chem. Soc. 1950, 72, 5228. 59 Bingham, A. C; Schleyer, P. v. R. J. Org. Chem. 1971, 36, 1198. 6° Doyle, M.; Hafier, R.; Parker, W. J. Chem. Soc., Perkin Trans 1 1977, 364. 61 Leonard, N. J.; Coll, J. C. J. Am. Chem. Soc. 1970, 92, 6685. ‘2 Schleyer, P. v. R; Isele, P. R.; Bingham, R. C. J. Org. Chem. 1968, 33, 1239. ‘3 Heumann, A; Kolshorn, H. Tetrahedron 1975, 31, 1571. 6‘ Bax, A; Grifl‘ey, R. H.; Hawkins, B. L. J. Magn. Reson. 1983, 55, 301. 6’ Bax, A T wo-Dimensional Nuclear Magnetic Resonance in Liquids; Delft University Press: Delfi, Holland, 1982; p 99. CHAPTER 6 HEATS OF FORMATION OF MEDIUM-RING STRAINED CYCLO- AND POLYCYCLOALKANES: COMPARISON OF AB INITIO GROUP EQUIVALENT SCHEMES WITH THE PM3 AND MMX METHODS Abstract: Optimized structures and energies were calculated for 57 small- and medium- ring strained polycyclic aliphatic hydrocarbons using ab initio HF/3-21G and HF/6-31G* as well as PM3 (semiempirical) and MMX (force field) methods. Best fit CH2, CH and C group increments relating ab initio energies to heats of formation were derived. The ab initio increments deviate little from those previously reported by \Vlberg and by Ibrahim and Schleyer, yielding the expected conclusion that the intrinsically isodesmic group increment approach extends efiiciently to medium-ling strained systems. For the present data set, the standard deviation between experimental and calculated heats of formation is 1.8 kcal/mol, and the correlation coefficient is 0.9994 for the RI-IF/6-31G*//RHF/6-31G* calculation. Less successfill results are obtained from the HF/3-21G, PM3 and MMX data. As expected, systems with fused small rings are especially problematic for the latter methods. 193 194 In the course of the study of hybridization and 13OH NMR coupling constants described in chapter 5,1 we recently obtained RHF/6-31G*//RHF/6-31G* wavefunctions and energies for a large number of small- and medium-ring strained polycyclic hydrocarbons. Roughly half of this number have had experimental heats of formation reported. It was of interest to examine the performance of the Wiberg2 and Ibrahim/Schleyer3 (IS) hydrocarbon group increments in calculating heats of formation from ab initio energies for these compounds, as most previous work has focused on unstrained or small-ring systems. This paper provides such an analysis for S7 hydrocarbons, of which several were beyond the range of practical computational tools when the above papers appeared. The new best fit for the CH2, CH and C fiagments are essentially unchanged from those previously reported, yielding the expected conclusion that the intrinsically isodesmic group increment approach extends effectively to medium- ring strained systems. 6.1 Results and Discussion The heat of formation of a compound is a useful characteristic, traditionally determined fiom combustion measurements. However, the accumulation of computational data at a consistent level for a wide variety of molecules and their correlation with experimental results allow an evaluation of their heats of formation from ab initio energies, as well. Molecular mechanics or semiempirical methods are not as generally usefill, since the former method needs good experimental data, not always available, for parametrization,4 195 while the latter approximates minimal basis set calculations which frequently handle strained small-ring compounds unevenlys. Conversion of ab initio calculated energies to heats of formation is commonly done by the use of isodesmic comparisons with closely related compounds of known thermochemistry, such that errors due to inadequacies of basis set or electron correlation treatment largely cancel out.6 As was pointed out by Wiberg’, group equivalent schemes can be viewed as a subset of isodesmic reactions in which the substitution levels of all C sites are maintained constant. Thus, Wrberg2 and subsequently, Ibrahim and Schleyer3, empirically determined sets of group and atom equivalents, which, when subtracted from a compound’s ab initio energy, yield its heat of formation, AHf(calcd). Accordingly, AHr(calcd) (in kcal/mol) is expressed as the difi‘erence between the molecule’s total energy and the summed increments of the component groups, as shown by the following relation: AH.(ca1cd) = 627.5 (ET - Z n. Er) where ET is the ab initio total energy, n represents the number of atoms or groups of each sort, and E is the corresponding atom or group equivalent. Following these reports, simplified schemes with reduced number of parameters have been proposed] and individualized atom or group parameters were developed for particular classes of compounds". Bond/group equivalents have also been derived for alkanes fi'om density functional calculations.9 In a series of recent articles, Allinger et al.10 outlined an alternative method which combines bond energy with group increments, while it includes terms to explicitly account for statistical mechanical effects of populating a molecule’s 196 higher energy conformations and low-lying vibrational states, as well as its translational and rotational motions. The present work confirms that group equivalent-based heats of formation can be calculated with an accuracy close to that from experiments (see Table 6.1). In addition, strain energies have been determined for all compounds recorded in Table 6.1.11 The wide variety of small- and medium-ring strained hydrocarbons provide a stringent test of the method. Experimental heats of formation, AHf(exp), and calculated RI-IF/3-ZIG12 and RI-lF/6-31G"‘l3 total energies for the compounds considered in this study are listed in Table 6.1. The calculated values refer only to the lowest energy conformation, although in several cases the compounds exist as a Boltzmann distribution of different conformational isomers with somewhat different energies. A least squares fit14 of experimental vs. calculated heats of formation with the increments for CH2, CH and C groups as adjustable parameters yielded AHr(ca1cd) values at the 3-21G and 63 1G* basis-set levels as listed in Table 1, along with the group increments in Table 6.2. For a range of AH,(exp) fi'om -60 to +150 kcal/mol, the standard deviation of experimental vs. calculated heats of formation is 1.8 kcanol for the RHF/6-31G*//RHF/6-31G* calculation (Table 6.2; Figure 6.1). The thermochemical measurements recently reported for a variety of Spirocyclopropanated cYCIOpropane and cyclobutane derivatives15 allowed to establish unambiguously an equivalent for the quaternary carbon atom, which was not available from the work of Ibl‘ahim and Schleyer’, while Wiberg’s value for this parameter is based only on neopentane and spiropentane. 197 . . . 8.5 Gas m 8.. ”so: 0%. 4 NM- $82 .wmm- 3.8 «.2. cadence- EB 888296 . . . 35 AZ: m on Emma mm. 0 3.. 8324.98- gm. 8:. stoned- 3v 888265 . . . :4: 8.3 5 am- 888 as. a S- $2863- 3m- mam- 3.18.8- 3V oesoooio . . . 9% $8 a am- 888 E- 0 mm- 5828. 8.0m- SN- 33mm- 5v ofiaaoeoefi . . Gd Ame _ mm- 888 as- e. _ m- 885mm- 0. _ m- new- Samoa- 3% sagaeeoefi . . . 2.8 c .8 c om- 822 5. w om- 5332- 8.8- 2:- SEE. 3v 4285838 . . . 688 688 m m 580 or. a m 288.3- mm- 08 Same ES 4233265 . . amu 3.5 m .2 838 S _- SN coma}: 2. n2 v.2 33.2 38 49888268 69 46.8 Aéessamv .Goaoemn 38am as... Godsend Seem any @6852 Geneva? .aaevamq 836.com .. been? .933: @385 95 x22 enough—om mo 38m 8332.80 use BacoEtoaxm a.» «Ban. 198 NS m.0m w.0m 0.: Ohm fimv 9mm mdm- mow. moofioodmmt mem©0Nmmr mv0m00Nmmr mhmuo0dmmr 8030.87 03.20.07 $2562- 8300on 0533.091 mm: on wém v.34 “new n.0v Nah m .0vt Ndm- 0mmvohgmmr 23450. _ mm- M0500. fl mm- mNZHCAmN- N©w0mwNOT ovSmwNOT $88062- 2.3 $69G- 3530.me mm #2 oAN mom wsm fimv «.00 ”no- “..er 21.9 cm 8.3 men 2.3 0% 8.ch Sm Sea mam A38 93. 33 was. 3.3 2w. 8.5 an... 3.3 Sumo $3 eating A33 3.38.2 6.43 T14: :83 snout? $.ch Noumea 8.ch 348.: :18 e333- :2: steamer <-> XE) X RE 0 8v oeexoaozdoecmm A48 rogaoaomdoeczm EB 85885388 A88 308020on 0.8 socscofio. _ $2965 88 vocficoaeam A48 recess. _ . 22665 33 8888263 QB 238n§2o>0 199 W03 2:. NA :0 w.0~ 0.0m 0N0 0N0 000m0m.00m- 00“ ~00.NON- m000m0NON- 0mv000.00m- E 0000.00N- v0m0m000m- mmmw _ 0.00m- 0N0 m00.00m.. 0mg 0.07 N0 0.00 0.3 v.0v m.m0 #0: 00300.03. 0mm00m.00m- wvaNm .00N- 00003000... 330000? 0000— 02000. 00002 .00N- m~0w00.w0m- 06: 0.2.. NA 0.00 0.0m N; 000 0.00_ 5.20 0.03 0.”: «.2- 3.3 S 2.80 0.00 3.30 mi 8.3 v.3 A020 v60 $.30 0.02 3%: E0: A30 .0.0HV.NT 0.30 0.30.0 0.80 50.0...umN0 $.90 38 A25 xv.0.l..0.mm €80 flow $.30 gag” 09 $5330 060 3550085 35300:: .Ndo—gomm 000 05302— -Hoa .Eofixomméu 3.00 8800: A00. H .3000me 060 05.000tto Z 50 053253. 0 i -ofiSthéng 000 033005330 30 OquOEHn.mO.v.NO.O. m .VH— 5308500 200 5.3. 5.2- ”0. Va- 04% odm hooNSA—m- 0530; g m- 0038.2? 3850.2? Symmmwdom- bmommwdom- 93.. 0.2- fin- 9v m.mm 9mm Gmwwmdooc... gvgmdow- mmmvmmmem- SZdeOM- _wmmwo.w0m- 338.8? 0.;. wa ”m7 o.m- Nov adv A3: 3a. 8.5 5.2- 8.5 2- 3.3 :V- $.83 vac A33 9% A2 a $3.8- 8 8 GE 33%? $5 3.3.0- 9.an A 333. 2.5V .8“va DP A23 .33.? _.||_ EB 238° -Hoadsofiméa 16v 888 -Sanoismég QB 2880 -Saioéamés QB 2880 -8. _ 32035-8 28 088°: .2 .381me 38 ofisofiadsofimfi 201 war.- 9m- ode meow mdm mém- $52;de- Nh 900.3 m- mommmodm m- 538.2%- camp-0.2%- mwgomdvm- 5532.“;- 06m- 5m Wm o.wo_ ”Mme as» mam.- 3.55.me- oomm: .me- gov: .me- 0088.me- 3800.me- mummamdvm- mfiommmdom- 0.0m- 9m. m6- mam Kama v65 wHN- 9m: 3m- 35 3. $3 2.- 3%: N62 3.2: 0.2: 8.2: #3 8+: Zn- 3.8 mdfiodm- 3N3 mdfiod- 8.3 N.0Hm.m- 2.x: .3.-awe 3.3: «.0302 at: Nfihm 3.3 .332- .0 A XX 95362.23 33 28:0: -3mioiommé3 QB 28:0: -8. fl .aoigmég GOV 05:0: -.oggoigmas 38 28:0: -8.~.o.m.ode_95 Q8 28:8 -:._.~.o.o.ueam§ QB 0593083.“ 38 ..ofisofiadgofim 202 9mm.- 5mm- mgm- 5N7 0.2. min.- :38.me- cowoswwm- mmmboowwm- v3 Moog-mm- moon: .Omm- 3216mm- 90v- odm- mwm- de mdm- 9mm.- Emmawwwm- wmnmhwdmm- mowmuwwwm- fimvmmwhwm- mommfimdg- 383de- QVm- owm- _.m~- flow- wdm- ohm.- 3.3 Em- Ana: EN- 6.8V SN. :35 n2- 8.3 new. 8.3 nom- at 33.:- 3.3 3.?ng- 3.0: 23.8. 8.3 3.04%.:- as 33.:- 8.3 odeWdMn K s A5 o§u§§v< 3: 65853523503 05823923 -o_o»othéu.§ 93 83583885 98 Aocowfiegfiuoo -m_-o§:88-n.$ 058233.63 -o_o>ot.—--oh§ Aofiugxfggv 38 05:0: -8.m.:o_&2m.§h cg 8.28:: .n. 22985 203 A3 2.8 $8 _ #- mooafiawm- new. $23.53- Sé- w?- 339. g Egon-Bub . . . as 8.3 W 38 p :1 £3: 9:- v $- 29333- 2v- 0.?- 333. Ease-E . . . A2: :5 8 5v V 3. 3%: gm- 0 mu- $38.an- mav- mom- .25.:- ofioozmieam . . . 32V 8.3 8 G8 280% N S. ~32 3m- v 2. 228.23.- Em- m.~m- 23?:- 2322055582“ . . . 8.2V AQNS ‘ G8 280% m mm- 38: an- v mm- «533%. 2:. N:- ..SHZm- -8.m.ao_&2m-§ 8.3 2.»: ”fl _ H M8 . . - . - . - 3N- ~.mu- 3.33a- ofioow $3. 20 E N mm. 5352 own m VN awvwoo 5mm 204 38 And 5 Ava-V mg oncofificaegfiom mom- vaflim- 3w- mmomflém- 2m- o.~m- an??? -§§-u2o§§.§§ 33 AS: 8.3 @Efififi Em- aowvwém- now- 8222mm- flwm- 3m- ,odfioem- 0338928 35 2.: 06 38 M.;. 3:39- Zm- $352-$- o.%. 9%. S-u-Nam- @2235 ES ogoowoc Aim: 5%: -8.~.o.~.o.N.od 32 883.8? 32 32.38? 32 3: 23.82 -oiméfi 6.8 Si 00 Ag .2? 03239- Sm- Nvfimwwmv- 2&- 3? b.353- 080830222223 ES 9an 2.5 @8535 «.8- 2832.3.- 33. 23:39- SM. 3&- .m.fim._- osoocgzgwsfim 205 0% ..S .305 60% .5250 .55 .5. .m .3 .8023- ”.A 1m .0250 M.0 d daunam ”a. .2 £825 M...— ._. 6.6-$32 .. 5858.. 0.03.:- :o egg .5 8005.8 :0 52 32950 2: 0o 80055208. 5 0.82 505%: .33: 80850588 05%: a a .55: 35 .5 .oom. .5250 5:5 .5. .5 .5 .5083“ ”a. .2 $020002 Md .0282 £- 485— ”..H. .2220 5 02.890: :32:an ma 80: 0: 0:: .0320.— fiofi 05 3:020:80 a? .2 .05 5555.5: .556 .5 :- 55 ”.0 2 .5: em 9 .5 £8.05 3 .2 .5520: 5: .m 6.583 a 3:82 8.5.588 :o .8: 2 as": .35 .5 .05 55.55.: .555 .5 .: 053 ”.3 05:00 ..> .3 .285 ”.3 :25: . .002 .883 ”:0: as. 58:20 .0: 05 5.5.2558 5.50655 53 58.555.55.55 .0 .m .35: 5: 2 0252 ..m .5 .528: . .28: 3: .8 .80 .525 .5... .5 .> .523 3 .z .5052 5 85.0%: 0580852220555 9 8:35.58. 0o .8: a. 8o: 8580.: _. .000 65% 535580 .5250 room. .5250 .5. 4 .2 5023—02 0: .0302 ”.5- 08; ..H .250 .. Hon .0 .55-3 5.925055 .5250 .5. 2 .3055 w> .3 .2005 ”.3 :0me o .8: m .5: .053 .80 .555 .55 .5 .m 2 05o: 5: .2 .0 .52: ”.0 .Eoe.§m Md 53202 ”.70 .0520 .. .82 .50 £92 .5250 05.55 .5. .5555 .m .2 $050M J .m .Bfimafiom M.> .O 08:30M 0m ..H- 0.6000552 Mm .> $82312 .w .0>o:0EE w< .> 05:05.23 5 +02 .55 :5: .555 53.5 .5 d 5520: ”.0 55309-520 ”.2 .m .5580 5: 2 .0500. .32 .3. .053 .5250 .500 .5. .0555 .4 .0 62:23.30 05031.; 40mg M.2 .m 56:25: ._ woo-o- .3. .33 59.50.50.505 .55-d 523:3 ”d .3 god ...m .::Em M.2 £2.50 .0:0ES:0:2050.n.m.~ 3 3:352:82: :00: 0: 80:5 00:08:09 .03: .55 5 .985 dam. .5250 5:5 .5. .< 20:02 21m .m .:E>0:0> M.z .> .>o_0m 2 .m ..Sxamafiom M.0 42222 ”dam .3228? d .00: 00m 0:05:20? 9 :o:::0wo€.3 mo .00: 0: Bob 022500”; a «when .35 .33 dew. .8250 .55 .5- .Bum .5853 ”dam .0809 MM .3 .532 25:08.03 9 802000.65: mo :00: m: 825 38:52 0 .23 23> B02 ”000E 280284.. 6505505580 0580585530 505: Bram-50 \0 55558050055055 .0 .220?— md .5 .080 2:088:32: 03000:: :80.» Eat 8:08:98 8 5:02.893 0o :00: 020880 :0 0003 8:258 mo 08m 0 .30 035- 0000 2053:: 8:203:00 95:» 88.: 0083030 2:03 0::- o .653 28> 302 ”50:3 :55 £0.50sz 5025050055055 .3 .m £858 3:033:00 95% 0.:8:0m 0:: 54 3:085:08 Bob ”520:0 :35 0 .008: 02.5050 82:: ..< .m .m :22 M.: d :93 5: .5 58.55 ”.0 .m .85 B .32 :35: .80 52:2 .8228: 0.5 85250 .2 9.853 883.5: 05:50 :22 as: . dengue-E8 35:0 6032 50 5:08am .. ._.. :23 09:2: 0:: AN .505 3:202:00 25:» 32:20 :8 20520:: :23030: 0.9255 352.8 05 E 020305 0:58:80 25- .20 65?: :05 88302.30 0.... 55:0 0&8 3:808 5a: 203 855.5. :< 5558.. 2.8 n m H 0822 a 8:058 39 0858: .5: 53 5 8058.: :o 08: . A500 3.8 8. : 8U 0:000:58865E0m Sm- mooo:.$m- :8- 052.22%- 0.5m- Sm- adflwm- -§5.5.§.8-§5 206 Table 6.2 Comparison of the Wiberg and Ibrahim/Schleyer Group Equivalents with Those Derived from the Ab Initio Energies of Table 6.1 Group equivalents in hartrees Group Wiberga Ibrahim/Schleyer" This work (via Table 1) HF/3-21G HF/6-31G* HF/3-21G HF/6-31G* HF/3-21G HF/6-31G“ CH2 -38.81054 -39.02662 -38.81150 -39.02684 -38.81108 -39.02614 CH -38.24087 -38.45350 -38.23954 -38.45338 38.24054 -38.45402 C -37.65633 -37.87895 - - -37.65544 -37.88182 s.d.c 7.3 2.3 7.5(1 2.5d 6.7 1.8 ‘ Ref. 2b. b Ref. 3. ° Standard deviation (in kcal/mol) of experimental vs. calculated heats of formation for the present data set (Table 6.1). 6 Standard deviations are based only on compounds with CH2 and CH groups; spiranes were excluded from correlations since Ibrahim/Schleyer do not provide an equivalent for the quaternary C atom Figure 6.1 207 150 . F 120 - I [kcal/mol] 8 8 ’a. 30 - 55 \z E o- -30 .1 -60 l l l -60 -30 0 30 6O 90 120 150 AH,(calcd) [kcal/mol] Plot of experimental heats of formation, AHKexp), vs. calculated values, AHf(calcd), from the HF/6-3 1G* group equivalents evaluated in this work, for the compounds in Table 6.1. Slope 1.00 was taken for the correlation line. 208 The equivalents derived in this work for the CH2, CH and C fi'agments (see Table 6.2) are essentially unchanged from those previously reported, supporting the consensus that errors due to incompleteness in basis set, correlation treatment, and vibrational contributions, scale linearly with the numbers of each group. They are absorbed in the group parameters, to yield calculated heats of formation of accuracy comparable to experimental measurements. The AH¢(calcd) values derived for the 3-21G basis set show large errors especially in the case of cyclopropane derivatives, where the flexibility afi‘orded by inclusion of polarization firnctions into the basis set is essential for a proper description of these compounds. Analogous values for the semiempirical PM3 method16 are included in Table 6.1 for comparison and the resulting heats of formation show, as expected, unacceptably large errors; the standard deviation for the best linear fit between PM3 calculated and experimental heats of formation for the compounds listed in Table 1 is 8.0 kcal/mol. The MMX method, derived fiom Allinger’sl7 MMZ force field, was also employed to compute heats of formation for the compounds included in Table 6.1.18 Usually, MM reproduces well the thermodynamic properties of hydrocarbons; e.g., the new MM4 force field applied to 56 alkanes and cycloalkanes, excluding small rings, calculate AH; with a standard deviation of 0.4 kcal/mol vs. experimental values. 19 However, the NHVIX results in Table 6.1 show that although most compounds have MMX calculated heats of formation within experimental accuracy, in some cases there are large discrepancies between experiment and calculation (5 compounds in Table 6.1 have MMX AH.(calcd) in error vs. AH¢(exp) by more than 10 kcal/mol). Thus, the performance of the MMX 209 method,20 although much better than that of PM3 or HF/3 -ZlG models, is not entirely consistent, leaving the ab initio I-[F/6-31G* group equivalent scheme as the most reliable when compared to experiment. The estimates of the enthalpies of formation using the 6—3 1G* basis-set are uniformly quite good. Hence, the group equivalents at the 6—3 1G* level successfirlly predict heats of formation of both small and medium-ring strained hydrocarbons fi'om ab initio energies, in rather good agreement with experimental measurements. The new group equivalents yield a modest improvement over those of Wiberg2 and Ibrahim and Schleyer3 . The essential message, however, is that Wiberg’s original set is quite adequate as expected and the principal enhancement ofi‘ered herein is an updated estimate of the quaternary carbon equivalent. Predictably, the equivalents at the unpolarized 3-21G basis set level cannot be used safely for strained compounds since polarization firnctions are known to be needed to properly describe small ring carbocyclics. Such calculations can be used when experimental results are unavailable, or as an independent check when an experimental result is in question. 6.2 References ‘ Craciun, L.; J. E. Jackson, J. E. submitted to J. Phys. Chem. 1997. 2 (a) Wiberg, K. B. J. Comput. Chem. 1984, 5, 197. (b) Wiberg, K. B. J. Org. Chem. 1985,50, 5285. 3 Ibrahim, M. R.; Schleyer, P. v. R. J. Comput. Chem. 1985, 6, 157. 4 Burkert, U.; Allinger, N. L. Molecular Mechanics, ACS Monograph 177, 1982. 210 5 Dewar, M. J. 3.; Thiel, w. J. Am. Chem. Soc. 1977, 99,4907. 6 (a) Hehre,W. J .; Ditchfield, R.; Radom, L.; Pople, J. A. J. Am. Chem. Soc. 1970, 92, 4796 and subsequent papers. (b) For an example of the use of homodesmic reactions to calculate the heats of formation of 15 representative strained and unstrained medium-sized hydrocarbons, see: Disch, R. L.; Schulrnan, J. M.; Sabio, M. L. J. Am. Chem. Soc. 1985, 107, 1904. 7 (a) Yala, 2. J. Mol. Struct; THEOCHEM. 1990, 207, 217. (b) Castro, E. A. J. Mol. Struct.: THEOCHEM. 1994, 304, 93. (c) Hemdon, W. C. Chem. Phys. Lett. 1995, 234, 82. (d) Smith, D. W. J. Chem. Soc., Faraday Trans. 1996, 92, 1141. 8 Schulman, J. M.; Peck, R C.; Disch, R. L. J. Am. Chem. Soc. 1989, 111, 5675. 9 Allinger, N. L.; Sakakibara, K.; Labanowski, J. J. Phys. Chem. 1995,99, 9603. 1° (a) Allinger, N. L.; Schmitz, L. R.; Motoc, I.; Bender, C.; Labanowski, J. K. J. Phys. Org. Chem. 1990, 3, 732. (b) Allinger, N. L.; Schmitz, L. R.; Motoc, 1.; Bender, C.; Labanowski, J. K. J. Am. Chem. Soc. 1992, 114, 2880. (c) Allinger, N. L.; Schmitz, L. R; Motoc, I.; Bender, C.; Labanowski, J. K. J. Comput. Chem. 1992, 13, 838. (d) Schmitz, L. R.; Motoc, I.; Bender, C.; Labanowski, J. K.; Allinger, N. L. J. Phys. Org. Chem. 1992, 5, 225. (e) Liu, R.; Allinger, N. L. J. Phys. Org. Chem. 1993, 6, 551. “ Strain energy is the difi‘erence between the experimental (calculated) heat of formation of a compound and that of a hypothetically “strainless” model, calculated here from Benson’s group equivalents (see Benson, S. W. Ihermochemical Kinetics; John Wiley: New York, 1976). ‘2 Binkley, J. 8.; Pople, J. A.; Hehre, w. J. J. Am. Chem. Soc. 1980, 102, 939. ‘3 Hariharan, P. C.; Pople, J. A. Chem. Phys. Lett. 1972, 66,217. 1‘ The analysis was done with Microsofi Excel version 5.0 employing the function LINEST, which uses the “least-squares” method to calculate simple or multiple linear regressions that best fit the input data. 1’ Beckhaus, H.-D.; Rfichardt, C.; Kozhushkov, S. 1.; Belov, V. N.; Verevkin, S.P.; de Meijere, A J. Am. Chem. Soc. 1995, 117, 11854. ‘6 Stewart, J. J. P. J. Comput. Chem. 1989, 10,209. ‘7 Allinger, N. L. J. Am. Chem. Soc. 1977, 99, 8127. ‘8 The MMX calculations were done using the interactive molecular modeling program 211 ‘7 Allinger, N. L. J. Am. Chem. Soc. 1977, 99, 8127. 18 The MMX calculations were done using the interactive molecular modeling program PCMODEL version 4.0 (Serena Software, Bloomington, IN). ‘9 Allinger, N. L.; Chen, K.; Lii, J.-H. J. Comput. Chem. 1996, 17, 642. 2° The best linear fit of AHt(exp) vs. AHr(calcd) gives a s.d. of 4.2 kcal/mol for the compounds listed in Table 6.1. APPENDIX 212 213 bah 000 035 a: 0308 00.00- 00.02 00.00 00.00 0005.0 0002.0- 0000.0 0000.0 000.0 0000; 0.00 a: 8 00 00.00- 00.02 00.000 $.00 000020 0000000- 0000.0 0000.0 0800.0 000.0 0.00 000 Q .3 00.00- 00.02 3.000 00.00 00002.0 0000.0- 0000.0 0000.0 00:00.0 0000.0 0000. 0.000 AHA «00 2.0- 0.02 0.000 00.00 «3020 00003.0- 0000.0 000.0 500.0 0003 0.00 000 AV 0 00.00- 00.02 00.0: 00.00 000030 3002.0- 0000.0 0000.0 00.0 0000.0 Se 00.0. Q 2 2.00- 0.0.0: 00.02 00.00 00002.0 000020. 00000 3000 $0000 0000.0 0.00 0R 4 0 00.00- 2.02 00.0: 00.00 00003.0 00200.0- 0000.0 0000.0 000000 0000._ 0.? 20 Q «a 00.00- 00.02 00.0: 00.00 00002.0 «00000.0- 0.0.00.0 000.0 000000 0003 0.? 000 Q 3 as: as s e a a .2 as as a... gees .2 020 i. an; 80 388880 83326 *2 0 214 00.00- 00.00- 0 .00- 00.0? 00.0. 00.0w- 0.0V- 00.00- v0.00- 00:00 00.03 00.02 00.000 00.03 00.03 ~02 00400 0v.0N_ .00.00 00.00 #000 00.00 00.00 00.00 00.00 0.00 0.00 00.00 v0.00 00.00 :A0 0.00 00.00 00.00 00 060 00002 .0 0002.0- _0v00_.0 0003—0- 00020 3003.0- 00~00~.0 00003.0- 20020 00000.0- 00002.0 30000.0- 0302.0 0200—0- 0000 .0 000000 .0- 0000~ ~.0 00002 .0- V0000 0000.0 0000.0 0000.0 0000.0 000.0 0000.0 0000.0 300.0 030.0 00.0.0 03.0 003.0 00v0.0 000.0 0000.0 300.0 _0v0.0 00000.0 00000.0 0000.0 00000.0 00000.0 00000.0 00000.0 00000.0 000.0 0000. 0 000. _ 0000; 0000; 0000; 0000A 0000. 0 0000._ 0000.0 0.00 0.00 0.00 0.00 0.00 0.00 N00 #00 ~00 ~00 ~00 v: 0: 0.000 5000 000 0.000 00_ awho®flfifih 3 a: «00 «00 n0 3 A: 215 0.00- 00.00- 00.00- 0—00- v.00- 0.00- 00.00- 30.00- 00.00- 0: 00.0: .0303 00.0: 0.303 0.003 00.00— 00.303 03.003 0.00 30.0“ 00.00 3.00 0.0V 00.0w 00.00 V0.00 3.00 v.00 00.00 30.30 0.00 _.00 30.00 03.00 00.00 0330 0000000 00003 .0- 0000030 0303 .0- 000003 .0 000030- 000000 00003 .0- 0000030 00303.0- 0000030 000003.0- 003003 .0 5.0003 .0- 0000030 000003 .0- 000003 .0 00000.0- 300.0 0000.0 000.0 300.0 V0000 0000.0 000.0 0000.0 0000.0 330.0 00V0.0 303.0 330.0 030 .0 30.0 00V0.0 030.0 030.0 00000.0 00000.0 00000.0 0V000.0 00000.0 00v00.0 3000.0 00000.0 5.00.0 3003.3 0000. 3 900.3 3003.3 0000.3 0000.3 3000.3 0000.3 0000.3 0.00 060 0.00 0.00 0.00 0.00 0.00 0.00 0.00 003 0.003 003 003 003 003 0.003 003 303 43:2?4 @quéé vii a0" {03 03 «03 00 a3 03 «03 003 216 00.00- 00.03- 00.0V- 00.0V- 00.0V- 0V._V- 0003- V00- 00 03.00- 00.003 00.0: 00.003 0V.003 30.003 00.003 00.0: 0.0: 0.303 00.0: 000V 0.00 0.00 0V.00 00.30 00.00 0.00 00.00 V0.30 V0.00 00.00 0.00 00 00.00 00.00 30.00 3.003 V0.00 0V.00 00 .00 0000000 _ V0000 .0 0003030 030030 00VV03 .0 000000 0000000 0000000 000003 .0 0000000 303000- 000000.0- 0003030- 00003 3 .0- 00003.0- 00000.0- 000003 .0- 00003 .0- 00000.0- 00003 .0- 00000 000.0 0V000 000.0 0000.0 0000.0 0000.0 0V00.0 0000.0 000.0 300.0 0000.0 00V00 00V00 0VV00 0_V0.0 0000.0 03V00 0VV00 :V00 3000.0 0000.0 03000.0 00000.0 00000.0 00V000 V00000 VOV000 00000.0 00000.0 0V00. _ 0V0_.3 0000.3 0000.3 0000.3 0000.3 V0033 0003.3 0000.3 030: 0.V0 0.V0 0.00 0.00 0.00 0.00 0.00 0.00 V00 0.00 003 0.V03 0.V03 003 0.003 0.003 003 303 003 V2 gng©®®0 . g2:- :03 :0 V0 «00 «00 90 .30 a: a0 .50 217 V0.00- 0.0- V0.V0- 0.00- 00.00- 00.00- 03.00- 00.00- VV.00- 00.033 0.333 00.033 033 30.003 00.033 30.003 00.003 00.303 V0.00 03.3 00.00 30.00 0.0V 0.00 00.00 03.0V V0.0V 00.303 00.V03 00.00 00.00 00 3.303 00.00 33.00 0V.V0 0000000 V000000 0000000 0000000 0003000 0000000 0000000 0000000 0000000 000000.0- 0000V00- 000000.0- 0330330- 0003000- 0030030- V00003 .0- 000000.0- 3 00000.0- 0000.0 0000.0 0000.0 3000.0 0000.0 000.0 0V000 V000 3000.0 0V000 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 VOV00 00000.0 V0000 00000.0 0VV000 00000.0 00000.0 0V0000 00000.0 0000.0 0003.3 0333.3 0000.3 0003.3 0V00.3 0000. 3 0000.3 0V00. 3 0300.3 0.00 0.30 V00 0.00 0.V0 0.00 3.V0 0.V0 0.00 0.VV3 03 03 0.03 03 03 0.03 0.303 003 @4®fl.CXJB$QZX 00 a0 000 00 00 A.03 A00 00 A.00 218 00.0.. V0.0- 00.0- 00.0- 30.03- 00.0- 00.0- 00.0- 00.03- 0300- V0.3: 00.033 00.003 00.033 00.033 00.033 00.333 03.333 30.33 30.033 00.0 00.0 V0.0 03.0 00.3 33.V V0.0 V.0 30.03 00.00 V0 .003 00.003 00.003 00.003 00.V03 03.003 00.003 0.003 00.003 00.303 0000000 3030000 V000000 0000000 00000.0 V0000 0003000 0000000 0VOV000 000000.0- 3300000 0000V00- 3 V0000 .0- V000000- 300000.0- 000000.0- 0003000- 0000000 000000.0- 000000.0- V0000 0000.0 000.0 0000.0 0300.0 0000.0 300.0 0300.0 V0000 0000.0 00.0 0000.0 0000.0 0000.0 0300.0 0000.0 0300.0 0300.0 0000.0 0000.0 V00000 003000 0V0000 0VV000 00300.0 00000.0 0000.0 03000.0 3000.0 00000.0 0333.3 0033.3 033.3 0033.3 0003.3 0333.3 0033.3 0033.3 0V03.3 0003.3 0.00 0 .00 3.00 0.00 0.30 0.00 0.00 3.00 0.30 0.00 0.003 0.V03 0.V03 003 003 0.003 003 003 33 V3 @4€@%£6Q£Q cvm A303 00 «00 «30 003 Q3 00 00 903 219 8.?- mmefl v: 0.: 0%de 388.0. 0806 $36 $30.0 88.“ 3m 20 Q 2 833 a? 083w 8.3 a a mg on .o.m 5c a .x. 3 @5888 .24 *030-00m33 2.8 3.83 3:- 30: 235.0 SS8? v3: 203 8035 «on: «.8 S: a 20m :2 2.02 «0:- 3%: ~33: ”$8.0- good SN; 0886 am: now 0.02 NMR 3». 2.2 8.2: 3.2- 0;: Said 0802.0- 38.0 0%; 033.0 02: EN «02 a «an». 2.: $02 :2- 8.0: $336 ”28.0. 8%.o SN; 208.0 82.0 mom ~02 D“ ..S 0 8+ 3:: www- omd: 8386 388.0. :86 SN; 0235 wt: 23 v.83 J“ «on 2.0- who: 8+ 2.? $886 ”808.c- wmwod 8m; 283 0:: man E Q» 3" 25 $.83 26 $.83 :woood 858.0. 033 SN; «$8.0 $2.3 cam v.02 Q 00 220 03 00.0V- 00.V03 00.00 00.00 V000000 00V3V00- 030.0 0000.0 30000.0 0000.3 0.00 003 00.00- 00.003 0.00 00 V000000 V000000- 030.0 3000.0 0000.0 0000.3 0.00 0.003 0 0.00- 3.003 30.V0 00.V0 0VOV00 VOV3V00- 0030.0 V0000 0V0000 0000.3 0.00 003 «0 00.00- V0003 00.00 0V.00 0000000 000VV00- 3330.0 0300.0 00000.0 0000.3 0.00 003 «0 000.3 0.00 003 «0 300000 030V00- 0030.0 V0000 00000.0 00.00- V0003 03.003 33.V0 0.00- 0.003 00.003 30.00 030000 0VOV00- 03300 003000 00V000 00000.3 00.00 0.000 «am 00.00- 00.003 0.003 3.00 VOV0000 0000V00- 0330.0 3300.0 0000.0 00.3 0.00 000 03.00- V0.V03 V033 0.00 3000V00 003V000- V0000 0000.0 30000.0 0000.3 0.V0 000 03 0.00- 0.003 00.003 V0.00 0000V00 3000000- 0000.0 0V000 00000.0 0000.3 V00 030 0 0.V0 030 00 Q4QQQQ®%« 30.V0- 00.V03 00.033 00.30 0000V00 V000000- 000.0 0000.0 00300.0 0000.3 221 03.00- 00.30- 00.00- 00.00- 00.00- 00.00- 3.00- 0V- 00.0V- 00.003 00.033 30.303 0V.003 03.003 0V.003 0.003 0.003 00.003 0.00 00.0V 00.00 00.0V 00.00 03.00 00.V0 00.00 00.00 0.00 30.00 3.00 00.00 V0.30 33.30 00.30 30.00 30.00 0000300 000000.0- 0000000 3300000- 033VV00 0000V00- 0030000 00000.0- 0000V00 0003V00- 030V00 000VV00- 0000V00 V000V00- 0000000 0000V00- 0000000 00000.0- 0300 030.0 0030.0 3030.0 0030.0 0030.0 0030.0 030.0 0030.0 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 3000.0 0000.0 030V00 00V00 000V00 0VOV00 330V0.0 000V00 300V00 003V00 000V00 0000. 3 3000.3 0000.3 000.3 0V00.3 0000.3 0V00.3 0V00.3 V0003 0.00 0.00 0.00 0.00 3.00 0.00 0.00 0.30 30 0.003 003 303 303 303 V03 003 0.003 3.003 fi®fl®®§®§4 \O v-q «03 :03 U3 «3 «03 «03 .1: [x I-l III-1 222 00.00- 00.00- 00.00- 0.00- 00.00- V0.00- 0.00- VV.30- 03 .00- 00.033 00.003 00.033 0.033 3.033 00.003 0.033 00.033 00.003 00.V0 30.0V 3V.V0 00.V0 00.0V V00 0V.V0 03.0V 00.0V 00.00 0.00 00.00 00.00 00.00 0.00 30.00 00.00 V0.V0 0000000 0000000 0000000 00V000 VO0V000 0000000 00000.0 0VOV00 0V00000 0300000- 00V0000- 00VOV00- 03300.0- 00VOV00- 000000.0- 000000.0- 0000V00- 0300V00- 0030.0 0030.0 0030.0 V0300 030.0 330.0 0030.0 0030.0 0030.0 0000.0 3000.0 0000.0 V0000 0000.0 0000.0 0000.0 V0000 V0000 000V0.0 00VV00 000V0.0 000V0.0 0VOV00 033V00 000V0.0 000V0.0 000V0.0 3000.3 0000. 3 0000.3 0000.3 0000.3 0000.3 0000. 3 0000.3 0000.3 0.00 0.00 0.00 0.00 0.00 V00 0.00 0.00 0.00 303 003 V03 003 0.003 003 003 003 003 100%Q©£€£% .303 A0 «00 «03 «03 3.3 «03 .30 .33 223 0.00- 00.00- 00.00- 00.00- 30.03- 30.0V- 00.0V- 0.0V- 30.00- V0.03- 0.033 00.003 00.303 0V.003 03.033 00.003 00.003 3.003 00.003 00.033 V0.00 00.00 3.0V 00.00 00.00 0.00 00.00 00.00 00.0V 0V.00 00.00 V0.00 0.00 0V.00 00.00 00 03.00 30.00 00.00 00003 003000 0003000- 30000 000000.0- 0000000 0030000- 00V0000 V00000- 3033000 000000.0- 00000.0 00000.0- 0VOV000 0300300- 3030000 330300- V000000 0300300- 0000000 VO0V00- 3000.0 0030.0 0030.0 0030.0 0030.0 0030.0 3030.0 3030.0 0030.0 0030.0 0V000 0000.0 0000.0 000.0 0000.0 0000.0 3000.0 0000.0 0000.0 0000.0 00300.0 00300.0 000V0.0 000V0.0 00300.0 00000.0 V00000 003V00 3VOV00 0000.3 0000.3 0300.3 0300.3 00.3 300.3 0000.3 3000.3 00.3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 00000.0 0V000.3 00.00 0.03 0.303 003 003 0.V03 0.V03 003 0.003 0.003 003 ©®®@ g§%4® A300 00 :00 003 00 V0 «00 «00 U0 «#0 224 00.V- 00.03- 30.03- 00.03- 03.0- 3.00- 00.00- 00.00- 00.00- 00.033 00.33 00.033 30.033 00.333 0.033 30.033 00.003 3V.033 00 .V 00.003 00.03 00.003 30.00 00.003 0V.30 00.003 V0.03 00.V03 00.00 00.003 30.00 0V.00 0 .0V 00 30.00 00.303 0000V00 0000V00- 0003000 000000.0- V000V00 003V000- V000000 03V0000- 0V00000 0V00000- 3 000 00.0 V000V00- 0330000 0300000- V030000 V030000- 0000000 0VO0V00- V300 0030.0 330.0 0000.0 0000.0 0030.0 0030.0 0030.0 0030.0 000.0 0000.0 0000.0 VV00.0 VV00.0 0000.0 0000.0 0000.0 0000.0 00000.0 0V0000 00000.0 00000.0 03000.0 00V00 000V0.0 00 0V00 00V000 0V00. 3 0000.3 0300.3 VV00. 3 0000.3 0000.3 0V00.3 3000.3 0300.3 0.V0 0.00 0.00 0.00 V00 0.00 0.00 0.00 0.00 003 33 V3 0.VV3 03 03 0.03 03 03 6.50119Q4Q‘QME 00 00 903 00 00 900 00 00 A303 225 00.0- 0.0- 00.0- 00.3- 00.0- 00.V- 00.03- 03.0- 30.V- 3V.033 0.003 03.333 03.033 00.003 03.333 V0033 00.033 00.333 00.0 0V0 00.0 03.0 V0.0 0.0 03.33 0V0 30.0 30.003 00.003 30.003 00.003 00.003 0.003 00.003 V0.003 00.003 0000000 0300V00- 3V00000 V30000- V000000 0030V00- 300V000 00V3V00- 3300.0 0V30V00- 0V00000 03V000- 00V0000 000000.0- 0000000 003V00- 3000000 0000V00- 0030.0 000.0 V0300 3000.0 0030.0 000.0 0030.0 0000.0 3030.0 0000.0 00300.0 00000.0 0300.0 0000.0 0030.0 0000.0 030.0 0000.0 0V0000 00000.0 00300.0 30000.0 00000.0 3000.3 0000.3 0000.3 0000.3 00003 30000.0 00000.3 03300.0 00000.0 00000.0 0000.3 0000.3 0000.3 0.V0 0.00 0.V0 0.V0 V0 3.V0 00 0.V0 0.V0 003 V003 0.003 0.V03 0.V03 003 003 0.003 003 0%QBQ0QQQ .300 00 «V0 003 00 «00 «30 903 9V3 226 00826 08 00: :0 39.8 000 38050925: 93 02.0 a 83 050.296 Became-8.359 coo: .38 z 888 ._. 23 5028032 009% - 032% 23 Soon - 032% 8 83% 33% a 88.9% 99% Ba 990V §§§§ .8 ea 2: some? 8:3“ Ba 863 ”Mk-99%.“; 2a $0999va 3 83% daemou a 3&5 98% 9a 89m :0 “3.3 395 0 so 8039 93:83 353: 03 m0 395 939 $603.83 00373 :80 3.3 05 0 no 35600009 8035300 8a :0 3:8 on ”0038 003:0 0:00 EOE-88¢ 593333333 95 :80 83:0 980 33-0 23.. mm .O.m 2.x 5 85.3% 03-0 on“ m_ :90 $7.038 Omz 800 2.5.03 8.3.23 95 00 808.35% 8982. mm 9m .x. 3:838 053959 5%.-5% 9-02 39:00-28 3583598 05 3:883. 983. . 0V.00 00.03 00.33 3V.03 V0 00.003 00.003 30.003 00.003 0.003 00.00- V0033 30.3. 00.33 0003- 00.033 00.03- 00.033 00.0- 03.333 0300000 000000.0- V000000 0003V00- 300003 .0 00 3 V000- 3000000 V000V00- 3V00000 300000- 33000 0000.0 0300.0 0000.3 V0000 0300.0 0300.0 0000.0 0000.0 0000.0 0000.0 3V00.0 00000.0 00000.0 00000.0 0V0000 0000.3 3000.3 0000.3 0000.3 3.00 0.333 30 0.003 0.30 V003 0.30 0.003 0.00 V003 40.2% «.300 «00 «000 «00 «0m