cum}... .;.. Lin. .. .h..r..:n.-:lt_..: £7... 1...... r1. 3.. 7:...r. :— .. .3. TIES In : . V ‘ ‘ 1v», ., an 5.? 4:... . .3. m .m.» Y:h2m.v.§!. Illillllillilillll’lllfilllllllfllll 3 1293 010191868 This is to certify that the dissertation entitled I. «FEpoxyalkyl Radical Fragmentation: A Computational Study II. Experimental and Computational Studies of Carbene Processes: Homologa iagntgqxgen, atom-transfer, and Alkene CycloadditionE e MingShi Lee has been accepted towards fulfillment of the requirements for l’hADA degree in __Chem_i§jLL_ Major professor Date MSU is an Affirmative Action/Equal Opportunity Imtitution 0- 12771 LIBRARY Michigan State University PLACE IN RETURN BOX to romovo this checkout Itom your record. To AVOID FINES return on or baton date duo. DATE DUE DATE DUE ‘ DATE DUE I. a-Epoxyalkyl Radical Fragmentation: A Computational Study 11. Experimental and Computational Studies of Carbene Processes: Homologation, Oxygen Atom-Transfer, and Alkene Cycloaddition By MingShi Lee A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1993 ARM 1. oc-Epoxyalkyl Radical Fragmentation: A Computational Study 11. Experimental and Computational Studies of Carbene Processes: Homologation, Oxygen Atom-Transfer, and Alkene Cycloaddition By MingShi Lee The unusual regioselectivity of a-epoxymethyl radical ring- opening was studied by ab initio methods. Computational results at the PMP4/6-3lG*//UMP2/6-BIG* level indicate that the barrier for C-O bond cleavage is lower by ~6—7 kcal/mol than that for C-C bond cleavage. The heat of reaction is ~3 kcallmol more exothermic C-C than for GO bond cleavage. Several substituted a-epoxymethyl radical ring-openings were examined at the MP4/6-3 lG*//UHF/6-3 lG“ level. In general, it electron- withdrawing substituents favor C-C cleavage whereas donating substituents favor C-O opening; substituent effects are stronger on the epoxide ring than on the (at-methyl. Reactions of fluorenylidene with terminal alkynes were studied by ESR and product analysis at 77 K. Due to steric interactions, the vinylcarbene intermediate derived from the addition of triplet fluorenylidene to a monosubstituted acetylene could not be intramolecularly trapped by its aryl ring, as occurs in the case of its triplet diphenylcarbene analog. Through ESR studies and the observation of a 1,2-chlorine atom shift in the addition of fluorenylidene to propargyl chloride at 77 K, it was shown that triplet fluorenylidene addition to alkynes proceeds through the same type of intermediate. A new type of carbene reaction has been examined: the oxygen atom-transfer from a carbonyl group to fluorenylidene to produce a more stable secondary carbene, such as dimethoxy or diamino carbene, which is not easily generated by a photolytic process. The relative rate constants for various oxygen atom-transfers have been determined by competitive trapping with methanol. Transition structures for the cycloadditions of singlet carbenes to various "push-pull" alkenes, HO(X)C=C(X)CN, were studied using the MNDO method. The calculations show an electronic orientation preference of the reaction paths. Carbenes prefer by 3-5 kcal/mol to ”add to" the electron-donor end over the electron-acceptor end of the olefin. Steric effects also make distinctions between rotameric transition structures, where X = CH3 or Br, the energy differences are 0.5-0.9 kcal/mol. A parallel experimental effort is described to find evidence for the asymmetric path as opposed to a more symmetrical approach of a singlet carbene. To my parents Thank God. I am finally done with this after exhausting my English vocabulary. Many people deserve thanks for making my stay at Michigan State University an enjoyable one. I had always wanted to be a synthetic chemist. However, Dr. James E. Jackson, my advisor, opened my narrow vision of chemistry and though me how to enjoy chemistry. I do not remember how many nights, or should I say early mornings, we spent together chatting about science until recently he finally has synthesized the best natural product he ever made, the bug, Kelvin Charles Parker Jackson. Here I have to mention another Dr. Jackson, Dr. Evelyn Jackson, who is also my Mrs. Advisor. Her generosity and consideration always made me feel at home. As a foreigner here, I have to thank my lab-mates Sei-Hum J ang, Theresa Wagner and Scott Stoudt, who made the cultural transition much easier for me. Especially, Scott, my personal English tutor, who put a lot of effort into correcting my "chinglish". Also, Einhard Schmidt, a true friend and big brother, never said no to my asking for help and also a long-time "lunch-mate". Dr. Peter J. Wagner's group was like my second research group. Bong-Set Park, Bob Smart, Kevin McMahon, and Kung-Lung Cheng, always shared the entertainment with me, either scientific or recreational. Finally, I have to thank my family for their spiritual and economic support. Though they still do not know what I am doing, the only question they ask is "what do you need?” I how I have a chance to repay some of their love in the future. ABSTRACT: LIST OF TABLES: viii LIST OF FIGURES: xii CHAPTER 1: a-Epoxyalkyl Radical Fragmentation: A Computational Study 1 PART I. The Parent System 2 PART II. Substituent Effects 18 CHAPTER 2: Low Temperature Carbene-to-Carbene Homologations 58 CHAPTER 3: Carbene-to-Carbene Oxygen Transfer Reactions 106 CHAPTER 4: On the Symmetry of Carbene-Alkene Cyclo- propanations: Computational and Experimental Studies 128 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 1.10 1.11 Total and Relative Energies of Calculated Structures 1. l A-l . 5 A ......................................................................... l6 Calculated Geometries (UHF/6316*) of Initial Substituted Radicals (1.1) ...................................................................... 36 Calculated Geometries CUHF/6-31G“) of Transition Structures for CC Bond Cleavage (1.2) .................................. 36 Calculated Geometries (UHF/6-31G“) of Transition Structures for CO Bond Cleavage (1.4) .................................. 37 Calculated Geometries (UHF/6-3lG“) of Substituted Product Radicals 1.3 from CC Bond Cleavage ....................... 37 Calculated Geometries (UHF/6-31G“) of Substituted Product Radicals 1.5 from C0 Bond Cleavage ........................ 38 Total Energies for LIB-1.53 (Amino Group on a-Methyl) ................................................. 39 Total Energies for l.1C-1.5C (Cyano Group on a-Methyl) .................................................. 40 Total Energies for LID-LSD (Boron Group on a-Methyl) .................................................. 41 Total Energies for l.1C'-1.5C' (Cyano Group on Epoxide Ring) ............................................ 42 Total Energies for l.1D'-l.5D' (Boron Group on Epoxide Ring) ............................................ 43 1.12 1.13 2.1. 3.1 3.2 4.1 4.2 4.3 4.4 4.5 4.6 Reaction Barriers and Heats of Reactions for Ring-Openings of Substituted a-Epoxymethyl Radicals ............ 44 Heats of Reaction for Ring-Opening of Various Substituted a-Epoxymethyl Radicals ....................................... 45 Irradiation of 9-Diazofluorene in Mixtures of Propargyl Chloride and 1,2-Dichloroethane ........................................... 76 Rate Constants for Oxygen Transfer to Fluorenylidene .......... 113 Heats of Formation of X=O and X: Species (AM 1 Calculated Values in Parentheses) ............................... 115 Selected Geometry Parameters and Activation Energies for Transition Structures for the Addition of Methylene to Hydroxyethylene ......................................... 138 Selected Geometry Parameters and Activation Energies for Transition Structures for the Addition of Chlorocarbene to Hydroxyethylene .................................. 139 Selected Geometry Parameters and Activation Energies for Transition Structures for the Addition of Methylcarbene to Hydroxyethylene .................................. 140 Selected Geometry Parameters and Activation Energies for Transition Structures for the Addition of Methyl Fluorocarbene to Hydroxyethylene ........................ 141 Selected Geometry Parameters and Activation Energies for Transition Structures for the Addition of Methylene to Cyanoethylene ............................................ 145 Selected Geometry Parameters and Activation Energies for Transition Structures for the Addition of Chlorocarbene to Cyanoethylene ...................................... 146 4.7 Selected Geometry Parameters and Activation Energies for Transition Structures for the Addition of Methylcarbene to Cyanoethylene ...................................... 147 4.8 Selected Geometry Parameters and Activation Energies for Transition Structures for the Addition of Methyl Fluorocarbene to Cyanoethylene ........................... 148 4.9 Heats of Formation and Activation Energies of Transition States of Methylcarbene Addition to Various Push-Pull Olefins ............................................................... 151 4.10 Heats of Formation and Activation Energies of Transition States of Chlorocarbene Addition to Various Push-Pull Olefins ............................................................... 152 4.11 Heats of Formation and Activation Energies of Transition States of Bromocarbene Addition to Various Push-Pull Olefins ............................................................... 153 4.12 Heats of Formation and Activation Energies of Transition States of Methyl Fluorocarbene Addition to Various Push-Pull Olefins ................................................... 154 4.13 Heats of Formation and Activation Energies of Transition States of Dichlorocarbene Addition to Various Push-Pull Olefins ................................................... 155 4.14 Selected Geometry Parameters for Transition States for the Addition of Methylcarbene to the Various Push-Pull Olefins ............................................................... 157 4.15 Selected Geometry Parameters for Transition States for the Addition of Chlorocarbene to the Various Push-Pull Olefins ............................................................... 159 4.16 Selected Geometry Parameters for Transition States for the Addition of Bromocarbene to the Various Push-Pull Olefins ............................................................... 161 4.17 4.18 4.19 Selected Geometry Parameters for Transition States for the Addition of Methyl Fluorocarbene to the Various Push-Pull Olefins ............................................................... 163 Selected Geometry Parameters for Transition States for the Addition of Dichlorocarbene to the Various Push-Pull Olef'ms ............................................................... 165 Frontier Orbital Energies (eV) for Olefins ........................... 170 1.1 1.2 1.3 1.4 1.5 1.6 1.7 2.1 2.2 2.3 2.4 HSIQEEIQURES Structures of 1.1A-3.1A calculated at the UHF/6-31G“ and (MPZI6-31G“) levels of theory ................................................ 6 The potential energy surface for C-O bond cleavage ................ 13 Structures of 1.1B-3.1B calculated at the UHF/6-31G“ level of theory ..................................................................... 21 Structures of 1.1C-l.5C calculated at the UHF/6-31G“ level of theory ..................................................................... 22 Structures of LID-LSD calculated at the UHF/6-31G“ level of theory ..................................................................... 23 Structures of l.1C'-1.5C' calculated at the UHF/6-31G* level of theory ..................................................................... 24 Structures of 1.1D'-1.5D' calculated at the UHF/6-31G" level of theory ..................................................................... 25 ESR Spectra for Irradiation of 9-Diazofluorene with Phenylacetylene and 5-Phenyl-3-spirofluorenyl-3H-pyrazole at 77 K ................................................................................ 69 ESR Spectrum for Irradiation of 9-Diazofluorene with l-Hexyne at 77 K ................................................................. 70 ESR Spectrum for Irradiation of 5-t—Butyl-3 -spirofluorenyl -3H-pyrazole at 77 K ............................................................ 71 Calculated Geometries for Pyrazoles by MNDO ...................... 80 4.1 4.2 4.3 4.4 Two possible transition states for carbene addition to a substituted olefin ................................................................ 131 The proposed preference of a singlet carbene reacting with a push-pull olefin ........................................................ 132 The interaction between a singlet carbene and an olefin at a long distance ................................................................ 168 A revised Figure 4.2 based on the computational results ......... 173 CHAPTER 1 l'E'fi' tfls‘l . 1' 1 Fri: '11-: ”.11. ‘ .11. _ -1!.n¢- -1... PART I. The Parent System PART II. Substituent Effects Abstract: The unusual regioselectivity of a-epoxymethyl radical ring-opening is studied by ab initio methods. Computational results at the PMP4/6-311G**//UMP2/6-31G* level indicate that the barrier for C-O bond cleavage is lower than that of C-C bond cleavage by about 7 kcal/mol. Interestingly, the heat of reaction for the C-C bond cleavage is ca. 3 kcal/mol more exothermic than for CO bond cleavage. This suggests that the regioselectivity of the ring-opening is kinetically controlled. Several substituted a-epoxymethyl radical ring-openings are examined at the MP4/6-31G*//UHF/6-31G* level. In general, a 1: electron-withdrawing substituent favors C-C cleavage whereas a 1r. electron-donating substituent favors C-O cleavage. Substituents on the epoxide ring show stronger substituent effects than on the (at-methyl. A further study extended to substituents with charge, CH2+, CNH+, NH3+, CH2-, and BH3- is also discussed. Part I. The Parent System 1.1 Introduction: Ring-opening of a-epoxyalkyl radicals has recently emerged as a useful tool in organic synthesis.1 In alkyl substituted cases, the epoxide ring opens at the C-0 bond, leading to an oxygen—centered radical rather than the or -oxygen-stabilized carbon-centered radical obtained by C-C cleavage. This regioselectivity is opposite to that predicted from simple bond dissociation energies.2 C-C opening can of course be favored by 1:- delocalizing groups on the 3-carbon (see numbering in Scheme 1.1).3 However, we wish to understand the fundamental origin of the unsubstituted system's seemingly anomalous regioselectivity. In light of the reaction’s synthetic value, a predictive understanding of the factors that control the choice of cleavage path could aid in the design and control of new synthetic schemes. mam-1.1 0 Ph3SnH OH . 0 Cl 70 // . 2 3 Thermochemical estimates for product radicals 1.3A and 1.5A, assuming no special role for the vinyl groups, yield a AHf difference of ~5 kcallmol in favor of vinyloxymethyl radical 1.3A.4 However, K. W. Krosley et al. have recently observed that the a-chloromethyl epoxides react with Ph3 SnH to give only allyl alcohols.5 No oxiranyhnethyl radicals were trapped at any Ph3SnH concentration, suggesting that the initial radical’s lifetime is vanishingly short. For comparison, the lifetime for rearrangement 3 of cyclopropylmethyl radical is 1.0 x 10'8 s at room temperature (AEa = 5.9 kcal/mol), and it can be trapped at high tin hydride concentrations.6 More recent work in which an intramolecular competition was set up between epoxide and cyclopropane ring cleavage showed only epoxide opening.7 A second issue of concern is the heat of reaction for the ring opening. From the above thermochemical estimates, a reaction exothermicity of only 4 kcal/mol is calculated for GO ring opening; C-C cleavage would then be 9 kcal/mol exothermic. Combined with the rapid ring opening rate indicated by the above intramolecular selectivity, these numbers suggest that radicals l.1A and 1.5A should equilibrate as in the cyclopropylmethyl system. Thus, even after opening, radical 1.5A could recyclize and access the more exothermic cleavage to 1.3A. The fact that this process is not seen hints at a substantial barrier to C-C cleavage. We now report an ab initio molecular orbital study of this unique reaction, carried out to examine energies and conformations in the parent C3H50 system. In this work, we have sought to address the following questions: 1) What controls the epoxide ring-opening regioselectivity— relative energetics of the products or activation barriers? 2) What is the overall reaction exothennicity, and are these radicals in a mobile equih‘brium as in the cyclopropylmethyl/prop—3-en-1-yl radical system? 3) What level of ab initio model is needed to descn‘be this open-shell system adequately? 4) Can our calculations suggest new modes of control for these reactions? 1.2 Methods: Ab initio calculations were performed on the C3H 50 system with the GAUSSIAN 868 and 909 series of programs. Structures were fully optimized at the UHF/3-21G, UHF/6-3lG“, MPZ/3-21G, and MPZ/6-3lG" levels. Optimizations and vibrational frequency calculations used the analytical first and second derivative methods in the GAUSSIAN packages. The effects of electron correlation were included by way of Moller-Plesset perturbation theory up to fourth order, including single, double, triple and quadruple excitations (MP4(SDTQ), frozen core) with projection corrections for spin contamination (denoted PMP4) in the unrestricted open-shell wavefunctions; these corrections were especially significant in transition structures.10 Following conventional notation, the PMP4/6- 311G**//MP2/6-31G* level represents PMP4 energies computed with the 6- 311G“ basis set, using geometries optimized with an MP2/6-3lG* wavefunction. Energies for l.1A-l.5A were computed at the PMP4/6—. 311G**l/MP2/6-31G*, PMP4/6-3lG‘l/MP2/6-31G“, PMP4/6- 3lG*"'//UHF/6-31G*, and PMP4/6-BIG‘WUHF/6-31G“ levels. The latter method was also used for all reference species. Unsealed UHF/6616* vibrational frequencies for all species were used to characterize stationary points and to calculate zero point energies and thermal energies to 298 K for all structures. Each of the optimized transition structures (1.2A or 1.4A) has just one negative eigenvalue of the Hessian Matrix. Activation energies were computed from the calculated total energy differences between the starting radical (LIA) and the transition structures (1.2A or 1.4A). Similarly, heats of reactions reflect total energy differences between the product radicals (1.3A or 1.5A) and the initial radical (1.1A) calculated at the vibrationally 5 corrected PMP4/6-311G**//MP2/6-BIG* level. A second approach to calculation of the energies of l.lA, 1.3A, and 1.5A was to estimate their heats of formation by combining the isodesmic reactions of Scheme 1.3 with the experimentally known heats of formation of the reference compounds propene oxide, methyl vinyl ether, allyl alcohol, ethyl radical, ethane, methoxymethyl radical, dimethyl ether, ethoxy radical, and ethanol. Most computations were run on the MSU Chemistry Department's VAX cluster, while the Convex or Titan computers were used for some larger jobs, and the Cray YMP at San Diego Supercomputing Center ran the MP2/6-31G* optimizations. Some preliminary optimizations were run on the SPARTAN program (Wavefunction, Inc.; Irvine, CA) on Silicon Graphics Indigos at MSU. A generous grant of Cray YMP time from Cray Research, Inc. enabled higher level single-point calculations, detailed explorations of potential energy surfaces, and preliminary substituent effect studies. W Transition AlarmRing-opening AB. Structures Products (kml/mOI) , y" (1313 C: a: .. x 0’ . 4\8~ r. -0.6 /_:3: 1.5A l.1A ° ,' .4' 1.470 1.459) t.- 16.3 (14.8) I 1.39 5‘ 1:335" ’ 1.412 (1.432 (1.454) Initial Radical l.1A ''''''''' ...... 1.372 1.673 (1.383) (1.827) 1.2A Transition Structures 1.4A cc 020,0 - -130.4° ’ (430.2) 4"!“- rrrrrrrrr ggggggggg ‘‘‘‘‘‘‘‘ l.3A Product Radicals 1.5A Structures of l.1A-l.5A calculated at the UHF/6316‘ and (MM/6316‘) levels of tl'reoriy,1 showing selected distances and torsion angles. For complete geometries, see appendrx . . 1.3 Results and Discussion: Figure 1.1 shows the structures calculated for the species in Scheme 1.2, while Chart 1.1 and Table 1.1 summarize our energetic findings. Unless otherwise indicated, relative energies for l.1A-l.5A cited in the following text denote 298 K vibrationally corrected PMP4/6- 311G**//UMP2/6-31G* energies. Heats of formation in Chart 1.1 for l.1A, 1.3A, 1 .SA and their conformers were derived from known heats of formation for reference compounds and the heats of the isodesmic reactions of Scheme 1.3 calculated at the vibrationally corrected PMP4/6- 31G*//UHF/6-31G* level. Explorations of the potential energy surface surrounding 1.4A are graphically presented in Figure 1.2. Product Structures and Energies: Product radicals 1.3A and 1.5A were optimized by slightly stretching the C-C and C-0 distances in transition structures 1.2A and 1.4A and allowing them to relax to energy minima. Neither 1.3A nor 1.5A are the global minimum energy structures for vinyloxymethyl (1.3A) or allyloxy (1.5A) radicals; they are, nonetheless, no more than 2 kcal/mol above the lowest energy conformers (see Chart 1.1) as calculated at the PMP4/6-31G*//UHF/6-3 1G* level. The corresponding hydrogen-added C3H60 isomers methyl vinyl ether and allyl alcohol also have several close-lying conformational minima. For 1.3A and methyl vinyl ether, the lowest energy form is nearly planar in an s-cis-like conformation, while the structure first accessed by C-C bond cleavage is s-trans—like. Similarly, s-cis-like conformations ('tCCCO ~ 0°) of 1.5A and allyl alcohol are lowest, but the initially accessed forms are staggered with tCCCO ~ 120°. In all cases these species are separated by barriers of less than 3 kcal/mol. 8 Ring opening by C-C bond cleavage to give 1.3A is calculated to be only 1.6 kcallmol exothermic. Radical 1.3A is nearly planar, showing slight pyramidalization at the radical carbon, similar to that found in the methoxymethyl radical. Thus, there appears to be no special relationship in 1.3A between the vinyl group and the radical center, and 1.3A may be described as an ordinary or-alkoxymethyl radical. This view is supported by the theoretical finding of nearly thermoneutral hydrogen atom transfer from dimethyl ether to 1.3A (see Chart 1.1). We could not locate experimental data on methyl vinyl ether or anisole for comparison with the 93.1 kcal/mol C-H BDE for dimethyl ether.2a Ring opening by C-O bond cleavage to give the allyloxy radical 1.5A is calculated to be 0.6 kcal/mol exothermic, placing 1.5A only 1 kcal/mol above the quite ordinary radical 1.3A. At first, this result appears to suggest that allyl alcohol should have an anomalously low O-H BDE. However, the published BDE for benzyl alcohol is the same as that for ethanol (104 kcal/mol), suggesting no special stabilization in that case.2a No compression of the OH + <3) on ,o 1.5A AHmGKml/mol)‘ PMP4 -15(-1.3)" 05(c5)° _\0 + /0\ ——> —\OH + (4) Alimkal/mol)'l PMP4 -l.9(-l.6) EXP. -11.1 a Valueshswdinpmenmeseshaveheencarectedformermalenergycmnibudona b Theseenergiesareesfimatedfimntheconformafionsobtainedbyallowingme carespmdingtansidonsmwswrdumpmdmandmecmmspondinghydmgenawd species. c Theseenagiesmesfimatedfiomthemostsnbkcmfamadonsfmndforaflspeciea H3CCH3 PMP4/6-31G" a 49.53234 Thermal energy P 0.08316 Corrected energy -79.44918 AHf (kcal/mol)° -20. l:l:0.05 We PMP4/6-31G“ a -154.55873 Thermal energy P 0.09014 Corrected energy -154.46859 AHf (kcal/mol)° -56.l:t0.1 PMP4/6316* a -192.52561 Thermal energy b 0.09614 Corrected energy -192.42947 AHr (kcal/mol)c [-22122] “\e. PMP4/6-31G* a -192.52859 Thermal energy b 0.09652 Corrected energy -192.43207 AHf (kcal/mol)¢ [-29.0:I:0.5] 10 C C I'I'J'vts '5 H3CCH2 -78.87326 0.06719 -78.80607 28.0105 _\>. ~153.90435 0.07474 -153.82961 -4.0i1.1 ()_‘° -19l.87415 0.08056 -191.79359 [19.l:l:3] -=\_6 -19l.87664 0.08150 -191.79514 [21.82117] 0 51' .‘Hllt at o :3: 1 /°\ /°\. -154.54635 0.09035 -154.45600 -44.0:I:0.l \A -192.52851 0.09650 -l92.43201 -22.6:I:0.1 //_‘i ~192.52982 0.09663 192.43319 -24:I:2 flu 192.53030 0.09666 -192.43364 -30.0:I:0.5 11 These values are calculatedatthe UHF/6316* geometries. b Thermal energies are sum of zero-point energy and thermal energy contributions at 298 K computed at the UHF/6316* level, frequency corrections were not applied as they makeverylittledifference(<0.2kcallmol)tothecomparisonsmadehere. -153.89497 0.07529 -153.81968 -3.0:l:1.0 \A -191.87114 0.08060 -191.79054 [24.4:I:0.6] Fl. -191.87558 0.08113 -l91.79445 [18513] We -l9l.87668 0.08125 -l9l.79543 [21611.7] cThevaluesin square bracketsareestimatedfmmtheisodemicreactions(l)—(3)in Scheme 1.3 and experimental heats of formation. 11 Since equations (1)-(3) are isodesmic, isogyric reactions, the errors therein should cancel. Thus, we can combine these isodesmic reaction energies with experimental heats of formation of reference compounds to estimate heats of formation for radicals l.1A, 1.3A, 1.5 A, and their conformers. The resulting energies, summarized in Chart 1.1, are close to those expected based on our original naive analogies.4 In other words, the vinyl groups in vinyloxymethyl radical and allyloxy radical do not significantly stabilize the radicals; in fact 1.3A appears to be slightly destabilized by the vinyl group. In equation (4), a nonisodesmic comparison of the BDEs for C-H of dimethyl ether and O-H of ethanol, there is a significant difference (> 9 kcal/mol) between computational and experimental estimations. This large discrepancy can be traced to two substantial errors: First, Hehre et al.12 have shown that ab initio calculations underestimate heteroatom-hydrogen BDEs (i.e. O—H bond) by 5-10 kcal/mol even at a high level (MP4/6-31G**//UHF/6-31G*); this difficulty remains severe even when spin-projected wavefunctions are used. Second, the MP4/6-31G*//HF/6-31G* energy difference between the closed-shell C2H50 isomers dimethyl ether and ethanol is 7.9 kcal/mol (12.1 expt.) whereas the PMP4/6-31G*//UHF/6-31G* energy difference between CszO methoxymethyl and ethoxy radicals is 6.2 kcal/mol (1.0 expt.). These two errors add to give a 9.4 kcal/mol error for the reaction of equation (4). Transition Structures and Activation Barriers: Transition structures 1.2A and 1.4A are shown in Figure 1.1. Our best calculated barrier for breaking the C-0 bond of the epoxide (4.8 kcal/mol) is lower by 6.7 kcal/mol than for breaking the C-C bond (11.5 kcal/mol). In the UHF/6- 31G* optimized structures, the shortening of the nascent double bond may 12 be used as a measure of reaction progress at the TS. At this level the vinyl groups in 1.311 and 1.5a are similar in length (1.317 and 1.320 A, respectively), so the ending points are comparable. By this analysis the TSs for C-C (12A) and C-0 (1.4A) bond cleavages are “late” and “early” (~57 % and ~39%) respectively. In both cases, the erstwhile radical center has rotated to optimize n-overlap with the methine carbon, while the breaking bonds have stretched by 41 and 26% of their final extension. For TS 1.2A, the 3-carbon has not begun to rotate to allow rt-stabilization by the neighboring oxygen; thus 1.2A does not yet “feel” the radical stabilizing effect of a-oxygen. As a probe of the effects of electron correlation on geometry, structures l.1A-l.5A were also optimized at the MP2/3-21G and MP216- 31G* levels; MP4(SDTQ)/6-311G** and MP4(SDTQ)/6-31G* single point calculations were then run on the MPZ/6-31G“ geometries. For the radical minima l.1A, 1.3A, and 1.5A, the geometry differences are minor between the UHF/3-21G, UHF/6—3lG“, MP2/3-21G, and MP2/6-31G“ levels (see Figure 1.1 and Appendix 1.1). However, both TSs 1.2A and 1.4A show significant geometrical changes on reoptirnization at the MP2 levels (see Figure 1.1). Specifically, the breaking bonds are more extended than in the UHF/6-31G“ TS structures, suggesting that the latter may not be adequate representations of the energy maxima for this reaction. Though the “late” vs. “early” timing of the TSs is retained, at the MP2/6-31G" level, both 1.2A and 1.4A are later (83% and 71%, respectively as defined previously), and their geometries are much more similar in reaction progress. The higher barriers obtained at the MP2/6-31G“ geometries suggest that these structures more accurately approximate the MP4 maxima Energy (H) -191. 13 .27 - .28 d I HF . o PUHF .29 . .30 ~ .31 .80 'i 1 a MP2 -191 -191 491 -191 -191 -191 -191 -191 -191 -191 .82 o I MP3 A MP4 .34 - O PMPZ D PMP3 A PMP4 .86 d .88 r l ‘ l ' t ' I f 1 ' 1.4 1.6 1.8 2.0 2.2 2.4 2.6 Distance rCO (A) The potential energy surface for GO bond cleavage generated by stepping the 00 length with reoptimizau'on of all other geometry variables at the MP2/6810" level. Energies for the various levels are all in harnees. 14 than do the UHF structures. To explore this interpretation, the potential energy surface for GO bond cleavage was examined in the vicinity of the MP2/6-31G* transition structure by stepping the 00 length with reoptimization of all other geometry variables at the MP2/6-31G“ level. Energy calculations at the MP4(ST'DQ) level were then run for each point. Results for UHF, MP2, MP3, MP4, and the corresponding spin polarization corrected PUHF, PMPZ, PMP3, and PMP4/6316* calculations are shown in Figure 1.2. Three noteworthy items were uncovered: First, the PMP4 maximum is broad and flat and coincides with the optimized MP2/6-31G"I maximum, so that the PMP4 energy barrier calculated at the MP2 geometry should be very close to the best value. Second, the HF energies calculated for MP2/6-31G* geometries peak close to the UHF/6-31G“ optimized maximum; thus partially optimized "step" structures are not significantly distorted by the unbalanced treatment of geometrical variables (i.e. fixing one and optimizing all others). Third, spin projection hugely affects energies and calculated TS positions along the C-0 distance; this last finding suggests that optimization with the spin-projected wavefunctions might be needed in some systems. We did not similarly analyze the potential energy along the C-C distance about 1.2A. The geometry and the AEa value for 1.2A change much less than for 1.4A between UHF/6-31G“ to MPZ/6-31G“ structures (see Figure 1.1 and Table 1.1). Also, the MP2/631G“ maximum had proven to be the proper choice for the higher level calculations on 1.4A. The calculated activation barriers, which clearly favor C-O over C-C cleavage, readily explain the experimental selectivities. The low barrier and near thermoneutrality of CO bond opening suggest that fast equilibrium of closed and open forms (l.1A and 1.5A) would be expected, while access 15 to 1.3A would be much slower. Thus, since all reports to date are based on product studies, the apparent selectivity for carbon vs. oxygen radical trapping may be affected by an imbalance in relative rates of hydrogen atom transfer to l.1A vs. 1.5A. Nonetheless, C-C bond opening, when assisted by appropriate delocalizing 3-substitution, can lead to trappable carbon- centered radicals. 1.4 Conclusion: Our calculations have established that (1) The apparent stability of the allyloxy radical 1.5A compared with vinyloxymethyl radical 1.3A results from over-estimation by the absolute energy calculations; (2) The vinyloxymethyl radical 1.3A, also a homoallylic species, gains no special stability other than that afforded by or-oxy substitution; (3) The C-O ring- opening of cr-epoxyalkyl radicals is a facile process whereas C-C cleavage is much more difficult; (4) The regioselectivity of a-epoxymethyl radical cleavage is controlled by relative rates for C-O vs. C-C bond breaking instead of the heats of reaction; (5) Spin contamination and electron correlation strongly affect the structures of TSs 1.2A and 1.4A, and the overall relative energetics in this system. We believe it may also be possible to control the C-C vs. CO bond cleavage choice via substituents on the initial radical center, an unexpected locus of regiocontrol. 16 =o' n.==nc{'=1" I'w‘e‘i = :‘a II! as .- 15A Saufing Thanfifimn anhun Radical Structures Radicals Basis set 1.111“ 1.2Ab 1.4141b 1.3ab 1.511b UHFB-ZIG 0.19959 102 02 45.7 -26.1 MP2/3—216 4157739 16.9 9.8 UHF/6310* 427916 15.4 6.3 -3.5 44.4 UHF/6-3lG"//UHF/6—3IG* 428777 152 6.3 -3.9 442 UI-lF/6-3llG“/IUMP216-3IG* 4.32562 13.6 15 -5.9 46.6 UMP2/6-3ro*//UHF/6-310* 4.81752 225 112 0.5 2.7 UMP2/6-3IG**/IUHF/6-3IG* 1.85757 22.6 112 0.1 3.0 UMP2/6-3lG*/IUMP216-31G‘ 4.81965 23.3 17.6 0.7 3.6 UMPQJ6-3IIG**/IUMP2/6-3IG* 4.93386 21.3 17.3 0.9 2.7 UMP3/6—3lG‘l/UHF/6-3lG‘ 4.84277 20.8 95 -0.6 -3.9 UMP3/6-3IG**//UHF/6-31G* 4.88548 20.9 9.5 4.0 -3.6 UWSI6-3IGV/UMP2/6-3IG“ 4.84412 21.3 12.6 -0.6 -3.9 UMP3/6—3llG“//UW2I6-3IG* 4.95940 19.1 12.1 -2.3 4.6 UMP4l6-3IG*INTIF/63IG* 4.86950 19.7 7.7 4.5 -2.8 UMP4l6-3IG**/IUHF/6-3lG* 4.91210 19.8 7.7 4.8 -24 UMP4/6-3lG"//UMP2/6-BIG* 4.87229 20.7 12.3 4.0 4.6 UMP4/6-3llG**//UMP2/6-3IG* 4.99323 19.0 12.6 -2.1 4.1 PUIIF/6—3lG‘l/UHF/6-3IG* 428306 62 4.9 -3.6 44.8 PUHF/6-3IG“//UHFI6-3IG* 429116 5.8 -22 .4.4 45.0 PUHFI6-3lG‘l/UMP216-3IG* 428112 6.7 -7.8 -2.9 45.3 PUHF/6-3llG“/NMP2l6-BIG* 4.32989 5.6 -8.2 4.9 46.3 PWGSIG‘IIUHF/6-3IG* 4.82017 14.3 4.0 0.1 2.1 PW6-3lG“/IUHF/6-3IG* 4.86017 14.4 4.0 -0.3 2.2 PMP216-3IG‘INMP2/6-BIG“ 4.82257 15.9 8.7 1.3 3.5 PNIP2J6-3l lG**/NMP216-3IG‘ 4.93679 14.1 8.5 -0.2 2.7 PWI6-3IGV/UHF/6-3lG’ 4.84441 14.4 4.3 4.0 .45 pMP3/6316**//Um=/6.31G* 4.88709 145 4.3 2.3 -4.4 PMP3/6-3IG*/NMP2l6-3IG‘ 4.84595 15.4 5.7 -0.3 -3.7 PMP3/6-3llG“//UMP2I6-3IG‘ 4.96122 135 5.3 4.9 -4.8 massroummmos 4.87114 13.3 25 4.9 -3.4 PMP4/63IG“IIUHF/6-3IG* 4.91370 13.7 25 -2.2 -32 l7 PMP4/6-31G*//UMP2/6-31G* C -1.87412 14.9(13.0) 5.4(4.5) -0.7(-0.7) -l.8(-1.2) PMP4/6-311G**//UMP2/6-31G* -1.99506 13.4(11.5) 5.7(4.8) -1.6(-l.6) -1.2(-0.6) Zero-PointEnergyd 0.07603 -1.6 -O.6 03 0.7 zrhermnl Energies d 0.08060 4.9 -0.9 0.0 0.6 Dipole (UHF/6-31G“ geometries) 2.28 D 1.72 D 2.35 D 1.39 D 1.96 D Dipole (MP2/6310* geometries) 2.47 D 1.84 D 2.10 D 1.45 D 1.97 D a Total energies for radical l.1A are given in Harnees (H) relative to -190.0 H (l Hartree = 627.5 kcal/mol). Zero-point energies are directly reported in Hartrees, and dipole values are in debye. b Energies for l.2A-l.SA are given in kcal/mol, relative to the energy of l.1A at the corresponding level. 9 Values listed in parentheses have been corrected for thermal energy contributions and represent best estimates for the experimental activation energies/reaction exothermlcrties. d Zero-point and thermal energies were computed for 298 K at the UHF/6—31G*//UHF/6— 31G* level and were used without scaling. 18 Part II: Substituent Effects 1.5 Introduction: Ring-opening of a-epoxyalkyl radicals has been widely used in organic synthesis. The epoxide ring is easy to build, and this methodology can be used to generate cis-fused bicyclic systems through a highly regioselective C-O bond cleavage. Consistent with the experimental observations of Murphy et al., we have shown, by ab initio calculations, that the high regioselectivity of C-0 bond cleavage of the epoxide is kinetically favored, even though CO bond cleavage is more exothermic than CO bond cleavage. However, Murphy et 81.13 and Stogryn and Gianni” have reported that substitution of an aryl group or simple vinyl group on the epoxide ring leads exclusively to products of C-C bond cleavage (Scheme 1.4). Additionally, Murphy and co-workers showed that C-C bond cleavage becomes competitive with C-O bond cleavage when a carbonyl group is placed on the epoxide ring, as shown in Scheme 1.5. Schema Br 0 O\/ Bu38n- I f V _..._, ”Le-J 19 O r-ani r-Bu Br BUgSll’ t-Bu O O OMJkt-Bu + t-Bqu/ONY 37 % 22 % 0 If our prediction by ab initio calculation is correct, CO bond cleavage of the parent epoxide is kinetically favored over C-C bond cleavage by ~ 6-7 kcallmol. Thus, the activation energy difference between C-C and C-0 bond cleavage can be decreased by a similar amount by putting proper substituents on the epoxide ring, as shown above. Here, we attempt to answer the following questions using ab initio calculations: ( 1) How does such a simple electronic effect reverse the expected regioselectivity? (2) How do substituents on the a-methyl position influence the regioselectivities of the epoxide-ring opening? (3) Can we separate a and 7t substituent effects? It is well-known that radicals are easily polarized.” It seems reasonable, then, that the transition states for C-C and C-0 cleavage of the epoxide should experience different degrees of polarization. Specifically, in the transition state for C-O bond cleavage (TS A), polarization may result in accumulation of negative charge on oxygen. In contrast, the transition state for C-C bond cleavage (TS B) would prefer positive charge build-up on C3, 8. .0 Y 3”, \/ Y (xcs/ C4—C3 l . I: 0 (40 5+ C' x/ 1 X/S: TSA 6- TSB which could be stabilized by the neighboring oxygen. If these models are correct, then electron-withdrawing groups on C1 and electron-donating groups on C3 should prefer TS B; electron-donating substituents on C1 should favor TS A, and those on C3 should not significantly affect TS A. 1.6 Results: Geometries: Three different substituents on either a-methyl or epoxide ring of a—epoxymethyl radical have been calculated, as shown in below. All structures were fully optimized at the UHF level using a 6-31G“ f” l 2 X A: X=H;Y=H B X=NH2; =H, B': X=H;Y=NH2 C X=CN;Y=H, C'. X=H;Y=CN D: X=BH2;Y=H, D': X=H;Y=BH2 21 1.45 ‘3 H .428. 1. Initial Radical 1.1B H H H 1.827 .46 1: = -21.9 - -> f' .. V\ H 1.378 1.378 .671 1.374 1.2B Transition Structures 1.48 H 1.390 H accoc = 426.2 N rccco = 122.3 1.317 H 2.31 H 1.323 1.498 H . 1.35 H 1.370 H 2.407 1.390 1.38 Product Radicals 1.5B W Structures of 1.13-1.51! calculated at the UHF/6-31G* level of theory, showing selected distances and torsion angles. 1.3C Product Radicals 1.5C W. Strucun'es of l.1C-1.5C calculated at the UHF/6316* level of theory, showing selected distances and torsion angles. 23 1.21) Transition Structures ’ 1.41) ICCOC = 178.5 . Structures of LID-LSD calculated at the UHF/6316* level of theory, showing selected distances and torsion angles. t “ 1.468 : '47 3': 18.3 1.460 a ’ 1.39 1. 1.135 Initial Radical l.1C' H H " l 1. 47 1.410 x 1.47 .15 ' '3-9 1.3 H I 1. 1.4C' 1.2C' Transition Structures 1.5C' 1.3C' Product Radicals W Structures of l.1C'-l.5C' calculated at the UHF/6316‘ level of theory, showing selected distances and torsion angles. 1.2D' Transition Structures 1.4D' 1.3D' Product Radicals 1.5D' Structures of l.1D'-l.SD' calculated at the UHF/6316" level of theory, showing selected distances and torsion angles. 26 basis set. Geometries of the initial radicals (1.1), transition states (1.2 for C- C and 1.3 for CO bond breaking), and final radicals (1.4 for C-C and 1.5 for C-O bond cleavage) incorporating different substituents, are summarized in Tables 1.2-1.6 and Figures 1.3-1.7. The starting geometries (1.1) of the or- epoxymethyl radicals basically are very similar, as shown in Table 1.2. Bond distances of futtue breaking-bonds, C2C3 or C20, are ca. 1.5 A and 1.4 A, respectively. The transition states for the C-C bond cleavage of the substituted a-epoxymethyl radicals are also similar to the transition state of the parent radical, as shown in Table 1.3. For all substituents at the X position, the rC2C3 for the transition states of C-C bond cleavage are about 1.8 A; they become just a little shorter (~ 1.76A) when Y = CN or BH2, the electron-acceptors. There is not much geometry variation among the transition states for GO bond cleavage either, as shown in Table 1.4. For the X or Y substituted radicals, the rCzO values are ca. 1.68 A, which is a little longer than that of the parent system (X = Y = H; 1.64 A). Unfortunately, we were unable to locate the initial radical or transition states for C-C and C-0 bond cleavages for a—epoxymethyl radical with an NHz substituent on the Y position; the initial radical fell monotonically down hill to generate the C-0 bond cleavage product radical. Repeated attempts at locating a minimum for the amino substituted radical failed, suggesting that with such a strong electron donor, C-O cleavage occurs without a barrier. The broken bond (rC2C3) distances for the product radicals obtained from c-c bond cleavage are all about 2.3 A (Table 1.5). The broken bond (rC20) distances for the CO bond cleavage product radical are around 2.4 A (Table 1.6) except that the distance is a little shorter for the 27 NHz-substituted epoxide ring (1.5B) (C-O cleavage product when Y = NH2). Generally speaking, comparing the corresponding radicals, the geometries of substituted a-epoxymethyl radical ring-opening transition structures are not changed much by changes in the nature—7t electron- donors or acceptors—or the position—X or Y—of the substituents. Even though the distances of the breaking bond at the transition states are slightly sensitive to substituents, the changes are trivial (< 0.1 A). Energies: Total energies were calculated at the MP4(SDTQ)/6-3lG*//UHF/6—3 16* level including projection correction for spin contamination (PMP4) for all substituted systems. The results are summarized in Tables 1.7-1.11 and heats of reaction and activation energies calculated at the PMP4/6-3IG"‘//UHF/6-31G* level are listed in Table 1.12. Activation Energies: For substituents NH2, CN, and BHz on the a-methyl position (X position) the activation energies of C-C bond cleavage increase by 2.2, 3.1, and 0.7 kcal/rnol, respectively, with respect to the 13.3 chmol activation barrier in the parent system (Table 1.12). However, these substituents have different effects on the energy baniers for C-O bond cleavage. With electron-donating groups like NH2, the energy barrier for GO bond breaking is just 0.8 kcal/mol, which is 1.7 kcal/mol less than for the parent system. However, electron-withdrawing groups CN and BH2 increase the barrier by ~ 3 kcal/mol with respect to the parent system. For substituents on the epoxide ring (Y position), electron- withdrawing groups dramatically decrease the activation energies for CC bond cleavage. With a CN group at the Y position, the activation energy for 28 C-C bond cleavage is 6.6 kcal/mol, 6.7 kcal/mol lower than that for the parent system, and the barrier for GO bond cleavage (3.1 kcal/mol) is essentially the same as the parent system. The activation energy difference between C-C and CO bond cleavage is then narrowed from 10.8 kcal/mol for parent system, to just 3.5 kcallmol for the a-cyanoepoxymethyl radical. In the case of BH2, an even stronger rt electron-withdrawing group (but sigmatropic donor), the activation energy for C-C bond cleavage is decreased to 2.3 kcal/mol, while the barrier to C-0 bond cleavage is increased. At this level of calculation, a BH2 group at the Y position of the or-epoxymethyl radical changes the selectivity of the ring opening to favor C—C bond cleavage over CO bond cleavage by 4.1 kcallmol. Energies of Reaction: The energies of reactions shown in Table 1.12 clearly show that the stronger the electron-withdrawing substituent at Y, the more C-C bond cleavage in favored. For substituents on the tit-methyl position (X), the energy differences between C-C bond cleavage (vinyloxymethyl radicals) and C-0 bond cleavage (allyloxy radicals) are 6.6, 1.5, -0.3, and -4.3 kcal/mol for NHz, H, CN, and BH2, respectively. Compared to substituents on C1 (X position), the same substituents at the Y position have stronger effects favoring C-C bond cleavage over C-O bond cleavage. When Y is CN or BH2, the C-C bond cleavages (1.30 and 1.3D') are calculated to be about 10 and 15 kcal/mol, respectively, more exothennic than CO bond cleavage (1.50 and 150'). Other Substituents: In order to broaden our understanding of the substituent effects, we pushed the edge of the electron-donor and 29 acceptor to charged groups with and without x-conjugation abilities at the UHF/3-21G*//UI-IF/3-21G* level, as shown in Table 1.13. For substituents at the X position, generally, the stronger the electron-acceptor, the more favored is C-C bond cleavage, relative to the parent system. Different from other substituents which increase the exothermicity for the C-C bond cleavage and slightly decrease the exothermicity for the C-0 bond cleavage, NH3+ dramatically increases the exothermicity for both cleavages, in which it turns out that C-O bond cleavage is favored over C-C bond cleavage by 11.3 kcallmol. Interestingly, the CN group has exactly the opposite effects for both bond cleavages; however, the results are the same as for NH3+; C-O bond cleavage is favored over C-C bond cleavage by 11.3 kcallmol. For electron-donors, compared to the parent system the heats of reaction for C-C bond cleavage are essentially the same, while the exothermicity for the CO bond cleavage is slightly increased (~ 2 kcallmol). For the substituents at the Y position, the trend is similar to those described above. The electron-acceptors favor C-C bond cleavage over C-O bond cleavage; however, the effects are stronger than for substituents at the X position: Except for NH3+, all electron-acceptors increase the exothermicities for the C-C bond cleavage by more than 10 kcallmol relative to that for parent system; meanwhile, the exothermicities for the C-0 bond cleavage are basically the same as for the parent system. For NH3+, again, the epoxide ring-opening still favors C-O bond cleavage over C-C bond cleavage by 3.0 kcallmol. For both NH3+ and CH2+, we cannot locate a minimum for the initial substituted radicals, which directly 30 fell down to bond cleavage products; C-O cleavage for NH3+ and C-C cleavage for CH2+. For electron-donors at the Y position, NH3+, CH3, and BH3" are the same as for the parent system, favoring C-O bond cleavage. But it is a different story for CHz", which shows the opposite trend compared with the other substituents. Again, unfortunately, we could not find a minimum for the initial CH2" substituted radical which directly fell down to the CO bond cleavage product radical. 1.7 Discussion: Basically, placing substituents either on the methyl (X position) or directly on the epoxide ring (Y position) does not change the geometries of the initial radicals, transition structures, or final radicals too much from the parent system (X = Y = H), regardless of the group's nature. This means that the substituents do not significantly change the relative location of the transition states along the reaction coordinates. Therefore, changes in barrier heights are hard to interpret in terms of early or late transition states using the Hammond postulate. Nevertheless, relative barriers and relative reaction exothermicities do tend vary together. This finding is not surprising since the more exothermic process already is seen to have the higher energy barrier (i.e. non-Hammond postulate behavior). Substituent effects for regioselectivity of the epoxide ring-opening just, then, result from the substituents having different effects on stability of the starting point (initial radical), mid-point (transition states), and fun] point (final radical) along the reaction coordinate, not the shapes of the molecule. 31 For substituents on the X position, the electron-withdrawing groups, CN and BH2, increase energy barriers for both C-C and CO bond cleavage; the electron-donating group, NH2, decreases the activation energy for C-O bond cleavage but increases the energy for C-C bond cleavage. This is different from what we postulated in the introduction — that the stronger the electron-withdrawing group, the more favorable C-C bond cleavage would become, even though relative activation energies (AAEa) still points in the predicted direction. Obviously, these substituents, especially the electron-withdrawing groups, can also stabilize the initial radicals. During bond breaking, either GO or C-C bond, the unpaired electron's locus moves away from the initial radical center, thereby losing the X-substituent‘s stabilization and increasing the barrier. For the same reason, the heats of reaction become more endothermic compared to the parent system, since the final radical center has completely lost the stabilization from the substituents. Nevertheless, the reactions still should favor C-C bond cleavage if only the exothermicities of the reactions are considered. Electron-acceptors at the X position in the vinyloxymethyl radical, as shown below, should stabilize the double bond between C1 and C2 via push-pull conjugation between the electron-acceptor and the electron-donor (alkoxy group). Y I C1 0 Y C vinyloxymethyl radical allyloxy radical Because of the compensation for losing the stabilization of the initial radical by the neighboring substituents, substituents on the X position 32 do not have much effect on the bond cleavage selectivity of a-epoxyrnethyl radical ring-opening. For substituents on the Y position, the same radical stabilization concept applies as well for the epoxide ring-opening. For C-C bond cleavage, the further bond-breaking proceeds, the more radical electron density shifts to C3, and the more stabilization the radical "feels". As shown in Table 1.12, for C-C bond cleavage, the activation energies are substantially decreased and the heats of reaction are also about 10 kcallmole more exothermic than those of C-0 bond cleavage. As shown in Table 1.13, the a: electron donors or acceptors are the major players here. The ionic species NH3+ and BH3', which do not have unshared p—orbitals, do not affect much about the preference and both substituents either X or Y substitutents are still in favor of C-0 over C-C bond cleavage of the a-epoxymethyl radical ring-opening. Even though the UHF/3-21G/IUHF/3-21G level of description may not be good enough, we still can observe the trend shown in AAEnj in Table 1.13: For either at X or Y positions, the 7: electron-acceptor substituted epoxide rings favors the C-C bond cleavage and a: electron-donor substituted epoxide rings favors the CO bond cleavage. As we discussed previously, though the regioselectivity of a-epoxymethyl radical ring-opening is determined by the preference between the transition states of C-0 and C-C bond cleavage, the heats of reaction are do respond to substituent effects. 33 1.8 Conclusion: We have clearly demonstrated that putting a substituent on the a-epoxyalkyl radical can change the ring-opening pathway of the epoxide. Electron-withdrawing substituents at C1 (X position) stabilize the initial radical (1.1); however, the net effect is not as strong as for substituents at C3 (Y position) which can form captodatively stabilized radicals after the epoxide ring—opening. Stabilization of the radical is mainly a result of 7c- conjugation between the 7t-electron-acceptor and the oxygen of the epoxide or its opened form. Polarization through the o-bond has little or no effect on the a-epoxyalkyl radical ring-opening. 1.9 References and Notes: 1 (a) Carlson, R. o; Huber, J. H. A.; Henton, D. E. J. Chem. Soc. Chem. Commun. 1973, 223. (b) Barton, D. H. R.; Motherwell, R. S. H.: Motherwell, W. B. J. Chem. Soc., Perkin Trans. I 1981, 2363. (0) Johns, A.; Murphy, J. A. Tetrahedron Lett. 1988, 29, 837. (d) Johns, A.; Murphy, J. A.; Patterson, C. W.; Wooster, N. F. J. Chem. Soc. Chem. Commun 1988, 294. (e) Rawal, V. H.: Newton, R. C.; Krishnamurthy, V. J. Org. Chem. 1990, 55, 5181. 2 (a) Lias, s. G.; Bartrness, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.; Mallard, W. 6.; “Gas Phase Ion and Neutral Thermochemistry” J. Phys. Chem. Rd. Data 1988, I 7, Suppl. 1. (b) Heusler, K.; Kalvoda, J. Angew., Chem. Int. Ed. Engl. 1964, 3, 525. (c) Brun, P.; Waegell, B. In 34 Reactive Intermediates; Abramovich, R. A., Ed.; Plenum Press: New York, 1983; vol. 3, p 367. (a) Strogryn, E. L.; Gianni, M. H.: Tetrahedron Lett. 1970, 34, 3025. (b) Cook, M.; Hares, 0.; Johns, A.; Murphy, J. A.; Patterson, C. W. J. Chem. Soc., Chem. Commun. 1986, 1419. We naively estimate heats of formation of 17 and 22 kcallmol for vinyloxymethyl radical 3.3A and allyloxy radical 3.5A by combining the AHf valuesZa for vinyl methyl ether (-24 kcallmol) and allyl alcohol (-30 kcallmol) with the appropriate BDEs (93.1 and 104.2 kcallmol, respectively) of dimethyl ether (OH) and ethanol (O-H) and subtracting the AHf of a hydrogen atom (52.1 kcallmol). Krosley, K. W.; Gleicher, G. J.; Clapp, G. B. J. Org. Chem. 1992, 57, 840. These authors note the possibility of epoxide opening concerted with stannyl radical attack on the B-chlorine on the grounds of the the enhanced reactivity of chloromethyl oxirane relative to chlorocyclohexane. However, various a—oxiranylalkyl radical systems have been generated via routes in which tin does not directly participate in the radical-forming step; see Sabatini, E. C.; Glitter, R. J. J. Org. Chem. 1963, 28, 3437. (See also refs. 1, 3, especially la, 38). Newcomb, M.; Glenn, A. G. J. Am. Chem. Soc. 1989, III, 275. Krosley, K. W.; Gleicher, G. J. J. phys Org. Chem. 1993, 6, 228-232. Frisch, M. J.; Binkley, J. S.; Schlegel, H. B.; Ragavachari; K.; Melius, C. F.; Martin, R. L.; Stewart, J. J. P.; Bobrowicz, F. W.; Rohlf'mg, C. M.; 10 11 12 13 14 15 35 Kahn, L. R.; Defrees, D. J .; Seeger, R.; Whiteside, R. A.; Fox, D. J .; Fleuder, E. M.; Pople, J. A. Gaussian 86; Carnegie-Mellon Quantum Chemistry Publishing Unit: Pittsburgh, PA, 1984; Revision C. Frisch, M. J.; Head-Gordon, M; Trucks, G. W; Foresman, J. B.; Schlegel, H. B.; Ragavachari, K.; Robb, M. A.; Binkley, J. S.; Gonzalez, C.; Defrees, D. J .; Fox, D. J .; Whiteside, R. A.; Seeger, R.; Melius, C. E; Baker, J.; Martin, R. L.; Kahn, L. R.; Stewart, J. J. P.; Topiol, S.; Pople, J. A. Gaussian Inc.; Pittsburgh, PA, 1990. Schlegel, H. 13., J. Chem. Phys. 1986, 84(8), 4530. Oxygen charges were -0.338 for 1.5A and -0.348 for ethoxy radical at the UHF/6-3lG“ level. Hehre, W. J.; Radom, L.; Schleyer, P. v. R.; Pople, J. A. Ab Initio Molecular Orbital Theory; John Wiley & Sons: New York, 1986, pp 272-279. Cook, M.; Hares, 0.; Johns, A.; Murphy, J. A.; Patterson, C. W. J. Chem. Soc, Chem. Commun., 1986, 1419-1420. Stogryn, E. L.; Giarmi, M. H. Tetrahedron Lett. 1970, 34, 3025-3028. Fleming, 1. Frontier Orbital: and Organic Chemical Reactions; Wiley: London, 1976, Chap. 5 36 -1 -.r. .4 a or. n. 42-..... .r «.8: 322- 82 $2 82 82 £2 2 2.2 22. :2 $2 $2 a2 20 2 I I I I I I £2 2 2.2 25:. 2.2 82 :2 222 2 2 22 38:- 22 £2 82 222 2 £2 28 222- :2 :2 82 222 2 20 ES 28:- £2 £2 82 £2 2 £2 2.2 222. :2 82 :2 $2 2 2 088 80% 0182 08. 68. £6. a x . . I.-. 23:“... . all . .._~ 32.2.... -.r. r 2: one- .22 $2 82 52 £2 2 2: 2.2- 32 82 :2 22 2o 2 I I I I I I £2 2 2: 32- 82 :2 $2 22 2 2 2: 38- 32 22 $2 $2 2 £2 2: 22- 82 82 S2 32 2 2e 22 4.8:- 82 32 $2 $2 2 £2 2: 22- 32 22 $2 22 2 2 089 80% one. 08. 86. 862 > x 37 :2- ...r.“ 1— c .r... u. .51}. 4 .l. 22 o... 22 82 22 22 £2 2 2.92 2 22 s2 22 22 20 2 22 2.9; 222 E2 22 82 £2 2 2: 2- 22 22 R2 :2 2 2 2: 2- $2 22 22 22 2 £2 2: 2- 82 22 22 22 2 20 22 2E- 22 2.2 22 :2 2 £2 2: 2- R2 22 R2 :2 2 2 cocoa 809 one. emu. 88. 86. > x :. . . ...t a .r. :L e . .4 a 2.... .2: .r. .r. r 2: 2a- .22 82 B2 82 £2 2 2: 2a- 32 322 2.2 22 zo 2 I I I I I I £2 2 2.: 2a- 22 2.82 82 22 2 2 3: 2a- 22 32 2.2 .22 2 £2 3.: 2a- 2.2 2%.: 22 £2 2 2e 2: 2a- :2 :2 s2 22 2 £2 3: 2a- 22 $2 $2 22 2 2 cocoa cocoa one. 08. «Mme. 66. a x 38 wéom mom; coed an." «an; «In «ans awn; thfi «34 fin.“ ZU @ch EMA mafia 8n; wan; N ham? 5m; wood New; can; cdwT Nwmé ontu wan; . 0mm; wfimT mam; nmmfi 8mg 2mm." mdfi can.“ SYN nae; mam; 5.23. gm.“ Btu «cm; can.“ 11:33:51: >EIEEEE 0009. One Cave mono.— «0202 .r n] ‘ r . . . .I/ r4 4.. 3.. r ..J.. 1.. .i ._—~.i.n.:.. .. 39 N336 #886 235938. #335242 588.55- onnmméva- 3559522 803.53- 958.542 2253- mm.“ 8826 6886 85854.n- emnmo.5¢a- «$48.53- macadam- 858.544? 8.48.5242 88952.n- 35003.n- EMA 23.3.. as... @886 vmvaod nmm5o.5&- «88.522? «38.53- NNQ medu- 598.52%- 558.542 35363- 353.32 5.4 §£§2023>== a... 8388 18.2.5 338:» Ewan-Pan {UHm-Qfimzm §O~m6\mm2m .mem-cgm avg-SEDA— «Cnm-QVAE ‘68-ng .036; flux-QED 43 $366 n5wwo6 onwmgufifia- 639153- NXNXSK- 33.363- 353.58- 323.53- 532.53- 53368- @9666 8366 535mm.5§- vmomcm.5_- 332.58- 89.368- 333.53- 5868.53- 333.53- 33363- $366 65366 9.603.53- SEEKS- $562.53- “Randa- 363153- n H58 153- awn-.315” 633003- .GQA cm 366 8366 838.53- 3562.58- $52.53- 5£omn6§- 833.58- 382.53- 688153- 368003- .9".— 83... 818a; _.44...w @6366 3586 358.53- 335153- 633153- 23.963- «82553- $35153- Sue-15g- «V8963- Gué 18%.... as axzuaaaanhwazzy>=2‘a BEBE 18.85. 83.88 «Eon-0.58 5986\vm2m *U—m-ngm {Ogm-egm .02“..ng Lug-33:2 am: 9ng in: 96% «03-6333 yawn-eE—fig gum swam .V....‘...u-..r. r. .,fi...<.;._u .,...rt...u .,n.....u.6.:nn..rn..u BEES— E 0.8 mead; :< .w *Dfim-©\§\*0~m6\wm2m * wéH- N4»- 0.3.. m4». 96 ON $5 I 2:- ad- GA 7 Wm. fim 06 20 3 m4». wd WT 0.x in 0.3 m 9.5 md- v6 3» We fin v.2 I 20 0.0 in. NM 93 wd 5.2 m ~32 mg in- m4- wdfi Wm mg I : Sag 00m Uum «mad 00m 00m .r N #5293 mfiomvam 33m *Aamdv 33% €33“; ficoaumnam a . Kr _ x\ 45 15.10;: 1' C ”:1! ! Y Substituents Final Radicals (any x Y 13cc BCO AAEm. AAEmlT H H -15.7 -26.1 10.4 (0.0) CN H -95 -20.8 11.3 0.9 3H2 H -l6.4 -24.1 7.7 -2.7 CNHGB H -223 -255 3.2 -7.2 CH2$ H — — .1715;t -27.6 NH3$ H -21.9 -332 11.3 0.9 NHz H — — — — CH3 H -152 .29.0 13.8 3.4 311:9 H .15.0 -28.3 13.3 2.9 H CN -24.4 -26.2 -l.8 -12.2 H BH2 -292 -27.2 .20 -124 H CNHe -35.4 -26.2 -92 -l9.6 H CH2$ — — .1341: -23.8 H NH3€B — — 13.41 3.0 H N112 — — 13.41 3.0 H CH3 -9.6 -27.2 17.6 7.2 H CH29 — — 5.01 -5.4 H 13ng -19.9 -321 12.2 1.8 * UHF/3-ZlG/IUHF/3-ZIG * All values are in kcallmol TEnagydiffaencebemeanAamfm-mcmbsdmmdcascmdAAanm=Y=m 1Notethatinthesecascs,mminimumcouldbelocawdformestardngradical1.1. '10 ca:- 3: 2.: «.8: 3.:- c.::- c.«:- 8.8:- «8:- «8:- c8:- «.«:- «166:9 «.8 «.8 N8 «.8 2: «.8: «.8: «.8: 8.8: «.8: «.8: «.8: «86629 3::- ¢.«: :8: 3: «.8: :3: «s:- «8:- «.8: :8: «.«:- o.«: «:«UGEV «.«« . c.«« «.8 «.«« :8 «.3: m8 «.8 8.8 «.8 «8 «.«v 06:89 3.: 3:: 8.3: c.«: «.3: «.8: «.8: «.3: «.«: «.«:: 8.8: 3: «866m . - . - .3. oda. .. .. -1 ...». - ..::. .- .... .. .. . - ... - .... - -. - .« u : V c.«:: «.«:: 9:: «.«:: «.:«: «.8: :..:«: «.2: 9:: «.«:: «.8: «.2: o. 8.8 «.:« «.«« «.8 3:... «.8 «18 «.:« «.8. can «.8 «.8 «060v 8.8 can «3 8.8 «.8 «.8 «.8 8.8 8.5 8.8 «.8 can «86» «.8 c.«« 8.8 «.8 «.8 «.8 «.8 En :..«« «8 «.8 8.8 9.08» «.««: «.««: «.««: 9:: «.8: «.8: «.8: «.8: «.:«: «.««: «.o«: «.8: «08:0» «.8: 3.: S: «...: o.««: «.:«: c.:«: «a: 3:: «.2: o.«:: 8.2: 06:0» «.5 3:: ....:: 3:: «.8: «.8: 3:: «a: «.2: «.2: «8:: 2:: Magmw .. ... ...: ... .. .. .. .. ... . . . «8.: :8.: «8.: «8.: 8.: «8.: «8.: 88.: «8.: «8.: «8.: «8.: o: 88.: «8.: 88.: «8.: «8.: «8.: :8.: «8.: «8.: «8.: «8.: «8.: £8: «8.: «8.: «8.: 88.: «8.: «8.: :.8.: 88.: 88.: «8.: «8.: :8.: «:«o. «8.: #8.: «8.: :8.: «8.: «8.: «8.: :8.: :8.: «8.: :8.: :8.: «:6: 38.: «8.: :.8.: «8.: «8.: 8.: «8.: :8.: «8.: «8.: «8.: «8.: 5:0: «8.: «8.: 2...: 88.: 8:: «8.: «8.: «8.: «8.: 8n: «:«.: «3.: 06: «8.: «8.: «8.: 82 «8.: 88.: «8.: «3.: 3.3 «::..: «««z: :3 06: «8.: 8:»: «on: 88.: «8.: 5.: «8.: «2.: «8.: 3:: 8n: «8.: «08: «««.: «::..: «8.: 3...: 3n: «8.: «on: 8«.: «9.: 88.: «8.: «8.: «06. «.95 85 «.95 85 «.85 :5 «85 :5 «.95 :5 «.95 85 2.3; ...-«ad a .93 9.3. .§ 9.3.. 838% ::.: 28:8 28:5: :53: .3: 23035.: 33.88:: .:.: 5:83. 1.3A 3:215 UHFUMPZUHF 1.314 2.404 1.380 1.386 1.070 1.070 1.074 1.069 1.074 123.0 118.6 121.5 151.0 29.6 120.7 29.7 121.2 7.9 -175.5 -89.5 180.0 ~175.6 -180.0 1.317 2.327 1.352 1.355 1.074 1.073 1.076 1.077 1.072 122.7 1 18.6 122.0 152.9 30.7 1 18.5 30.8 1 19.6 -2.0 178.9 -90.0 180.0 179.0 180.0 47 t 1.320 2.316 1.382 1.377 1.083 1.082 1.089 1.087 1.081 123.6 1 18.8 121.4 150.6 33.0 1 14.2 32.9 120.2 37.7 -159.4 -87.5 -180.0 -159.6 180.0 ___1.5A__ 8215 (2119:“ UHF UMPZ UHF UMPZ 1.316 1.320 1.312 1.510 1.502 1.502 2.411 2.404 2.420 1.456 1.387 1.390 1.074 1.077 1.087 1.073 1.075 1.084 1.074 1.078 1.088 1.082 1.087 1.100 1.082 1.091 1.104 120.6 120.5 121.1 1 16.4 1 16.5 1 16.4 137.0 139.0 138.6 123.9 124.0 123.3 108.7 1 12.6 113.5 36.4 35.2 34.7 34.9 32.2 31.8 109.6 107 .2 106.3 - 125.4 - 129.7 -130.2 83.7 76.7 75.1 48.2 52.4 53.7 180.0 180.0 179.7 55.6 50.5 49.7 -180.0 - 179.5 ~179.2 48 Appendix 1.2. Cartesian Coordinates for Figure 1.3. 333132310000:- 1.213 O C C mammmzmo EIEEIZEO 1.071732 0.278564 1 .730026 0.487915 0.018555 -1 .854080 0.256653 2.280991 2.234268 -2.254929 -2.105850 1.213654 0.1 18149 1 .903214 0.563080 0.142303 -l.656281 0.337250 2.405136 2.310700 -2.310700 -2.138107 0.052795 -1 .876068 0.123810 2.261307 2.350662 -2.350677 -2.281600 -l.285583 0.671371 0.684174 0.531235 1.416916 0.393555 -1 .390808 0.21 1060 -1.359300 1.170853 0.458694 -1.l71478 0.715424 0.677353 0.41 1789 1.022964 0.955093 -1 .427994 0.261246 -1.406616 1 .406631 0.291977 -1.l44012 0.522400 0.5461 12 0.598465 1.507141 0.578064 -1.360229 0.397095 -1.196854 1.196854 0.333679 0.834793 0.181 870 0.243546 0.169479 0.508240 0.435257 0.780548 0.004990 0.914780 0.914764 0.895355 1.059860 0.358795 0.026321 0.789810 1.56581 1 0.099289 0.307236 0.118179 0.7041 17 0.7041 17 0.4731 14 1.1 16104 0.171 143 0.024796 0.1 14288 0.484009 0.01 1002 0.676254 0.217560 0.61 1633 0.611633 0.033234 33333230000g 333332300001; 1.172653 0.042100 2.184570 -1.1 19431 -1.032745 -2.403015 0.044205 1.938660 3.1 18271 -3.1 18256 -2.497650 -2.574753 0.337845 -1.7l3837 0.746552 0.657928 2.063629 0.262512 -1 .686188 -2.202728 2.668121 2.166840 0.508438 0.076965 0.338486 0.3 10944 -1 .080887 0.232742 0.841934 1 .191879 0.169098 0.464676 0.834229 0.506943 0.191376 0.150436 0.437408 -1.243744 0.125854 0.992630 0.932220 0.712784 0.916840 0.467743 49 0.372345 0.614990 0.058380 0.036057 0.781250 0.103729 -1 .372650 0.551010 0.090439 0.127533 0.896300 0.835007 0.286377 0.196762 0.138229 0.847290 0.1 80008 1.004150 0.952316 0.651 154 0.244751 0.976227 50 Appendix 1.3. Cartesian Coordinates for Figure 1.4. 0 333203000Oi— 0.702530 0.004349 -1.456696 0.715759 0.368759 1 .803650 2.707367 0.543533 -2.028763 -1 .95 5643 § 3332030000 -1.31961 1 0.256882 -2.07 3502 0.399902 0.026657 1 .542007 2.471558 0.165848 -2.566040 -2.471558 5% 3332030000' -1 .435760 0.626617 -2.070358 0.090652 0.402176 1.473312 2.599136 0.120453 -2.635620 -2.599136 --1 .042618 0.554703 0.617004 0.724000 1.395432 1 . 102646 1.408752 -l.301666 0.281830 -1 .396729 -1.008636 0.602783 0.573212 0.585236 1.237640 0.994476 1.328918 -1.338364 0.372925 -1.328918 0.925354 0.277863 0.412918 0.897064 1.715897 1.018890 1.132309 -1.07 8018 0.493591 -l.132309 1.151 108 0.042068 0.052322 0.256058 1.018677 0.529022 -1.185730 0.514954 0.206238 0.49681 1 1 .752502 0.987808 0.686264 1.278275 2.044662 0.549683 0.035065 0.329346 0.802124 0.035065 1 .590897 0.245789 0.488480 0.577667 1.067017 0.336868 0.120132 0.254990 0.656342 0.120148 2330033000 2033030300 8 2.134613 1.216324 0.083969 1.604935 0.490417 0.948181 -2.263809 -2.871017 -2.641739 2.871017 8 0.649185 -1 .369125 -2.477432 -1.796097 -2.372787 0.580048 0.135071 -1.343155 0.51 1261 1.369125 0.188950 0.253815 0.016357 0.805389 0.572205 0.399353 0.069687 0.537628 0.080353 0.537628 0.442596 -1 .497421 0.074081 0.890060 -1 .659622 0.770752 -1.161 148 1.51 1063 51 0.826172 0.177704 0.07 1518 1.01 1780 0.754578 0.995880 0.875229 1 .622162 0.120377 -1.622162 0.022690 2.048996 1 . 105530 0.882263 0.366882 0.531 189 0.139191 0.37 1552 1.169159 -1.374756 1 .497421 -2.048981 Appendix 1.4. Cartesian Coordinates for Figure 1.5. 1.1D 33330030303 0.303909 -1.670654 0.123703 0.204544 -2.225952 0.940277 0.534653 1 .765 366 -2.182419 2.21 1700 2.284485 1.2D O C C C H B H H H H H -1 .093369 0.161 102 -1.985748 0.478775 0.133621 1 .597 641 0.183640 -2.497559 -2.435303 2.152298 1.928787 1.4D 0 33333000 H H -1 .474579 0.471359 -1.938019 0.206894 0.416200 1 .733871 0.091949 -2.527374 -2.364594 2.253983 2.405106 -1.415176 0.702835 1.188019 0.667267 0.200226 -1 . 196060 0.585754 1 .03 3905 -1 .442383 0.335693 2.079926 -1 .377258 0.963867 -1.005219 0.265366 0.920639 0.660675 -1.734970 0.073151 -1 .793594 1.709854 0.091415 -1 .362778 0.638916 0.785172 0.497100 1.236282 0.677048 -1.420074 0.11 1800 -1 .467056 1.670380 0.178300 52 0.503189 0.084915 1.121323 0.082123 0.275300 1 .165024 0.325958 0.471573 0.507858 -1.329224 0.242706 1 .124466 0.197067 0.150421 0.285416 1.069489 0.685669 0.468048 0.295578 0.425 156 0.589584 -1 .5507 05 0.937042 0.222061 0.214325 0.290558 0.769073 0.197266 0.693802 0.081085 0.937042 0.595428 0.290771 33330033003!» 33330033003 3 2.073074 1 .022522 0.272156 1.266342 0.600510 -1.251 190 -2.545303 -2.756546 3.240600 -3.240585 1.723190 3 1 .961029 0.508621 0.527695 0.302872 0.362686 -1.976776 -2.856461 -2.178970 -2.239838 2.157410 2.894424 0.862732 0.066132 0.290009 0.675888 1.017578 0.365570 0.053391 0.322876 0.703064 0.703064 1.670029 0.061234 0.250229 0.155838 0.715820 0.623962 0.017456 0.518814 0.532547 1.026184 0.455322 0.419327 53 0.773804 0.002945 0.238159 0.745544 0.961044 0.388565 0.106857 0.878433 0.597931 0.597931 4.582581 0.030914 0.464264 0.272568 -1.418945 1 .229065 0.120377 0.820908 -1.058029 0.284668 1.087494 0.617400 54 Appendix 1.5. Cartesian Coordinates for Figure 1.6. 1 Q 2033330000i— 0.069794 0.567963 0.875504 1 .264908 1.955338 1.035278 0.993744 4 .445007 4.661606 -2.279648 1.2C' 2033330000 1.4C 2033330000 0.089310 0.715942 4.045685 1 .374832 2.045090 1.162888 1.029694 4 .530350 4 .752000 -2.300446 I 0.097656 0.586000 0.826981 1.392960 2.428146 1.015396 0.959686 4.318100 4.721802 -2.417648 0.786026 0.137375 0.283432 1.138763 1.498413 1.746140 0.809006 0.558731 4 .266769 -2.008774 0.721283 0.176163 0.245193 1 .03 3508 1.413437 1.632080 0.865738 0.672897 4 . 179688 4 .898926 0.665421 0.11 1053 0.145966 1.036179 1.010040 1.920715 0.970856 0.795273 4.082153 4.786621 1.104553 0.041885 0.216263 0.157532 0.579926 1.012100 0.767227 0.565475 0.523102 4 .1 18652 1 .05 3665 0.070236 0.194763 0.278854 0.467500 1. 144394 0.690933 0.460022 0.621979 4.323532 1 .294861 0. 140808 0.265869 0.001434 0.278992 0.478760 0.658707 0.46321 1 0.436417 0.989258 2333030000 32003300038 1.262085 1 .27 5070 2.378738 0.0551 15 3.326920 2.356491 4.037781 -2.278015 -3.326920 0.946136 8 4.543304 0.652039 0.725647 4.226410 0.558334 1.149078 1.364670 0.528259 2.151978 -2.219604 55 1.282120 0.422516 0.272705 0.053009 0.367096 0.270782 0.320419 0.063538 0.123535 0.169266 4.375946 0.637161 0.371078 0.278107 0.250610 0.152679 0.729400 0.063766 1.375931 0.637161 0.690598 0.033630 0.205400 0.750336 0.361496 0.147736 4.446381 0.947937 0.223100 1.747177 0.206848 0.963196 4.008621 0.722153 0.850540 4.559097 0.040787 4.311096 1.383820 0.623566 56 Appendix 1.6. Cartesian Coordinates for Figure 1.7. 1. H D! 33330030303 1.226883 0.768738 4 .481 1 10 0.731339 4 .213608 0.002609 1.494125 2.286285 -2.260742 1.182343 4 . 199265 1.2D' 33333330000 333333300003:- 0.137573 0.780304 0.975342 1.397888 2.172958 1.045929 1.23001 1 4 .381866 4.445480 -2.202057 4 .026169 0.1 10809 0.724747 0.741669 1 .523468 2.584396 1.1061 10 1.169022 4.166275 4.556213 -2. 179626 4 .505737 0.817261 0.050598 4.064438 0.054900 0.909409 0.544327 1 . 182419 1.447479 0.674622 1 .884537 -2.217682 0.622314 0.186432 0.222794 1 .063187 1.355850 1.772507 0.956300 0.762787 4.197723 0.819260 -2.310623 0.545883 0.090408 0.100098 1 .057419 1.009689 1.976517 0.985748 0.893097 4.121521 0.720596 -2.285950 0.374985 0.1951 14 0.745728 0.204575 0.404007 1.278534 0.405029 0.268677 4 .555817 1.154953 0.697006 1 .654419 0.736923 0.614990 0.869812 0.188019 1.594650 0.137451 0.763077 0.462341 4 .299210 0.487381 1 .762985 0.365570 0.614273 0.509964 0.351257 0.876236 0.023666 0.715378 0.263824 4 . 194733 0.036438 1.3D' 33330033003 33330033003 ‘63 3.307983 2.519730 1 .262833 2.779083 0.969620 0.241592 4.026932 -2.l72485 4.139175 -3.284439 4 .962997 1 .970490 0.957550 0.067841 0.824081 0.090744 4 .524857 -2.384933 4.727798 4.492645 4 .222412 4 .656357 0.433456 0.11 1298 0.356979 0.408400 0.874710 0.035217 0.224396 0.210876 0.753265 0.041626 0.790710 0.476547 0.323959 0.601227 0.045563 0.964462 0.353989 4 .022858 0.572205 1.204605 1 .9565 12 1.586900 57 4.061600 0.409256 0.724548 0.493637 4.620590 0.076538 0.263779 0.625946 4 . 197891 0.283585 1 .641830 0.270920 0.596176 0.191620 4.598495 1.192535 0.168304 0.660858 4.212067 0.151230 0.725250 1.261642 CHAPTER 2 Abstract: Since alkynes have higher symmetry than Olefins, it is not easy to infer the mechanism of a triplet carbene's addition to an alkyne by traditional product analysis studies. Specifically, no stereochemical information which might offer insight into the carbene's spin state can be extracted from the cyclopropene products. In 1971, Hendrick, Baron, and Jones showed that diphenylcarbene reacts with terminal alkynes in solution to produce indenes via a "self-trapping" vinylcarbene. They also examined the diphenylcarbene reaction with disubstituted alkynes and found at most trace amounts of the "self-trapping" indene product. In this work, we report the direct observation by organic matrix EPR of the vinylcarbenes generated from triplet fluorenylidene and terminal alkynes, our attempts at their structural confirmation by independent synthesis, and trapping the intermediate by another "self- trapping" method — halogen-migration. 58 59 2.1 Introduction: Reactions of triplet diphenylcarbene with monosubstituted acetylenes (Scheme 2.1) were reported 20 years ago by Baron, Hendrick and Jones. They found that the addition of triplet diphenylcarbene to acetylenes went through a self-trapping 2,2-diphenylvinylcarbene, described as its 1.3-diradical resonance form, to generate the corresponding substituted phenyl indene. Furthermore, by studying deuterium isotope effects, they showed that the formation of indene proceeded via addition to the phenyl ring to give 2.1 followed by a 1,5- hydrogen shift.l WW4» mum diphenylcarbene P R R - H . H .- . ——> .. J H R H [1.5-H]. 2 2 60 Similarities in the reactivities of diphenylcarbene and fluorenylidene has drawn a lot of attention?" The reactions of triplet fluorenylidene and monosubsitituted acetylenes, however, have not been addressed. Diphenylcarbene and fluorenylidene are both known to have ground-state triplet multiplicities7, but their chemistries are quite different. The reactions of fluorenylidene in solution show more "singlet" character than the reactions of diphenylcarbene. Two arguments have been advanced to explain the differing behaviors. One is steric; fluorenylidene is planar and "tied back" but diphenylcarbene, being bent less than fluorenylidene and also twisted, may experience more steric interactions in the transition states of addition reactions.‘ The other argument centers on the equilibrium population of singlet and triplet states of fluorenylidene and diphenylcarbene. The singlet-triplet gaps of diphenylcarbene and fluorenylidene are approximately 5 kcallmol" and l kcallmol,‘ respectively. It is generally believed that the singlet and triplet states of diphenylcarbene are in rapid equilibrium before being trapped by other species. However, because of its small singlet-triplet energy gap and the greater reactivity of singlet compared to triplet carbene, the chemistry of fluorenylidene shows primarily singlet behavior. From laser flash photolysis studies of 9-diazofluorene with methanol, Schuster et a1. calculated that ~ 5% or more singlet fluorenylidene is present at equilibrium at room temperature.8 Moss and Joyce also estimated that the ratio of singlet to triplet fluorenylidene was ~l .2 by product analysis of the reaction of fluorenylidene and cis-butene at room temperature. However, this earlier estimate was based on same flawed assumptions concerning the relative reactivities of singlet and triplet fluorenylidene. 61 As Grasse, Brauer, Zupancic, Kaufmann, and Schuster have pointed out, it is difficult to assign a particular chemical behavior to a specific electronic state of a carbene such as fluorenylidene, which has such a small energy difference between the lowest states. Then, if one wants to investigate the chemistry of triplet fluorenylidene, how can one be sure the species examined is the pure triplet state? At least, how can one be confident that the triplet dominates the observed chemistry? In 1978, Moss and Joyce convincingly demonstrated that fluorenylidene reacts with 13C labeled isobutene at 77 K through a triplet carbene mechanism hydrogen atom abstraction/radical recombination9 (Scheme 2.2). The reaction of fluorenylidene with 13CH2=C(CH3)2 at 77 K gave fluorenylalkene (2.3) in which the label was equally distributed between C(1) and C(3). Furthermore, those workers examined the reaction of fluorenylidene with various butene isomers over a wide temperature range, from solution experiments at 0 'C down to solid 62 mixtures at 77 K. The yield of cyclopropane decreased dramatically in the solid phase, from ~ 85% at 273 K to ~ 20% at 77 K, but the cyclopropane stereoisomer ratio from cis-2-butene changed very little. They believed that the cold temperature enhanced the triplet abstraction-recombination by decreasing the reaction rate of the singlet addition and also moving the singlet - triplet equilibrium over to the triplet state. Tomioka found similar changes in the reactions of arylcarbenes with alcohols and alkanes at low temperature and showed that this phenomenon applies only to those carbenes which have triplet ground states. Later, Platz and co-workers suggested that the changing chemistry of fluorenylidene at low temperature in frozen organic matrices would be better interpreted in terms of matrix effects; he pointed out that even if singlet fluorenylidene reacts with the organic glass at a diffusion-controlled rate (kdiff ~ 105 M'ls‘l, 120 K relatively warm, soft glass), this is much too slow to compete with intersystem crossing (ksr s 1010 3-1 ).4.10 In this work we have tried to examine the reaction of triplet fluorenylidene with terminal alkynes by direct observation of organic matrix EPR and analysis of the various trapping products in low temperature glasses. 2.2 Results: Reaction of Fluorenylidene with Phenylacetylene, l- Hexyne and 3,3-Dimethyl-l-butyne. Degassed solutions (0.1 M) of diazofluorene in the monosubstituted acetylenes were irradiated at 77 K with light from a 500W high pressure mercury arc lamp, filtered with 63 uranium and pyrex glasses as described in the Experimental Section. The product mixtures were separated by flash column chromatography over silica gel using hexane as eluent. Besides the major product cyclopropene (2.7), three other products were indentified: 9-fluorenone (2.4), bifluorenylidene (2.5) and fluorenone ketazine (2.6); no compound 2.8, which corresponds to 2.2 obtained from diphenylcarbene, was found. All cyclopropenes were identified by 1H NMR, 13C NMR and mass a.R=Ph b.R=t—Bu c.R=n-Bu spectrometry. For comparison, the above photolyses were repeated at room temperature; the product distributions were essentially the same as in the low temperature matrix photolyses. o 6 S e V humam o o Anm§ 0% GE 0. o N: on. n5 6% e do z 0 652.3.— .3u S .5 a w a ancfllm 0 ant» “M“ a" 65 Synthesis of Pyrazoles. Pyrazoles were prepared from 9- fluorenone (2.4) according to the reaction sequence shown in Scheme 2.4. Treatment of 9-fluorenone with phenyl ethynlmagnesium bromide in diethyl ether gave the ynol 2.9. Exposure of 2.9 to dilute sulfuric acid in THF resulted in yneol isomerization to enone 2.10. Then, the ketone reacted with toluenesulfonhydrazide in ethanol, followed by base-induced detosylation to afford pyrazole 2.13. We failed to generate the n-butyl pyrazole 2.13c, presumably due to the a-hydrogens next to the hydrazone 2. 1 1 c. M he 77 K PhCI'I3 01' PhC! CH a. R = Ph b. R = t-Bu Photolyses of Pyrazole (2.13). Photolysis of pyrazole in low temperature toluene or phenylacetylene (0.1 M) matrices using light from a 500W mercury arc lamp, filtered with uranium and pyrex glasses, 66 was a very clean reaction and only cyclopropene (2.7) was identified (Scheme 2.5). Photolyses of pyrazoles in benzene were carried out at room temperature and the results were the same as in the low temperature matrix. No indene (2.8) was found. Photolyses of 5-Phenyl-3-spirofluorenyl-3H-pyrazole (2.13a) with Various Trapping Reagents. A degassed solution of the pyrazole (0.05 M) in cyclohexene containing 30% (by volume) of toluene was photolyzed at 77 K and gave only 1-phenylspiro[cyclopropene-3,9'- fiuorene] (2.7a). The same photolyses were carried out in methanol, isopropanol, and acetic acid with the same results as in the cyclohexene matrix. (Scheme 2.6) In order to use oxygen as a trapping reagent, a solution of the pyrazole in methylene chloride (0.08 M) was oxygenated by bubbling with oxygen gas for 10 minutes before being frozen at liquid nitrogen temperature and photolyzed. Photolyses of the pyrazole in the oxygen-saturated methylene chloride gave only the cyclopropene (2.7a). The above reactions were also carried out in solution at room temperature and still the cyclopropene (2.7a) was the only product. Thermolyses of 5-Phenyl-3-spirofluorenyl-3H- pyrazole. A solution of the pyrazole (14.7 mg) in 4 ml of acetic acid was refluxed for one and a half hours until no starting material remained, as assessed by TLC. The solution was diluted with 100 ml of benzene, washed with saturated aqueous sodium bicarbonate solution and dried over magnesium sulfate. After removal of solvent, the residue was separated by flash chromatography over silica gel eluting with hexane. Only the cyclopropene (2.7a) could be identified. 67 36. Am a a R .5 .2 ‘ a g mommo S a. M 2. .8 .53 . «d .3... “on A a R s .2 moooamoé M R. .8 .53 ll 30 .9: \./ M R. .5 .5— 10 ... 68 Electron Spin Resonance Studies. Irradiation of a dilute degassed solution (~ 5 x 10-3 M) of 9-diazofluorene in phenylacetylene cooled to 77 K in the microwave cavity of an ESR spectrometer gave rise not to the well-known ESR spectrum of triplet fluorenylidene formed in other matrices but to the triplet spectrum shown in Figure 2.1. The lifetime of the triplet species was at least several hours in the low temperature matrix. A similar triplet ESR spectrum was obtained from irradiation of a dilute degassed solution (~ 5 x 10-3 M) of 5-Phenyl-3— spirofluorenyl-3H-pyrazole in toluene/pentane (l :1) or phenylacetylene matrices at 77 K, as shown in Figure 2.1. Irradiations of 9-diazofluorene in l-hexyne were carried out as described above in the neat alkyne or in dry toluene/pentane (1 :1), yielding the ESR spectra shown in Figure 2.2. Irradiations of 9-diazofluorene in 3,3-dirnethylbutyne were carried out as described above in the neat alkyne or in dry toluene/pentane (1:1). A triplet ESR spectrum was obtained which also showed significant radical contamination; however, this spectrum, judging from non-free- radical region, is significantly different from the ESR spectrum of triplet fluorenylidene. A procedure analogous to that given above was used for irradiation of 5-t-butyl-3-spirofluorenyl-3H-pyrazole in toluene/pentane (1:1) or 3,3-dimethyl4-butyne matrices at 77 K; the resulting spectrum are shown in Figure 2.3. 69 oceausfufi .5 A meme: ME 3. ox F 52; _ _ _ _ 8mm 83 Bow moon ©um 3: can I? oaseosfufi .30 0 e 1 V252. .E + g v— : am 23.“.— -3 .. 5.. 5:9: :33 .:. .-3.. 30-..? :32... . . z. ...: ... : .4 -1. 70 mmov hmmv <\/\ cascaemui .5 30m 053-: AIJEfiII =m-= ovom oovm omom mhmm mm: mom _. OS. 71 vmhm 8353.55 .5 . :umosmé womv mmov 3-..!“ + 0.6 mvmw ommw 72 on." ‘W a PM 8 M E. .5 73 Irradiation of 1-phenylspiro[cyclopropene-3,9'-fluorene] (2.7a) in phenylacetylene or dry toluene/pentane (1 :1) matrices at 77 K gave only a featureless free-radical signal. Only the triplet fluorenylidene ESR spectra were obtained from the irradiation of 9-diazofluorene in neat propargyl chloride or propargyl bromide (80% w/w in toluene) at 77 K following similar procedures described above. Photolyses of 9-Diazofluorene in 1,5-Hexadiyne and 1,7-Octaadiyne. Photolyses of 9-diazofluorene in 1,5-hexadiyne and 1,7-octadiyne were carried out as above, and giving only cyclopropene addition products 2.14 and 2.15, respectively (Scheme 2.7). Both the cyclopropenes were identified by 1H NMR, 13C NMR and mass spectrometry. Photolyses of 9-Diazofluorene in 6-Hexyn-l-ol. Photolysis of 9-diazofluorene in 6-hexyn4-ol were carried out similar to previous procedure, and gave only one product ether 2.16, which was identified by 1H NMR, 13C NMR and mass spectra (Scheme 2.7). Photolyses of 9-Diazofluorene in Propargyl Bromide. Photolyses of 9-diazofluorene in propargyl bromide containing 20% (by weight) of toluene were carried out similarly to the previous procedure and gave 9-fluorenone, bifluorenylidene and 2-bromo-3,9'- fluorenylidenylpropene (2.18) (Scheme 2.8). 75 Photolyses of 9-Diazofluorene in Propargyl Chloride. These experiments were carried out as in the case of propargyl bromide; the reaction mixture was separated by flash column chromatography over silica gel to give 9-fluorenone, bifluorenylidene, 9-chlorofluorene (2.23), 2-chloro-3 ,9'-fluorenylidenyl-propene (2.2 1), l-(chloromethyl)- spiro[cyclopropene-3,9'-fluorene] (2.22) and 3-chloro-3,9'-fluorenyl- propyne (2.24) (Scheme 2.9). Bromide Exchange of 1-(Chloromethyl)- spiro[cyclopropene-3,9'-fluorene] (2.22). A solution of 2.22 and lithium bromide in 10 ml of methanol was refluxed for 10 hours and the solvent was removed. The residue was directly subjected to 1H NMR. Only 1-(chloromethyl)-spiro[cyclopropene-3,9'-fluorene] (2.22) and 2- bromo-3,9'-fluorenylidenyl-propene (2.18) were identified and in product mixture, as shown in Scheme 2.10. Product Ratio Studies. Product distributions of irradiation of 9-diazofluorene in propargyl chloride and 1,2-dichloroethane are summarized in Table 2.1. All product yields were absolute yields and were determined by 1H NMR using toluene as the internal standard which was added just prior to analysis. 76 E w h a a S 8 ma 3 2 2 a a : 8 ma 3 w 3 an a m on wan as m n mu 2. 2 o ”a 3 a n a .2 M: on E :52." n2 2." 2." 3." ma «5&6 uses... @283 s .203 662g 8 as s 77 R = Ph, t-Bu, n-Bu. 2.3 Discussion Just as a triplet carbene reacts with an alkene via a stepwise mechanism, it should similarly react with an alkyne in a stepwise manner via an intermediate biradical. Jones et al. demonstrated the stepwise mechanism of the reaction of triplet diphenylcarbene and monosubstituted alkynes by intramolecularly trapping the vinylcarbene intermediate with one of the phenyl rings of the original diphenylcarbene.l However, our attempt to follow the same strategy failed, as shown in Scheme 2.11. The only adduct of triplet fluorenylidene and acetylene was cyclopropene (2.7) both in solution and in the low temperature matrices. As mentioned in the introduction, the reported literature rules out the explanation that the indene (2.8) was not observed because the only reactive species in the 78 reaction is singlet fluorenylidene. Additionally, the ESR spectrum of an irradiated sample of 9-diazofluorene in a phenylacetylene matrix at 77 K shows a triplet species that is not triplet fluorenylidene. Since similar ESR spectra were obtained from photolyses of 5-phenyl-3-spirofluorenyl-3H- pyrazole (2.13a) in toluene/pentane or phenylacetylene matrices (Figure 2.1), it seems reasonable that a similar or even the same type of intermediate was generated during these photolyses. As shown in Figures 2.2 and 2.3, it is clear that there is a triplet carbene other than simple triplet fluorenylidene generated in the reactions, although we were unsuccessful in making direct comparisons of the corresponding ESR spectra for irradiation of 9-diazofluorene with 3,3- dimethyl-butyne or 5-n-butyl-3-spirofluorenyl-3H-pyrazole at 77 K. ‘2 82 2.13 R = Ph, t-Bu, n—Bu. 79 The ESR studies obviously show that a secondary triplet species was generated in organic alkyne matrices. Then the question becomes what is the triplet carbene? Is it the analogue to the vinylcarbene intermediate proposed by Jones et a1? Actually, we can not think of any intermediate other than the vinylcarbene (2.26) or its diradical resonance form, as shown in Scheme 2.12. Unfortunately, as shown in Scheme 2.6, attempts to trap 2.26 were not very successful. Even thermolyses of 5-phenyl-3-spirofluorenyl- 3H-pyrazole (2.13a) in acetic acid did not trap the intermediate. Assuming the loss of N2 must yield an intermediate before product, the rate of ring-closure of the intermediate (2.26) to form cyclopropene (2.17) must exceed the rate of intermolecular reaction even though most carbenes react with carboxylic acids at rates near the diffusion limit. Some insight may be gained by comparing the geometries of 3,3 ,5-triphenyl-3H- pyrazole and 5-phenyl-3-spirofluorenyl-3H-pyrazole (2.13a) calculated by the MNDO method (Figure 2.4).11 The plane of the fluorenyl group is perpendicular to the plane of the pyrazole ring and the whole molecule is rigid. After extrusion of nitrogen, the orbitals on the carbene center can not overlap with the p—orbitals of the fluorenyl ring because of the rigidity of the ring. However, for triphenyl-pyrazole, the dihedral angle between the planes of the phenyl groups at C3 and the plane of the pyrazole are around 120 degree. More important, after extrusion of the nitrogen in the pyrazole ring, those two phenyl rings can free rotately to develop overlap between the orbitals on the carbene center and on the phenyl ring. The overlap of those orbitals is even better during the addition of diphenylcarbene and acetylene. Because of the twisted and bent 80 F' re 2.4. Calculated Geometries for les b MNDO 5-Phenyl-3-spirofluorenyl-3H-pyrazole 81 geometry of diphenyl carbene, the approaching acetylene can interact with the carbene in any orientation without suffering much steric hindrance from the ortho hydrogens on the phenyl rings of the carbene. However, in the case of the planar fluorenylidene, because of blocking by the hydrogens on the 2 and 8 positions, the acetylene can only approach the carbene in a perpendicular orientation. Jackson and O'Brien found that, Fluorenylidene Diphenylcarbene by MNDO and ab initio calculations, there is no rotational preference at the transition state distance of the addition reaction of triplet methylene to acetylene. ‘2 So, the absence of indene (2.8) product formation analogous that found from the reaction of diphenylcarbene, is the result of the the planar geometry of fluorenylidene. In addition, Jones et al. have shown that an entirely different course is followed by the reaction of triplet diphenylcarbene with disubstituted acetylenes. The cyclopropene becomes the major product and the corresponding indene appears in only trace amounts (< 2%) or is not observed, findings which these authors also attribute to steric effects. 83 Since the failures of the self-trapping and intermolecular trapping reaction of the vinylcarbene were due to steric effects, we attempted to get around this problem by using another intramolecular trap in the form of another hydroxy or alkyne group at the tail of the acetylene substrate. It was our hope that the vinylcarbene would be quenched by either O-H insertion or addition to the second alkyne, as shown in Scheme 2.13. It turned out that the rate of cyclopropene ring-closure is still much faster than the rate at which the tail wraps around the secondary carbene produced in the photolyses of 9-diazofluorene in 1,5-hexadiyne or 1,7- octadiyne matrices, as shown in Scheme 2.7. And the O-H insertions found on photolyses of 9-diazofluorene in S-pentyn-l-ol or 6-hexyn4-ol matrix were not surprising; triplet arylcarbenes favor O-H insertion in soft matrices like methanol and ethanol,‘-13 in which the diffusion rate is still faster than the rate of the triplet carbene addition to the alkyne. In 1984, Gaspar et al. studied the reaction of fluorenylidene with cis- and trans—1,2-dichloroethylene.“ With fluorenylidene, they observed three products containing the elements of dichoroethylene: Cir-cyclopropane, trans-cyclopropane and 9-(2,2-dichloroethylidene)fluorene, which was presumably formed via rearrangement of a diradical intermediate, as shown in Scheme 2.14. In the presence of styrene or butadiene, which were believed to be triplet carbene quenchers, the stereoselectivity of cyclopropanation increased, and the yield of chlorine-migration product decreased. An opposite trend was found when the reaction was diluted with hexafluorobenzene. Those workers concluded that the butadiene was efficient in trapping the triplet carbene, while the dilution in hexafluorobenzene was believed to enhance the triplet carbene formation. 34 So, 9-(2,2-dichloroethylidene)fluorene was generated solely from the diradical intermediate of triplet fluorenylidene addition to 1,2- dichloroethylene. W14 \=\ . ., We modified the dichloroethylene methology based on the same idea - chlorine migration. Here is the experiment: Photolyze 9- diazofluorene in a propargyl chloride matrix. If addition of triplet fluorenylidene to propargyl chloride indeed generates a triplet vinylcarbene intermediate (2.27), then the intermediate should be trapped by chlorine migration to give 2-chloro-3,9'-fluorenylidenyl-propene (2.21) (Scheme 2.15). The results are summarized in Scheme 2.9 and Table 2.1. Three products of the photolyses contained the propargyl chloride fragment: 2-chloro-3,9'-fluorenylidene-propene (2.21), 1- (chloromethyl)-spiro[cyclopropene-3,9'-fluorene] (2.22) and 3-chloro- 3,9'-fluorenyl-propyne (2.24). Although 2.22 and 2.24 can be generated from either the singlet or triplet states of fluorenylidene, 2.21 must arise through a stepwise process, in which triplet fluorenylidene adds to propargyl chloride followed by 1,2-chlorine atom shift (see Table 2.1). Further, the ratio (2.22/2.21) is sensitive to temperature and dilution with 1,2-dichloroethane. The yield of 2-chloro-3,9'-fluorenylidene (2.21) increases with increasing concentration of 1,2-dichloroethane or at lower temperature. In both cases the increase in 2.21 yield is because singlet-to- triplet intersystem crossing of fluorenylidene is enhanced relative to 86 reaction with propargyl chloride. However, only the triplet fluorenylidene ESR spectrum was obtained from irradiation of 9-diazofluorene and propargyl chloride or propargyl bromide at 77 K. It is possible that propargyl chloride or propargyl bromide, like other haloalkenes (i.e. 1,2- dichloroethane) are more inert than the other alkynes we examined at the low temperature. In the reaction of fluorenylidene and propargyl bromide, 2- bromo-3,9'-fluorenylidene-propene (2.18) was the only adduct. Apparently, the cyclopropene (2.19) was not thermally stable and rearranged to 2.18 (Scheme 2.15). To further test the stability of cyclopropene 2.19, the corresponding chloride 2.22, was exposed to LiBr in refluxing methanol. This reaction yielded only 2.18 and unreacted 2.22; no cyclopropene 2.19 nor rearranged product (2.21) from 2.22 was found (Scheme 2.10). We do not quite understand the process of this rearrangement, and also we can not rule out the possibility that the rate of cyclopropene ring closure can not compete with that of bromine migration. 2.4 Conclusion: As originally suggested by Jones et al., it appears that steric interactions prevent the vinylcarbene intermediate derived from the addition of triplet fluorenylidene to monosubstituted acetylene from being intramolecularly trapped by its aryl ring as in the case of its analogue, triplet diphenylcarbene. We have demonstrated through ESR studies and 1,2-chlorine atom shift that triplet fluorenylidene addition to alkynes 87 proceeds through the same type of intermediate, vinylcarbene or its resonance 1.3-diradical resonance form. W 2.5 Experimental: General Procedures. All 1H NMR and 13C NMR spectra were obtained by using a 300 MHz Varian Gemini, a 300 MHz Varian VXR-300 or a 500 MHz Varian VXR-SOO instrument. UV spectra were recorded on a Shirnadzu UV-160 spectrometer kindly shared by the group of Prof. Peter Wagner. Mass spectra were recorded on a Fisons VG Trion-l mass spectrometer. High resolution mass spectra were recorded 88 on a JEOL JMS-HXl 10 high resolution double-focusing mass spectrometer. Gas chromatographic analyses were performed on Perkin-Elmer 8500 equipped with a flame ionization detector. Unless specified, concentration of mixtures after workup was performed using a Biichi rotary evaporator. Dry solvents, benzene, toluene, tetrahydrofuran, and diethyl ether were distilled from sodium/benzophenone immediately prior to use. Methylene chloride and acetonitrile were heated at reflux over calcium hydride and distilled immediately prior to use. Photolytic Procedures in Organic Matrices. In the general procedure, a solution of 9-diazofluorene (0.1 M) in the neat appropriate alkyne was placed in an NMR tube (8 in). The sample was then degassed using three freeze-pump-thaw cycles. The sample tube was transferred to a quartz-tailed Dewar flask which contained liquid nitrogen. Samples were irradiated by placing the tail of the Dewar in front of a 500W mercury arc lamp shielded with a water jacket and a uranium glass filter. Matrices were irradiated for 942 hours, with thawing and shaking every 2 hours for 540 minutes in order to homogenize them. After irradiation, samples were brought to room temperature and the tubes were opened. After the alkyne had evaporated, the residue was separated by flash chromatography over silica gel (10 x 100 mm) eluted with hexane. Besides the major product, cyclopropene, unreacted diazofluorene, and trace amounts of fluorenone ketazine and fluorenone were visible in the 1H NMR spectrum before separation. 89 Irradiation of 9-Diazofluorene in Phenylacetylene Matrix. The only addition product of fluorenylidene with phenylacetylene is 1-phenylspiro[cyclopropene-3,9'-fluorene] (2.7a): 1H NMR (300 MHz, acetone-d6) 8 7.17 (dt, 2 H, J = 7.5, 1.1 Hz), 7.26 (td, 2 H, J = 7.4, 1.1 Hz), 7.30-7.42 (complex, 7 H), 7.81 (s, 1 H), 7.97 (dt, 2 H, J = 7.7, 0.8 Hz); 13C NMR (75.5 MHz, acetone-d5) 8 148.4, 139.9, 129.7 (doublet), 128.7, 126.4, 126.2, 125.8, 120.7, 119.8, 119.6, 103.1, 36.3; MS (EI) m/e 266 (M+), 265, 181, 180, 152, 105; UV Xmax [nrn] (log 6) 258.9 (5.78), 221.4 (4.96). The cyclopropene was dimerized at room temperature after a few days. 1H NMR (300 MHz, CDC13) 8 5.28 (d, 2 H, J = 7.5 Hz), 6.28 (t, 2 H, .I = 7.5 Hz), 6.59 (d, 4 H, J = 7.5 Hz), 6.75-7.00 (complex, 10 H), 7.10 (d, 2 H, J = 7.5 Hz), 7.25-7.41 (m, 4 H), 7.45 (d, 2 H), 8.32 (d, 2 H, J =7.5 Hz); 13C NMR (75.5 MHz, CDC13) 8 146.6, 144.5, 143.5, 142.0, 141.0, 140.8, 130.0, 129.2, 128.3, 127.4, 126.7, 126.6, 126.5, 126.4, 126.3, 126.1, 125.9, 125.1, 119.9, 118.1; MS (EI) m/e 532 (M+), 460, 367, 289, 265, 252, 165, 105. Irradiation of 9-Diazofluorene in t-Butylacetylene Matrix. The only addition product of fluorenylidene with :- butylacetylene is 1-t-butylspiro[cyclopropene-3,9'-fluorene] (2.7b): 1H NMR (300 MHz, CDC13) 8 1.17 (s, 9 H), 6.79 (s, 1 H), 7.18 (d, 2 H, J = 7.3 Hz), 7.28 (tt, 2 H, J = 7.3, 1.3 Hz), 7.34 (tt, 2 H, J = 7.3, 1.3 Hz), 7.85 (d, 2 H, J = 7.3 Hz); 13C NMR (75.5 MHz, CDC13) 8 149.8, 139.5, 130.0, 126.8, 126.2, 120.7, 119.9, 97.2, 37.1, 33.6, 29.1; MS (ED m/e 246 (M+), 231, 216, 215, 203, 202, 190, 189, 165. Irradiation of 9-Diazofluorene In l-Hexyne Matrix. The major products were 1-n-butylspiro[cyclopropene-3,9'-fluorene] 90 (2.7c) and the C-H insertion products (relative ratio < 10% calculated from the integral ratio in NMR spectrum). Data for l-n-butyl- spiro[cyclopropene-3,9'-fluorene] (2.7c): 1H NMR (300 MHz, CDC13) 8 0.85 (t, 3 H, J = 6.8 Hz), 1.25-1.70 (m, 4 H), 2.55 (t, 2 H, J = 7.5 Hz), 6.91 (s, 1 H), 7.16 (d, 2 H, J = 6.6 Hz), 7.31 (td, 2 H, J = 7.3, 1.1 Hz), 7.35 (td, 2 H, J = 7.3, 1.1 Hz), 7.85 (d, 2H, J =7.4 Hz); 13C NMR (75.5 MHz, CDC13) 8150.4, 139.8, 126.4, 125.9, 122.7, 120.6, 119.8, 114.0, 99.9. 29.3, 24.1, 22.1, 13.4; MS (BI) m/e 247 (M++1), 246 (M+), 231, 217, 203, 189. Photolytic Procedures at Room Temperature. Basic procedures are the same as described above except that the sample tube was directly put into the water jacket without using the liquid nitrogen Dewar. Irradiation of 9-Diazofluorene in Alkynes (Phenylacetylene, t-Butylacetylene and l-Hexyne) at Room Temperature. The major products are the same as in low temperature matrices, except that the relative ratio for the C-H insertion products is higher for the l-hexyne (< 10% calculated from the integral ratio in NMR spectrum) 9-Diazofluorene (9-DAF) was prepared by literature procedures.“15 1,9'-Fluorenyl-3-Phenylprop-2-yn-1-ol (2.9a). In an oven-dried 300 mL three-neck round-bottom flask equipped with pressure equalizing dropping funnel, reflux condenser, magnetic stirrer, and an argon inlet, a solution of ethylmagnesium bromide in diethyl ether (ca. 15 mL), was generated from magnesium turning (0.66 g, 27.3 mmol) in dry 91 diethyl ether (5 mL), ethyl bromide (2.95 g, 27.1 mL) in dry diethyl ether (10 mL) and a trace of iodine. A solution of phenylacetylene (2.79 g, 27.3 mmol) in dry diethyl ether (10 mL) was added into the flask slowly through the dropping funnel. The reaction mixture was gently refluxed for 2 hours and then cooled to room temperature. After the stirrer was started, a solution of 9-fluorenone (4.92 g, 27.3 mmol) in dry ether (10 mL) was slowly added; the reaction was stirred at room temperature for 1.5 hours. Finally, it was refluxed for 1 hour and cooled in an ice bath. The reaction was quenched by adding ammonium chloride (8 g) as a saturated aqueous solution. The aqueous layer was extracted with diethyl ether (3 x 30 mL). The combined organic layers were dried over magnesium sulfate and the solvent was removed under vacuum to afford a yellow oil. The yellow oil was crystallized from methanol/hexane (1 :1), affording a yellow solid (5.36 g, 69%): 1H NMR (300 MHz, CDC13) 8 2.8 (s, l H), 7.15-7.42 (m, 9 H), 7.61 (d, 2 H, J = 7.5 Hz), 7.75 (d, 2 H, J = 7.5 Hz); 13C NMR (75.5 MHz, CDC13) 8 147.5, 139.3, 132.1, 129.9, 129.8, 128.7, 128.4, 124.6, 122.6, 120.4, 89.0, 83.0, 75.1; MS (EI) m/e 282 (M+), 265, 252. Preparation of 2.9b. A procedure analogous to that given for 2.9a was followed. Reaction of 9-fluorenone (5.86 g, 32.5 mmol) in dry diethyl ether with 3,3-dimethyl-l-butynylmagnesium bromide, generated from t—butyl acetylene (2.67 g, 32.5 mmol) and ethyl magnesium bormide (32.5 mmol), gave lb (2.7 g, 32%) after recrystallization from pentane and a few drops of benzene: 1H NMR (300 MHz, CDC13) 8 1.18 (s, 12 H), 2.43 (s, 1 H), 7.26-7.40 (m, 4 H), 7.59 (d, 2 H, J = 6.9 Hz), 7.67 92 (d, 2 H, J = 6.9 Hz); 13C NMR (75.5 MHz, CDC13) 8 147.7, 139.0, 129.4, 128.4, 124.2, 120.0, 92.4, 78.2, 74.8, 30.8, 27.4. Preparation of 2.9c. A procedure analogous to that given for 2.9a was followed. Reaction of 9-fluorenone (6.27 g, 34.8 mmol) in dry diethyl ether with l-hexynylmagnesium bromide, generated from n- butyl acetylene (2.86 g, 34.8 mmol) and ethylmagnesium bromide (34.8 mmol), gave 1c (8.13 g, 89%) after recrystallization from pentane and a few drops of benzene: 1H NMR (300 MHz, CDC13) 8 0.87 (t, 3 H, J = 7.5 Hz), 1.2745 (complex, 4 H), 2.18 (t, 2 H, J = 7.5 Hz), 2.43 (s, l H), 7.27- 7.4 (m, 4 H), 7.58 (d, 2 H, J = 6.6 Hz), 7.67 (d, 2 H, J = 6.6 Hz); 13C NMR (75.5 MHz, CDC13) 8 147.6, 138.9, 134.7, 129.4, 129.1, 128.5, 124.3, 124.2, 120.3, 120.1, 84.4, 79.8, 74.9, 30.5, 21.9, 18.5, 13.6. Isomerization of 2.9a. A solution of 2.9a (4.02 g, 14.3 mmol) and dilute sulfuric acid (1 mL, 25% v/v in H20) in THF (20 mL) was refluxed for 5 hours. The mixture was cooled to room temperature and quenched by addition of saturated aqueous sodium bicarbonate solution (80 mL). The aqueous layer was extracted with benzene (3 x 30 mL) and the combined organic solution was washed with water, brine, and dried over magnesium sulfate. The solvent was removed under vacuum and the residue was crystallized from THF/benzene (1:1) to give a yellow solid (2.10a) was obtained (3.14 g, 78%): 1H NMR (300 MHz, acetone-ds) 8 7.26 (tt, 1 H, J = 7.5, 1.5 Hz), 7.36 (tt, 1 H, J = 7.5, 1.5 Hz), 7.46 (m, 2 H), 7.59 (t, 2 H), 7.69 (t, l H), 7.81 ( d, 2 H, J = 7.5 Hz), 7.97 (s, 1 H), 8.08 (d, 1 H, J = 7.5 Hz), 8.20 (dm, 2 H), 8.37 (d, 1 H, J = 7.5 Hz); 13C NMR (75.5 MHz, CDC13) 8 193.0, 146.4, 142.5, 141.1, 139.0, 138.5, 135.6, 93 133.5, 131.1, 130.7, 129.3, 129.0, 128.1, 127.9, 127.8, 127.6, 121.3, 120.1, 120.0, 119.9, 119.7, 119.6. Isomerization of 2.9b. A procedure analogous to that given for 2.9a was followed. Treatment of 2.9b (2.33 g, 8.9 mmol) with dilute sulfuric acid (1 mL, 25% v/v in H2O) in THF (20 mL) gave 2.10b (1.66 g, 71%) as a yellow solid: 1H NMR (300 MHz, CDC13) 8 1.30 (s, 9 H), 7.25 (t, 2 H, J = 7.5 Hz), 7.38 (t, 2 H, J = 7.5 Hz), 7.61 (d, 2 H, J = 7.5 Hz), 7.68 (d, 1 H, J = 7.5 Hz), 8.51 (d, 1 H, J = 7.5 Hz). 13C NMR (75.5 MHz, CDC13) 8 207.1, 145.0, 142.2, 140.8, 139.0, 135.4, 130.7, 130.3. 128.0, 127.8, 127.3, 120.9, 119.8, 119.6, 118.1, 44.7, 26.7. Isomerization of 2.9c. A procedure analogous to that given for 2.9a was followed. Treatment of 2.9c (11.6 g, 14.4 mmol) with dilute sulfuric acid (10 mL, 25% v/v in H2O) in THF (110 mL) gave 2.10c (7.88 g, 68%) as a yellow solid: 1H NMR (300 MHz, CDC13) 8 0.93 (t, 3 H, J = 7.5 Hz), 1.47 (m, 2 H), 1.69 (m, 2 H), 2.70 (t, 2 H, J = 7.5 Hz), 7.01 (s, 1 H), 7.23 (m, 2 H), 7.30-7.40 (m, 2 H), 7.56 (d, 2 H), 7.63 (d, 1 H, J = 7.5 Hz), 8.73 (d, 1 H, J = 7.5 Hz); 13C NMR (75.5 MHz, CDC13) 8 201.5, 145.7, 142.3, 141.2, 139.0, 135.4, 131.1, 130.6, 128.4, 127.3, 121.0, 120.8, 119.9, 119.5, 44.7, 26.4, 22.4, 14.0. Preparation of 2.11a. A mixture of 2.10a (2.37 g, 8.4 mmol), p-toluenesulfonhydrazide (2.36 g, 1.27 mmol) and a few drops of acetic acid in absolute ethyl alcohol (50 mL) was refluxed for 36 hours. The reaction mixture was then poured into methanol (150 mL) and a yellow precipitate formed while the solution cooled. The yellow precipitate was collected by suction filtration, air-dried and recrystallized 94 from methanol to afford needle-shaped yellow crystals (1.9 g, 50%): 1H NMR (300 MHz, CDC13) 8 2.42 (s, 3 H), 6.70 (td, 1 H, J = 8.5, 0.6 Hz), 6.77 (s, 1 H), 6.95 (d, 1 H, J = 8.5 Hz), 7.22-7.38 (complex, 7 H), 7.43 (td, J = 8.5, 0.6 Hz), 7.61 (d, 1 H, J = 8.6 Hz), 7.67 (d, 1 H, J = 8.6 Hz), 7.75 (m, 4 H), 8.00 (br, 1 H). Preparation of 2.11b. A solution of 2.10b (4.87 g, 18.4 mmol), a few drops of acetic acid and p-toluenesulfonhydrazide in absolute ethyl alcohol (60 mL) was refluxed for 2 days. The solvent was evaporated to about 25 mL, and the reaction mixture was diluted with pentane (80 mL) and put in a freezer (~ -20 °C) overnight. A pale yellow solid was precipitated and collected by suction filtration (1.0 g, 13%): 1H NMR (300 MHz, CDC13) 8 1.18 (s, 9 H), 2.48 (s , 3 H), 6.61 (s, 1 H), 6.76 (t, 1H, J = 7.5 Hz), 6.89 (d, l H, J = 7.5 Hz), 7.31 (d, 2 H, J =7.5 Hz), 7.39 (t, 1 H, J = 7.5 Hz), 7.64 (m, 2 H), 7.76 (d, 1 H, J = 8.6 Hz); 13C NMR (75.5 MHz, CDC13) 8 159.9, 143.8, 141.6, 141.2, 139.8, 137.1, 135.5. 134.6, 129.9, 129.7, 129.6, 129.3, 128.2, 127.9, 127.5, 123.9, 120.8, 119.9, 112.6, 39.7, 28.0, 21.7. Preparation of 5-Phenyl-3-spirof1uorenyl-3H- pyrazole. (2.13a) To a solution of the tosyl hydrazone (2.11a) (1.0 g, 2.22 mmol) in pyridine (10 mL) was slowly added sodium methoxide (0.18 g, 3.33 mmol). After the addition was complete, the dark-red reaction mixture was protected with a drying tube filled with calcium chloride and warmed gently on a water bath to 60-65 °C for 20 min. The reaction was quenched with ice-water (50 mL) and the aqueous layer extracted with pentane (3 x 20 mL). The combined organic layers were washed with aqueous copper(Il) sulfate until no more color change was observed, water, 95 and dried with magnesium sulfate. The solvent was removed under vacuum to give a pale-pink solid, which was recrystallized from methylene chloride and pentane to afford 2.13a (0.4 g, 61%): 1H NMR (300 MHz, acetone-(15) 8 6.85 (d, 2 H, J = 8.0 Hz), 7.28 (td, 2 H, J = 8.0, 1.6 Hz), 7.36 (s, 1 H), 7.51 (td, 2 H, J = 8.0, 1.6 Hz), 7.58 (m, 3 H), 8.00 (d, 2 H, J = 8.0 Hz), 8.24 (dm, 2 H, J = 8.0 Hz); 13C NMR (75.5 MHz, CDC13) 8 159.5. 143.2, 135.6, 130.3, 129.7, 129.6, 129.0, 127.9, 127.4, 123.7, 120.9, 108.1; MS (ED m/e 293.8 (Mi), 265.3, 239.2, 190.1, 163.2, 132.9; UV (pentane) Mnax [nm] (log 8) 304.2 (3.69), 235.6 nm (4.58). Preparation of S-t-Butyl-3-spirof1uorenyl-3H- pyrazole (2.13b). The procedure as described above for (2.13a) was followed. Treatment of the tosyl hydrazone (2.11b) (1.0 g, 2.3 mmol) with sodium methoxide (0.2 g, 3.7 mmol) in pyridine (8 mL) afforded 2.13b (25 mg, 3.9%): 1H NMR (300 MHz, CDC13) 8 1.52 (s, 9H), 6.28 (s, 1 H), 6.68 (d, 2 H, J = 7.5 Hz), 7.20 (t, 2 H, J = 7.5 Hz), 7.40 (t, 2 H, J =7.5 Hz), 7.78 (d, 2 H, J =7.5 Hz); 13C NMR (75.5 MHz, CDC13) 8 172.0, 143.3, 135.7, 129.4, 127.9, 127.8, 123.3, 120.9, 106.5, 32.5, 29.1; MS (ED m/e 274.2 (Mi), 259.4, 232.3. Irradiation of S-Phenyl-3-spirofluorenyl-3H-pyrazole (2.13a) at Room Temperature. The pyrazole 2.13a was irradiated in toluene according to the general photolytic procedures described above. The reaction is very clean and the only product found was 1- phenylspiroIcyclopropene-B,9'-fluorene] (2.7a). Irradiation of 5-Phenyl-3-spirofluorenyl-3H-pyrazole (2.13a) in Toluene Matrix at 77 K. The procedures are as described 96 above in general photolytic procedures; 1-phenylspiro[cyclopropene-3,9'- fluorene] (2.7a) was the only product isolated. Irradiation of S-t-Butyl-3-spirofluorenyl-3H- pyrazole (2.13b) at Room Temperature. Basically, the procedures are as described above in general photolytic procedures except that the pyrazole was irradiated in toluene. The reaction is very clean and the only product found was 1-t-butylspiro[cyclopropene-3,9'-fluorene] (2.7b). Irradiation of 5-t-Butyl-3-spirofluorenyl-3H- pyrazole (2.13b) in Toluene Matrix at 77 K. The procedures are as described above in general photolytic procedures; l-t- butylspiro[cyclopropene-3,9'-fluorene] (2.7b) was the only product. Irradiation of 5-Phenyl-3-spirofluorenyl-BH-pyrazole (2.13a) in a Mixture of Cyclohexene and Toluene at Room Temperature or 77 K. A solution of 5-phenyl-3-spirofluorenyl-3H- pyrazole (2.13a) (0.1 M) was prepared in a mixture of cyclohexene and toluene (7:3, v/v). The solution was degassed, irradiated and worked up as described above. The reaction was very clean and the only product found is l-phenylspiro[cyclopropene-3,9'-fluorene] (2.7a). Irradiation of 5-Phenyl-3-spirofluorenyl-3H-pyrazole (2.13a) in a Mixture of Methanol and Toluene at Room Temperature or 77 K. A solution of 5-phenyl-3-spirofluorenyl-3H- pyrazole (2.13a) (0.1 M) was prepared in a mixture of methanol and toluene (7:3, v/v). The solution was degassed, irradiated and worked up as described above. The reaction is clean and the only product found was 1- Phenylspiro[cyclopropene-3,9'-Fluorene] (2.7a). 97 Irradiation of S-Phenyl-3-spirofluorenyl-3H-pyrazole (2.13a) in a Mixture of 2-Propanol and Toluene at Room Temperature or 77 K. A solution of 5-phenyl-3—spirofluorenyl-3H- pyrazole (2.13a) (0.1 M) was prepared in a mixture of 2-propanol and toluene (7:3, v/v). The solution was degassed, irradiated and worked up as described above. The reaction is clean and the only product found is 1- Phenylspiro[cyclopropene-3,9'-Fluorene] (2.7a). Irradiation of 5-Pheny1-3-spirofluorenyl-3H-pyrazole (2.13a) in a Mixture of Acetic Acid, Methanol and Toluene at Room Temperature or 77 K. A solution of 5-phenyl-3-spirofluorenyl- 3H-pyrazole (2.13a) (0.1 M) in a mixture of acetic acid, methanol and toluene (0.5:8:2, v/v) was degassed, irradiated and worked up as described above. The reaction is clean and the only product found was 1- Phenylspiro[cyclopropene-3,9'-Fluorene] (2.7a). Thermolysis of S-Phenyl-3-spirofluorenyl-3H- pyrazole (2.13a) in Acetic Acid. A solution of 5-phenyl-3- spirofluorenyl—3H-pyrazole (2.13a) (14.7 mg, 0.05 mmol) in acetic acid (4 mL) was refluxed for 1.5 hours until all starting material was gone, as shown by TLC. The solution was poured into benzene (100 mL). The benzene solution was washed with aqueous saturated sodium bicarbonate solution three times and dried over magnesium sulfate. After removal of solvent, the residue was directly subjected to NMR analysis. Only 1- phenylspirolcyclopropene-3,9'-fluorene] (2.7a) was detected. Electron Spin Resonance Studies. Irradiation of a dilute degassed solution (~ 5 x 10'3 M) of 9-diazofluorene in polycrystalline 98 phenylacetylene cooled to 77 K in the microwave cavity of an ESR spectrometer gave rise not to the well-known ESR spectrum of triplet fluorenylidene formed in other matrices but to the triplet spectrum shown in Figure 2.1. The lifetime of the triplet species was at least several hours in the low temperature matrix. A similar triplet ESR spectrum was obtained from irradiation of a dilute degassed solution (~ 5 x 10'3 M) of 5- phenyl-3-spirofluorenyl-3H-pyrazole (2.13a) in toluene/pentane (1 :1) or phenylacetylene matrices at 77 K, as shown in Figure 2.1. The similar procedures described above for irradiation of 9- diazofluorene in l-hexyne and hex-1-yne-6-ol were carried out in the neat alkyne or in dry toluene/pentane (1:1). The ESR spectrum for 9- diazofluorene reacting with l-hexyne was shown in Figure 2.2 and a similar spectrum was obtained for 9-diazofluorene reacting with 5-hexyn— l-ol under the same condition. Procedures similar to those described above for irradiation of 9—diazofluorene in 3,3—dimethylbutyne were carried out in the neat alkyne or in dry toluene/pentane (1 :1). A triplet ESR spectrum was obtained but was not considered very successful due to significant radical contamination; however, this spectrum, judged from the non-free-radical region, was still different from the ESR spectrum of triplet fluorenylidene. A procedure analogous to that given for irradiation of 5+ butyl-3-spirofluorenyl-3H-pyrazole (2.13b) was carried out in toluene/pentane (1:1) or 3,3-dimethyl4-butyne matrices at 77 K and the ESR spectrum was shown in Figure 2.3. 99 Irradiation of 1-phenyl-spiro[cyclopropene-3,9'-fluorene] (2.7a) in phenylacetylene or dry toluene/pentane (1 :1) matrices at 77 K gave only a featureless free-radical signal. Only the triplet fluorenylidene ESR spectrum was obtained from the irradiation of 9-diazofluorene in neat propargyl chloride or propargyl bromide (80% w/w in toluene) at 77 K followed the above similar procedures. Irradiation of 9-Diazofluorene in 1,5-Hexadiyne at Room Temperature or 77 K. The procedures were as described above in general photolytic procedures. Only one major product, 1-(3- butynyl)spiro[cyclopropene-3,9'-fluorene] (2.14), was found for both photolysis at room temperature and at 77 K: 1H NMR (300 MHz, CDC13) 8 1.98 (t, 1 H, J = 2.5 Hz), 2.43 (td, 2 H, J = 7.5, 2.5 Hz), 2.82 (td, 2 H, J = 7.5 Hz), 7.08 (s, 1 H), 7.19 (d, 2 H, J = 7.6 Hz), 7.30 (td, 2 H, J = 7.6, 1.5 Hz), 7.37 (td, 2 H, J = 7.6, 1.5 Hz), 7.84 (d, 2 H, J = 7.6 Hz); 13C NMR (75.5 MHz, CDC13) 8 149.3, 139.7, 126.3, 126.0, 120.8, 120.5, 119.7, 102.1, 82.7, 69.3, 36.2, 24.4, 17.4, 17.2; MS (EI) m/e 242 (M+), 241, 203, 202. Irradiation of 9-Diazofluorene in 1,7-0ctadiyne at Room Temperature or 77 K. The procedures were the same as described above in general photolytic procedures. Only one major product, 1-(5-hexynyl)-spiro[cyclopropene-3,9'-fluorene] (2.15), was found for both photolysis in solution and in 1,7-hexadiyne matrix at 77 K: 1H NMR (75.5 MHz, CDC13) 8 1.5-1.78 (complex, 4 H), 1.92 (t, 1 H, J = 3.0 Hz), 2.15 (td, 2 H, J = 6.7, 3.0 Hz), 2.60 (td, 2 H, J = 7.5 Hz), 6.95 (s, 100 1 H), 7.14 (d, 2 H, J = 7.0 Hz), 7.29 (td, 2 H, J = 7.0, 1.7 Hz), 7.35 (td, 2 H, J = 7.0, 1.7 Hz), 7.84 (d, 2 H, J = 7.0 Hz); 13C NMR (75.5 MHz, CDC13) 8 149.6, 139.6, 126.3, 125.8, 122.2, 120.4, 119.7, 100.4, 84.0, 68.5, 36.2, 27.8, 26.5, 24.1, 18.0; MS (El) m/e 270 (M+), 269, 242, 241, 216, 215, 203, 202, 189, 178, 165. Irradiation of 9-Diazofluorene in S-Hexyn-l-ol at Room Temperature or 77 K. The procedures were as described above in general photolytic procedures. Only one major product, 9-fluorenyl 5- hexynyl ether (2.16) was found: 1H NMR (300 MHz, CDC13) 8 1.60 (m, 4 H), 1.89 (t, 1 H, J = 2.6 Hz), 2.12 (m, 2 H), 3.16 (t, 2 H, J = 6.0 Hz), 5.61 (s, 1 H), 7.31 (dd, 2 H, J = 7.3, 1.2 Hz), 7.38 (trn, 2 H, J =7.3 Hz), 7.60 (d, 2 H, J = 7.3 Hz), 7.66 (d, 2 H, J =7.3 Hz); 13C NMR (75.5 MHz, CDC13) 8 143.0, 140.8, 128.9, 127.5, 125.4, 119.9, 84.3, 80.6, 68.3, 63.9, 29.1, 25.1, 18.1; IR cm'1 3299 (s), 2941(8), 2869(3), 2050(w); MS (131) m/e 262 (W), 182, 181, 180, 165, 152, 139. Irradiation of 9-Diazofluorene in Propargyl Bromide at 77 K. A solution of 9-diazofluorene (16.3 mg, 0.085 mmol) in propargyl bromide (0.5 mL, 80% w/w in toluene) was prepared in an NMR tube. The solution was deaerated by three freeze-pump-thaw cycles. The setup of irradiation was same as described above. The solution was irradiated for 11 h, while the solution was warmed to room temperature for every 1.5 h. After removal of solvent, the residue was separated by flash column chromatography over silica gel eluting with methylene chloride/hexane (20%). Besides trace amount of fluorenone, 2-bromo- 3,9'-fluorenylidenyl-propene (2.18) was the only major product: 1H NMR (300 MHz, CDC13) 8 5.98 (t, 1 H,J =1.5 Hz), 6.15 (t, 1 H, J = 1.9 101 Hz), 7.00 (8,1 H), 7.18-7.25 (m, 2 H), 7.35 (t, 2 H, J = 7.6 Hz), 7.58 (m, 3 H), 8.22 (d, 1 H, J = 7.7 Hz); Selected decoupling 1H NMR 5 5.98, 6.15 and 7.00 were coupling each other; 13c NMR (75.5 MHz, CDC13) 5 157.0, 138.3, 129.4, 127.8, 127.2, 127.1, 125.6, 125.3, 124.4, 121.8. 120.7, 119.8, 119.7; MS (EI) rule 284 (M++2), 282 (M+), 204, 203, 202, 200,101. Irradiation of 9-Diazofluorene in Propargyl Bromide at Room Temperature. A solution of 9-diazofluorene (101.2 mg, 0.53 mmol) in propargyl bromide (5 mL, 80% w/w in toluene) was prepared in an NMR tube. The solution was deaerated by three freeze-pump-thaw cycles. The setup of irradiation was the same as described above and the solution was irradiated for 2 h. After removal of solvent, the residue was separated with flash chromatography eluting over silica gel with methylene chloride/hexane (20%). Only 2-bromo-3,9'-fluorenylidenyl-propene (2.18) has been isolated. Irradiation of 9-Diazofluorene in Propargyl Chloride at 77 K. A solution of 9-diazofluorene (19.0 mg, 0.099 mmol) in propargyl chloride (0.5 mL) was prepared in an NMR tube. The solution was deaerated by three freeze-pump-thaw cycles. The setup of irradiation was the same described as above. The solution was irradiated for 11 h, with warming to room temperature and mixing every 1.5 h. After removal of solvent, the residue was separated by flash chromatography over silica gel eluting with methylene chloride/hexane (20%). Five compounds were isolated: 2-chloro-3,9'-fluorenylidenyl-propene (2.21): 1H NMR (300 MHz, CDC13) 8 5.76 (s, 1 H), 5.79 (s, 1 H), 6.91 (s, 1 H), 7.21 (d, 1 H, J = 7.8 Hz), 7.24-7.46 (m, 3 H), 7.68 (m, 3 H), 8.23 (d, 1 H, J 102 = 7.8 Hz); 13C NMR (75.5 MHz, CDC13) 8 143.8, 140.0, 139.5, 138.5, 137.7, 135.7, 135.3, 129.3, 129.2, 128.0, 127.2, 127.1, 125.8, 125.5. 122.5, 120.6, 120.1, 119.8, 119.7, 117.9; MS (EI) m/e 240 (M++2), 238 (M+), 236, 204, 203, 202, 200, 101. 1-(chloromethyl)-spiro[cyclopropene- 3,9'-fluorene] (2.22) 1H NMR (300 MHz, CDC13) 8 4.55 (d, 2 H, J =l.5 Hz), 7.22 (d, 2 H, J = 7.2 Hz), 7.34 (td, 2 H, J = 7.3, 1.1 Hz), 7.42 (m, 3 H), 7.87 (d, 2 H, J = 8.7 Hz); 13C NMR (75.5 MHz, CDC13) 8 147.8, 139.9, 126.7, 126.6, 120.8, 119.9, 119.1, 106.1, 39.1, 36.1; MS (EI) m/e 240 (M++2), 238 (M+), 204, 203, 202, 200, 101. 9-chloro-fluorene (2.23): 1H NMR (300 MHz, CDC13) 8 5.75 (s, 1 H), 7.33 (td, 2 H, J = 7.8, 1.5 Hz), 7.46 (td, 2 H, J = 7.8, 1.5 Hz), 7.66 (d, 2 H, J = 7.8 Hz), 7.82 (d, 2 H, J = 7.8 Hz); 13C NMR (75.5 MHz, CDC13) 8 157.0, 140.0, 129.3, 128.0, 125.8, 120.1, 57.5; MS (EI) m/e 202 (M++2), 200 (M+), 180, 166, 165, 163, 83, 82. bifluorenylidene (2.5): 1H NMR (300 MHz, CDC13) 8 7.19 (t, 2 H, J = 7.1 Hz), 7.31 (t, 2 H, J = 7.1 Hz), 7.68 (d, 2 H, J = 8.1 Hz), 8.38 (d, 2 H, 8.1 Hz); 13C NMR (75.5 MHz, CDC13) 8 141.3, 138.2, 129.1, 126.8, 126.7, 119.9; MS (EI) m/e 329 (M++1), 328 (M+), 327, 326, 324, 164, 163, 162, 150, 149. And there was trace amount of 9- fluorenone (2.4). Irradiation of 9-Diazofluorene in Propargyl Chloride at Room Temperature. A solution of 9-diazofluorene (56.1 mg, 0.29 mmol) in propargyl chloride (5 mL) was prepared in an NMR tube. The solution was deaerated by three freeze-pump thaw cycles. The solution was irradiated as described above for 2 h. After removal of solvent, the residue was separated by flash chromatography over silica gel eluting with methylene chloride/hexane (20%). Besides the same five products as 103 above, there was another product: 3-chloro-3,9'-fluorenyl-propyne (2.24): 1H NMR (300 MHz, CDC13) 8 2.49 (d, 1 H, J = 2.4 Hz), 4.39 (d, 1 H, J = 4.1 Hz), 5.20 (dd, 1 H, J = 4.0, 2.4 Hz), 7.38 (t, 2 H, J = 7.5 Hz), 7.43 (t, 2 H, J = 7.5 Hz), 7.72-7.82 (m, 4 H); 13C NMR (75.5 MHz, CDC13) 8 141.8, 138.6, 128.6, 128.3 (doublet), 127.2 (doublet), 125.1, 119.9 (doublet), 80.0, 75.5, 52.8, 49.8; MS (EI) m/e 240 (MM-2), 238(M+), 203, 202, 200, 166, 165. IR cm'1 3291 (m), 1450 (s). Bromide Exchange of l-(Chloromethyl)- spiro[cyclopropene-3,9'-fluorene] (2.22). A solution of the cyclopropene (2.22) (9.6 mg, 0.04 mmol) and lithium bromide (86.9 mg, 1 mmol) in methanol (10 mL) was refluxed for 10 hours. After removal of the solvent, the residue was directly subjected to NMR. Only 1- (chloromethyl)-spiro[cyclopropene-3,9'-fluorene] (2.22) and 2-bromo- 3,9'-fluorenylidenyl-propene (2.18) were found. Yield Studies. A solution of 9-diazofluorene (14.6 g, 0.076 mmol), propargyl chloride (3 mL) and 1,2-dichloroethane (3 mL) was prepared. The degassing and irradiation procedures were the same as described above for 77 K studies. After removal of the solvent, the residue was transferred to an NMR tube and toluene (2 11L) was injected into the tube prior to analysis. The yields were determined by 1H NMR integration relative to toluene (8 2.36 ppm) in CDC13. The results are summarized in Table 2.1. A solution of 9-diazofluorene (11.2 g, 0.058 mmol) with various ratios of propargyl chloride and 1,2-dichloroethane were prepared. The degassing and irradiation procedures were the same as described above 104 for room temperature studies. After removal of the solvent, the residue was transferred to an NMR tube and toluene (2 uL) was injected into the tube prior to analysis. The yields were determined by 1H NMR integration relative to toluene (8 2.36 ppm) in CDC13. The results are summarized in Table 2.1. 2.6 References and Notes: (1) Baron, W. J.; Hendrick, M. E.; Jones, M., Jr. J. Am. Chem. Soc. 1973, 95, 6286-6294. (2) Baron, W. J.; DeCamp, M. R.; Hendrick, M. B.; Jones, M., Jr.; Levin, R. H.; Sohn, M. B. In Carbenes; M. Jones Jr. and R. A. Moss, Ed.; Wiley: New York, 1973; Vol. I; pp 73-84. (3) Griller, D.; Nazran, A. S.; Scaiano, J. C. Acc. Chem. Res. 1984, I 7, 283-289. (4) Platz, M. S. In Kinetics and Spectroscopy of Carbenes and Biradr'cals; M. S. Platz, Ed.; Plenum: New York, 1990; pp 143. (5) Savino, T. G.; Senthilnathan, V. P.; Platz, M. S. Tetrahedron 1986, 42, 2167-2180. (6) Wright, B. B.; Platz, M. S. J. Am. Chem. Soc. 1984, 106, 4175- 4180. (7) Turro, N. J. Tetrahedron 1982, 38, 809. 105 (8) Grasse, P. B.; Brauer, B.-E.; Zupancic, J. J.; Kaufmann, K. J.; Schuster, G. B. J. Am. Chem. Soc. 1983, 105, 6833-6845. (9) Moss, R. A.; Joyce, M. A. J. Am. Chem. Soc. 1978, 100, 4475- 4480. (10) Ruzicka, J.; Leyva, E.; Platz, M. S. J. Am. Chem. Soc. 1992, 114, 897-905. (11) We have obtained the crystal structure for compound 2.13a which is quite similar to the optimized structure by the MNDO method but, we were unable to crystallize the 3,3-diphenyl-5 -phenyl-3H-pyrazole. In order to make a fair comparison, both geometries shown in Figure 2.4 are optimized by MNDO method. (12) Jackson, J. B.; O'Brien, T., Jr. J. Phys. Chem. 1988, 92, 2686-2696. (13) Leyva, B.; Barcus, R. L.; Platz, M. S. J. Am. Chem. Soc. 1986, 108, 7786-7788. (14) Gaspar, P. P.; Lin, C.-T.; Dunbar, W.; Mack, D. P.; Balasubramanian, P. J. Am. Chem. Soc. 1984, 106, 2128-2139. (15) Moss, R. A.; Dolling, U.-H. J. Am. Chem. Soc. 1971, 93, 954. CHAPTER 3 We Abstract: Our initial work on oxygen atom transfers to fluorenylidene from various donors is described. Product studies in benzene and acetonitrile have established that fluorenylidene reacts with pyridine-N-oxide to make 9-fluorenone in high yield. Competition experiments using methanol in acetonitrile over a wide concentration range have shown that the rate constant for oxygen atom transfer to fluorenylidene from pyridine-N-oxide (PNO) exceeds that for O-H insertion into methanol (kpNo/kMeoH) by a factor of 1.7 i 0.4; together with the literature value for koon, this result yields an absolute rate constant of kpNo = 1.5 :1: 0.5 x 109 M'ls'l. Fluorenylidene is also oxygenated by other substrates: N-methyl morpholine-N-oxide, sulfolane, trimethyl phosphate, and dimethyl carbonate. Unlike the first three, dimethyl carbonate (DMC) should yield not a stable deoxygenated product, but rather dimethoxycarbene. Competition with methanol yields a rate ratio of kDMC/kMeOH = 1.1 i 0.3 x 10'3 in dimethyl carbonate solvent. We have not yet demonstrated the presence of products from the stabilized carbene, but the new observations are consistent with the known reactivity of fluorenylidene with carbonyl compounds to form ylides, and with the strong stabilizing effects of donor groups such as methoxy on carbene stability. The net oxygen atom-transfer from one carbene to another represents a new mode of carbene reactivity. 106 107 3.1 Introduction: Atom abstraction is an important type of carbene reaction,1 with hydrogen and chlorine atom-transfer being the two most familiar examples. It is well accepted that a triplet carbene is responsible for the hydrogen atom-transfer,"3 but this is not always true for chlorine atom- transfer. By using CIDNP to study the reactions of many types of carbenes with chloroform, Roth clearly showed that hydrogen abstraction is due to a triplet precursor, whereas the corresponding singlet carbenes preferentially abstract chlorine.‘ Roth also demonstrated that both singlet and triplet states of methylene can abstract chlorine atoms from suitable donors. More recently, Platz et al. have examined the reactions of singlet phenylchlorocarbene and triplet diphenylcarbene with carbon-chlorine bonds using laser flash photolysis techniques.l These workers indicated that the transition states of both reactions have considerable carbene-chlorine bond formation and charge development. They also showed that the rate A 1' 5_ Ph g—R —> Ph 0 ..... R C! O constants for triplet diphenylcarbene reacting with chlorine donors are roughly one order of magnitude faster than those for. singlet phenylchlorocarbene. Extensive reviews have been written about ylide formation resulting from the interaction of a carbene with the lone-pair electrons of heteroatoms such as oxygen, nitrogen, phosphorus, and sulfur. However, there are few cases in which a carbene abstracts a divalent atom or group, 108 such as oxygen, to produce a closed-shell species having a double bond between the former carbenic center and the transferred fragment.” Reactions of carbenes with molecular oxygen have been well studied owing to interest in the intermediate carbonyl oxide (3.1), which plays a role in the mechanism of ozonolysis, and in its isomeric form, the dioxirane (3.2).7 It has also been known for some time that carbon atoms ' 0 R. R 3.2 R 3.1a 3.1b R. 3.1c abstract oxygen from a wide variety of carbonyl compounds to produce carbon monoxide and carbenes.8 These reactions are not understood in full detail but they are enormously exothermic, producing "hot" carbenes which can undergo a variety of thermally activated processes.9 However, there are just a few reported cases in which a free carbene abstracts oxygen from a donor other than molecular oxygen. In 1984, Scaiano and co-workers studied the reactions of triplet diphenylcarbene with nitroxideslo (Scheme 3.1). These reactions involve oxygen-transfer, leading to quantitative yields of benzophenone. In the case of 4-hydroxy-2,2,6,6-tetramethylpiperidine-N-oxide, the triplet diphenylcarbene deoxygenates the nitroxide center instead of inserting into the O-H bond. Field and Schuster later showed that triplet anthronylidene is oxidized by molecular oxygen at a nearly diffusion-controlled rate which is two orders of magnitude faster than oxidation of the same carbene by pyridine-N-oxide.11 In a study of the reactions of epoxides with fluorenylidene, Shields and Schuster observed equimolar amounts of .. o- /“ if mesa 9-fluorenone and the respective alkene resulting from the stereospecific deoxygenation of the epoxide.12 These workers also found that diphenylcarbene does not efficiently abstract oxygen from epoxides. They concluded that the deoxygenation of epoxides was proceeding by addition of the singlet state of the carbene to the epoxides to form an oxonium ylide which rearranges to give a carbonyl compound and an olefin with retained stereochemistry (Scheme 3.2). W ccogA: \Z ”8801: 1 10 In this work, other oxygen donors were sought which could be deoxygenated by fluorenylidene. As shown in earlier studies, different spin states of carbenes show different preferences for oxygen donors. For example, triplet diphenylcarbene abstracts an oxygen atom from doublet TEMPO or triplet molecular oxygen, whereas singlet fluorenylidene reacts with closed-shell oxygen-donors such as epoxides. This bias suggests that oxygen atom-transfer could be used to determine singlet versus triplet pathways in reactions of arylcarbenes which have small singlet-triplet energy gaps. 3.2 Results: Product Studies. Photolyses of 9-diazofluorene with pyridine-N-oxide or 4-picoline-N-oxide in dry, degassed acetonitrile afforded 9-fluorenone as the major product (> 70%), and bifluorenyl and bifluorenylidene as minor byproducts. Similar results were obtained with other oxygen donors, such as sulfolane and N-methyl morpholine-N-oxide, examined similarly or used neat as solvents (sulfolane and trimethyl phosphate). Reactions carried out in dry benzene gave two additional minor byproducts (< 5%): triphenylene (3.3)13 and 9,9':9,9"-terfluorene (3.4).14'16 111 In light of the stability of dialkoxy- (3.5) and diamino- (3.6) carbenes, we extended our study to include dimethyl carbonate, tetramethylurea, and 1,3—dimethylimidazolin-2-one. As with the more traditional oxidants, fluorenylidene reacted with these substrates to give 9- fluorenone; the corresponding stabilized carbenes are implied as by- products. Control Experiments. To show that pyridine-N-oxide does not oxidize the fluorenylidene - acetonitrile ylide (3.7) to give 9- fluorenone, we independently synthesized 2,2-(2,2'-biphenylene)-3-methyl- 2H-azirine (3.11),17 a direct photolytic precursor to 3.7, by the route shown in Scheme 3.4. No 9-fluorenone was detected from photolysis of the azirine in acetonitrile in the presence of pyridine-N-oxide18 (Scheme 3.5). Rate Studies. Estimates of the absolute rates of reaction of fluorenylidene with all oxygen-donors were made using the competition method with methanol and the literature value for the methanol quenching i. CH3CH(Br)C02(‘4H5, Zn, benzene-ether (4:1). ii. TsOH, benzene. iii. KOH, H20-EtOH (1:1). iv. Brz, CC14. v. KOH (0.2 N). vi. NaN3,DMF. vii. benzene, reflux. rate of fluorenylidene in acetonitrile (ko0H= 8.95 x 103 M-1 s' 1).19 Ratios of 9-fluorenone to 9-methoxyfluorene products were determined either by NMR or GC analysis. For all oxygen-donors, except dimethyl carbonate and trimethyl phosphate, the studies were run in dry acetonitrile solvent. A range of absolute concentrations and concentration ratios of oxygen- donor to methanol were examined to ensure that the observed product ratios reflected only direct reactions of fluorenylidene with the substrates and not the same products formed via indirect pathways. The results are summarized in Table 3.1. 113 WWW Elunmnxlirlens Substrate ksub/kumw of evaluations) ksybm-l s-l) methanol (1.0) 8.95 x 108 pyridine-N-oxide 1.7 :t 0.4 (30) 1.5 x 109 4-picoline-N-oxide 1.7 :1: 0.1 (3) 1.5 x 109 1,3-dimethylimidazolin-2—one 5.2 :1: 0.2 x 10'1 (9) 4.7 x 108 tetramethylurea 4.9 :1: 0.2 x 10'1 (9) 4.4 x 108 trimethyl phosphate 1.4 :1: 0.1 x 10'2 (3) 1.3 x 107 dimethyl carbonate 1.2 :l: 0.2 x 10'3 (3) 1.0 x 106 cis-2,3-epoxybutane12 3.1 x 108 trans-2,3-epoxybutane12 9.3 x 108 Photolyses of diphenyldiazomethane with pyridine-N-oxide in acetonitrile were carried out similarly, and the rate constant of oxygen transfer to diphenylcarbene was estimated from the ketone/ether product ratio (or oxygen transfer over O-H insertion) obtained from the 1H NMR spectra, along with the literature value for the methanol quenching rate of diphenylcarbene in acetonitrile (kMe0H= 2.4 x 107 M“1 8'1).20 The rate constant for oxygen transfer from pyridine -N-oxide to diphenylcarbene is essentially the same as that for O-H insertion (k = 2.4 x 107 M-1 8'1). 3.3 Discussion: Despite the fact that the ground state of fluorenylidene is a triplet, it exhibits significant singlet reactivity under most experimental conditions at room temperature due to a small singlet-triplet energy gap 114 (~1.1 kcallmol in CH3CN).“W'23 Therefore, it is assumed that the singlet fluorenylidene is responsible in the spin-allowed deoxygenation reactions decribed here. Although it is well-known that singlet carbenes often react with non-bonding electron pairs to form ylide?” none of these ylide species are known to undergo the atom-transfer process. Oxygen atom abstraction from pyridine-N-oxide, which is believed to be an extraordinary singlet carbene quencher,25-27 has been known for years. In this work, a rate constant of 1.5 x 109 M'ls'1 was obtained for the reaction between fluorenylidene and pyridine-N-oxide. This was the only substrate studied that was found to be faster than methanol in reaction with fluorenylidene. As shown in Table 3.2, the oxygen atom-transfer from pyridine-N-oxide to fluorenylidene is quite exothermic (ca. 124 kcallmol) (Scheme 3.6). This is not surprising, since the C=O double bond in 9-fluorenone is much stronger than the N+-O'bond in pyridine-N-oxide. Since it is known that pyridine-N-oxide can also Schema (5g 1 *9“:- function as an oxidizing agent, we have photolyzed azirine (3.7) with pyridine-N-oxide in acetonitrile in order to rule out oxidation of the acetonitrile ylide 3.8 as a source of 9-fluorenone. No 9-fluorenone was found in these reactions (Scheme 3.5). So, there is no 0 . r r . - - O I .t .: a run: it t ,0: :H . I (AMl Calculated Values in Parentheses) x: AHrrxoa AHf(XO)a Differenceb Aern 1F]: 1560 (154) 13(1 (34) 143,121d(120) 0.0 C5H5N: 33 (32) 14 (40) 19 (-8) -124 (428) (MeO)2C: -55,-61° (-53) -139 (-137) 80 (84) e3 (-36) (Me2N)2C: (40.0) (-24) (64) (-56) (MeO)3P: 467 (489) -265 (059) 98 (70) -45 5-50) a Unless otherwise noted, these are AHf values at 298 K, taken from Lias, S. G.; Bartrness, J. B.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.; Mallard, W. G.; “Gas Phase Ion and Neutral Thermochemistry” J. Phys. Chem. Ref. Data 1988, I 7, Suppl. 1. We thank Professor J. A. Allison for generously sharing his copy of this essential document with us. b Note that the AHf of oxygen atom is 59.6 kcallmol at 298 K; the c=o bond strengths can then be calculated by adding this quantity + 3/2RT to the values listed in the difference column. Thus, for example, the C=O bond strength in CH20 is 128 + 60 + 1 = 189 kcallmol. C Li, Y.; Schuster, G. B. J. Org. Chem. 1988, 53, 1273. This number is based on computational estimates, so it cannot be consider an independent experimental value. <1 Sabbah, R.; Watik, L. B.; Minadakis, C. Comptes Rendus de I’Academie des Sciences de Paris, 307, Serie II, 239 (1988). 0 El-Saidi, M.; Kassam, K.; Pole, D. L.; Tadey, T.; Warkentin, J. J. Am. Chem. Soc. 1992, 114, 8751. 1 16 doubt that the atom-transfer from pyridine-N-oxide to fluorenylidene is the major source of 9-fluorenone. Although carbonyl ylides have been known for some time,” no one has reported the formation of a secondary carbene by fragmentation of a carbonyl ylide intermediate generated by carbene addition to a carbonyl compound. However, fragmentations of oxiranes through carbonyl ylide intermediates, as depicted in Scheme 3.7, have been extensively studied experimentally”32 and theoretically.33'37 Contradictions are found between experimental results32 obtained for C-O bond cleavage pathways in substituted oxiranes and predictions based on theoretical calculations.34 The calculations suggest that the :‘A: &:‘/§\ 4.12 ' 4.10. 653%., 55¢ a R = H 0": I”: b R = CI'I3 .00. '1 0’0 : c R=Br .' ' ' ' Rur—w' vCN Rue—wen H0 4.1313 R H0 4. 1311' R Transrtinnm mumm 4.133. 4.13131 olefins AHf AHf AEa AHf AEa AHf AEa AHf AEa ethylene 96.4 104.8 8.4 4.1031 74.0 84.0 10.0 84.2 10.2 86.9 12.9 86.8 12.8 4.10b 61.7 76.3 14.6 76.2 14.5 79.8 18.1 79.6 17.9 4.10c 85.6 98.9 13.3 98.2 12.6 103.6 18.0 103.0 17.4 * All values are in kcallmol. 21“ .‘:‘1 111 11.11.51 11 ItflK‘1 :1‘H11 '14 ' 1 31111 11 .11 111 '1 W3 Br H e H 3,31’ C.’ C. | I ‘. I, R I 0. "CN H0 4.15A R H0 4 15A' R . / Rh=M|CN ’C‘Br + H0 R ' 12 BK . 4 4.10 BI‘KC H‘C a' R = H 0": ’00: b. R = CH3 o." '1 o" '1 c. R = Br . ' ' .o' ' lily—v RH!" T°'S wmm 41.53. 4153'. olefins AH: AH: 113:. AH: AEa AH: AEa AH: 1133 ethylene 113.4 121.2 7.8 4.1011 90.7 100.7 10.0 100.4 9.6 103.0 12.3 103.7 13.0 4.101) 78.3 92.8 14.5 92.3 14.0 96.2 17.9 96.0 17.7 4.10c 102.3 114.7 12.4 113.8 11.5 119.3 17.0 118.8 16.5 * All values are in kcallmol. :1-1 .-‘:~1 111:11“ 11.11 11 1'2'4‘1 :VI‘H11 ‘ "1 \1‘1 11 1 ‘;11”11 1 W CH F e 39 ”CH: C.’ c... 1 ' .‘ 11 ' .‘ 11CN H0 4.17A R H0 4.17A' R F Rh'=‘n||CN . CTGCH: HO 4 10 R Ft ”3‘; . . H3C‘C Pic a. R = H b. R = CH3 0". : ’0' '1 C. R = Br 0.. ' o" ' RIH" RH"! H0 4.17B R H0 4.17B' R Tmnsifinnmm olefins AHf ethylene 23.6 39.0 15.4 4.1011. 1.6 16.6 15.0 4.10b -11.5 11.0 22.5 4.10c 13.1 29.8 16.7 * All values are in kcallmol. 17.6 16.0 12.5 24.0 32.3 19.2 22.6 21.0 21.2 32.7 39.3 26.2 AHf AEa AHf AEa AHf AEa AHf AEa 22.7 21.2 18.1 29.6 39.6 26.3 21' I" 1 111.11* 11; 11 I J-tx’1 :1”11 1 1111 11',11 111 1 W C1 3’0 ole-13C C1 . / Rn 11111CN 1 .. + ,o 1 0 CV (31+ HOP‘ R —' 1 .0. e" 1' 4.18 4.10 RlyL—‘WCN ‘ ' 1101 a. R = H H0 4.1% R H0 4.1913 R b. R = CH3 0. R = Br I . . S 5.1.6.1443 AJ2L 412B. olefins AHf AHf AB; AHf AEL ethylene 73.0 85.7 12.7 4.1011 50.3 68.4 16.1 68.2 15.9 4.1011 38.7 62.6 25.9 63.3 24.6 4.10c 62.0 84.2 22.2 87.5 24.5 * All values are in kcallmol 156 For the addition of methyl fluorocarbene to various olefins (see Table 4.12), energy differences between carbene attacks on the two ends of the olefins exceed those for the above monosubstituted carbenes. 4.17A is preferred by 6.0, 9.8, and 9.5 kcallmol over 4.17B for 4.10a, 4.10b, and 4.10c, respectively. Sterically, the energy gaps between 4.17A and 4.17A' increase with alkene bulky group size, yielding barrier differences of 1.0, 1.5, and 2.5 kcal/mol for R = H (4.10a), CH3 (4.10b), and Br (4.10c), respectively, in favor of 4.17A over 4.17A'. In contrast to all previous carbenes, dichlorocarbene does not show a significant preference between the two ends of olefins (see Table 4.13). This behavior is similar to that found for :CH2; this similarity is sensible, as these two carbenes are the most electrophilic of those studied. Transition Structure Geometries. The selected geometric parameters for transition structures for the addition of methylcarbene, chlorocarbene, bromocarbene, methyl fluorocarbene, and dichlorocarbene are listed in Tables 4.14-4.18, respectively. For each carbene addition to the different olefins we examine here, transition structures A and A' are almost identical; likewise B and B' are also very similar. The shortest C- C bond distances from carbene center to olefin in all transition structures (parameter a in the Tables) decrease generally along the series :CHCH3 > :CHBr > :CHCl > :CFCH3, from 2.2311 down to 2.0311. Along the same series, the distance from the carbene center to the middle point of olefin (parameter r in the Tables) also decreases from 2.38A to 2.22/X. For a given carbene, the distances also gradually decrease with increasing size of the bulky groups on the olefins. C313 ‘5. 13 4% as Rufllc—fi-ICN R11 ' ' "ICN H0 4911 R H0 4.9A' R ..9....I-£ .63.; H c H ‘ ~ 3 3:19 15 .39 ”of”. l: a bo’i‘. . a. a RV'LéyWCN H0 4.9B R “0 4.913' R Push-pull olefin 4.1011 (R = H) arameters ethylene 4.9A 4.9A' 4.9B 4.9B' a (A) 2.266 2.233 2.229 2.175 2.183 b (A) 2.683 2.694 2.680 2.725 2.725 c (A) 1.348 1.383 1.379 1.376 1.376 r (A) 2.390 2.376 2.366 2.367 2.371 a (degree) 71.3 69.8 70.2 65.6 65.9 13 (degree) 120.8 138.8 129.4 122.7 121.3 Push-pull olefin 4.10b (R = CH3 ) arameters eth lene 4.9A 4.9A' 4.9B 4.9B' a (X) 2.266 2.210 2.208 2.172 2.169 b (A) 2.683 2.604 2.618 2.646 2.654 c (A) 1.348 1.405 1.406 1.397 1.397 r (A) 2.390 2.310 2.317 2.318 2.321 01 (degree) 71.3 73.0 72.4 69.4 68.9 13 (degree) 120.8 137.0 137.5 123.7 124.5 158 Push-pull olefin 4.10c ( R = Br) arameters eth lene 4.9A 4.9A' 4.9B 4.9B' a (X) 2.266 2.219 2.227 2.173 2.171 b (A) 2.683 2.745 2.713 2.711 2.695 c (A) 1.348 1.386 1.381 1.376 1.379 r (A) 2.390 2.398 2.384 2.358 2.348 01 (degree) 71.3 66.8 68.6 66.0 66.8 [3 (degree) 120.8 150.2 137.6 137.4 129.9 '1 Q E E a" 0 g . U 0 O -----a O a ‘1’ t. a .b Rwich'Q"CN R“ ‘. WCN H0 4.13A R H0 4.13A'R 3.1-C 13 Inc bf: :' a "1;" {'a RWWCN RHE'VWCN H0 4.133 R “0 4.133' R Push-pull olefin 4.10a (R = H) parameters ethylene 4.13A 4.13A' 4.13B 4.13B' a (A) 2.205 2.162 2.166 2.126 2.111 b (A) 2.644 2.592 2.572 2.708 2.722 c (A) 1.350 1.382 1.382 1.381 1.382 r (A) 2.339 2.284 2.275 2.335 2.336 01 (degree) 70.3 71.0 72.2 64.0 62.8 13 (degree) 114.5 127.0 125.9 116.7 113.0 Push-pull olefin 4.1011 (R = CH3) parameters ethylene 4.13A 4.13A' 4.13B 4.13B' a (A) 2.205 2.145 2.143 2.104 2.100 b (A) 2.644 2.516 2.524 2.634 2.646 c (A) 1.350 1.404 1.404 1.403 1.403 r (A) 2.339 2.230 2.233 2.278 2.283 a (degree) 70.3 74.0 73.6 66.9 66.0 [3 (degree) 114.5 127.5 128.0 116.2 115.9 160 Push-pull olefin 4.10c ( R = Br ) arameters eth lene 4.13A 4.13A' 4.133 4.133' a (A) 2.205 2.142 2.140 2.096 2.097 b (A) 2.644 2.596 2.615 2.624 2.643 c (A) 1.350 1.386 1.386 1.382 1.381 r (A) 2.339 2.276 2.287 2.272 2.283 a (degree) 70.3 70.0 69.0 66.6 65.8 13 (degree) 114.5 132.8 134.8 130.5 128.6 ‘I 1 ‘ ‘ 1 1 ' . 1 1 1 1 1 1 1 1 1 W gB' ’0 H go; Br 15be as ‘01» R‘y'fi‘fi—‘QWCN R11 ' c WCN H0 4.15A R H0 4.15A' R .‘ss....I-k .. 96.93%. Br ‘ ° H ~ I}; 13 one o 0"; .00 {a b.,'.:‘.‘.. :a Ry-‘EVWCN RVL-c-Qg'v-UCN H0 4.153 R H0 4153 R Push-pull olefin 4.1011 (R = H) arameters etthlene 4.15A 4.15A' 4.15B 4.15B' a ( ) 2.244 2.207 2.209 2.155 2.153 b (A) 2.637 2.585 2.572 2.751 2.735 c (A) 1.348 1.384 1.384 1.382 1.381 r (A) 2.374 2.302 2.295 2.372 2.362 on (degree) 70.7 73.5 74.2 63.5 64.2 [3 (degree) 115.2 130.9 128.8 111.6 112.2 Push-pull olefin 4.10b (R = CH3 ) parameters egylene 4.15A 4.15A' 4.15B 4.15B' a (A) 2.244 2.186 2.181 2.136 2.140 b (A) 2.637 2.520 2.539 2.659 2.648 c (A) 1.348 1.406 1.406 1.400 1.400 r (A) 2.374 2.251 2.260 2.308 2.303 01 (degree) 70.7 75.6 74.6 67.2 67.9 13 (degree) 115.2 131.7 131.6 116.0 115.4 162 Push-pull olefin 4.10c LR = Br) arameters eth lene 4.15A 4.15A' 4.15B 4.15B' a (X) 2.244 2.174 2.168 2.130 2.132 b (A) 2.637 2.637 2.644 2.649 2.663 c (A) 1.348 1.388 1.387 1.379 1.379 r (A) 2.374 2.315 2.316 2.303 2.312 a (degree) 70.7 69.7 69.0 67.0 66.4 3 (degree) 115.2 139.7 139.7 129.8 129.9 '1 ..---, ‘ ‘93 u '13 O f-1 ‘6‘ 0 gm.“ '0: e .33 ‘ I .0 fl 0 8"?" [a bf?" :31 RWRE'WCN RH?“ H0 4.173 R “0 4.1731 R Push-pull olefin 4.10a (R = H) arameters ethLlene 4.17A 4.17A' 4.173 4.1731 a ( ) 2.048 2.032 2.029 1.991 2.000 b (A) 2.603 2.594 2.573 2.657 2.634 c (A) 1.360 1.396 1.393 1.391 1.390 r (A) 2.241 2.223 2.210 2.242 2.233 a (degree) 65.0 65.2 66.0 60.2 61.7 13 (degree) 126.0 140.0 134.8 125.4 126.0 Push-pull olefin 4.10b (R = CH3 ) parameters ethylene 4.17A 4.17A' 4.17B 4.17B' a (A) 2.048 2.017 2.014 1.906 1.993 b (A) 2.603 2.542 2.541 2.657 2.584 c (A) 1.360 1.421 1.421 1.431 1.412 r (A) 2.241 2.182 2.180 2.200 2.197 a (degree) 65.0 67.3 67.2 57.0 64.2 13 (degree) 126.0 139.8 140.3 130.2 127.3 164 Push-pull olefin 4.10c (R = Br) arameters eth lene 4.17A 4.17A' 4.17B 4.17B' a (X) 2.048 2.025 2.030 1.981 1.995 b (A) 2.603 2.663 2.603 2.643 2.611 c (A) 1.360 1.397 1.395 1.392 1.392 r (A) 2.241 2.260 2.227 2.224 2.217 a (degree) 65.0 61.7 64.7 60.9 62.6 13 (degree) 126.0 151.2 140.6 135.7 132.6 Push-pull olefin 4.10a (R = H) parameters ethylene 4.19A 4.19B a (A) 2.040 2.004 2.008 b (A) 2.640 2.509 2.756 c (A) 1.360 1.397 1.400 r (A) 2.259 2.161 2.307 a (degree) 62.8 67.8 56.5 B (degree) 120.2 133.7 113.0 Push-pull olefin 4.10b (R = CH3) parameters ethylene 4.19A 4.19B a (A) 2.040 2.003 1.980 b (A) 2.640 2.516 2.669 c (A) 1.360 1.414 1.420 r (A) 2.259 2.161 2.240 a (degree) 62.8 67.7 59.8 B (degree) 120.2 132.3 117.5 Push-pull olefin 4.10c Q = Br) parameters ethylene 4.19A 4.19B a (A) 2.040 1.993 1.964 b (A) 2.640 2.546 2.683 c (A) 1.360 1.400 1.404 r (A) 2.259 2.177 2.244 a (degree) 62.8 65.7 58.0 B (degree) 120.2 140.0 120.0 166 The positions of the carbene centers in the transition structures are (angle a) all off—center by 15-30‘. For monosubstituted carbenes, the angles a for transition structures A and A' are always 4-10' larger than those for the corresponding transition structures B and B'; however, the difference decreases as the size of the R groups on the olefins increases. For disubstituted carbenes, the carbene centers are still off-center by 25- 30', but the angles are generally more acute than for the monosubstituted carbenes, and the differences between transition structures A and A' and transition structures B and B' are also general by 2-3' less than those of transition structures of monosubstituted carbenes. For the transition structures of the carbenes adding to the hydroxy substituted carbon atom of 4.10 (TS A) , the tilt angles (B) of the carbene planes, in general, are garound 10-20' larger than the B for the carbenes adding to the cyano substituted carbon atom (TS B), and the angle also slowly increases along the series of carbenes :CHCl < :CHBr < :CC12 < :CHCH3 < :CFCH3 . For the transition structures of each carbene, the angle B also increases with the size of the R group on the olefins H < CH3 < Br. For the T88 B of carbene addition to olefins 4.10a and 4.10b, the angles B are quite similar to the corresponding parent unsubstituted olefin system, however, the angles for addition to 4.10c increase by 10- 15'. 4.4 Discussion: The reported theoretical investigations of carbene cycloaddition to ethylene all agree that the reaction path is nonsymmetric, 167 only C. symmetry (or pseudo symmetry in e.g. propene rxn) is preserved, and the reaction starts with an electrophilic phase and ends with a nucleophilic phase. The energy maximum occurs around the transition between the two phases. A monosubstituted olefin with electron-donor or electron-acceptor, which will increases the energy of HOMO or decreases the energy of LUMO of the olefin, respectively, may influence the reaction path of the carbene cycloaddition to the olefin differently. That is the initial reason why we set out to examine the transition structures of cycloaddition of carbenes to cyanoethylene and hydroxyethylene first. Not surprisingly, the barrier for methylene, the most electrophilic carbene, adding to hydroxyethylene, the electron-rich olefin, is lower than the barrier for addition to plain ethylene for both orientations (4.1A and 4.1B). But for the cycloaddition of other carbenes to the olefin, the orientation of carbenes in transition structures either have no effect, as in the case of :CHCl, or are opposite to our expectation, which is that the carbene should add to the substituted carbon atom, allowing its empty p orbital to interact with the larger x-orbital coefficient on the unsubstituted end. Rondan and Houk have also found that the interaction between the LUMO of donor-substituted olefin and the HOMO of a carbene has no influence on the orientation of the carbene at the transition structure.8 They conclude that the LUMO of an unsymmetrically substituted alkene is polarized insignificantly by donors. In other word, as mentioned previously, the semiempirical reaction path starts with pure carbene LUMO - olefin HOMO interaction at large a, as shown in Figure 4.3, in which the p orbital of the carbene sits right above one of orbitals of 168 the olefin's 1: system; then the carbene center begins to slide toward the center of olefin as the reaction progresses.” So, for hydroxyethylene, in which the unsubstituted carbon atom has the larger HOMO coefficient, the early carbene LUMO - olefin HOMO interaction already determines the favored reaction path and orientation at the transition structures. Another way to illustrate this idea is to note that the geometries of the favored transition structures, in this case TSs B, are quite similar to the geometry of the corresponding transition structure of the carbene addition to plain ethylene; moreover, the more nucleophilic the carbene is, the smaller is the energy gap between TSs A and TSs B, as shown in Tables 4.1- 4.4 (i.e. the alkene's LUMO may begin to participate a little). If the olefin's HOMO polarization does not determine the orientation of a carbene at the transition structure, it must be controlled by the different coefficients between the two ends in the olefin's LUMO. Except for the transition structures for methylene addition to cyanoethylene, the orientation of the carbene in transition structures of all other carbenes adding to cyanoethylene follows our predictions, by which the carbene adds toward the unsubstituted carbon atom. Obviously, the 169 carbene LUMO - olefin HOMO interaction still has some influence on the orientation of carbene in the transition structures. The more electrophilic the carbene is, the smaller is the gap between TSs A and T88 B, as shown in Tables 4.5 - 4.8: Nevertheless, the transition structures of methylene addition to cyanoethylene are still affected by the interaction: TS B is favored over TS A by 3.3 kcallmol, as shown in Table 4.5. Another interesting feature should be noted: All the carbenes examined here are electrophilic carbenes, yet the barriers for almost all favored transition structures, TSs A, of carbenes adding to cyanoethylene, an electron-poor olefin, are lower than the corresponding barriers for carbene addition to plain ethylene. Meanwhile, the transition structures for those carbenes adding to hydroxyethylene, an electron-rich olefin, do not show much lowering relative to the ethylene barriers. It is clear that the carbene HOMO - olefin LUMO interaction, or the energy gap between these two orbitals, is a more important factor in determining the location of the transition structure (includes the barrier and the orientation of a carbene) than is the carbene LUMO - olefin HOMO interaction, at least for the electrophilic carbenes examined here. We have shown that individual electron-donor and electron- acceptor substituents on an olefin have different effects on the transition structure of cycloaddition of a carbene to the olefin. But such monosubstitution also changes the character of ethylene to electron-rich or electron-poor, plus, they do not polarize the HOMO and LUMO of ethylene to equal extents. So, we further investigated the transition structures of cycloadditions of carbenes to donor, acceptor di-substituted or "push-pull" olefins, as shown in Figures 4.1 and 4.2. In order to test the 170 "size" effect, we also examined three different size bulky groups, H, CH3, and Br, on the olefin. For the same reason, we added another monosubstituted carbene, :CHBr, besides those carbenes discussed above. Obviously, it would be meaningless to further pursue the reaction of methylene, which is extremely electrophilic and so dominated by the carbene LUMO - olefin HOMO interaction. For transition structures for all carbenes, strong electronic preferences are clearly seen, as shown in Tables 4.9 - 4.13. The differences of energy barriers between TS As and TS Bs increase as the electrophilicity of the carbenes decreases due to the reduced influence of carbene LUMO - olefin HOMO interactions, as discussed previously. This is the same reason why the energy barriers for transition structures of these carbenes adding to the olefin are extraordinarily high when R = CH3, which is almost the same size as Br. The methyl groups (4.10b), which are weak electron-donors, raise the energy of the olefin's HOMO, whereas bromine (4.10c), a weak electron acceptor, decreases the LUMO energy of the olefin, as shown in Table 4.13. Even though CH3 and Br have roughly the same size, electronically, they change the character of the olefin in opposite directions. It is also well known that the addition of singlet carbenes to olefins is sensitive to steric effects.17"3 That is why both bulky substituents increase the energy barriers in comparison with the corresponding parent system. : 1 ' ' ' 1 1 . 01. : 1 ‘ " ‘ ' 1 0 ‘ C=C C=OOH C=CCN 4.10a 4.10b 4.10c LUMO +1.318 +1288 40.022 +0.049 -0.110 0.996 HOMO -10.l76 -9.298 -10.612 -9.785 -9.509 -10.083 ' n 171 However, the steric effects between TS Arum (or A in Tables) and the corresponding TS Asyn (or A' in Tables) are weak or actually opposite to our prediction as shown in Figure 4.2. Nevertheless, for the transition structures of bromocarbene, the carbene with the largest size difference between its two substituents among those calculated, the energy barriers do show the steric effects and the energy difference between 4.15A and 4.15A' increases along the size of the R groups, H, CH3, and Br, on the olefins, but in the direction opposite to that predicted. Further, in examining the geometries of all transition structures, the distance b is always seen to be more than 0.4 A longer than distance a. In addition, the angles B, specially for TSs A, are larger than 130'. In other words, the substituents on the carbene at the transition structure experience more steric effects from the bulky group on the near carbon atom of an olefin than from the bulky group on the other end of the olefin. The substituent on chlorocarbene and methylcarbene may not be big enough to show this steric hindrance. However, this interpretation is difficult to apply on transition structures of :CFCH3, 4.17A and 4.17A', which agree with the prediction shown in Figure 4.2. The size difference between F and CH3 is not as large as the difference between H and Br or H and CH3, but the transition structures of neither of the latter carbenes show the same phenomenon as between 4.17A and 4.17A'. There must be another effect involving in the transition structures of cycloaddition of :CFCH3 to push- pull olefins. It is not very difficult to imagine that the source of the problem must come from the fluorine on the carbene. One possible explanation is that the 4.17A gain extra stabilization from the dipole - 172 dipole interaction between the methyl fluorocarbene and the push-pull olefin, as shown in Scheme 4.1. Since fluoride is the most electronegative atom, it can polarize the carbene to form a significant dipole which is in the opposite direction to the dipole generated in the resonance structure of the push-pull olefin. Plus, when R = Br, an electron-withdrawing group, it also enhances the dipole moment of the olefin. This interpretation is just a postulate which needs more detailed investigation; however, examination of though the individual pieces does show that the direction of their dipole moments are the same as shown inScheme 4.1. N R R H H 3 F 3 3‘ OH 18’ 8H W 4.17A Dichlorocarbene, which is the most electrophilic carbene we have examined here at the MNDO level, is not sensitive to the polarization effect, but does show an electronic effect resulting from the R groups on the olefins which change them into relatively electron-rich (R = CH3) or electron-poor (R = Br) n-systems. 4.5 Conclusion: The MNDO results suggest that a carbene, especially an electrophilic carbene except for methylene, shows an orientational bias in 173 adding between the two ends of a polarized olefin. For a push-pull olefin, a carbene prefers adding to the electron-donor end over the electron- acceptor end of the olefin. The intention to use steric effects to separate between the two polarization favored transition structures (A and A') shows some promise but the prediction needs to be reversed. Overall the picture of the fours TSs appears to need revision to Figure 4.4 from Figure 4.2. W TS AiselectronicallypreferredoverTS B Syn TSs are sterically preferred over anti 174 Part II. Experimental: The previous calculation results suggested that an experiment which make use of a preference for one of the four possible cyclopropanation reaction paths should be possible: the calculated differences of energy barriers between TS A and TS B, as shown in Figure 4.4, are enough to give observably large product ratios. Obviously, the size of the bulky group (R) should be increased in order to enhance the steric preference for syn vs. anti cyclopropanation products (from TS A in Figure 4.4). So, we have designed a push-pull substituted alkene 4.20, with more bulky R groups, gem-dimethyl, that should electronically prefer one carbene approach path (TS A) over the other (TS B). Steric interactions would favor syn-substituted cyclopropane products from the T88 A (as defined in Figure 4.4), so study of the reactions of 4.20 with unsymmetrical carbenes :CXY should reveal the electronic orientation preference in the regiochemistry of the carbene's addition, making a clear distinction between paths. In 4.20, the donor and acceptor groups are made sterically uniform by being moved away from the olefin center and placed on phenyl ring spacers. 175 4.6 Results: Synthesis of 3-Amino-5,5,10,lO-dimethyl-lz-Nitro- [2,l,a] Indene (4.25). The desired push-pull disubstituted olefin (4.25) was synthesized from 2,5-dimethyl-3-hexyne-2,5-diol (4.21) in four steps, as shown in Scheme 4.2. Alkene 4.23 was generated by the condensation of 4.21 and benzene in the presence of sulfuric acid to make 4.22, followed by dehydrogenation with DDQ. 4.23, a white solid with a very strong fluoresence, was then treated with nitric acid in acetic acid followed by mono-reduction with Pd/C and cyclohexene in ethanol to afford the push-pull disubstituted olefin 4.25. W H c H38 Ho 8 H2804, benzene. b DDQ, benzene. c AcOH, HNO3. d Pd/C (10 %), cyclohexene, EtOH. 176 Schema F :ccrL ’ :c(¢ C‘ r NO RXN. .C,C(=O)oc2115 O \H V 10 , Reaction of 4.23 with Various Carbenes (Scheme 4.3) Dimethyl vinylidene was generated from 2,2-dimethylvinyl triflate according to previous literature.19 Dichlorocarbene was generated from the Seyferth reagent20 and from trichloroacetic acid sodium salt. Phenylchlorocarbene was generated from phenylchlorodiazirine.21 Carboethoxycarbene was generated from ethyl diazoacetate by thermolysis in benzene. All reactions of 4.23 with above carbenes were refluxed in dry benzene and products were analyzed either by NMR or GC analysis. Only 4.23 was identified from the reaction mixtures. Reaction of Benzenediazonium-2-Carboxylate-HC1 with 4.23. The benzyne precursor, propylene oxide, and 4.23 were refluxed in bromobenzene for 2 hours, when N2 ceased coming out from the reaction. After the removal of the solvent, the residue was separated by flash chromatography. No adduct from benzyne addition to 4.23 was isolated. 177 Epoxidation of 4.23. A solution of 4.23 and 3- chloroperbenzoic acid (MCPB A) in methylene chloride was ultrasonicated for 4 hours until no starting material remained, as determined by a TLC. The solution was diluted with hexane, washed with saturated aqueous sodium bicarbonate solution and brine, and dried over magnesium sulfate. After the removal of solvent, a white solid (4.26) was obtained in 82% yield and a reasonable purity without further purification. 4.7 Discussion: In spite of numerous efforts, the indenoindene (4.23) fails to react with any of the carbenes tried. However, the reaction of the olefin with MCPBA gives a reasonable yield. Since the whole molecule of MCPBA is on the same plane approaching the double bond of 4.2332»23 it may not "feel" the steric hindrance from the gem-dimethyl groups next to the double bond. However, those bulky groups are evidently too big to let the carbenes reach the double bond even when sterically "small" such as dimethylvinylidene are examined. So, we have to make some modifications to those bulky groups if this approach is going to reveal the symmetry of the reaction path of a carbene. A variety of synthetic efforts aimed at 178 building a simplified, less sterically demanding alkene led to no further success. Cl 4.8 Experimental: General Methods. General experimental procedures are the same as described in Chapter 2. 5,5,10,10-Tetramcthyl-4b,5,9b,10-tetrahydroindeno [2,l,a] indene (4.22). The procedure was modified from Hancock's method.24 To a mixture of cone. sulfuric acid (30 mL) and benzene (48 mL) was added slowly 2,5odimethyl-3-hexyne-2,5-diol (10 g, 70.3 mol) at 0' C over a period of 40 min. After the reaction mixture was refluxed for 7 h, the reaction mixture was cooled to room temperature and poured slowly onto ice (100 g). The aqueous layer was extracted with hexane (3 x 100 mL) and the combined organic extracts were washed with saturated aqueous sodium bicarbonate solution and brine, dried over magnesium sulfate, and filtered. The organic solvents was removed to give the crude 179 brown product, which was purified by flash chromatography over silica gel eluting with neat hexane to afford (4.22) (1.7 g, 9%): 1H NMR (300 MHz, CDC13) 5 0.87 (s, 6 H), 1.53 (s, 6 H), 7.17-7.37 (m, 8 H); 13C NMR (75.5MHz, CDC13) 5 28, 46, 61, 123, 126, 127, 128, 142, 154; MS (El) We 77, 91, 247, (M+) 262. Dehydrogenation of 5,5,10,10-Tetramethyl- 4b,5,9b,lO-tetrahydroindenoI2,l,alindene. A solution of (4.22) (3.8 g, 14.6 mmol), 2,3-dibromo-5,6-dicyano-l,4-benzoquinone (14.0 g, 17 .6 mmol) and a few drops of acetic acid in benzene (15 mL) was refluxed for 38 h. The benzene was removed and the residue was purified by flash chromatography eluting with neat hexane to afford a white solid. The white solid was recrystallized from methanol to give (4.23) (1.6 g, 41%): 1H (300 MHz, acetone-ds) 5 1.5 (s, 6 H), 7.21 (t, 2 H, J = 7.5 Hz), 7.28 (t, 2 H, J = 7.5 Hz), 7.48 (m, 4 H); 13C NMR (75.5 MHz, aceton-da) 5 159.7, 156.3, 138.5, 127.4, 125.7, 122.7, 120.1, 45.2, 24.2; MS (EI) m/e 260.5 (M+), 245.4, 230.3, 215.4. Nitration of 4.23. A suspension of 4.23 (1.0 g, 3.8 mmol) in acetic acid (25 mL) was heated to 50 °C. Concentrated nitric acid (3 mL, 200 mmol) was added dropwise over a period of 15 min and the resulting mixture was then warmed up to 60 - 65 ’C and stirred for another 2 h at that temperature. The mixture was allowed to cool down to room temperature and a yellow solid was precipitated. The yellow solid was collected by suction filtration, washed with cold glacial acetic acid (2 x 10 mL) containing potassium acetate (0.25 g) and washed with water several times. The yellow solid was further dried under vacuum over phosphorus pentoxide to afford 4.24 (1.2 g, 81%): 1H (300 MHz, CDC13) 180 5 1.6 (s, 12 H), 7.51 (dd, 2 H, J = 6.9, 2.2 Hz), 8.26 (dd, 2 H, J = 7.0, 2.2 Hz); 13C NMR (75.5 MHz, CDC13) 5159.9, 159.5, 146.2, 142.8, 123.8, 119.9, 117.8, 45.8, 23.9; MS (131) 111/e 350 (M+), 335, 289, 274, 226. Reduction of 4.24. A solution of 4.24 (0.762 g, 2.2 mmol) and 10% Pd/C (0.3 g) in absolute ethanol (9 mL) was heated to gentle reflux under argon atmosphere and cyclohexene was then added slowly dropwise. The resulting mixture was refluxed for 12 h. The suspension solution was then filtered hot through a celite pad, and the pad was washed with several portions of ethyl acetate (3 x 10 mL). The filtrate was allowed to cool to room temperature and solvent was under vacuum. The residue was purified by flash chromatography over silica gel eluting with methylene chloride/hexane (30%) to afford a white solid 4.25 (0.49 g, 70%). The white solid was recrystallized from ethanol for further purification: 1H NMR (300 MHz, CDC13) 5 0.97 (s, 6 H), 1.68 (s, 6 H), 7.21 (m, 2 H), 7.30 - 7.60 (m, 5 H); 13C NMR (75.5 MHz, CDC13) 5 154.2, 151.6, 135.6, 131.0, 128.2, 125.3, 123.7, 104.9, 48.1, 29.9, 25.2, 25.1; MS (E1) m/e 320 (M4), 305, 261; HRMS: calcd for M+ 320.1525, found 320.1525. Reaction of 2,2-Dimethylvinyl Triflate with Potassium tert-Butoxide in Presence of 4.23. Into a 25 mL round- bottom flask equipped with a magnetic stirrer and argon inlet was added dry CH2C12 (3 mL), 4.23 (0.15 g, 0.57 mmol) and potassium tert- butoxide. The mixture was cooled to -30 °C and 2.2-dimethyl triflate (0.67 mL) was added slowly. The resulting mixture was allowed to warm up to 0 °C and stirred for another 3 h. The reaction was quenched by adding water (20 mL) and the aqueous solution was extracted with benzene (3 x 181 10 mL). The combined organic layer was washed with water and dried with magnesium sulfate. The solvent was removed by vacuum and the residue was directly analyzed by NMR. No addition product was found. Thermolyses of Trichloroacetic Acid Sodium Salt in Presence of 4.23. A solution of 4.23 (0.26 g, 1.0 mmol) and trichloroacetic acid sodium salt in dry ethyl acetate (15 mL) was refluxed for 2 days under argon atmosphere and at mean time a lot of salt was precipitated out. The mixture was allowed to cool to room temperature, filtered out the white precipitate, and analyzed directly by GC directly. Reaction of Seyferth Reagent (Phenyl Bromo- dichloromethyl Mercury) in Presence of 4.23. A tube containing 4.23 (50 mg, 0.2 mmol) and phenyl (bromo-dichloromethyl) mercury (0.1 g, 2.3 mmol) in dry benzene (3 mL) was freeze-pump-thawed three times before sealing. The tube was warmed up to 70 'C with a water bath held for another 24 h. After cooling to room temperature, the mixture was diluted with hexane (30 mL) and filtered through a short column of silica gel. The hexane solution was directly analyzed by GC. Only 4.23 was detected. Thermolyses of Phenylchlorodiazirine in Presence of 4.23. A solution of phenylchlorodiazirine (0.42 g, 2.71 mmol) in dry benzene (25 mL) was added slowly into a refluxing solution of 4.23 (0.31 g, 1.18 mmol) in benzene (5 mL) under a nitrogen atomosphere. After the addition was complete, the mixture was refluxed for another 3 h and then cooled to room temperature. The mixture was concentrated on a rotary evaporator and separated by flash chromatography over silica gel eluting 182 with hexane. Only 4.23 was recovered. No addition product was obtained. Thermolyses of Ethyl Diazoacetate in Presence of 4.23. A solution of 4.23 (0.26 g, 4.33 mmol) and ethyl diazoactate (1.52 g, 1.33 mmol) in benzene (5 mL) was refluxed for 24 h. After the solvent was removed, the residue was purified by flash chromatography over silica gel. Only 4.23 could be detected. Reaction of Benzenediazonium-2-carboxylate-HCI (Benzyne Precursor) with 4.23. A solution of 4.23 (0.26 g, 1.0 mmol), Benzenediazonium-2-carboxylate-HCl (0.2 g, 1.1 mmol) and propylene oxide (0.4 mL) in bromobenzene (10 mL) was refluxed for 2 h. The solution was cooled down to room temperature and the solvent was removed by vacuum. The residue was purified flash chromatography over silica gel . No addition product was found. Epoxidation of 4.23. A solution of 4.23 (0.32 g, 1.24 mmol) and meta-chloroperoxybenzoic acid ( 0.32 g, 1.57 mmol) was ultrasonicated for 4 h, and then the solution was diluted with hexane (80 mL), washed with sat. aqueous sodium bicarbonate solution (3 x 20 mL) and brine, and dried with magnesium sulfate. The solvent was removed under vacuum and a white solid was obtained (0.3 g, 83%): 1H NMR (300 MHz, CDC13) 5 1.4 (s, 6 H), 1.67 (s, 6 H), 7.15 - 7.32 (m, 6 H), 7.52 (d, 1 H, J = 6.6 Hz); 13C NMR (75.5 MHz, CDC13) 5 158.3, 137.6, 128.5, 126.3, 125.3, 124.3, 79.8, 43.6, 28.2, 23.3; MS (BI) m/e 276 (M+), 261, 246, 228, 215, 202. 183 4.9 References And Notes: (1) Moreno, M.; Lluch, J. M.; Oliva, A.; Bertran, J. J. Mol. Struct., THEOCHEM 1984, 107, 227-232. (2) Apeloig, Y.; Kami, M.; Stang, P. J.; Fox, D. P. J. Am. Chem. Soc. 1983. 105. 4781-1792. (3) Bodor, N.; Dewar, M. J. S.; Wasson, J. S. J. Am. Chem. Soc. 1972, 94, 9095-9102. (4) Hoffmann, R. J. Am. Chem. Soc. 1968, 90, 1475-1485. (5) Hoffmann, R.; Hayes, D. M.; Skell, P. S. J. Phys. Chem. 1972, 76, 664-669. (6) Moreno, M.; Lluch, J. M.; Oliva, A.; Bertran, J. J. Chem. Soc., Perkin Trans. 2 1985, 131-134. (7) Rondan, N. G.; Houk, K. N.; Moss, R. A. J. Am. Chem. Soc. 1980, 102, 1770-1716. (8) Rondan, N. G.; Houk, K. N. Tetrahedron Lett. 1984, 25, 5965- 5968. (9) Zurawski, B.; Kutzelnigg, W. J. Am. Chem. Soc. 1978, 100, 2554- 2659. (10) Fox, D. P.; Stang, P. J.; Apeloig, Y.; Kami, M. J. Am. Chem. Soc. 1986, 108, 750-756. (11) Skell, P. S.; Cholod, M. S. J. Am. Chem. Soc. 1969. 91, 7131-7137. 184 (12) Houk, K. N.; Rondan, N. G.; Mareda, J. Tetrahedron 1985, 41, 1555-1563. (13) Moss, R. A.; Mallon, C. B.; Ho, C.-T. J. Am. Chem. Soc. 1977, 99, 4105-4110. (14) Moss, R. A.; Guo, W.; Krogh-Jesperson, K. Tetrahedron Lett. 1982,23, 15-18. (15) Fleming, 1. Frontier Orbitals and Organic Chemical Reactions; Wiley: London, 1976. (16) These two compounds should properly be called vinyl alcohol and acrylonitrile, but for the purposes of this discussion, we have elected to name them as substituted ethylenes. (17) Carbene Chemistry; Kirrnse, W., Ed.; Academic Press: New York, 1971; Vol. 2., Chapter 8, Part I and II. (18) Moss, R. A. In Carbenes; M. J. Jones and R. A. Moss, Ed.; John Wiley & Sons: New York, 197 3; Vol. 1.; Chapter 2. (l9) Stang, J. P.; Mangum, M. G.; Haak, P. J. Am. Chem. Soc. 1974, 96, 4562-4569. (20) Seyferth, D.; Burlitch, J. M. J. Am. Chem. Soc. 1964, 86, 2730- 2731. (21) Moss, R. A.; Whittle, J. R.; Freidenreich, P. J. Org. Chem. 1969, 34, 2220-2224. (22) Woods, K. W.; Beak, P. J. Am. Chem. Soc. 1991, 113, 6281-6283. 185 (23) Hanzlik, R. P.; Shearer, G. O. J. Am. Chem. Soc. 1975, 97, 5231- 5233. (24) Hancock, J. E. H.: Pavia, D. L. J. Org. Chem. 1961, 26, 4350-4352.