The Chemistry of 1,4-Dehydrobenzenes Citation Lockhart, Thomas Paul (1981) The Chemistry of 1,4-Dehydrobenzenes. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/35fh-q724. https://resolver.caltech.edu/CaltechTHESIS:10072025-210733150 Abstract Upon heating, alkyl substituted cis 1,2-diethynyl olefins undergo cyclization to yield reactive 1,4-dehydrobenzenes; the products isolated may be derived from either unimolecular or bimolecular reactions of the intermediate. Z-4,5-Diethynyl-4- octene (19) undergoes rearrangement to yield 2,3-di-n-propyl-1,4- dehydrobenzene (33). Solution pyrolysis of 19 in inert aromatic solvents produces three unimolecular products, (Z-dodeca-4,8- diyn-6-ene (23), benzocyclooctene (25) and o-allyl-n-propylbenzene (26)) in high yield. When 1,4-cyclohexadiene is added to the pyrolysis solution as a trapping agent, high yields of the reduced product o-di-n-propylbenzene (28) are obtained. The kinetics of solution pyrolysis of 19 in the presence and absence of trapping agent establish that 2,3-di-n-propyl-1,4- dehydrobenzene is a discrete intermediate on the pathway leading to products. When the reaction was run in the heated probe of an NMR spectrometer, CIDNP was observed in 26. This observation, along with kinetic and chemical trapping evidence, indicates the presence of two additional intermediates, formed from 33 by sequential intramolecular [1,5] hydrogen transfer, on the pathway to products. The observation of CIDNP, coupled with the reactivity exhibited by 33 and the other two intermediates, implicates a biradical description of these molecules. Two approaches have been used to determine the spin state(s) of 1,4-dehydrobenzenes produced in the solution reaction of diethynyl olefins. The first method relies on the "spin correlation effect" which postulates a relationship between the spin state of a caged radical pair and the ratio of cage and escape reactions (C/E) which may occur in the pair. When the 2,3-di-n-propyl-1,4-dehydrobenzene biradical abstracts hydrogen from 1,4-cyclohexadiene, a radical pair is generated. If a mixture of l,4-cyclohexadiene-d o and -d 4 is employed it is possible, by performing a VPC-MS analysis, to determine the ratio C/E leading from the radical pair to the reduced product, 28. Applying this method to the reaction of 19, C/E was found to be 0.6, independent of the concentration of 1,4-cyclohexadiene (between 0.1 and 10 M) in the chlorobenzene reaction solution. This result suggests the presence of the singlet state of 33 in the reaction of 19. Independent support for this analysis came from the reaction of 3,4-dimethyl-1,5-diyn-3-ene (38) in hexachloroacetone sol vent in an NMR probe. The major product, l,4-dichloro-2,3-dimethylbenzene (39), obtained by chlorine abstraction from the sol vent, showed polarization (emission) in the aromatic protons. The interpretation of this result is straightforward and indicates solvent trapping of the singlet state of the intermediate 2,3-dimethyl-1,4-dehydrobenzene. Both of these experiments indicate that only the singlet state of 1,4- dehydrobenzenes is generated upon thermal reaction of diethynyl olefins. The failure to observe evidence for the triplet state of the 1,4-dehydrobenzenes under the reaction conditions requires that, if the triplet is the ground electronic state, the rate of intersystem crossing from the singlet must be <10 9 sec -1 . Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Chemistry) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Bergman, Robert G. Thesis Committee: Unknown, Unknown Defense Date: 22 August 1980 Record Number: CaltechTHESIS:10072025-210733150 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:10072025-210733150 DOI: 10.7907/35fh-q724 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 17715 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 27 Oct 2025 18:51 Last Modified: 27 Oct 2025 18:51 Thesis Files PDF - Final Version See Usage Policy. 42MB Repository Staff Only: item control page

THE CHEMISTRY OF 1,4-DEHYDROBENZENES Thesis by Thomas P. Lockhart In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1981 (Submitted August 22, 1980) ii ... "I must have fallen asleep, for all of a sudden there was the moon, a huge moon framed in the window. Two bars divided it in three segments, of which the middle remained constant, while little by little the right gained what the left lost. For the moon was moving from left to right, or the room was moving from right to left, or both together perhaps, or both were moving from left to right, but the room not so fast as the moon, or from right to left, but the moon not so fast as the room. But can one speak of right and left in such circumstances? That movements of an extreme complexity were taking place seemed certain, and yet what a simple thing it seemed, that vast yellow light sailing slowly behind my bars and which little by little the dense wall devoured, and finally eclipsed ... " -from Molloy, by Samuel Beckett iii Acknowledgements My first comments must express the appreciation I feel for the guidance and insight Bob Bergman has provided me. As a scientific mentor he served as a continuous and exciting example of the dynamic application of scientific method. I also gratefully acknowledge the small but tenacious "POC" unit that existed within the research group during my tenure at Caltech and Berkeley. Within this group noble efforts were made to nurse the flame of inquiry into the closets of mechanistic organic chemistry. I have benef itted enormously, also, from the positive, helpful attitudes which run rampant within the research group. This character has survived, very much intact, being uprooted and transplanted into the cooler Berkeley environment. While these compositions hung suspended in the precarious state of "Works in Progess", the innumerable comments and insights offered by my colleagues were invaluable. In an extra-academic vein, I want to give my respects to Dr. H. E. Young, DDS, who capped-off my formal education with a dramatic demonstration that, contrary to widely held opinion, graduate school is not nearly as bad as having teeth pulled. Also to David Hume, that w iley fox, who must be the true, if 1 i ttle recognized, ancestor of that branch of inquiry known as mechanistic organic chemistry. Lastly, and most sincerely, I acknowledge the loving support of my parents, siblings, and friends. Abstract Upon heating, alkyl substituted cis 1,2-diethynyl olefins undergo cyclization to yield reactive 1,4-dehydrobenzenes; the products isolated may be derived from either unimolecular or bimolecular reactions of the intermediate. Z-4,5-Diethynyl-4- octene (19) undergoes rearrangement to yield 2,3-di-n-propyl-1,4dehydrobenzene (33). Solution pyrolysis of 19 in inert aromatic solvents produces three unimolecular products, diyn-6-ene (23), benzocyclooctene propylbenzene (26)) in high yield. (25) (Z-dodeca-4,8- and o-allyl-n- When 1,4-cyclohexadiene is added to the pyrolysis solution as a trapping agent, high yields of the reduced product o-di-n-propylbenzene (28) are obtained. The kinetics of solution pyrolysis of 19 in the presence and absence of trapping agent establish that 2,3-di-n-propyl-1,4dehydrobenzene is a discrete intermediate on the pathway leading to products. When the reaction was run in the heated probe of an NMR spectrometer, CIDNP was observed in 26. This observation, along with kinetic and chemical trapping evidence, presence of two additional intermediates, indicates the formed from 33 by sequential intramolecular [1,5] hydrogen transfer, on the pathway to products. The observation of CIDNP, coupled with the reactivity exhibited by 33 and the other two intermediates, implicates a biradical description of these molecules. Two approaches have been used to determine the spin state(s) of 1,4-dehydrobenzenes produced in the solution reaction diethynyl olefins. The first of method relies on the "spin V correlation effect" which postulates a relationship between the spin state of a caged radical pair and the ratio of cage and escape reactions (C/E) which may occur in the pair. When the 2,3-di-n-propyl-1,4-dehydrobenzene biradical abstracts hydrogen from 1,4-cyclohexadiene, mixture of a radical pair is generated . l,4-cyclohexadiene-d 0 If a and -d 4 is employed it is possible, by performing a VPC-MS analysis, to determine the ratio C/E leading from the radical pair to the reduced product, 28 . Applying this method to the reaction of 19, C/E was found to be 0.6, independent of the concentration of 1,4-cyclohexadiene (between 0.1 and 10 M) in the chlorobenzene reaction solution. This result suggests the presence of the singlet state of 33 in the reaction of 19. Independent support for this analysis came from the reaction of 3,4-dimethyl-1,5-diyn-3-ene hexachloroacetone sol vent in an NMR probe. l,4-dichloro-2,3-dimethylbenzene abstraction from the sol vent, the aromatic protons. (39), (38) in The major product, obtained by chlorine showed polarization (emission) in The interpretation of this result is straightforward and indicates solvent trapping of the singlet state of the intermediate 2,3-dimethyl-1,4-dehydrobenzene. Both of these experiments indicate that only the singlet state of 1,4dehydrobenzenes is generated upon thermal reaction of diethynyl olef ins. The failure to observe evidence for the triplet state of the 1,4-dehydrobenzenes under the reaction conditions requires that, if the triplet is the ground electronic state, the rate of intersystem crossing from the singlet must be <10 9 sec- 1 . vi Table of Contents The Chemistry of 1,4-Dehydrobenzenes I. II. Introduction A. General 2 B. Literature Survey 3 C. Goals 15 Chemistry of 2,3-Di-n-propyl-1,4-dehydrobenzene A. Introduction 16 B. Results 19 c. Discussion 31 D. Conclusions 50 E. Experimental 52 III. Determination of the Reactive Spin State of 1,4-Dehydrobenzenes A. Introduction 66 B. Results and Discussion 68 C. Experimental 84 References 90 Proposition Abstracts 95 1. Generation of Free and Metal-Bound Oxirenes 96 2. Selective Pyrolysis ' by Multiphoton Infrared Excitation (MIRE) 105 _ 3. CIDNP as a Tool for the Study of the Photochemistry of Organosilicon Compounds 117 4. Investigations in Organometallic Photochemistry 129 5. A Novel Synthetic Route to Binuclear Organometallic Compounds Containing a Bridging Methylene Function 141 1 THE CHEMISTRY OF 1,4-DEHYDROBENZENES 2 Chapter .I. INTRODUCTION General The highly reactive group of isomeric dehydrobenzenes, or benzynes, have provided challenging synthetic, mechanistic and theoretical targets for a number of years. 1 Of particular interest in these molecules is the extent of interaction between the dehydro-centers. characterized The o-benzyne isomer (1) has been well experimentally; spectroscopically 2 in a matrix at it has been studied s° Kand its reactivity toward a variety of substrates examined . la,b These studies indicate that substantial pi-bonding exists between the dehydro-centers. 1 The 1,3- 3 and l,4-dehydrobenzenes 4 - 12 have been much less yielding to experimental investigation. For both 1,3- and 1,4- dehydrobenzene bicyclic and biradical structures (2a,b; 3a,b) must be considered. Offsetting the energetic gain of forming a bond between the dehydro-centers is the substantial strain energy associated with the bicyclic structures. In addition, bicyclic may 1,4-dehydrobenzene (butalene) be the further destabilized due to antiaromatic cyclobutadi ene resonance. The 1,3- and 1,4-dehydrobenzenes may be molecules, then, in which the 3 gain in energy due to bonding of the unpaired electrons is more than offset by the increase in strain energy. Thus they may belong to a small, unusual class of organic molecules containing a negative bond dissociation energy.13 An additional consideration regarding the biradical structures is whether the lowest energy open-shell electronic state is a singlet or a triplet. The relative energies of these spin states will depend on the extent of interaction between the dehydro-centers. 14 • o. 0 2.o. The • II 1=1 • 2b possibility that 0 3b 2 and 3 may possess several energetically similar structures has made them challenging subjects for study. 1,4-Dehydrobenzene is the subject of the investigations reported in this dissertation. Literature Survey The first reported attempt to generate 1,4-dehydrobenzene was made by Fischer and Lossing 4 in 1963, who examined the pyrolysis of 1,4-diiodobenzene (they had previously produced obenzyne by the similar pyrolysis of 1,2-diiodobenzene; Scheme I). Mass spectroscopic analysis of the pyrolysate showed the formation of a compound with a highest m/e peak at 76 and an 4 Scheme I ~ o: 0 I I 01 6 -I2 I 6 [1,4- dehydrobenzene] -12 I HC ! = C - C = C - C :=CH 4 Scheme Il ~ co0 2 h ZI X ► Y+m/e =76 5 ionization potential of 9.46 ev. From these data they assigned the product of the pyrolysis to diethynyl olefin 4. Berry and coworkersS attempted to generate dehydrobenzene by the photolysis of compound 5 (Scheme 1 , 4II). Monitoring the reaction by flash-absorption optical spectroscopy and time-resolved mass spectroscopy, they observed a signal at m/e 76. From time-of-flight experiments they estimated the lifetime of the species! X, giving rise to the m/e 76 peak to be greater than 2 minutes under their experimental conditions (high vacuum, temperature unspecified). No evidence was obtained for the formation of 4 in this reaction. From these experiments it is not possible to distinguish the case where X is 1,4dehydrobenzene from that where Xis a meta-stable compound which falls apart under ionizing current to a compound (not necessarily 1,4-dehydrobenzene) having m/e 76. Further work in this area was not forthcoming until Jones and Bergman6 performed the experiments summarized in Scheme II. The equilibration of 4a and 4b took place in the gas phase at 2000 C and only products containing two deuteria per molecule were observed. unimolecular. This suggested that the reaction The observation that neither 6a nor 6b was were produced in the reaction indicated the presence of a transition state or intermediate (a 1,4-dehydrobenzene) containing a new C2 axis of symmetry. provided Pyrolysis of 4 in various solvents (Scheme IV) evidence that 1,4-dehydrobenzene was intermediate of finite lifetime. indeed an Furthermore, the abstraction of hydrogen from a hydrocarbon solvent (RH) and chlorine from CCl4 6 Scheme m gas phase 4b ,.....o / D H D ¢:, X ~ ¢;, D X ~ ~"-H H 'H not observed 6a 6b __ Scheme "-""'-- n[, ~ 4 solution [ ¢] RH • 0 3b CH70H OOH but not o-Q CH 3 tS~ ¢ Cl Cl - 0 Q-cH2 7 strongly suggested a biradical structure (3b) for at least the reactive form of the intermediate. methanol The trapping observed with argued against a zwitterionic description of 1,4- dehydrobenzene, which would have been expected to produce anisole. Experiments reported by Chapman and Masamune on related 1,4dehydroaromatics also provided support for biradical structures. Masamune, et al. 7 , performed the elimination of two equivalents of methanesulfonic acid from dimesylate 7 to give 8 (Scheme IV) which contains the elements of a diethynyl olefin within a ten membered ring. The strain in compound 8, presumably, is responsible for the mild temperature (ca. 25° C) at which further rearrangement occurs. isolated. In fact, 8 was never successfully Abstraction of hydrogen from the solvent is consistent with the formation of 9,1O-dehydroanthracene 9 in this reaction, in analogy with the results of Jones and Bergman. Chapman and coworkers8 studied the photolysis of 10 in both a hydrocarbon glass and an argon matrix (Scheme V). When the photolysis was carried out in a hydrocarbon glass, anthracene was obtained upon warming the photolysate to room temperature. Doping the hydrocarbon glass dichloroanthracene upon warming. with CCl4 produced 9,10- These results suggest the formation of a 9,10-dehydroaromatic intermediate which reacted as a biradical even at subambient temperatures. The low temperature photolysis of 10 was also monitored by ESR; failure to observe an ESR signal in these experiments provides negative evidence in 8 Scheme nr ~ O ~jOMs ~ -2 HOMs () ~) ~ OMs # 8 7 l H (D) RH (D) H (D) [©] 9 Scheme "'SL ~ 0 II C h11 77° or 8°K -2 co C II 0 oco • • ~ 9 m, Cl 4 10 ~ H Cl Cl 9 favor of a singlet ground state in 9. A provocative Johnson from result was recently reported by Gilbert this laboratory. 9 The products obtained from thermolysis of diethynyl olefins 11 indicated that at elevated temperatures a novel trimethylsilyl migration in the substituted 1,4-dehydrobenzene 12 occurred to produce 1,3-dehydrobenzene isomer 13 (Scheme VII). This is the only reported example of interconversion of dehydrobenzene isomers.10 In contrast to the studies cited above, in which the 1,4dehydroaroma tics cl early di splay bi radical reactivity, Breslow, et ai. 11 , have reported evidence which suggests that they have generated and trapped the bicyclic isomer, butalene (3a). Their approach was to employ a base-induced elimination of HCl from Dewar benzene 14 (Scheme VIII). When the reaction was performed in the presence of 1,3-diphenylisobenzofuran, 15, Diels-Alder trapping of (presumably) l-deuterio-4-dimethylaminebicyclo- [2.2.0l-2,5-hexadiene (17) produced a modest yield of 16. In the absence of trapping agent dimethylaniline was obtained. These results, especially the formation of 16 in which the 1,4-bond remains intact, are consistent with a mechanism in which butalene is an intermediate. In a subsequent study by Breslow and coworkers,12 l-chloro- 2-methylbicyclo-[2.2.0l-2,5-hexadiene 18 was treated with strong base in a deuterated solvent (DNEt 2 ). Both the position of methyl substitution and the amount of deuterium incorporation (do to d 3 ) in the N,N-diethyltoluidine products indicated that the reaction is considerably more complicated than it was originally 10 Scheme lZil. ~ 0 4soc. -vSiM • II SiMe 3 12 13 Products Products 11 Scheme 'SlIII ~ cp Cl ctJ LiNMe2 35°C 11 I I 14 NMe2 DNMe2 cp cq ::,.._ D o cp cp 3a 16 15 HNM•2j [tj3] 62 17 w NMe2 Cl Li NMe '62 2 DNMe 2 18 relotiveyield : D 80% 20% % dn %dn do 53 21 d1 31 42 d2 14 27 d3 2 9 12 thought to be. It was necessary to conclude from this study that the butalene intermediate may not be on the major pathway leading to the N,N-dimethylaniline products, although it lias still presumed to be the precursor of 16. The complexities encountered in this work arose in large part from the relatively harsh conditions (strong base) employed to induce the elimination reaction. Coincident with the efforts of experimentalists to generate and characterize the reactivity of 1,4-dehydrobenzene, a number of theoretical treatments have been carried out (Table 1).15-19 Wilhite and Whittenl6 reported a detailed ab initio study in which three calculations were performed: a full SCF-MO treatment of both the singlet and triplet electronic states, a limited configuration interaction determinant CI treatment. (CI) calculation, and a many- The simplest calculation predicted that the energy of the triplet biradical lies well below that of the singlet. Inclusion of CI in the calculations, however, led to a much smaller predicted difference in the singlet and triplet biradical energies. The smallest energy difference was predicted in the full CI calculation which placed the triplet state 3.5 kcal/mole below the singlet. In these calculations, the geometry of 1,4-dehydrobenzene was somewhat arbitrarily taken to be that of benzene. Whilhite and Whitten were careful to point out that, given the small singlet-triplet energy difference found, a calculation performed at the equilibrium geometry might lead to an inverted ordering of the electronic states. Because the Noell and Newton (1979) Ab initio GVB; 4-31G (18) Mueller (1973)< 19 > Modified MINDOl2 (no CI) Washburn et al. (1979) Ab initio, 4-31 G (no CI) (3a) +1.4 0 0 +5 (17) Dewar et al. ( 1974) MINDOl3; lim. CI 0 (16) Triplet ~ o:) 1 1 ? ~ 0 (+24) (+82) O(~Hf = +117) +3.45 Singlet 11 +18 +94 +36 Singlet -- o:)1 -- I I ~ ? Relative Energy of Structures (kcal/mol) Calculated Energies of 1,4-Dehydrobenzene Structures Wilhite, Whitten (1971) SCF-MO-CI Table f w I-' 14 geometry was fixed in their treatment, no prediction was made concerning the position of butalene (which would be expected to have a much shorter 1,4 distance than benzene) on the energy surface. Dewar and Lil7 reported a MIND0/3 study in which geometry optimization was carried out for the singlet and triplet electronic states. The singlet biradical was predicted to be 6.2 kcal/mole more stable than the triplet. An investigation of the singlet surface led to the prediction that butalene lies in a relative energy minimum, biradical. 35.9 kcal/mole above the singlet The transannular bond in butalene was predicted to be 1.667 A long . A generalized valence bond (GVB) calculation of the 1,4dehydrobenzene energy surface was recently reported by Noell and Newton.18 These authors performed limited geometry optimization for the singlet and triplet states. They concluded that the lowest energy structure of 1,4-dehydrobenzene is the singlet biradical and that the bicyclic butalene structure lies in a local energy minimum very roughly estimated to be 77 kcal/mole higher. The triplet biradical was calculated to have an energy slightly above that of the singlet (1.4 kcal/mole), though the difference calculated for the two biradicals appears to be less than the uncertainty of the calculations. To summarize the experimental and theoretical work on 1,4dehydroaromatics, the most convincing cases for its generation suggest that a biradical description is appropriate. The two important geometry-optimized theoretical studies of the 1,4- 15 dehydrobenzene energy surface are in agreement in predicting that the bicyclic isomer lies substantially higher in energy than the bi radicals. The possibility exists, however, that the bicyclic structure is lowest in energy but undergoes either facile 1,4bond cleavage or suffers 1,4-cleavage in concert with atom abstraction reactions. Convincing experimental evidence for the existence of butalene as a meta-stable species must await its generation and characterization under extremely benign conditions, such as in an argon matrix. Goals One of the goals of my research has been to obtain kinetic evidence for the existence of a discrete 1,4-dehydrobenzene intermediate in the thermal reaction of diethynyl olefins. We were also interested in further characterizing the reactivity of the 1,4-dehydrobenzene intermediate produced in these reactions. Chapter II of this dissertation describes a mechanistic investigation of the thermal reaction of Z-4,5-diethynyl-4-octene (19) which provides compelling evidence for the occurrence of a true 1,4-dehydrobenzene intermediate on the pathway leading to products. It was also hoped that evidence for the spin state of the 1,4-dehydrobenzene intermediate(s) present in these pyrolyses could be obtained; Chapter III details the results of our efforts in this area. 16 Chapter ll CHEMISTRY OF 2,3 - DI-n-PROPYL-1,4-DEHYDROBENZENE Introduction As mentioned in Chapter I, there has been continuing interest in this laboratory in characterizing the reactivity of 1, 4-dehydrobenz ene. We wished especially to obtain conclusive evidence6 that 1,4-dehydrobenzene is a discrete species of finite lifetime. Several approaches to test this hypothesis suggested themselves. The first was to generate 1,4-dehydrobenzene under conditions where it could be directly observed. was to obtain kinetic evidence for A second method the occurrence of 1,4- dehydrobenzene as a reactive intermediate on a pathway leading to observed products. temperature, Duncan Brown in this laboratory pursued a low photochemical route to the butalene/1,4- dehydrobenzene energy surface that was intended to allow the direct observation of these species.20 My approach has involved a kinetic study of the generation of a substituted 1,4dehydrobenzene by the thermal reaction of a diethynyl olefin. There are several drawbacks to the use of diethynyl olefin 4 as a thermal precursor of 1,4-dehydrobenzene. The yield of aromatic products in solution pyrolyses of 4 is generally quite low (<50%).6,21 In addition, the sensitivity of 4 toward air oxidation and rapid thermal polymerization (even at subambient temperatures) makes it rather inconvenient to work with. It was hoped that substitution of the diethynyl olefin framework would lead to improved stability at ambient temperature and to kinetic 17 stabilization against adventitious side reactions during thermal reaction in solution. Ideally, the substituents should be ones that have an insignificant effect on the electronic structure of 1,4-dehydrobenzene. Toward this end, a number of compounds (20a-w, Table 2) were prepared in this laboratory2 2• The thermal reactions of some of these molecules proved to be quite interesting.9,23 investigations indicated that alkyl and These trimethylsilyl substituted diethynyl olefins are appreciably easier to work with than the unsubstituted compound, 4, roughly in proportion to the number and steric bulk of the substituents. Unfortunately, only modest improvements in the yield of aromatic products were realized through these modifications. Substitution at the acetylenic carbons greatly raised the temperature required to effect cyclization to the substituted 1,4-dehydrobenzenes. is an undesired effect, This since more vigorous reaction conditions are expected to increase the mechanistic complexity. these investigations an additional problem was During identified: interaction of 1,4-dehydrobenzene biradicals with solvent molecules can produce free radicals which may subsequently attack unreacted diacetylene. Observations by Charles Mallon of the interesting unimolecular thermal chemistry of compound 20n prompted us to further investigate the chemistry of di-n-propyl substituted diethynyl olefins.23 promising molecule Compound 19 was identified as the most to study; because of the unsubstituted 18 acetylenic positions 19 was expected to rearrange at temperatures considerably lower than 20n.21 20 Table 2. Diethynyl Olefins That Have Been Prepared in this Laboratory. 20a H H methyl methyl b H H phenyl phenyl C H H .!_-butyl H d H .!_-butyl H H e H .!_-butyl SiMe f H .!_-butyl SiMe g H h 3 H SiMe 3 ethyl 3 ethyl H H E_-propyl E_-propyl H i H ethyl ethyl SiMe j H n-propyl E_-propyl SiMe k methyl methyl H H 1 methyl methyl methyl methyl m methyl methyl E_-propyl E_-propyl n E_-propyl n-propyl methyl methyl 0 methyl methyl phenyl phenyl p .!_-butyl SiMe H H q Rl = R2 = -(CH2)4- methyl methyl 3 3 3 r " E_-propyl E_-propyl s " phenyl phenyl H H t u H H SiMe V .!_-butyl H SiMe H SiMe w SiMe 3 3 3 3 SiMe SiMe 3 3 .!_-butyl 19 Results Synthesis Four diethynyl olefins were prepared in the course of the mechanistic investigations reported in this section. Two contained alkyl substituents at both vinyl positions (19, 21). The synthetic route to these compounds followed a general method previously employed by Gilbert Johnson and John Stofko in this laboratory (Scheme IX). 24 The key step (equation 2) is an olefination reaction by the method of Pederson, et ai. 25 Anion 22 undergoes both addition to the carbonyl and acetylenic H abstraction; this led to only modest yields of the desired diethynyl olefins, which were produced in roughly a 40:60 ratio of £1§ and £~~Il§ isomers. Separation of the isomers was conveniently effected by column chromatography on silica gel. Photolysis of the trans olefin in alkane solvents led to cistrans isomerization. By this method the £rans product of the olefination reaction was converted to a mixture (ca. 1:1) of the £is and trans isomers. Finally, removal of the acetylenic trimethylsilyl (TMS) group was accomplished in high yield by the method of Arens and Schmidt (equation 3). 26 We also required two compounds, 23 and 24, substituted at the acetylenic positions. Compounds of this type are most conveniently prepared by the coupling of copper acetylides27 with trans-l,2-diiodoethylene 28 (Scheme X). Photoisomerization of the trans isomer followed by column chromatography gave the desired cis isomer. R- =-H Scheme :X: cis-8 A NH 3, H2 0 Cul 2) KCN, H20 I) Ag N03 R 22 (80 %) R R,...... ?/' ✓.: ,......R )~ 60% .,.....TMS h ll --- ~ ~ r( ¢- B (cis+trons) 35% 19 , R = n Pr 21 , R = Et I ¼I A RXH TMS ef::, 'R ...,.R 24, R= Et 23,R=nPr Hc=c-CH(R)= CH(R)-C= C-TMS 2) TMS-CI I) n Bu Li, TMEDA A~'H 0 Pyridine ~,,,,TMS R- =-cu ~ ~ x~ R R = R- CH2-TMS 80% TM'?' 2) Me 3SiCI (TMS-CI} I} n Buli n Buli, TMEDA RCH 2 -C!!!C-H Scheme IX. (4 ) (3) (2) (I) N 0 21 Although the diethynyl olefins obtained after chromatography on silica gel were >95% pure, before pyrolysis they were usually further purified by preparative gas chromatography. provided the diacetylenes in >99% purity. This method The neat diacetylenes could be handled briefly at room temperature in the air but yellow coloration appeared after several minutes under these conditions. When stored in solution (ca. 1-5% v/v) the lifetime of these compounds was greatly improved, although temperatures of -60° C were required to effectively eliminate discoloration due, presumably, to polymerization. Thermal Reactions Gas phase experiments were performed by passing the diacetylenes through a heated quartz tube either under a stream of N2 (1 atm pressure) or at reduced pressure. The products were collected on a cold finger at -196° C. Solution reactions were carried out in sealed glass tubes. The concentration of the diacetylenes was usually 10-2 Mor less and the samples were subjected to four freeze-pump-thaw cycles to remove oxygen. Compound 19 was pyrolyzed in the gas phase and in a number of solvents. The thermal reaction of 19 in the gas phase CN 2 flow, 320° C) produced a quantitative yield of three products: isomeric diyne 23, benzocyclooctene (25), and o-allyl-npropylbenzene (26, Scheme XI). The three products were isolated by preparative VPC and characterized by their NMR, IR and high resolution mass spectra.29 Compound 23 was identified 22 Scheme XI ~ 19 26 23 Cl 27 23 additionally by independent synthesis. When 23 was heated in the gas phase at 400° c (N 2 flow, contact time ca. 1.5 min), greater than 95% conversion to 25 and 26 was observed. We studied the thermal chemistry of 19 in solution in order to obtain accurate kinetic data for its cyclization. Aromatic solvents were expected to be unreactive toward free radical hydrogen atom abstraction and, indeed, proved to be almost completely inert toward the intermediates produced during the reaction of 19. Heating 19 at 196° c for 15 minutes in diphenyl ether, chlorobenzene or benzene led to complete conversion of the starting material to 23, 25 and 26 in high yield (Table 3). Notably, at this temperature diacetylene 23 was quite stable and did not react detectably. In addition to unimolecular products, several isomeric compounds (total yield <5%) of empirical formula C1aH21Cl were observed by VPC-mass spectroscopy (27). compounds of appear to be formed by addition These reactive intermediates to the solvent. We also wanted to find a suitable trapping agent for the intermediate(s) produced in the reaction of First 19. we investigated the reaction of 19 with simple alkanes and alkyl substituted aromatic solvents which, we reasoned, effectively transfer intermediates. hydrogen to biradical and would radical Unfortunately, the yield of tractable products was quite low (<40%) when these solvents were employed. The reacted solutions were badly discolored which suggests the occurrence of competitive polymerization reactions. Subsequently, we found that the addition of a small amount (<15% 14 5 132 156 156 ( 6) ( 7) ( 8) ( 9) 0.0 0.0 0.0 6 . 9 X 10 -4 -4 9.3 X 10 -5 6.4 7.1 X X 0.19 0.38 -4 -4 10 10 2.9 X 10 X 0.0 10.6 -10- 3 1.3 0.4 (M) -- ) 0.0 -1 -- kb (sec 0 S 5. 2 8.7 7.9 9.8 11 . 8 13.5 ~1.0 10.1 20.3 24 6.7 13.5 38.5 37.3 35 . 8 38. 5 <l 8.9 36.9 26 3.1 8.7 16.4 17.6 17 . 2 20.0 <l 5.4 20.8 25 ( '?,) b 47.4 27.6 -- -- -- -- 76 48 -- 28 Absolute Yield 62 58 63 65 65 72 <79 71 78 Total (24 - 28) bYields determined b y digital integration of FID vpc trace and r e ference to an internal standard. 0. 01 M 156 ( 5) = 166 ( 4) ~ 196 ( 3) [1q] 196 ( 2) a 196 ( 1) ( o C) T 1,4-cycloh e xadi e ne Product Yields and Ra te Constants in the Solution Py rolysis of 19. a Run Table 3. ""' I\J 25 by volume) of a better hydrogen atom donor to solutions of 19 in one of the inert solvents produced a high yield of products. Both 1,4-cyclohexadiene and 9,10-dihydroanthracene were extremely effective as hydrogen donors; 1,4-cyclohexadiene was convenient to use because of its solubility properities. most In the presence of these trapping agents a new product, o-di-npropylbenzene (28) was obtained in high yield. The yield of 28 increased with added trapping agent at the expense of the three unimolecular products (see Table 3, runs (1)-(3), and figure 1). Several higher molecular weight products (total yield roughly 1/4 that of 28) were also observed in these pyrolyses. By VPC-mass spectroscopy these compounds were found to be isomers of molecular formula c 1 aH 22 (29) and c 18 H24 (30) and are believed to have the structures shown below. In . addition, several products with the molecular formula modest yield. c 12 H14 and c 12 H12 were formed in These are believed to be dimers formed by the combination of cyclohexadienyl radicals. 29 z 0 ...E 0 N Cl) -,:, >- Cl) -,:, ~ 0 ..._,, - 0 0.8 1.0 amount of 1,4-cyclohexadiene in the reaction solution. I, 4-C y c Io he x ad i en e , M Plot of normalized product yields as a function of the 0 .6 Figure 1. 0 .4 0.2 0 40 80r------T---~---.---,------.--,--------- Iv O'\ 27 When 2,2,5,5-tetradeuterio-l,4-cyclohexadiene was employed as the trapping agent, the 28 formed contained two deuteria per molecule. The yield of tractable products in this reaction, however, was substantially lower than when undeuterated cyclohexadiene was used. reactivity of biradicals. 30 This is believed to reflect reduced the deuterated trap toward the intermediate When cyclohexadiene-d4 was employed as trapping agent the high molecular weight products formed, 29 and 30, contained 2 and 4 deuteria per molecule, determined by mass spectroscopy. As expected, respectively, as the unimolecular products formed showed no incorporation of deuterium. The location of deuterium in 28 was determined in the following way: using the method of werstiuk and Kadai31 the aromatic deuteria were selectively exchanged for protons through acid catalysis. After the exchange was completed, 28 was examined by mass spectroscopy and the molecular ions (corresponding to 0, 1 and 2 residual deuteria per molecule) were measured. The results indicate that, when 19 was heated in chlorobenzene with l,4cyclohexadiene-d4 (0.8 M), 66% of the product 28 formed contained two aromatic deuteria, 33% contained one aromatic and one aliphatic deuteria and about 1% of product 28 contained two aliphatic deuteria. The ramifications of this result are discussed further in the next section (C). The kinetics of disappearance of 19 in chlorobenzene solution (10-2 M) were measured both in the presence and absence of added cyclohexadiene (Table 3). Linear first order plots for three reaction half-lives were obtained at four temperatures 28 spanning a range of 34° C. Clean first order kinetics were also observed for the reaction of 19 in 0.19 and 0.38 M solutions of 1,4-cyclohexadiene with chlorobenzene sol vent. of added trapping agent, In the presence the reaction rate was unchanged within experimental error (Table 3, runs (5), (8) and (9) ). The thermal reactions of 21 and 24 in the gas phase were also investigated. 3 2 At temperatures below 425° C quantitative conversion of 21 to 24 was the only process observed. At higher temperatures additional unimolecular products appeared. predominant products were o-ethylstyrene, benzocyclobutene and styrene (Table 4). The tetralin, These products were also observed when 24 was allowed to react under similar conditions. Thermal reaction of 21 in benzene solution cio- 2 M, see Table 4) gave diacetylene isomer 24 plus a small amount of odiethylbenzene, biben2yl, and a product (31) of molecular formula c 16 H18 (identified by VPC-mass spectroscopy) postulated to be that shown below. in a solution of benzene-d 6 , deuterium, in the ratio whose structure is When 21 was heated at 190° C o~diethylbenzene containing d 0 :d 1 :d 2 = 3:3.4:1, was formed, demonstrating that significant abstraction of deuterium from benzene took place. The bibenzyl produced was found to be fully deuterated and is believed to have been formed as a result of this reaction channel. The source of hydrogen is assumed to have been the diacetylenes 21 and 24 and the protio reaction products. Reaction of 21 in a benzene solution containing 1,4cyclohexadiene gave greatly reduced yields of unimolecular gas benzene . C so 1ution benzene+ 1,4-cyclohexadiene (2) (3) 195 6 51 76.8 100 0 0 1. 3 0 5 75 0 0 0 0 0 0 1.6 0 0 8.8 0 0 7.8 0 o-diethyl tetralin benzene N2 flCM c[21] = 0.01 M b ~ields detennined by digital integration of FID VPC trace and reference to internal standard (1.1 M) c b (4) 500 gas b (1) 195 400 Phase Run 24 T(°C)(c+t) Absolute Yield (%)a benzoo-allyl cyclostyrene butane styrene Table 4. Proouct Yields in the Reaction of 21. N I..O 30 products; o-diethylbenzene was obtained in good yield (Table 4) . Several high molecular weight products formed were detected by VPC-mass spectroscopy and have the molecular formula 31 32 c 16 H20 (32). 31 Discussion Proposed Mechanisms To account for the results obtained in the thermolysis of 19 we propose the mechanism outlined in Scheme XII which involves initial ring closure of 19 to produce the 2,3-di-n-propyl-1,4dehydrobenzene biradical 33. In the absence of trapping agent two unimolecular pathways are available to 33. form rearranged diacetylene 23 {or Ring opening to return to the starting compound) may take place in analogy with the results observed by Jones and Bergman for the deuterium labelled diacetylene 4a {Scheme III). 6 Additionally, biradical 33 may transfer hydrogen from the terminus of a propyl group to the nearest aromatic radical site to produce biradical 34. This is expected to be a facile favorable process transition since state is a kinetically involved. 33 six-membered Furthermore, the heat of formation of biradical 34 is estimated to be about 12 kcal/mole less than that of 33, the difference in strength of primary alkyl and aryl C-H bonds.34 Ring closure in 34 to give a [3Jmetacyclophane is expected to be unfavorable as the smallest known [n]metacyclophane has a pentamethylene bridge, {n rearranges to indane at 150° c. 35 Instead, a = 5) and second intramolecular hydrogen transfer to produce biradical 35 may take place. Unimolecular products 25 and 26 argue very forcefully for the presence of 35 on the reaction pathway and strongly suggest the location of the radical sites at the termini of the n-propyl subs ti tuents. In fact, the failure to observe smaller ring 26 I 25 3 CH 3 CH 2 CH 3 CH co oc::: 19 x::: ¢:, ~ Scheme XlI. ks k-1 kl !k3 34 !k2 33 CH 3 35 T"-/"'-CH2· H ~--- k4 ( /4H] k [SH] 6 ~ CH 2• CH 3 ~ C H2' H • . I CH 3 o:::: • 28 T-.........,,,,. "-CH3 H ~ C H3 23 ~ C H3 H ~ /'--/CH 3 ¢:' N w 33 ben z ocycles or o-propyl-p-methylstyrene argues against the occurrence of other intramolecular hydrogen transfer modes <ll. [1,4)) in 33 or 3 4. The presence of 1,4-cyclohexadiene in the reaction solution leads to the formation of 28 which may in principle come from trapping of any of the three biradical intermediates. The results of the deuterium labelling study mentioned previously (see the Results section) indicate that trapping occurs predominantly from biradicals 33 and 34 (Scheme XIII). Only 1% of 28 formed was found to contain two deuteria in the alkyl side chains when the pyrolysis solution was 0. 8 M in cyclohexadiened4. The formation of high molecular weight products in the solution pyrolyses of 19 is readily understood in terms of the proposed mechanism. Product 27 appears have been formed from attack on the chlorobenzene solvent by intermediate biradicals 33 and 3 4. By our mechanism (Scheme XIV), transfer of hydrogen from cyclohexadiene to one of the biradical intermediates generates a radical pair; transfer of a second hydrogen within the solvent cage led to the 28 formed. formation of 28. Cage escape may also have led to the Combination of cage-escaped cyclohexadienyl radicals appears to have been responsible for the formation of products with the molecular formula C12H 14 and c 12 H12 (in the reaction with cyclohexadiene-d 4 these products have new molecular weights corresponding to c12 o12 and c 12 o14 >. Cage combination of the radical pair may be responsible for the several isomers with 34 Scheme XIII. cc:: cc:: . • • RD (= Q RD o.sM) RD D D D D D D D 66% D 33% 1% 35 0 -~ - + -en .. C .. N + C 0 0 C 0 0 ~ 0 a. 0 ~ a. 0 C ., ·.0 N 0 If) E 0 ·-,:,V> 0-0 C 0 Q,) ·- Cl- o u 0 0 Q,) ~ ~j @t I u~ _ _Cl>:__ _ lo IC) IC) Q,) .c 0 Cl) 0+ 10 GO N 36 the formula C1aH24 (30) found. Hydrogen loss from some of these combination products (either under the reaction conditions or in the mass spectrometer) accounts for the formation of products of formula c1aH2 2 observed. (2 9) Supporting the identification of 29 and 30 are the observations that (1) these compounds only appeared in reaction solutions containing cyclohexadiene; (2) the yields of 29 and 30 were directly related to the yield of trapping product 2 8; (3) mass spectral fragmentation gives large peaks due to the phenyl cation, cyclohexadienyl radical cation and dipropylbenzene cation cc 12 H18 >; and (4) when cyclohexadiened 4 was used, 29 was formed containing two deuteria per molecule, 30 with four deuteria per molecule. It is informative to consider here the results of the thermal reactions of diacetylenes 21 and 24 (Table 4). The products obtained in the gas phase and in solution pyrolyses in benzene suggest the mechanism presented in Scheme XV. Intermediate 36 appears to undergo intramolecular [1,41 hydrogen transfer slowly relative to ring opening to diacetylene 21 or 24. Thus, 21 may be converted quantitatively to 24 in the gas phase (4000 C) without appreciable intramolecular trapping of the 1,4dehydrobenzene intermediate. (ca . soo° C), Only at much higher temperatures where 24 is repeatedly converted to 36, were products of intramolecular hydrogen transfer found. 36 In benzene solution 21 was similarly converted to 24. No intramolecular hydrogen transfer was observed, though in the presence of added cyclohexadiene the 1,4-dehydrobenzene intermediate was efficiently trapped to give o-diethylbenzene. 37 Scheme X2: ~ • J) • 36 21 2-H1/ / (slow) 37 +-- 38 At the temperatures employed in the solution studies, the rearrangement of 36 to 24 is essentially irreversible. Each molecule of 21, therefore, can generate 36 only a single time in the course of this reaction. Our failure to detect even trace amounts of tetralin or o-ethylstyrene indicates that the rate of [1,41 hydrogen transfer must be several orders of magnitude slower than ring opening to 24. we may reasonably assume that the rate of rearrangement of 36 to 24 is comparable to that of 33 rearranging to 23. This provides a reasonable explanation for the failure to detect [1,4] hydrogen transfer in 33, where [1,51 transfer is faster than ring opening to 23. Kinetic Studies In accord with the mechanism presented in Scheme XII, the disappearance of 19 shows first-order kinetics. From the rate data measured over a 34° c range, the activation parameters for the first step in the reaction mechanism can be determined from an Arrhenius plot (figure 2, Ea= 27.4 ± 0.5 kcal/mole and log 10 A = 10.8 ± 0.3 sec -1, . The Arrhenius parameters obtained for the reaction of 19 are valid only if the return of 1,4-dehydrobenzene 33 to 19 (k_1> is slow with respect to the other reaction rates Ck 4 , k 2 , k 5 [SH]). we may determine the importance of k-1 and obtain kinetic evidence for the intermediacy of 33 by testing the following hypothesis: if biradical 33 is a true intermediate on the 39 3 .0 4.0 0.22 0 .23 0.24 0 .25 1/T Figure 2. ~.rrhenius plot of unimolecular rate constants observed in the reaction of 19 (0.01 M) in chlorobenzene. 40 reaction pathway leading from 19 to products, and if the unimolecular rate constants (k 4 and k 2 ) and the bimolecular rate constant Ck5[SH]) are rapid with respect to k_ 1 , then the rate of disappearance of 19 must be independent of the presence of cyclohexadiene in the reaction solution. However, the product distribution the will be dependent on cyclohexadiene concentration if 33 is an intermediate and k 5 [sHl is of comparable magnitude to the unimolecular reaction rates. rate constants obtained, The when 0.19 and 0.38 Madded 1,4- cyclohexadiene were heated with 19 at 156° C, are identical, within experimental error, to that obtained in the absence of trapping agent (Table 3, runs (5), (8) and (9)). As predicted by our mechanism, while the rate of reaction of 19 was unchanged, the increase in cyclohexadiene concentration changed the yield of 28 from 0 to 47%. The mechanism in Scheme XII makes several other specific predictions about the dependence concentration of trapping Application of the agent of product yields on the in the steady-state reaction solution. approximation to the concentration of 33 gives equation (5). Scheme XII predicts that the yield of 23 will be inversely concentration of 1,4-cyclohexadiene proportional (equation to the (5)). The experimental data are plotted in figure 3 and show good agreement with the mechanism proposed. (normalized yield of 23)-l ( s) 4.0 6.0 8.0 10.0 0 0.8 0.2 0.4 0.6 [1,4- Cyclohexadiene], M 1.0 1.2 Figure 3. Plot of (normalized yield of 23)-l versus concentration of 1,4-cyclohexadiene. Cone. 19 = 0.01 M, T = 190° c. {23f 1 12.0 14.0 16.0 of::> I-' 42 The absolute yields of the unimolecular products 23, 25 and 26 are strongly dependent on the concentration of trapping agent (figure 1). The yields of 25 and 26 show a more pronounced decrease upon the addition of cyclohexadiene than does the yield of 23 (the ratio 25/26, however, remains constant). We may explain the nature of this dependence using the proposed mechanism if we make use of the following simplifying assumption: from experiments with cyclohexadiene-d4 we know that a negligible amount of 28 arises from trapping of biradical 35. d[25 + 26]/dt = k 3 [34J. Therefore, Equation (7) follows from equation (6) by application of the steady state approximation to intermediates 33 and 34. Yield of 23 k (33) 4 = ( 6) Yield of 25 + 26 Yield of 23 Yield of 25 + 26 k (34) 3 = k4 + k2 k4 k6 [SH] ( 7) k2 k3 Equation (7) relates the ratio of 23 to 25 and 26 as a function of added cyclohexadiene. The linear plot obtained by applying this function to the experimental data (figure 4) confirms the existence of a second intermediate which undergoes trapping with cyclohexadiene cyclohexadiene-d4 experiments). (in agreement with the The intercept of that plot gives the value of k 4 /k 2 as 0.36 ± 0.01; this is the ratio of unimolecular products in the absence of added trapping agent. Similarly, dividing the slope of the line by the intercept gives Since the intercept of the line 0.6 Plot of the ratio of the observed yields of Concentration of 19 = 0.01 M, T = 190° C. 23/(25 + 26) as a function of 1,4-cyclohexadiene concentration. Figure 4. [1,4 -Cyclohexadiene], M 0.2L-----JL-----'----'---1--..L..----JL...----L---L--~--'---L------1 0 0.2 0.4 0.6 0.8 1.0 1.2 (25+26) 23 0.8 1.0 1.2 .I'.:> w 44 agrees with the ratio of unimolecular products observed in the absence of trapping agent (0.35), the presence of cyclohexadiene does not significantly affect the reaction rates at the concentrations studied. CIDNP We also have obtained good evidence for the presence of a third biradical on the reaction pathway of 19. When a solution of 19 in o-dibromobenzene or diphenyl ether was heated at 160° C in the probe of an NMR spectrometer, several emissive signals were observed (figure 5). These signals are assigned to the vinyl protons and, tentatively, to the alkyl protons (terminal methyl and methylene) in polarized 26. The observation that only polarization of the hydrogens at the two end carbons of the propyl and propenyl chains occurred, and that all enhancements were emissive strongly implicate biradical 35 as the polarizing species. Thus, al though 1,4-cyclohexadiene is not sufficiently reactive to trap biradical 35 to an appreciable extent, the CIDNP process is rapid enough to provide evidence for the presence of that intermediate. 37 Summary of Arguments in Support of the Proposed Mechanism It may be helpful at this point to summarize the evidence supporting the mechanism proposed in Scheme XII: Cl) the formation of diacetylene 23 from 19 and the observation that 23 also rearranged at higher temperatures to generate products 25 45 8 Figure 5. 6.0 5.0 4.0 The upper spectrum (90 MHz 1 ppm H) shows the vinyl region of a purified sample of o-allyln-propylbenzene recorded at 30° C. The lower spectrum shows the ernissive signals observed during reaction of~ at 160° Cina 90 MHz NMR probe. The signals appearing in absorption on the left side of the lower spectrum are spinning side bands of the solvent (o-dibromobenzene). 46 and 26 argues for an intermediate or transition state with the symmetry of 1,4-dehydrobenzene 33.6 (2) The yield of 23 was substantially reduced by added trapping agent while the decomposition rate of 19 was unaffected. The dependence of the yield of 23 on added 1,4-cyclohexadiene displayed the behavior expected if the 1,4-dehydrobenzene exists as an intermediate of finite lifetime. Together these observations establish the formation of 1,4-dehydrobenzene intermediates in the thermal reaction of diethynyl olefins. (3) The observation that two 28- d2 isomers were formed when cyclohexadiene-d4 was used as a trapping agent indicates the presence of at least one additional intermediate. Experiments, which showed that one isomer was labelled at both the aromatic ring and the aliphatic side chain, support the postulate that 34 is formed in the reaction. evidence for the second intermediate was Kinetic also obtained by examining the ratio of unimolecular products formed as a function of added cyclohexadiene. (4) The reactivity of (hydrogen abstraction from 1,4-cyclohexadiene, and 34 followed by 33 combination and disproportionation) and intramolecular hydrogen transfer to produce biradical 35 clearly demonstrate the biradical nature of the intermediates. strongly suggest intramolecular (5) Products 25 and 26 combination and disproportionation from a common biradical precursor. The observation of CIDNP in 26 confirms the intermediacy of biradical 35. 47 Estimated Absolute Rate Constants and Reaction Energetics It is possible to estimate the absolute rate constants and activation energies for the reactions steps in Scheme XII. A reasonable model for cyclohexadiene trapping of 34 is the rate constant for abstraction of hydrogen from diphenylmethane by phenyl radicals (estimated to be 7.7 x 10 6 M-l sec- 1 at 60° c>. 38 The unimolecular rate constants k 2 , k 3 and k 4 are thus expected to lie between 10 6 and 10 7 sec- 1 at 60° C; ks must be at least one or two orders of magnitude slower (vide supra). A marked dependence of the ratio 23 /(25 + 26) on the reaction temperature is observed (Table 3, runs (1), (4), (5), (6) and (7)). This is convincing evidence that at least one of the intramolecular processes leading from 33 is activated. The difference in activation energies and A factors of the steps involving k 4 and k 2 can be obtained by a plot of ln[23/(25 + 26)] versus 1/T (equation (8)). A linear relationship is observed over a range of 64° c (figure 6); from the slope of the line the difference in activation energies, Ea Ck 4 ) - Ea Ck 2 ) was found to be 5.2 ± 0.4 kcal/mole. The ratio ACk 4 )/ACk 2 ) was determined from the intercept to be 10 2. The Ea for [1,5] hydrogen transfer in 33 should be similar to that for the exothermic [1,5] hydrogen transfer observed in the 2 ,2-dimethylpentoxyl radical (Ea = 5.0 kcal/mole). 39 The conversion of 33 to 23 should therefore have an Ea of about 10 kcal/mole. The absolute magnitude of the A factor for the rearrangement of 33 to 34 may also be of similar magnitude to that of the rearrangment of the 2,2-dimethylpentoxyl radical noll.5 sec- 1 ). This seems reasonable since A(k 4 ) is consequently predicted to 10 13 , 5 sec- 1 , be an appropriate magnitude for the ring opening reaction. ln 23 = ( 8) + 25 + 26 Surprisingly, perhaps, a small temperature dependence was also observed on the relative yields of 25 and 26. Treatment of the yield of 25 versus 26 as in equation (8) gives a linear plot from which the difference in the activation energies leading to 25 and 26 is found to be 1.6 kcal/mole, favoring rearrangement to 26. The ratio of the frequency factors favors rearrangement to 25 by a factor of 3. The absolute magnitudes of the activation energies for ring closure and disproportionation of biradical 35 are expected to be very close to zero, and certainly less than 5 kcal/mole . we may combine the activation energies estimated above with group additivity estimates of the heats of formation 34 of the discrete molecular species in Scheme XII to produce an energy surface for the reaction of 19 (figure 6). 40 The activation enthalpy predicted by the energy diagram for the reaction of 19 to give 33 is 27.4 kcal/mole (experimental) while the value for the conversion of 23 to 33 is predicted to be 24 kcal/mole. It has been observed experimentally that in order to obtain a rate of rearrangement of 23 equal to that of 19, much higher temperatures are required. The difference in reactivity must be due to a difference in A factors.40 The most convenient H 87- 19 ~......r- (f ?' Coordinate Enthalpy diagram for the reaction of 19 (all units in kcal/mole). Figure 6. Reaction _.;_;_--'-"-C=-:::~-'-----J --------76 - - - - - - - - - - - - - - 89 O.__________________________ 20 60 80 ~ ✓)(:::: 1oot91-,-·- - - ' 120~------------------------, ,l'> I.D 50 explanation of this difference is that alkyl substitution at the acetylenic positions lowers the frequency factor for cyclization due to steric crowding in the transition state. Conclusions In Chapter II we have presented observations that constitute strong support for the mechanism outlined in Scheme XII, and that provide information about the relative rates of the fast reactions of intermediates formed in the thermal reaction of 19 at elevated temperatures. The question concerning the reactive spin state of di-n-propy 1-1, 4-dehydrobenzene, addressed. remains to be Our efforts in this area are presented in Chapter III of this dissertation. In relating the data obtained for 33 to the parent 1,4dehydrobenzene, 3, it seems reasonable to postulate that 1,4dehydrobenzene lies in an energy minimum as does 33. The absence of an intramolecular hydrogen transfer pathway such as that available to 33 should make the fastest reaction channel available to 1,4-dehydrobenzene the ring opening back to diacetylene 4. The barrier to ring opening is likely to be within the range observed for opening of 33 to 19 and 23 (ca. 10 and 16 kcal/mole, respectively) and log 10 A is probably on the order of 13. The short lifetime predicted for 1,4-dehydrobenzene by this analysis stands in sharp contrast to the observations reported by Berry, _g_t al.5 (Chapter I). They claim to have generated 1,4-dehydrobenzene photochemically and, on the basis of 51 time-of-flight experiments, reported it to have a lifetime in excess of two minutes under their reaction conditions. Assuming the activation parameters for 1,4-dehydrobenzene suggested above, their estimate of the lifetime of 3 is at least several orders of magnitude too large. Their failure to detect evidence for the formation of diethynyl olefin 4 further suggests that 1,4dehydrobenzene was not, in fact, generated under their experimental conditions. A final question that needs to be addressed is whether or not the data presented here rule out the possibility that the lowest energy state of 1,4-dehydrobenzene may correspond to the bicyclic butalene structure. While the reactivity demonstrated by the 1,4-dehydrobenzene intermediate 33 is clearly that of a biradical, the possibility that a bicyclic ground state may be in equilibrium with the biradical or that the reactivity of butalene may be identical with the reactivity expected of biradical cannot be rigorously ruled out. the 1,4- The only other argument which bears on this point is that theoretical treatments have found the biradical structure to be substantially higher in energy than the "open" or biradical form. 52 Experimental Section General Pyridine was distilled from CaH 2 after heating at reflux for several hours. Dry diethyl ether was obtained from a commercial source (Mallinckrodt, anhydrous) and was used fresh from the container without further purification or drying. Absolute ethanol was also commercially available and was used without further purification. Reagent grade petroleum ether (bp 35-60° C) was purified by passing it through a column of activity I Alumina. Pyrolysis solvents were purified by repeated fractional distillation through a glass helices-packed column until only trace amounts of impurities (<0.1%) were detected by analytical VPC. IR spectra were obtained on a Perkin-Elmer Model 237 or model 257 grating spectrophotometer. 1 H-NMR spectra were recorded on an EM-390 spectrometer. Chemical shifts are expressed in ppm downf ield from tetramethylsilane. High resolution mass spectra (HRMS) were obtained on an AEI-MS12 spectrometer. VPC-mass spectral (VPC-MS) analyses were carried out using a Finnigan 4000 GC-mass spectrometer. Elemental analyses were performed by the Microanalytical Laboratory, operated by the College of Chemistry, University of California, Berkeley, California. Melting and boiling points are reported uncorrected. Preparative VPC was performed on a Varian 90P instrument. Analytical VPC was conducted on either a Perkin-Elmer 3920 or a 53 Perkin-Elmer Sigma 3 chromatograph. Both were equipped with flame ionization detectors (FID) and were interfaced with a Spectra Physics Autolab System 1 computing integrator. The VPC columns used in the work reported here were the following: preparative VPC-Column A: Chrom W; Column B: for 10' x 1/4" glass 10% SF-96 on 60/80 12' x 1/4" glass 10% SE-30 on 60/80 Chrom W- AW/DMCS; for analytical VPC- Column C: 9' x 1/8" stainless steel 10% SF-96 on 100/120 Chrom W-AW/DMCS; for VPC-MS analyses- 30 m glass capillary SP2100 wall coated open tubular (WC0T) column. Gas phase pyrolyses Manufacturing Company Type were performed with a Hoskins FD 303A Electric Furnace. The pyrolysis tube was made of quartz tubing 35 cm long x 12 mm diameter and was fitted with 14/20 outer joints on the end. Flow pyrolyses were performed by passing a stream of N2 gas over a magnetically stirred sample of the material to be heated; the gas flow was then passed through a liquid N2 cooled trap and finally through a meter with which the flow rate was determined. Vacuum pyrolyses were performed with the same apparatus except that a vacuum (regulated by a manostat) was applied after the cold trap. Because of low volatility, the compounds studied were usually heated gently in order to increase their vapor pressure. Solution pyrolyses were performed in the following way: the diethynyl olefins were isolated >99% pure by preparative VPC and promptly dissolved in the pyrolysis solvent in order to prevent discoloration due to polymerization. The solutions were syringed into hexamethyldisilazane treated glass tubes fitted with 14/20 54 female joints. After four freeze-pump-thaw cycles (to 0.02 torr) the tubes were sealed under vacuum. Samples for the CIDNP experiment were prepared in the same way except that NMR tubes fitted with 14/20 female joints were used, hexamethyldisilazane treatment was eliminated. and the Solutions were allowed to react by submersing them in an oil bath heated to the desired temperature. The concentration of the diacetylenes was determined by comparing the integrated peak area observed by FIDequipped VPC with the area of a known amount of an internal standard which had been added to the solution. The internal standards were n-alkanes of carbon number similar to the starting di acetylene. The error in the diacetylene concentration estimated by this method is expected to be small. The yield of unimolecular products was determined using the assumption that the response factors of compounds of the same molecular formula are equal. 41 The yield of higher molecular weight products was estimated by assuming that the response factors of these hydrocarbons, relative to those of the unimolecular products, was proportional to the number of carbon atoms in each. 41 Samples for kinetic experiments were prepared in the same manner. The reaction temperature was controlled by submersing the reaction tubes in a vigorously bromobenzene, 1560 C). bp = refluxing solvent (~. Care was taken to minimize the contact of the sample tube with the walls of the solvent flask since this could have introduced an error into the reaction temperature. Data points were derived from the mean value of three VPC analyses of each sample. 55 Synthese q_ The synthet i c schemes employed in the preparation of the pairs of diacetylenes 19 and 21, and 23 and 24, were identical. Procedure s are described in detail for and 19 Only 23. the the syntheses leading to properties of the corresponding intermediates leading to 21 and 24 are given. Pro py l ethynyl ketone and ethyl ethynyl ketone: These compounds were prepared by the method of Bowden, tl al. 42 Propyl ethynyl ketone: Bp = 52-530 Cat 45 torr (literature 42 : 6 5 - 6 6 o C a t 1 0 0 to r r ) . (sextet,2H,J=7), 2.52 J O. 9 5 ( t , 3 H, J = 7 ) , 1 . 6 7 1 H- NM R ( CDC 1 3 ) : (t,2H,J=7), Bp = 3.20 (s,lH). Ethyl ethynyl ketone (a strong lachrymator and sternutator) was isolated in 58% yield. Bp = 108-110° Cat 1 atm. lH-NMR (t,3H,J=8), 2.56 (q,2H,J=8), 3.15 (s,lH). (acetylenic C-H stretch), (CDCl3): IR (thin film): Jl.12 3260 2790-2990 (alkyl C-H stretch), 2090 (C- C triple bond stretch), 1690 cm-1 (carbonyl). This compound was too easily air oxidized to allow satisfactory elemental analysis; the HRMS, however, was obtained: molecular weight calc. for C5H60 = 82.0419; found 82.0422. 1-Trimethylsilylhexyne: To an oven-dried 1 L three-neck flask fitted with two addition funnels, a condenser cooled to 5° C and an N2 inlet was added 500 mL anhydrous ethyl ether and 1hexyne (30 g, 0.37 rnol). The solution was cooled to -20° C under an atmosphere of N2 and stirred rapidly with a magnetic stirring bar. 165 rnL of a 2.42 M (0.40 rnol) hexane solution of nBuLi was added over 1 h. to r.t. over After addition the solution was allowed to warm 1 h. A white ppt. (the alkynyl 1 i thi urn salt) 56 rapidly f or med. The solution was cooled to -20° c and trimethylsilyl c hloride (45.7 g, 0.42 mol) was added over 20 min. Reaction was complete after the solution was stirred at r.t. 4 h. for The reaction was worked up by pouring 400 mL H2 0 into the flask and then separating the mixture. washed with 250 mL ethyl ether. The aqueous phase was The combined organic phases were washed with 150 mL H20 and then dried over Na 2 so 4 . The ethyl ether solution was concentrated on a rotary evaporator and the product was isolated by fractional distillation through a Ta wire column. The product was obtained as a colorless liquid, bp 70- 71° C (35 torr); isolated yield 44.1 g (77%), 99% pure (by VPC analysis) . The NMR and bp agreed with those reported in the literature. 43 1-Trimethylsilylpentyne : The isolated yield was preceding preparation), 38g, 74% (96% pure by VPC). at 35 torr. (see Bp 55-57° C The physical properties agreed with those reported in the 1 i tera ture. 44 1,3-lti..§.(trimetbylsilyl)hexyne: An oven dried 500 mL three neck flask was fitted with two addition funnels and an N2 inlet. 200 mL anhydrous ethyl ether and 15.4 g (0.10 mol) 1- trimethylsilylhexyne were added and stirred magnetically at -20° C under an N2 atmosphere. solution, 0.11 mol) (11.6g, 0.10 mol). nBuLi (45 mL of a 2.42 M hexane was added over 15 min followed by TMEDA The solution was stirred for 2 hat -20° c and then for 1 hat 10° C. After cooling the solution to -20° C again trimethylsilyl chloride (12.0 g, 0.11 moU was added over 57 15 min. A white ppt (LiCl) formed immediately. The reaction solution was stirred for 1 hat -20° C and then warmed to r.t. for 1 h. To work up the reaction, 100 mL H2 o was added and the mixture separated. ethyl ether . The aqueous phase was washed with 100 mL The combined organic phases were washed with 100 mL H20 and then dried over Na 2 so 4 . After concentrating the product solution on a rotary evaporator the products were distilled through a Ta wire column (isolated yield, 57%, >95% pure). The product was a colorless liquid but developed a pink color on short exposure to air at r.t; Bp 92-97° C (20 torr). (CDCl3): cf 0.09 (s,9H), 0.14 (m, SH). IR (thin film): 960, 760 cm- 1 . 835, (s,9H), 0.88 (t,3H,J=7), 1 H-NMR 1.82-1.18 2950, 2150, 1460, 1400, 1250, 1055, HRMS : Calcd. for c12 H26 si 2 , 226.1573; found, 226.1578. 1,3-Bis Ctr imethylsilyl) pentyne: described above for the hexynyl Prepared by the method isomer. The product was a colorless liquid (bp 86-90° c, 30 torr) which rapidly turned pink on exposure to air at r.t. by VPC analysis). Isolated yield, 18g (62%) (>95% pure Preparative VPC gave samples of high purity for the foll owing analyses: (s,9H), 1.05 1 H-NMR (CDCl 3): Jo.lo (s, 9H), 0.16 (t,3H,J=6), 1.23-1.68 (m,3H). IR (thin film): 2980, 2943, 2920, 2168, 1258, 1075, 1034, 990, 904, 850, 763, 702, 640, 618 cm-l. HRMS: Calcd. for c 11 H24 si 2 , 212.1416; found, 212.1419. E- and Z-4-Ethynyl-5-Ctrimethylsilyl}ethynyl-4-octene: To an oven dried 100 mL round bottom flask capped with a rubber septum and flushed with N2 was added 25 mL anhydrous ethyl ether 58 and 1,3-bis(trimethylsilyl)hexyne (3.0 g, 13 mmol) . magnetic stirring at -20° c, With rapid nBuLi (5.2 mL of a 2.4 M hexane solution, 13 mmol) was added over 10 min followed by TMEDA Cl.5 g, 13 mmol). After stirring for 2 h the solution was cooled to -10° C and propyl ethynyl ketone (1.25 g, 13 mmol) was added in less than 3 sec (in order to minimize abstr a ction of the acetylenic hydrogen). The solution was slowly warmed to r.t. over 2 hand then stirred for an additional 1 h. The reaction solution was poured into 40 mL of an aqueous solution of NH4Cl and then the organics were separated and washed 2 times with 40 mL H2 o. The organics were dried over Na 2 so 4 and then passed through 5 g silica gel to remove polymeric materials. Concentration on a rotary evaporator gave a light brown oil. Chromatography on 80 g silica gel using pet ether eluent gave satisfactory separation of the reaction products. (trimethylsilyl)ethynyl-4-octene: 0.61 g (22%) isolated yield, >98% pure (determined by VPC analysis). (s,9H), 0.87 3 . 13 21 4o, (s,lH) . 1 46o, C15H24Si: 1 H-NMR CDCCl3): (1 5 %) • J0.16 (t,6H,J=7), 1.43 (quintet,4H,J=7), 2.12 (t,4H,J=7), IR (thin film): 1 248, 87 o, 8 4o, c, 77.54; H,10 . 41. 3315, 3283, 7 5 6 c m- 1 . Found: l H- NM R (CDC 1 3 ) : cf 0 .16 2960, 2932, An.sh ca 1 c d. c, 77 . 72; H, 10.34. Ethynyl-5-(trimethylsilyl)ethynyl-4-octene: g Z-4-Ethynyl-5- ( s , 9 H) , isolated yield, 0. 8 5 (quintet,4H,J=7), 2.30 (t,4H,J=7), 3.30 (s,lH). ( t, 6 H, J = 7) , 2875, f or E-40.50 1. 4 6 IR (thin film): 3315, 2962, 2932, 2876, 2138, 1460, 1250, 1167, 964, 872, 840, 756 cm- 1 . An.sh Two attempts were made to obtain an elemental 59 analysis of this compound. The results were unsatisfactory due to decomposition of the sample during handling. mass calcd. for c 15 H24 si, 232.1647; found, HRMS: precise 232.1648. E- and Z-hexa-l-trimethylsilyl-2,3-diethyl-l,5-diyn-3-ene: These compounds were prepared by the procedure described immediately above. 3-ene: Z-hexa-l-trimethylsilyl-2,3-diethyl-l,5-diyn- Isolated yield, 1.23 g 1 H-NMR Cl 7%). (CDC1 3 ): cf O.21 (s,9H), 1.07 (t,6H,J=7), 2.19 (quartet,4H,S=7), 3.12 (s,lH). (thin film): 3320, 3295, 2980, 2944, 2880, 2140, 1468, 1546, 1251, 1161, 1048, 984, 958, 905, 842, 759 cm- 1 . mass calcd. for HRMS: c 13 H2O si, 204.1334; found, 204.1335. trimethylsilyl-2,3-diethyl-l,5-diyn-3-ene: (9%). IR 1 H-NMR (CDCl3): {quartet,4H,J=6), 3.32 O.6 g (t,6H,J=6), 2980, 2942, 2880, 2135, 1466, 1255, 987, 894, 850, 759 cm -1 . HRMS: c 13 H2O si, 204.1334; Z-4,5-Diethynyl-4-octene (19): from {thin film): 2.36 3320, precise mass calcd. for IR E-hexa-l- Isolated yield, Jo.21 {s,9H), 1.07 (s,4H). precise found, 204.1335. This compound was prepared Z-4-ethynyl-5-(trimethylsilyl)ethynyl-4-octene in 81% yield by the method of Arens and Schmidt. 26 The reaction products were passed through a short column of silica gel after workup and 0.41 g 19 was obtained >98% pure (determined by VPC analysis). Product 19 was a clear liquid which discolored rapidly upon standing at r.t. 1 H-NMR (sextet,4H,J=7), 2.06 {CDCl3): (t,4H,J=7), 3.05 Jo.Bo (s,2H). (t,6H,J=7), 1.42 IR (thin film): 3316, 3292, 2964, 2938, 2878, 2097, 1460, 1380, 1250, 1110, 1090, 842, 792, 736 cm -1 . Anak The sensitivity of 19 to air and thermal decomposition resulted in an unsatisfactory elemental 60 analysis. HRMS: precise mass calcd. for c 12 H16 , 160.1252; found, 160.1253. Z-Hexa-2,3-diethyl-l,5-diyn-3-ene 1 H-NMR above. 3.17 (s,2H}. (CDCl3}: IR J 1.11 (21): (t,6H,J=8}, (thin film}: 3292, Prepared as for 2.22 2980, 1466, 1380, 1250, 1055, 952, 887, 630 cm- 1 . calcd. for c 10H12 , 132.0939; found, 19 (quartet,4H,J=8}, 2943, 2882, HRMS: 2097, precise mass 132.0939. E-Dodeca-4,8-diyn-6-ene and E-deca-3,7-diyn-5-ene: The procedure reported by Ukhin and coworkers for the preparation of E-hexa-l,6-diphenyl-1,5-diyn-3-ene, except for the workup, was employed.45 The workup was changed as follows: the reaction solution was cooled to r.t. and filtered to remove the copper salts. The solids were washed with petroleum ether and the combined organic solutions were washed twice with aqueous 10% HCl and once with H2 o. After drying over Na 2 so 4 , the solution was concentrated on a rotary evaporator to produce a brown oil. The oil was chromatographed on silica gel using pet ether as eluent and the desired product was obtained in >95% purity. Both diacetylenes crystallized from pet ether solution at -20° c and could be further purified by drawing off the supernatant liquid with a pipet. E-dodeca-4,8-diyn-6-ene: 1 H- NM R (CDC 1 } : 3 cJ O. 9 8 ( t, 6 H, J = 7} , (t,4H,J=7}, (s, 5.87 2H}. 1. 5 4 isolated yield, 75%. (sextet , 4 H, J = 7} , IR (thin film}: 3035, 2962, 2. 31 2936, 2875, 2817, 2220, 1753, 1460, 1428, 1380, 1338, 1327, 1278, 935 cm- 1 . Anal. Calcd. for 89.84; H,9.99. E-Deca-3, 7-diyn-5-ene: c 10 H12 : C,89.93; H, 10.07. Found: isolated yield, 71%. C, 1 H- 61 NMR J 1.13 (t ; 6H,J= 8 ), (CC1 4 ): (s,2H). IR (thin film): 3030, 2 . 28 2980, (quartet,4H,J=8), 5.70 2940, 2842, 2910, 2220, 1455, 1435, 1320, 1185, 1060, 938 cm- 1 • mass calcd. for c 10 H12 , 13 2.093 9; Z-Dodeca- 4,8-diyn-6-ene (23) 2878, HRMS: precise found, 132.093 8. and Z-deca-3,7-diyn-5-ene (24): Trans 23 and trans 24 were photoisomerized as follows: deaeration, after a solution of the diacetylene in pet ether (<2% v/v) was photolyzed in a quartz vessel with a medium pressure Hg lamp. The photolysis was monitored by VPC (Column A) and stopped when the photostationary ratio of the cis and reached (ca. 40:60). The mixture column chromatography of trans isomers was geometric isomers was separated by on silica eluent). The cis isomer was obtained in high purity Compound 23: cfl.04 (sextet,4H,J=7), 2.41 gel (pentane (>95%). (t,6H,J=7), (t,4H,J=7), 5.74 (s,2H). 1.62 IR (thin film): 2960, 2930, 2870, 2210, 1675, 1575, 1461, 1452, 1427, 1392, 1378, 1334, 1324, 1274, 740 cm- 1 . H, 10.07. Found: C, Anal, 90.03; ccc1 4 ): d · l.19 (s,2H). IR (thin film): H, (t,6H,J=7.5), Calcd. for c 12 H16 : 9.99. 2.36 Compound for c 10 H12 , 3026, 2978, 2940, 132.0939; 2 4: (quartet,4H,J=7.5), 1556, 1396, 1160, 1100, 1056, 740 cm- 1 . calcd. c, 89.93; 2916, HRMS: 2778, 5.57 2208, precise mass found 132.0938. 2,2,5,5-tetradeuterio-l,4-cyclohexadiene: l,4- Cyclohexadiene-d 4 was prepared by base catalyzed exchange of the allylic protons for deuterium by treatment with d 5 -dimsyl anion in DMS0-d 6 : to oil free 46 NaH (24 mmol) was added dry DMS0-a 6 (Merck and Co., 99.5% D, 0.29 mol) in a flask fitted with a 62 condenser (5° C) and flushed with N2· The mixture was heated to 75° C for 45 min to generate the dimsyl anion. 46 The solution was cooled to 23° C and, with rapid stirring, 1,4-cyclohexadiene (43 mmol) was added as fast as possible. A red color rapidly developed and the reaction was quenched by addition of D2 0 (50 mmol, 99.7% D) after 1 min. Hexadecane (15 mL) and ice water (30 mL) were added and, after stirring for several minutes, the mixture was forced through a coarse-frit filter to remove solids. The organic phase was removed and washed twice with cold H2 o. The aqueous phase was washed twice with hexadecane and the organic phases were combined and dried over Na 2 S04• The volatiles were isolated by bulb to bulb distillation at 0.05 torr. Preparative VPC (Column A, 30° C) gave 21% isolated yield of 1,4-cyclohexadiene with 92% deuterium incorporation in the After additional drying over Na 2so 4 the exchange was repeated a second time; the isolated yield after 2 allylic positions. exchanges was 9% and by 1 H-NMR and MS the product was observed to have 97.7% deuterium incorporation in the allylic positions. The primary complication encountered with this procedure was the presence of competing reactions which generated benzene and cyclohexene. Exchange was the (slightly) faster process; reaction times minimized side product formation. short Perhaps the best way to improve this reaction would be to use an additional solvent such as diglyme which may be cooled to lower temperatures lower than those obtainable using neat DMSO. 63 Thermal Reactions Pyrolysis of Z-4,5-diethynyl-4-octene (19): The gas phase pyrolysis was performed under a flow of N2 with a contact time of ca. 2 min and an oven temperature of 320° composed of 23, 25 and 26, c. The pyrolysate, was a yellow liquid at r.t. The isolated yield of products was 76%. Preparative VPC (Column B, 70° C) yielded the pure products. Thermal reactions of 19 in solution were analyzed temperature program: on Column C using the following initial temp. 150° C for 15 min; at 5° C/min; hold at 220° C for 20 min. increase The injector temperature was kept <235° C to prevent significant injector port reaction of 19. Compound 28 was obtained pure for analysis from the thermal reaction of 5 mL of a 3xlo- 2 M solution cyclohexadiene plus chlorobenzene of (10% v/v), 19 in 1,4- followed by preparative VPC on Column A Cat 125° C). Benzocyclooctene (25): 29 1 H-NMR (CC1 ): 4 Jl.34 Cbr m,4H), 1.66 (br m,4H), 2.67 (d of d,4H,J=5), 6.93 (s). IR (thin film): 3000, 2910, 2836, 1486, 1463, 1445, 1353, 1110, 748, 704 cm- 1 . HRMS: precise mass calcd. for c 12 tt 16 , o-Allyl-n-propylbenzene (t,3H,J=7), 1.60 160.1252; found, (26): (sextet,2H,J=7), 160.1258. 1 H-NMR (CC1 ): 4 2.54 J0.97 (T,2H,J=7), 3.32 (d,2H,J=7), 4.74-5.12 (m,2H), 5.65-6.18 (m,lH), 7.00 (s,4H). (CC1 4 ): 2940, 2915, 2855, 1635, 1435, precise mass calcd. for c 12 H16 , o-Dipropylbenzene (28): 1.60 film): (sextet,4H,J=7), 2.56 990, 915 cm-l. IR HRMS: 160.1252; found 160.1250. 1 H-NMR (CC1 ): 4 (t,4H,J=7), 6.98 cf0.97 (t,6H,J=7), (s,4H). IR (thin 3062, 3018, 2962, 2936, 2874, 1488, 1466, 1452, 1376, 746 64 cm- 1 . HRMS: precise mass calcd. for c 12 H18 , 162.1408; found, 162.1414. Acid catalyzed exchange of aromatic hydrogen in 28: 31 A solution of 19 in chlorobenzene plus cyclohexadiene-d 4 (8.3% v/v) was heated at 190° C for 15 min in a sealed glass tube. The solution was then concentrated to 0.3 mL total volume by static transfer (O.O5 torr) of solvent. The solution was divided into two samples which were each treated in the following way: The chlorobenzene solution was placed in a glass tube with O.5 mL of an aqueous solution of HCl (4% v/v). After freeze-pump-thawing to remove oxygen the tubes were sealed and heated at 260° C for 42 h. The tubes were opened and the organic phases were removed by pipet. After VPC-MS analysis of 28 for deuterium content the solution was sealed in a tube as before with fresh aqueous HCl. After a second period of heating, VPC-MS analysis of 28 indicated no further change in the deuterium content (Scheme XIII). Pyrolysis of Z-hexa-2,3-diethyl-l,5-diyn-3-ene (21): Gas phase and solution thermal reaction of 21 gave the products shown in Table 3. The solution and gas phase reaction mixtures were analyzed by VPC-MS. o-Ethylstyrene and tetralin were isolated from the gas phase reaction mixture by preparative VPC (Column A, 900 C) and identified by comparison of their NMR spectra with authentic samples. 47 Benzocyclobutene and o-diethylbenzene were identified by VPC retention time (Column C, initial temperature= 105° C for 15 min, temperature program = 6°/min, final temperature= 200°) and by comparison with the mass spectra 65 obtained from authentic samples under identical conditions of analysis. CIDNP experiment using z-4,5-diethynyl-4-octene (19): Solutions of 19 (O.l-0.5M) in diphenyl ether and o-dibromobenzene gave identical CIDNP signals upon reaction in the heated probe of a Varian EM-390 1 H NMR spectrometer (155-170° C). At the concentrations employed, reduction to yield o-dipropylbenzene occurred to the extent of ca. 5% of the yield of the unimolecular products; otherwise, the product distributions were the same as that observed in more dilute solution pyrolyses. 66 Chapter ll.l. DETERMINATION OF THE REACTIVE SPIN STATE OF 1,4-DEHYDROBENZENES Introduction In spite of the efforts of numerous investigators to generate and study the chemistry of 1,4-dehydrobenzene, the spin states populated under the reaction conditions have yet to be characterized. This is a particularly intriguing problem because the singlet and triplet states are presumed to be close in energy and because of the failure of theoretical treatments to reach a consensus in predicting the ground electronic state (Table 1, Chapter I). To date, the only reported experimental attempt to determine the spin state of a 1,4-dehydroaromatic is that of Chapman and coworkers 8 , who generated 9,10-dehydroanthracene (9, Chapter I) in a matrix at 8° Kand searched, without success, for an ESR signal which would have indicated population of the triplet state. In this chapter we detail our efforts to determine the number and description of the reactive spin states of the intermediates generated by the thermal diethynyl olef ins in solution. reaction of Our approach includes both chemically induced dynamic nuclear polarization (CIDNP) and chemical trapping experiments. As described in the preceding chapter, when 19 (cf. Scheme XII) was heated in a 1 H-NMR probe at 160° C, emissive signals were observed in 26 both in the vinyl protons and the terminal methyl and methylene protons of the alkyl side chain (figure 5). The location of the protons in 26 which showed emission indicates 67 that biradical 35 is the molecule in which the CIDNP effects arose. The following observations are inconsistent with S-T 0 mixing and indicate instead an S-T_ mixing mechanism: 48 (1) The protons alpha and beta to the radical centers showed the same polarization; this indicates that the sign of the hyperfine interaction has no effect on the spectrum. (2) All of the polarized signals were for emissive; normally, polarizing radicals with a g-value difference of zero, a mixture of enhanced absorption and emission (multiplet effect) is observed. The S-T_ mechanism has been observed at high magnetic fields only in small biradicals (unpaired electrons separated by fewer than ca. 10 carbon atoms). 48149 Closs 48 has delineated two mechanisms by which the CIDNP effects observed in 26 may be explained: if singlet 35 is present and is higher in energy than triplet 35, emissive signals may be observed if there exists a bimolecular reaction channel which drains off the triplet biradical formed by magnetic fieldinduced intersystem crossing (isc, Scheme XVI). On the other hand, if triplet 35 is produced in the reaction (by magnetic field-independent intersystem crossing in either 33, 34 or 35) and the ground state of biradical 35 is a triplet, then T_-s mixing can produce the observed signals, even in the absence of a spin-selective reaction channel. This analysis indicates that either may singlet polarizations; or triplet furthermore, 35 produce the observed it is difficult to distinguish between these possibilities on the basis of the experimental 68 obs e rv a d .~,.t1 so If th e tr i plet of 35 is produced in the reaction of 19 , th r 2e di s t inct modes of population of the triplet manifold are poss i bl e. Scheme XVII indicates that intersystem crossing in the 1,4 - de hydrobenzene biradical, 34 or 35 could all have led to the fo rmat i on of some fraction of 35 in the triplet state. even if triplet 35 is present it cannot be Thus, determined unambiguously in which of the three biradicals intersystem system cross i ng occurred. Scheme XSZI ~ magnetic field- induced intersystem crossing (isc) bimolecular products Results and Discussion CIDNP A more straightforward CIDNP analysis may be obtained by looking for polarization effects in the products of bimolecular reaction of the 1,4-dehydrobenzene biradical . This approach has been successfully applied reaction in the dimethyl-hexa-1,5-diyn-3-ene (38). thermal of 2,3- When a solution of 38 in hexachloroacetone (0.1 M) was heated to 160° C in the probe of a 69 Scheme :XW. ~ . :::=x ~ ~ ~ isc ~ ~ 19 ► -4 ~ ~ 33-S 33-T ! l ◄ isc ~ ... ◄ ~ 34-S 34-T l l isc ► 1 ◄ ~ 35-S l emission { "":: 35-T 70 90 MHz 1 H-NMR spectrometer, the spectrum obtained showed an emission in the aromatic region {figure 7). VPC analysis of the solution after reaction showed the formation of l,4-dichloro-2,3dimethylbenzene (39) and (relative yields, ca. 3:1). were detected bv VPC. l-chloro-2,3-dimethylbenzene (4 0 ) Only minor amounts of other products The emission observed during thermal reaction of 38 is assigned to the aromatic protons of 39. The broad proton absorption in the alkyl region (figure 7, spectrum (C)) is attributed to polymerization products formed as a result of the relatively high concentration of 38 in the NMR experiment. Thermal reaction of dilute solutions of 38 (0.01 M) in hexachloroacetone and cc1 4 gave results (Table 5) similar to the NMR experiment. Product 39 was isolated from a CC1 4 solution reaction by preparative VPC and characterized by IR, 1 H-NMR and HRMS. Table 5. Reaction of 38a in Solution at 190°C. Absolute Yield (%) Solvent 39 40 hexachloroacetone 17 5 20 5 a [38] = 10- 2M The mechanism shown in Scheme XVIII is proposed to explain the reaction of 38 in hexachloroacetone. thermal chemistry of 19, By analogy to the cyclization of 38 gives the 2,3- 71 (A) (Bl ~w.~~ ~?~~~~~ t,~ (D) Figure 7. CIDNP observed during reaction of a he.xachloro- acetone solution of 38. (A) NMR of solution before reaction . (B) Signals observed during reaction at 160° C. (C) Room temperature spectrum after complete reaction of 38. (D) Spectrum of 39 in cc1 . 4 72 dimethyl-1,4-dehydrobenzene biradical (41) which may abstract chlorine from solvent to produce a solvent-caged radical pair. Cage escape of the aryl radical (42) and abstraction of a second chlorine atom from the solvent gives 39. reactions of 41 and 42 lead to 40. Hydrogen abstraction The pentachloroacetonyl radicals generated by loss of chlorine may attack 38; this is presumed to be responsible for the modest yield of aromatic products. The observed polarizations can be readily interpreted by application of Kaptein's rules. 50 or (-) The analysis requires that (+) values be assigned to the parameters in equation 9. Because the aromatic signal appeared in emission, the product of the four parameters must be (-). Compound 39 was formed by escape from the polarizing pair, therefore The g value of 42 should be less is negative (-). f than that of the pentachloroacetonyl radical (cf. g values of phenyl (2.0025) and dichloromethyl (2.0080)) 51 so Ag = (-). The hyperfine coupling constant in phenyl radicals is positive for the ortho, meta and para hydrogens, so Ai= (+). The remaining parameter, _µ. , must be assigned a value which makes the product of the right hand side negative since the polarization in 39 was emissive; the sign of _µ, therefore, is (-) which indicates that the spin state Q.f. .tfil polarizing radical pair ~Mg singl.e.t. The singlet spin state of 1,4-dehydrobenzene 41 therefore must be the predominant source of product. The CIDNP observed in 26 may be reinterpreted in light of the findings for 41 by postulating that the polarizations either 73 Scheme XSZII1 ~ X: [:OJ I:::,. ... 38 41 ! (CCl 3 )2 CO • . )Q Cl 0 II CCl 2 CC Cl 3 l escape Cl :¢ ◄ (CCl 3 ) CO 2 )Q Cl Cl 39 42 Kaptein' s Rule: H • RH f7observed polarization) r 1-J == + }A {: {: AS i: t :¢ Cl 40 =fa€ .i13 Ai. absorption emission triplet radical pair singlet radical pair cage product escape product sign of difference between g-factors A~ {: sign of hyperfine coupling constant (9) 74 arose from singlet 35 or that triplet 35 was produced as a result of intersystem crossing in biradical 34 or 35 but not in the 1,4dehydrobenzene biradical. Chemical Trapping Studies A second approach to determining the spin states present in solution involves an attempt to distinguish between the chemical reactivity of the dehydrobenzene 33. triplet and singlet Because radical states of 1,4- pairs are generated by abstraction reactions of 33, the task reduces to finding a way to differentiate the reactivity of singlet and triplet radical pairs. 52 The spin correlation effect (SCE)53 postulates that radical pair reactivity is related to the spin state of the pair: singlet radical pairs may undergo both cage 54 and escape reactions but a spin prohibition against cage reactions limits triplet radical pairs to cage escape intersystem crossing). 55 (in the absence of In order to detect the presence of singlet and triplet radical pairs generated by trapping of singlet and triplet 33, we must distinguish between the cage and escape pathways leading to product formation; the magnitude of the ratio of cage/escape reactions will reflect the spin state of the radical pair generated from 33. Scheme XIX illustrates the cage and escape reactions that can occur in the radical pair generated by hydrogen transfer from 1,4-cyclohexadiene to biradical 33. While combination products 29 and 30 are unique to cage reaction, 56 28 is produced both by ~ Scheme XIX. ~Cl ~ 0 29+30 l + rxn. cage 0 escape 28 (d 0 + d 2 ) 0::::: H H ~~ H 28 (d0 +d 1+d 2 ) 0::::: -..J u, 76 cage disproportionation and cage escape; it is necessary, therefore, to determine the extent to which the cage and escape reaction pathways contribute to the yield of 28. It is possible to perform this analysis if a mixture of l,4-cyclohexadiene-d 0 and -a 4 is used in the reaction solution. reaction of 33 with cyclohexadiene-d 0 • Consider first the The cage reactions which the radical pair (formed by hydrogen transfer to 33) may undergo include tr~nsfer of a second hydrogen atom to give 28-d 0 and combination to produce products 29 and 30. Escape of the aryl radical from the solvent cage, followed by abstraction of hydrogen or deuterium from trapping agent will give 28-d 0 and -a 1 in the ratio 1:1 in the absence of a deuterium isotope effect. By a similar analysis, if 33 initially interacts with deuterated trapping agent, 28 formed by cage reaction will contain two deuteria while cage escape will lead to 28-d 1 and -d 2 in the ratio 1:1. In summary, cage reaction will lead to only 28-d 0 and -d 2 (1:1 ratio) and escape reactions of the radical pair should give 28-d 0 , -a 1 and -d 2 in the ratio 1:2:1. Because 28-a 1 is unique to the cage escape reaction channel, it is possible to dissect the experimentally observed ratio of 28-d 0 , (obtained by mass spectroscopic analysis) -d 1 and -d 2 into the relative contributions of the cage and escape pathways; when the yield of 29 and 30 is added to the yield of 28 produced by cage reaction, the ratio of cage to escape products (C/E) is obtained. It is impossible to predict,~ priori, the relative amounts of cage and escape reaction for a given singlet radical pair (the SCE postulates that C/E for a triplet pair is zero). For this 77 Scheme XX. ~ ~ 19 • 6 • kisc ~ ► ◄ • • 33-S 33-T 28 (d 0 : d 1: d2 = I: 2 : I ) 28 + 29, 30 78 reason an experimentally observed ratio of C/E by itself will provide limited quantitative information about the relative amounts of singlet and triplet 33 present in solution. analysis is broadened, relationships however, in Scheme XX. The by consideration of the kinetic Conservation of spin (in the cyclization reaction) requires that biradical 33 is initially generated in the singlet state. The ratio of trapping of the singlet biradical to intersystem crossing to triplet 33 will depend on the concentration of trapping agent in solution. 57 At low concentrations of cyclohexadiene intersystem crossing should be at its maximum value whereas a high concentration of the trapping agent will increase kT[SH] and the amount of intersystem crossing observed should be at a minimum. Compound 19 (0.01 M) was allowed to react at 195° c in a chlorobenzene solution which contained added cyclohexadiene (d 0 :d 4 = 1:4) ranging in concentration from 0.01 to 10.6 M. The data obtained by combined VPC and VPC-MS analysis are plotted in figure 8. The ratio C/E (0.55) did not vary, within experimental error, over the range 0.2 to 1.6 M cyclohexadiene. The ratio of C/E was experimentally difficult to determine for the entire product spectrum (28, 29 and 30) at low concentrations of cyclohexadiene; 58 the ratio of C/E for product 28, however, may be readily determined by VPC-MS analysis alone. As the lower plot in figure 8 shows, the ratio of C/E for product 28 (0.20) was independent of cyclohexadiene concentration from 0.01 to 10.6 M. The relatively large, constant value of C/E for the complete 79 0.7 0.6 (C/E \otol 0 0 0 0 0 0.5 0.4 (C/E)2a O O 0.2 0 0 0 0 0 0 0.4 0.8 1.2 1.6 [1,4-Cyclohexadiene -d0 td 4 ],M Figure 8. 10.6 Ratio of C/E observed in the reaction of 19 (0.01 M, 195° C) as a function of 1,4-cyctohexadiene concentration. and 30. Upper plot shows C/E for products 28, 29 Lower plot shows C/E for product 28 alone. 80 product spectrum (ca. 0.55) indicates the presence of the singlet radical pair generated from the singlet state of 33. If the ground state of 33 is a triplet but intersystem crossing from the singlet is slow relative to unimolecular and bimolecular reaction, exclusive trapping of the singlet state will be observed at all concentrations of cyclohexadiene. One way to increase kisc is to perform the reaction in a brominated solvent; the presence of bromine either in a reacting substrate or in the solvent is known from excited state chemistry to increase intersystem crossing rates (the heavy-atom effect). 59 When 19 was allowed to react in bromobenzene solution, the ratio of C/E was found again to be independent of cyclohexadiene concentration (figure 9). As before, the large value of C/E (ca. 0.64) suggests the exclusive formation of the singlet radical pair. The ratio C/E for product 28 (0.20) was also independent, within experimental error, of the concentration of cyclohexadiene. As experiments described in Chapter II demonstrated, both 33 and 34 are trapped by 1,4-cyclohexadiene. In the spin state study described above two radical pairs were generated (from 33 and 34) at low concentrations of cyclohexadiene and a single pair (from 33) presumably at high concentrations. produced multiplicity, from 33 However, without a because 34 is change in spin the spin state analysis above may still be appropriate; this is strongly supported by the observation that the ratio C/E was independent of cyclohexadiene concentration even though the relative amounts of trapping of biradical 33 and 81 0 .8 (C/E )total 0 0 0 0 .6 0 0.4 0.4 (C/E )28 0 0.2 0 0 0 0 0 1..----L.--..L.-----L.-----'-------''---_.__ _.,____ ___.__ 1.6 I. 2 0.8 0.4 0 [1,4 -Cyclohe xod iene -d 0 + d 4 ], M Figure 9. ___, Ratio of C/E observed in the reaction of 19 (0.01M, 195° C) in bromobenzene solution as a function of added 1,4-cyclohexadiene. 28, 29, and 30. Upper plot shows C/E for products Lower plot shows C/E for product 28 alone. 34 varied considerably. There are several alternatives to the conclusions drawn from the chemical trapping data. First, although the ratio C/E was found to have a value 0.6) (ca. which strongly suggests the presence of singlet 33, the same value could have been obtained if extremely rapid intersystem crossing produced an equilibrium ratio of singlet and triplet 33 at all concentrations of cyclohexadiene. If this were the case, the ratio C/E would reflect a component of both the singlet (C/E > 0.6) and triplet (C/E ca. 0) radical pairs. Another way to explain the results is that kisc and k 8 [SHJ are competitive, but kT[SHJ is very slow and failed to generate an appreciable amount of the triplet radical pair. It is difficult to imagine a factor that would lead to such an appreciable difference in the reactivity of singlet and triplet 33 toward cylohexadiene. cannot be ruled out, A final possibility, which is that kisc is too slow to produce an observable amount of triplet 33 under the reaction conditions, even if the triplet state is equal to or lower in energy than the singlet. Unfortunately, the lifetime of 1,4-dehydrobenzene biradicals are limited by ring opening and other unimolecular reactions, even in the absence of bimolecular reaction channels. Generation of 1,4-dehydrobenzenes at lower temperature might favor intersystem crossing over other reaction pathways; 60 the rapid cyclization observed for cyclic diethynyl olefin 8 (Scheme V, Chapter I) at ambient temperature suggests one synthetic approach. 83 Conclusions The unambiguous CIDNP result obtained in the reaction of 38 compares well with the chemical trapping study which indicates the presence of singlet state intermediates in the reaction of 19. The failure to detect evidence for the population of triplet 33 or 41 in these studies may be due to the short lifetime of the 1,4-dehydrobenzene intermediates under the reaction conditions. A slow intersystem crossing rate will preclude observation of the triplet biradical even if it is the lowest energy state. The lifetime of 2,3-dialkyl substituted 1,4-dehydrobenzenes may be estimated from the parameters obtained for 33 in Chapter II; at 200° c, unimolecular ring opening occurs with a half-life of ca. 10-8 to 10- 9 sec. Therefore, if the ground electronic state of 1,4-dehydrobenzenes 33 and 34 is the triplet, population from the singlet must occur with a rate constant <10 9 sec. Experimental Section General See Chapter spectrometers, II for descriptions of NMR, IR and mass VPC chromatographs a nd columns used. The method of preparation of solution and CIDNP reaction samples has also been d.escr ibed. Synthesis 2-Bromo-3-butyne: A procedure for a similar reaction has been outlined by Ashworth, Whitham and Whiting. 61 A dry ether solution (18 mL) of l-butyn-3-ol (22.2 g, 0.317 mol) and pyridine (0.22 mL) were placed in a 3-neck flask fitted with an addition funnel, an argon inlet and a reflux condenser. To the argon- flushed flask cooled to s° C was added an ether solution (12 mL) of freshly distilled PBr 3 (42.5 g, 0.317 mol) over 4 h. Reaction was complete at the end of the addition (VPC analysis on Column A, 45° C). water was carefully added to the clear orange-brown reaction solution until fuming ceased. The organic phase was separated and washed with 50 mL of an aqueous NaHco 3 solution followed by a wash with 50 mL brine. The ether solution was dried over Mgso 4 ; distillation through a Ta wire column at 1 atm gave 13.6 g (32% yield) of 2-bromo-3-butyne (>98% pure, as determined by VPC) bp 83-90° C. (d,1H,J=2.5), 4.51 1 H-NMR (CC1 4 ): Cd of q,1H,J=7 ,2.5). cf 1.90 (d,3H,J=7), IR (thin film): 2.53 3300, 2950, 2120, 1430, 1370, 1300, 1180, 1090, 1060, 990, 970, 855 cm 1. .8.Il-9.h Calcd. for c 4 H5 Br: C, 36.13; H, 3.75. Found: C, 85 35.95, H, 3.76. 3,4-Dimethyl-hexan-1,5-diyn-3-ol: An oven dried 3-neck flask was fitted with an addition funnel, condenser, thermometer and an argon inlet. ethyl ether Magnesium turnings (2.45 g, 0.10 mol) dry (10 mL) and a small amount of HgCl2 were added. After the solution became cloudy the reaction flask was cooled to 10° C and an ethyl ether (65 mL) solution of 2-bromo-3-butyne (13.6 g, 0.10 mol) was added over 1.5 h. A clear, faint yellow solution resulted. The solution was cooled to -15° C and methyl ethynyl ketone (6.9 g, 0.10 mol) dissolved in ethyl ether (70 mL) was added over 1 h. A white ppt. for med after the addition was one-half complete. After the addition, the solution was warmed to room temperature over 0.5 hand poured into a cold, saturated aqueous NH 4 Cl solution. The organics were isolated and the aqueous phase was washed three times with ethyl ether. The combined ethyl ether solutions were washed with 150 mL brine and then dried over Na 2 so 4 . Concentration on a rotary evaporator gave a somewhat volatile reddish oil. The crude product was purified by passing through a pad of silica gel with a mixture of pet ether and ethyl ether as eluent (7:3 v/v). Static vacuum distillation (>98% pure as (0.03 torr) of the oil gave 8.4 g determined by VPC) of 3,4-dimethyl-hexan-1,5-diyn-3-ol (70%) as a colorless oil. The alcohol was formed as a mixture of diastereomers (5:1 ratio) which could be separated by preparative VPC 1.33 (column B, 100° C). (d,3H,J=7.5), 1.54 NMR of major (s,3H), 2.19 diastereomer (CDCl3): (d,1H,J=2.5), 2.46 86 (s,2H,0H,acetylenic H), 2.65 (m,1H,J=7.5,2.5). IR (thin film): 3250, 3220, 2980, 2110, 1710, 1440, 1370, 1250, 1090, 1030, 1000, 970, 720 cm- 1 . Anal, Calcd. for c 8 H10 o: c, 78.65; H, 82.5. Found: C, 78.36; H, 8.23. A 25 mL 3-neck flask 2,3-Dimethyl-hexa-1,5-diyn-3-ene (38): was fitted with an argon inlet, addition funnel and a stopcock through which aliquots could be removed. After flushing with argon, 3,4-dimethyl-hexan-1,5-diyn-3-ol (1.65 g, 0.011 mol) and pyridine (3.5 mL) were added. The mixture was cooled to 5° C and a solution of P0C1 3 (1.98 g, 1.2 mL, 0.013 mol) in pyridine (1.7 mL) was added with stirring over 30 min. monitored by VPC (Column B, 98° C). The reaction was After the addition of P0C1 3 , little reaction had occurred, so the mixture was warmed slowly to r.t. After 1 hat room temperature the reaction was complete. The dark reaction mixture was poured over ice with pet ether and The organic layer was separated and the aqueous phase was washed 3 times with pet ether. The combined organics were washed with 10% aqueous HCl until acidic by litmus and then washed with H2 o to pH 4. The pet ether solution was dried over Na 2 so 4 and concentrated to give 0.63 g (50% yield) of an orange oil (>95% pure by VPC). Crystallization of cis and trans 2,3- dimethyl-hexa-1,5-diyn-3-ene purified product. (ratio 1:3) at -70° C gave further The geometric isomers were isolated by column chromatography on silica gel (pet ether eluent) . NMR (CDC1 3 ): IR (thin film): 3310, 3020, 2935, 2874, 2108, 1445, 1386, 1250, 1156, 1103 cm- 1 . Trans IR (thin isomfil: NMR cfl.88 (s,6H), 3.17 (s,2H). .ci..fi llQIDil: (CDC1 3 ): cf2.02 (s,6H), 3.36 (s,2H). 87 film): 3318, 2970, 2940, 2872, 2002, 1735, 1445, 1378, 1265, 1235, 1166, 1098, 800 cm- 1 . 104.0625; found for ~i~ HRMS: precise mass calcd. for 38, 104.0623; found for c8 H8 , ~~~n~ 38, 104.0622. Thermal Reactions Reaction of 2,3-dimetbyl-hexa-1,5-diyn-3-ene (38) chlorinated solvents: in cc1 4 (MCB spectral quality) was used without additional purification. Hexachloroacetone was purified by repeated distillation at 50 torr through a vacuum-jacketed glass helices-packed column. (Column A, 85° C). 38 was purified by preparative VPC n-Octane was used as an internal standard in the thermal reactions. Yields of products were determined by reference to the internal standard with the assumption that the response factor of the starting material and products were the same. 41 Pyrolyzed solutions were dark, guinness-brown, suggesting the occurrence of substantial polymerization. Reaction product 39 was isolated from the cc1 4 reaction mixture by preparative VPC (Column A, 135° C) and identified by NMR, IR and HRMS (see below). (m/e 140). 40 was identified by its mass spectrum No evidence for high molecular weight products, formed by combination of the radical pairs generated by transfer of chlorine (~. Scheme XVIII), was detected by VPC-MS analysis of the hexachloroacetone solution reaction. In the cc1 4 reaction a product with m/e 256, 258, 260, 262 was detected; this may have been formed by the reaction of aryl radical 42 with the 88 trichloromethyl radical. l,4-dichloro-2,3-dimethyl-benzene (39): NMR (CCl4): J2.33 (s,6H), 6.98 (s,2H). IR (CCl4): 3020, 2940, 1875, 1458, 1413, 1389, 1264, 1161, 1136, 1030, 833, 590 cm- 1 . HMRS: precise mass calcd. for c 8 H8 c1 2 , 174.0003; found, 174.9999. Reaction of 4,5-diethynyl-oct-4-ene d 0 and -a 4 : (19) in cyclohexadiene- Compound 19 was isolated >99% pure by preparative VPC and piomptly dissolved in chlorobenzene to give a solution 0.0lM in 19. n-Undecane was added as an internal standard and the initial concentration of 19 was determined by comparison of the integrated analytical VPC peak areas. Cyclohexadiene-d 0 and - a 4 we r e a a a ea to the ch 1 or ob en z en e so 1 u t i on of 19 in a py r o 1 y s i s tube which was promptly sealed. After reaction for 15 min at 195° c, the solutions were very light yellow colored. The reacted solutions were analyzed by VPC (Column C, initial temp. 150° C for 15 min; increase at 5° C/min; hold at 220° C for 20 min) and the product yields determined by reference to the internal standard. The relative yields of 28-d 0 , determined by VPC-MS analysis -a 1 and -a 2 were (WCOT capillary column (see Experimental Section, Chapter II} initial temp. 110° C; increase at 4° C/min; hold at 220° C). -a 4 used was 1:4. The ratio of cyclohexadiene-d 0 and With this ratio, 28-d 0 and -a 2 were formed in nearly equal yield (kH/k O is 4). The ratio of C/E was determined as follows: cage escape (E) is the only source of 28-d 1 ; escape Therefore, the gives 28-d 0 , -a 1 and -a 2 in the ratio 1:2:1. escape component of 28-d 0 and -a 2 is 1/2 the yield of 28-a 1 . The remainder of 28-d 0 and -a 2 was formed by cage reaction (C) and 89 was added to the yield 0 £ 29 and 30 to give the yield of cage products. 90 REFERENCES 1. For reviews, see (a) Hoffman, R. W.; "Dehydrobenzenes and Cy c 1 o a 1 k y n es , " Ac ad e rn i c Pr es s , New Yo r k , 1 9 6 7 ; ( b ) Fi e 1 d s , E. K., in "Organic Reactive Intermediates," McManus, s. P., Ed., pp. 449-508, Academic Press, New York, 1973, Levin, R. H., "Arynes", in "Reactive Intermediates", Jones, M.; Moss, R. A., Eds., John Wiley and Sons, New York, V. 1, 1978. 2. (a) Chapman, O. L.; Mattes, K.; McIntosh, C. L.; Pacansky, J.; Calder, G. V.; Orr, G.; .Ji. bID...t- ~he.m_._ ~Q.h, 1973, .2..5., 6134; (b) Chapman, o. L; Chang, c.-C.; Kole, J.; Rosenquist, N. R.; Tornioka, H.; .Ji. bID...t- ~MID..t- ~ , 1975, l i , 6586. 3. (a) For experimental evidence for bicyclic 1,3dehydrobenzene 2a see Washburn, w. N.; Zahler, R.; .Ji. bID...t~.m_._ ~ 1978, 100, 5873, and references therein; (b) evidence for the biradical isomer 2b is presented in Billups, W. E.; Buynak, J. D.; Butler, D.; I L L ~ ~.m_._, 1979, l i , 4218. 4. Fischer, I. P.; and Lossing, F. P.; ILL bID...t- ~filID..t- ~ , 1963 .!l.2, 1018. 5. Berry, R. S.; Clardy, Lett,, 1965, 1003. 6. (a) J.; Schaefer, M. E.; Tetrghedron Jones, R. R.; Bergman, R. G.; Ji. bID...t- ~filill..L ~ , 1972, l l , 660; (b) Bergman, R. G.; ~ ~ill..L ~ - 1973, .Q., 25. 7. Darby, N.; Kirn, C. U.; Salaun, J. A.; Shelton, Takada, S.; Masarnune, S.; Chern, Commun,, 1971, 1516. 8. Chapman, O. L.; Chang, C. C.; Kole, J.; ILL bill.L ~hefil.L ~ , 1976, 98, 5703; for a similar reaction sequence see Wong, H. N. c. ; s on d h e i rn e r , F . ; Te t r ah e d r on .LtlL., 1 9 8 o , 2 1 7 . 9. Johnson, G. C.; Stofko, J. J.; Lockhart, T. P.; Brown, W.; Bergman, R. G.; ILL Q.I.£.L ~filill..L, 1979, l i , 4215. 10. The observations of Fischer and Lossing, however, may represent an example of 1,3-dehydrobenzene rearranging to 1,4-dehydrobenzene at extremely high ternperature. 4 11. K. w.; D. (a) Breslow, R.; Napier ski, J.; Schmidt, A. H.; ILL bill.L ~filill..L l 9 7 2 , ~.i , 5 9 0 6 ; ( b ) B r e s 1 o w , R. ; N a p i e r s k i , J . ; Clarke, T. C.; ibid., 1975, l i , 6275. ~Q..Q.i., 12. Breslow, R.; Khanna, P. L.; Tetrahedron .L~~, 1977, 3429. 13. Rule, M.; Lazzara, M. G.; Berson, J. A.; ILL bill.L ~filID..t- ~_g_,_, 91 1979, l.Ql, 7091. The [1.1.lJ and [2.1.l]propellanes may also belong to this unhappy club; for a discussion see Greenberg, A.; Liebman, J. F.; "Strained Organic Molecules," V.38 of "Organic Chemistry", Academic Press, New York, 1978, pp. 344-351. 14. Salem, L.; Rowland, C.; ~ l ' h ~!!h. l..nL.. Ed, Engl,, 1972, ll, 92. l 5. For early theoretical treatments see (a) Gheorghiu, M. D.; Hoffman, R.; Rev, ROU!!h. .c.hi!!h., 1969, li, 947; Cb) Roberts, J. D.; Streitweiser, A.; Regan, C. M.; lL.. h!!h. ~!!h. ~ , 1952, IA, 4579. 16. Wilhite, D. L.; Whitten, J. L.; lL_ A!!h. .Che!!h. ~ , 1971, il, 2858. 17. Dewar, 5569. 18. Noell, J. O.; Newton, M. D.; lL_ h!!h. .Cfil!!h. ~ , 1979, l.Ql, 51. 19. Mueller, K., private communication to R. G. Bergman. 20. Brown, D. W., Ph.D. Technology, 1980. 21. Lockhart, T. P., Research Report submitted in partial fullfilment of the requirements for admission to the Ph.D. program, California Institute of Technology, 1978. 22. Shea, K. J.; Stofko, J. J.; Mallon, C. B.; Johnson, G. C.; Brown, D. W.; Lockhart, T. P.; Bergman, R. G.; unpublished results. 23. Mallon, C. B.; Bergman, R. G.; unpublished results. 24. Johnson, G. C.; Ph.D. Technology, 197 9. 25. Pederson, D. J.; lL_ ~ ~!!h., 1968, ll, 780. 26. Schmidt, H. M.; Arens, J. F.; 1141. 27. Castro, C. E.; Gaughan, E. J.; Owsley, D. C.; J:...t. Q.r~ .Cfil!!h., 1966, ll, 4071. 28. Ellis, C. P . .iL.. ~!!h. ~ , 1934, 726. 29. Waugh, J. S.; Fessender, R. W.; .iL.. A!!h. .CJ1g!!h. Q.Q.£....., 1957, ll, M. J. S.; Li, W.-K.; lL_ Thesis, h!!h. .Cfil!!h. ~ , 1974, il, California Thesis, California ~ Institute of Institute of Trav, .Chi!!h., 1967, .a&., 92 846. 30. kH/k O may be determined by comparing the ratio of D/H found in 28 with the amount of residual hydrogen in the trapping agent (2.3%). This analysis gives an effective kH/k n of 4.3. 31. Werstiuk, N. H.; Kadai, T.; 32. Performed in collaboration with Paul B. Comi ta University of Californ i a, Berkeley, California. 33. Wilt, J. W.; in "Free Radicals," Vol 1, Kechi, J. K.; Ed., J. Wiley and Sons, New York, 1973, p. 384. 34. Benson, S. W.; "Thermochemical Kine tics," 2nd Ed., J. Wiley and Sons, New York, 197 6 , Table A.22, p. 309. 35. van Straten, J. w.; de Wolf, w. H.; Bickelhaupt, F.; Tetrahedron Lett,, 1977, 4667. See also Greenberg, A.; Liebman, J. F.; "Strained Organic Molecules," Academic Press, San Francisco, 1978, p. 160. 36. The mechanism of formation of benzocyclobutene and styrene is uncertain. Pathways from biradicals 36 and 37 (possibly via 1,3-dehydrobenzenes) may be imagined. 37. Kaptein, R., in "Advances in Free Radical Chemistry," 5, 1975, p. 338. 38. Ingold, K. u., in "Free Radicals," Vol. I, Kechi, Ed., J. Wiley and Sons, New York, 1973, Chapter 2. 39. ibid., p. 94. 40. The heat of formation of the 1,4-dehydroaromatic was taken to be that of o-di-n-propylbenzene less the bond energy in a molecule of H2 • 41. Walker, J. Q.; Jackson, M. T.; Maynard, J. B.; "Chromatographic Systems," Academic Press, New York, 1972, p. 161. 42. Bowden, K.; Heilbron, I. M.; Jones, E. R. H.; Weedon, B. C. L.; iL.. .CMilh. ~ , 1946, 39. 43 • s k i n n e r , D. L. ; P e t e r s o n , Chellh., 1967, .12., 103. 44. Beeckman, R. K.; Blum, D. M.; .iL.. Q.£Sh .cMilh., 1974, 1.2., 3307. 45. Ukhin, L. Yu.; Sladkov, .Khillh., 1968, i, 25. A. ~ iL.. ~Ilh., D. M.; J. ; 1973, ll, 148 5 . Lo g a n , Gorskov, V. T. J.; I.; h of the Vol. J. K., J...t.. Q.£.9.L Q.[gan, 93 46. Greenwald, R.; Chaykovsky, M.; Corey, E. J.; iL.. QLg;,._ ~l!h., 1963, 28, 1128. 47. o-Ethylstyrene was identified by comparison with the NMR spectrum reported in the literature: de Meijere, A.; Chem. ~ , 1974, l.Ql, 1684. 48. Closs, G. L., in "Chemically Induced Magnetic Polarization", Muus, L. T., et .9..d:..L, Eds., D. Reidel Co., Boston, 1977, pp. 225-257; see also reference (36) pp. 370-373. 49. Kaptein, R.; van Leeuwen, P. W. N. M.; Huis, R.; .C.h.§J!l Phys, Lett., 1976, il, 264; Kaptein, R.; Frater-Schroeder, M.; Oosterhoff, L. J.; ~l!h. Phys. Lett., 1971, 12, 16. 50. Reference (37) pp. 327, 328. 51. ibid., pp. 332. 52. The spin state of the radical pair should accurately reflect the spin state of the biradical. A similar approach has been employed in the study of the singlet and triplet carbenes; see Roth, H. D.; ~ ~.ill.t.. ~ , 197 7, l.Q, 85. 53. For a recent study see Engel, P. S.; Bishop, D. J.; Page, M. A.; .J........ Anh. .C.h~I!h. ~ , 1978, l..Q.Q, 7009. Perhaps the most convincing demonstration of a SCE is that of Closs, G. L.; Trifnac, A. D.; .J........ AI!h. ~I!h. ~ , 1969, ll, 4554. For a less supportive discussion see Koenig, T., in "Free Radicals", Vol. I, Kechi, J. K., Ed., J. Wiley and Sons, New York, 1973, Chapter 3. 54. The term "cage reaction" as applied in this discussion is meant to denote bimolecular reactions which take place within the geminate radical pair, ie., combination and disproportionation. 55. Since the lifetime of solvent-caged radical pairs is ca. 1010 and the rate of intersystem crossing is ca, 108 sec- 1 , 37 triplet-singlet interconversion should have an insignificant effect on the product distibution. 56. That 29 and 30 are formed solely by cage combination is demonstrated by the observation that, when 19 was allowed to react in the presence of cyclohexadi ene-d 0 and -d 4 , 2 9 and 30 contained only 0, 1 and 2 deuteria/molecule. Random combination of the aryl and cylohexadienyl radicals would have led to appreciable formation of the d1 and d3 isomers. 57. The ratio of cyclohexadiene-d 0 and -d 4 was held constant but the total concentration of trapping agent was varied. 94 58. At low concentrations of trapping agent, the yield of 28, 29 and 30 was very small. Because 28 overlaps slightly with unimolecular product 26 in the VPC trace, small amounts of 28 were difficult to accurately measure. Likewise, because they refer to several isomeric, long retention time products, VPC integration of 29 and 30 was not accurate when low yields were produced in the reaction. 59. Turro, N. J., in "Modern Molecular Photochemistry", Benjamin/Cummings, Co., Menlo Park, 1978, pp. 191-193. 60. ibid., pp. 187-190. 61 . Ash w or th, P. J. ; Whi th am , G. W. ; Whi ting , M. C. ; J:..t. ~ & ~ , IV, 1957, 4635. 95 Proposition Abstracts A synthetic route to oxirenes via l-halo-2-trimethylsilylepoxides is proposed. Experimental methods which could be employed to establish the presence of oxirenes are described. A synthetic route to metalbound oxirenes is also discussed. Generation of Free and Metal-Bound Oxirenes. Selective Pyrolysis by Multiphoton Infrared Excitation (MIRE). The use of tunable pulsed IR lasers as a heat source capable of inducing thermal reaction of selected components of a reaction mixture is proposed. Two specific examples where the method may be applied to unique advantage are described. In addition, a design for a preparative IR laser photolysis cell is suggested. CIDNP as a Tool for the Study of the Photochemistry of Organosilicon Compounds. CIDNP holds great promise as a technique which may elucidate the details of silyl-radical generating photochemical reactions. To demonstrate the power of the method, the application of CIDNP to the study of four different photochemical reactions is discussed. Investigations in Organometallic Photochemistry. An examination of the photochemistry of five metallocyclopentanes is proposed; these compounds have been selected because their structures may lead to unusual photochemical reactivity. A Novel Synthetic Route to Binuclear Organometallic Compounds Containing a Bridging Methylene Function. A novel, and potentially general, method for preparing bridging methylene binuclear organometallic compounds is outlined. 96 GENERATION OF FREE AND METAL-BOUND OXIRENES Introduction Oxirenes (1) have been invoked as intermediates in a number of reactions. Their formation in the reaction of alkynes with peracids 1 and in the reaction of methylene with carbon monoxide 2 has been proposed. Convincing evidence for the existence of oxirenes as intermediates (or transition states) comes from labelling results observed in the Wolff rearrangement of o(-diazo ketones 3 (2, Scheme I). Nevertheless, to date there has been only a single, unpublished study in which a molecule containing the oxirene moiety appears to have been trapped (Scheme II). 4 Other efforts to trap or isolate oxirenes have been unsuccessfui.4,5 Oxirene also has been of interest to theoretical chemists. Dewar and Ramsden6 studied the rearrangement of formylmethylene ( 3, R1 ,R 2 = H) to oxi rene using MIND0/ 3 and NDDO semi-empirical methods. that The results of their calculations led them to predict oxirene is 18 to 20 kcal/mole more stable than formylmethylene. They also examined the energetics of motion along the energy surface and predicted an activation energy for the rearrangement of 1 to 3 of about 20 kcal/mole. (LCAO-MO-SCF) calculation by Hopkinson 7 An ab initio indicated essentially isoenergetic with formylmethylene. 1 to be In this calculation geometry was fixed at assumed parameters and no prediction was made concerning the energy surface connecting 3 and I. A more extensive At ini~iQ (SCF-MO) study of the 97 ? \\t3D1-\ ~~3 0 l 98 rearrangement was performed by Strausz geometry optimization was carried out. and coworkers 8 and Oxirene was predicted to be 12 ± 5 kcal/mole less stable than formylmethylene and the activation energy for the rearrangement of formylmethylene was calculated to be 7 kcal/mole. 9 oxirene to From these results it was predicted that oxirene has a lifetime of about 10 8 seconds at room temperature. In contrast to Dewar's studies, Strausz' result suggests that trapping of oxirene in solution will be possible only with high concentrations of a reactive trapping agent. Proposal Generation Qf ~ oxiranes A novel and effective route to oxirenes (1) by treating l-halo-2-trimethylsilylepoxides may be possible (4) with cesium fluoride or tetraalkylammonium fluorides (R 4NF) in solution at, or somewhat below, room temperature (Scheme III). If oxirenes are successfully generated by this reaction, ring opening to ot..- keto carbenes followed by Wolff rearrangement to ketenes (7) should be observed. 3 The presence of a substantial concentration of cyclopentadiene or 1,3-diphenylisobenzofuran in the reaction solution may produce Diels-Alder trapping products (.e.g., 5). Alternatively, it may be possible to trap the oxirene by performing the reaction in a solvent, such as an alcohol, with which it may rapidly react. 4 A kinetic analysis of the dependence of unimolecular reaction products (or products derived 99 )(-= 'Br ,C-1 1M ~-= S.: M~\3 1- ~Al< Cl 1='c.Cco11 ;> C\ ~ i=;,cco).1 Cy f;Q k. (C.0)3 1<~ \-\ > aJ l\ l 8 """ - Fe_l-(c.o )q ~ f\h Co(e_o 8 hv c( 'l' I\ 11 tJ \t ~ Ro C'c (to\ Cc(Co')3 ('.)1 $ft Crl__Q;\_ Co(Co\, 100 from 7 on the concentration of trapping agent may allow further demonstration of the intermediacy of 1 in these reactions. If 1 is sufficiently long-lived, it may be possible to A nitrogen gas purge of the reaction isolate the oxirene. solution might liberate the oxirene which could be trapped in a low temperature matrix. Clearly, it will be possible to isolate 1 by this method only if the energy barrier separating the oxirene and o<'..-keto carbene is larger than that predicted by Strausz and coworkers. Preparation of Metal-Bound Oxirenes Metal coordination has been employed to stabilize several highly reactive organic molecules which otherwise have only transient existence at ambient temperature (cyclobutadiene 11 and trimethylenemethane 12 have both been isolated as metalcoordinated ligands). By a similar approach it may be possible to produce isolable metal-bound oxirenes. A synthetic route via 1,2-dichloro epoxides, which may be prepared by mCPBA epoxidation of 1,2-dichloro olefins, is possible. 13 Reaction of the 1,2- dichloro epoxides with a metal dianion such as Collman's reagent 14 (tetracarbonyl ferrate) may lead to displacement of both chlorines, with the generation of the metal-complexed oxirene (Scheme IV). Alternatively, reaction with Fe 2 (co) 91 la or NaCo(C0) 4 , 11 b which have been employed to produce cyclobutadiene complexes from 3,4-dichlorocyclobutene, oxirenes (Scheme IV). may produce coordinated Behavior of the complexed ligand as a 2 or 4 electron donor could provide information on the amount of 101 electron delocalization in the oxirene moiety. Synthetic route to l-chloro-2-trimethylsilylepoxides Alkene synthesis via ,13-halo organosilicon compounds has been reviewed by Chan. 15 This exceptionally mild method has been used to generate several unstable olefins including allene oxides 16 , strained bicycloalkenes 17 and anti-Bredt olefins. 18 The usefulness of this synthetic method for the preparation of high energy molecules is derived from the strength of the silicon fluorine bond, which is about 129 kcal/mole. Two synthetic routes to l-halo-2-trimethylsilylepoxides are proposed in Scheme v. A procedure for the hydrobromination of trimethylsilylpentyne has been described and should be readily adaptable to the preparation of various E- and Z-l-bromo-2trimethylsilyl olefins. 19 Separation of the olefins followed by epoxidation with m-chloroperbenzoic acid (mCPBA) 2 0 should yield the respective syn and anti epoxides 9. A second route to 9 involves the dichlorination of acetaldehydes in acidic medium. 21 Addition of trimethylsilyl lithium is expected to produce a mixture of the E and z epoxides. Both methods may allow preparation of the bromo and chloro trimethylsilyl epoxides. 102 (I) R-=-TMS 'R:: \-\) o.\ki I, 1hen'\\ ~>==( J-J HBc ;;, TMS R(" l< \_/ i-t /\ 1?r i"Ms E. 9 ---1/ L~ o 'K .\-~Cl H·''l \C\ 1MS 1 ( £ i r.) 103 References 1. Ibne-Rasa, K. M.; Peter, R. H.; Ciabattoni, G.; Edwards, J. O.; .J..,_ Afu .c.h.g.m..,_ .£Q£..._, 1973, -2..2., 7894; Concannon, P. W.; Ciabatoni, J.; ibid., 3284; Ogata, Y.; Sawaki, Y.; Inoue, H.; .J..,_ ~ ~.m..,_, 19 7 3 , 1-a, 10 4 4. 2. Montagne, D. C.; Rowland, F. S.; i,L_ A.m..... .c.h.g.m..,_ .£Q£..._, 1971, ~ , 5381; Russell, R. L.; Rowland, F. S.; iQid., 1970, 22., 7508. 3. ( a ) Fen w i c k , J. ; Fr a t er , G. ; Og i , K. ; st r au s z , O. P. ; i,L_ A.m..... ,CheI!Lt. ~ , 1973, 95, 124. (b) Matlin, s. A.; Sammes, P. G.; .c.h..§..m..,_ ,C.Q.m.m.Y.n....., 1972, 11. (c) Frater, G.; Strausz, O. P.; .J..,_ A.m..,_ .ChgI!Lt. ~ , 19 7 4, 9 2 , 6 6 5 4. 4. Meier, H.; Zeller, K.-P.; .l.!, p. 40. ~~ .Cfil!!l.L..t. .I.11ti. .EJL_ ~ , 1975, 5. For a review of published attempts, see Hart, H.; Jiang, J. B.-C.; Sasaoko, M.; .J..,_ ~ ,Chgfu, 1977, !2., 3840; see also reference (4). 6. Dewar, M. J. S.; Ramsden, C. A.; ~l!h..,C.Qffiffi.Y.Il..t.., 1973, 688. 7. Hopkinson, C.; .J..,_ .ci_ ~ ]2_erkin ll, 1973, 794. 8. Strausz, o. P.; Gosavi, R. K.; Denes A. S.; Csizmadia, G. ; JL.!.. AID..!.. ,Che .m..,_ .S. o c. , 1 9 7 6 , ..211, 4 7 8 4. 9. The calculation of Strausz and coworkers 8 suggests that the rearrangement of formylmethylene to oxirene will have an activation energy of about 19 kcal/mole. This appears to contradict experiments3b which suggest that oxirene formation from thermally generated alkyl substituted .1...-keto carbenes takes place in competition with Wolff rearrangement. In fact, the temperature dependence on product yields indicates a difference in E of 1.5 kcal/mole, favoring Wolff rearrangement. If t~e Wolff rearrangem 8t has an activation energy of 6 kcal/mole as calculated for formylmethylene, then 3 and 1 should be roughly equienergetic. This analysis assumes only small effects on the reaction energetics due to alkyl substitution. I. 1 R. Strausz, O. P.; Gosavi, R.h.Y.§.i., 1977, fl, 3057. 11. (a) Emerson, G. F.; watts, L.; Pettit, R.; ~ AI!Lt. ,Che.m..,_ .Q.Q.£......, 1 9 6 5 , !il., 13 1. (b) Amie n t, R. G.; Pettit, R.; .iQi..9..L, 1968, ll, 1059. K.; Gunning, H. E.; ~ 10. .Cfil.!!h_ 104 12. Ward, J. S.; Pettit, R.; .c..h.g.m.... ~ID.mun., 1970, 1419. 13. Griesbaum, K.; Kibar, R.; Pfeffer, R.; Liebig§ Ann. ~film....., 1975, 214. 14. Collman, J. P.; 15. Chan, T. H.; A£.fu Chem..... Res., 1977, J...Q, 442. 16. Chan, T. H.; Ong, Lett., 1976, 3252. 17. Chan, T. H.; Massuda, D.; T.tl..[ahedron Iitl.t....., 1975, 3383. 18. Chan, T. H.; Massuda, D.; JL.. .8.m.... ~.m.... .SQ£...., 1977, .2....2., 936. 19. Boeckman, R. K.; Blum, C. M.; .J..:... ~ ~.m...., 1974, ll, 3307. 20. Both (bis)trimethylsilyl and dichloro olefins have been epoxidized by this method; Hudrlik, P. F.; Hudrlik, A. M.; Rona, R. H.; Misra, R. N.; Withers, G. P.; Jl,_ .8.m.... ~.ID..... ~ , 1977, .2....2,, 1993; see also reference (11). 21. ~ .c..h.g.m.... ~ , B. s.; 1975, ,a, 342. Mychajlowskij, W.; Tetrahedron Astor, J. G.; Newkirk, G.; Jenkins, D. M.; Dorsky, J.; Q.I.g_._ Coll, YQ.h J., 1955, 538. ~ 105 SELECTIVE PYROLYSIS BY MULTIPHOTON INFRARED EXCITATION (MIRE) Introduction In the past decade organic chemists became interested in the pulsed co 2 laser and were quite excited by the prospect of performing mode-selective multiphoton infrared excitation (MIRE). 1 In spite of early claims, 2 these hopes have now for the most part fallen by the wayside; if some examples of non-ergodic, mode-specific MIRE-induced reaction are discovered in the future, 3 their generality and usefulness for the organic chemist is likely to be quite limited. At the nadir, an organic chemist surveying the field might have lamented that the IR laser looked like little more than an expensive "bunsen burner". Out of the ashes of this fallen phoenix have recently arisen a number chemistry. of extremely promising applications for organic While mode-specific chemistry is not likely to be realized, the IR laser may prove to be a bunsen burner which offers a number of unique features. Because the diameter of the laser beam can be made smaller than that of the sample cell, it is possible to perform essentially "wall-less" thermal reactions: 4 when the sample is heated either by direct photolysis or by energy transfer from a sensitiser (such as NH 3 , 5 SiF 4 6 or SF 6 7 >, the energetically excited molecules lose their excess energy in collisions with other gas molecules before collision with the photolysis cell walls takes place. Recently, this feature has been taken advantage of in order to examine the "true" thermal chemistry of molecules which, in ordinary 106 reactors, have decomposed primarily via wall-catalyzed reactions. 8 Golden and coworkers are presently preparing a new generation of the VLPP apparatus 9 which takes advantage of this particular feature of MIRE.lo An especially intriguing feature of the MIRE method is its ability to produce non thermodynamic product ratios by selective irradiation of the components of a gas mixture. of this have appeared in the literature. Several examples In 1975 Yogev and coworkers 11 reported that isomerization of compound 1 to produce isomer 2 could be accomplished by selective irradiation of the vinyl C-H out-of-plane bending mode of 1 (Scheme I); the wavelength of this absorption band does not overlap with any absorptions in 2. At low pressures of 1 (5 torr), the formation of 2 from 1 was essentially quantitative even though Keq is about 1 at elevated temperature. This indicates that significant heating of all of the components of the reaction mixture did not occur during photolysis. In other studies 12 simple ~/trans olefin isomerization reactions were induced by MIRE. At low pressure, conversion of the irradiated isomer to non absorbing molecules was observed; by this method it was possible to produce gas mixtures enriched in the thermodynamically less stable (irradiation transparent) component. Thus, although the chemistry displayed by the radiation absorbing species was that observed in the normal thermal reactions of those compounds, the overall reaction mixture showed distinctly non-thermodynamic, or non-equilibrium, product ratios. 107 tet ( 1I!) = o .--=,.. l ( 2 5 °c ) D, 81 ( A0D 0 C.-J E 108 Proposal It is proposed that the ability to selectively "heat" one component of a reaction mixture be utilized as part of a general scheme to produce detectable quantities of molecules (stable at near ambient temperatures (ca. 100°)) which, if prepared from the same precursors at elevated temperature by the usual methods (flow tube or solution reaction), would be unstable to the reaction conditions and undergo secondary reaction. A second proposal is made for the design of a preparative scale MIRE photolysis cell in which samples sufficiently large to permit isolation and characterization by normal methods (NMR, IR, mass spectroscopy) may be irradiated. The energy diagram in figure 1 serves to illustrate a problem frequently encountered in organic chemistry: a molecule of synthetic or mechanistic interest (B) is the intended product of thermal reaction of a precursor molecule (A); however, because of the temperature required to effect reaction of A, B itself undergoes rapid (secondary) reaction to produce a different product (or products, C). This situation is particularly frustrating to the experimenter when it prevents the isolation of reaction intermediates or target synthetic molecules which are suspected, or known, to possess reasonable stability at room temperature. For problems of this type, pulsed IR lasers may offer special advantages. MIRE, at low pressures of reactants (ca. 1 to 10 torr), of an IR absorption band unique to A should induce reaction of A to produce B under conditions where the temperature of non absorbing species is moderate. 13 Temperature 10 9 buildup in the non absorbing gases, can be controlled by varying overall reactant pressure, the energy fluence of the incident irradiation and by adding inert bath gases to the reaction mixture. Factors limiting the application of this method include insufficient volatility of the compound to be photolyzed (which will prevent the irradiation of quantities large enough for analysis), the presence of overlapping IR absorption bands in A and B, and a low energy barrier to secondary reaction of B. The two examples cited below illustrate specific problems to which the unique qualities of this method might be applied. To employ the proposed selective pyrolysis method it is critical that the temperature experienced by irradiation- transparent gases be too low to cause appreciable secondary reactions. It appears that no investigations have sought to determine the gas temperature generated upon direct irradiation of one (reacting) component of a gas mixture. This question bears directly on the minimum depth of the energy well in which the product of photolysis CB) must lie if secondary reactions are to be avoided. A way to evaluate this factor would be to perform the direct irradiation-induced reaction of compound A in the presence of a transparent molecule CD) of known, lower activation energy. The ratio of A reacted to the reaction of non absorbing molecule D would be an empirical measure of the importance of energy transfer from A to D; thus D will serve as an internal thermometer in the irradiation. Any molecule whose thermal chemistry is well characterized and which has an Ea in the range 110 50 to 70 kcal/mole would be suitable for direct irradiationinduced reaction; molecules which might be employed as internal thermometers include 1,4-cyclohexadiene (log 10 A = 12, Ea= 43 kcal/mole for loss of H2 ) and cyclobutene (log 10 A = 13.3, Ea= 33 kcal/mol for ring opening to 1,3-butadiene). The r ela ti ve importance of various laser parameters to the proposed selective pyrolysis technique may be evaluated by this method. It is anticipated that the presence of inert bath gases such as He, Ar or CF 4 will help to lower the temperature experienced by irradiation-transparent molecules in the irradiated zone. With an elegent labelling experiment and kinetic analysis, Walsh and coworkers have convincingly demonstrated the intermediacy of cyclopropene in the thermal interconversion of allene and propyne. 14 , 15 However, because of the relative heats of formation of the molecules and the shallowness of the well in which cyclopropene lies (figure 2), cyclopropene has never been detected in the final product mixture. The energy diagram in figure 2 quite closely resembles that in figure 1 where cyclopropene (B), while stable to unimolecular reaction up to ca. 150° c, is unstable to the conditions of allene and propyne isomerization. By the proposed MIRE method, it may be possible to isolate the cyclopropene intermediate by direct irradiation of propyne or allene. 16 Another problem which recently appeared in the literature occurred during the effort of de Meijere and coworkers 17 to prepare a theoretically interesting molecule (3, Scheme II). The synthetic route to 3 involved gas phase thermal rearrangement of 111 -------- 104 --·- reacf,on Coo rd,~o.te_ F; 5t-t ;-e. ~ . Ert +\,'\L>J p1 d 1'(A3. r-o..Vl'\ ~ r -pr- 0 p<-p1e_ , . +' o.. ll~Yle - Ii/ I ':,DW\e r '3°' I DV) • Sc_heme_ ][ J4DOOC... ~ N t.. \ \o«l 4 ~ V ?" (D> s 3"'-- ""-' 0) -" I ........ 112 4 as the final step. Reaction of 4 however, yielded none of the desired compound; instead, unimolecular product 6 was obtained. On the basis of labelling results de Meijere argued that 3 was formed in the reaction but was unstable to the vigorous reaction conditions. MIRE of the vinyl H absorption band of 4 should lead to rearrangement. Compound 3 is not likely to have an absorption band at the irradiation wavelength since it lacks the vinyl H moiety. As further required by the proposed selective pyrolysis technique, 3 is expected to possess reasonable thermal stability. 17 This system thus meets the basic requirements for study by this method. In the first example discussed above the desired product, cyclopropene, is a known molecule and can be identified by VPC coinjection and mass spectral analysis. De Meijere's study, however, represents a different kind of problem, since the goal was to isolate and determine the physical properties of the target molecule. milligram Because it is a tedious process to obtain even quantities of products by MIRE using currently available techniques, there is a need for a preparative MIRE photolysis cell. One possible design for such a reactor is illustrated in figure 3. This cell is designed to allow the continuous static-vacuum transfer of a reactant zone irradiated by the laser. (A} through the The variables of total pressure in the cell, fluence of the beam, frequency of laser pulsing and the residence time of A in the cell (which may be controlled by adjusting the aperture leading from the cell to the cold trap} 113 \ J FiJlAre. 3 . Des ~5V\ f ~ otti \~ s; s ~ e_ II. 0.. 114 may be varied in order to optimize conversion of A. With this apparatus it should be possible to subject comparatively large amounts of sample to irradiation. This design may also be used for the sensitized photolysis of large amounts of sample: cold trap is sensitizer (for SiF 4 , possible to kept pass slightly above the the boiling point of -86° cat 760 torr), reactant through if the the the then it should be cell transfer) and observe the sensitized reaction. (by static In sensitized photolyses the energy fluence can be turned up to fairly high levels in order to obtain substantial conversion per pulse. At the same time, the advantage of "cold reactor walls" will still be enjoyed. 115 References 1. Grunwald, E.; Dever, D. F.; Keehn, P. M.; in "Megawatt Infrared Laser Chemistry", J. Wiley and Sons, New York, 1978; Kimel, s.; Speiser, s.; ~.ID..t... ~ , 1977, TI, 437. 2. Hill, G. A.; Grunwald, E.; Keehn, P.; .J.,_ .8.ID..t... Q}__gI!l.,_ ..S.Q£.u 1 9 7 7 , i.2. , 6 5 1 5 ; K a 1 do r , A. ; H a 1 1 , R. B. ; Cox, D. M.; Horsley, J. A.; Rabinovitz, P.; Kramer, G. M.; .J.,_ .8.ID..t... Q!_g.ID..t... ~ , 1979, 101, 4465; Hall, R. B.; Kaldor, A.; .iL.. .c..h~.ID..t... Phys,, 1979, 70, 4027. 3. Reddy, K. V.; Berry, M. J.; ~.ID..t... Phys. LetL., 1979, Q.Q, 223; Moore, C. B.; Smith, I. W. M.; Faraday Discussion Nih. il, introductory paper to section 2, 1979. 4. See reference (1), p.2. 5. Steel, C.; Starov, V.; Leo, ~ill.i:. Phys, _Lg.th, 197 9, R.; John, R.; Harrison, R. G.; .2.2-, 121. 6. Olszyna, K. J.; Grunwald, E.; Keehn, P. M.; Anderson, S. P.; Tetrahedron Lett., 1977, 1609. 7. Pazendeh, H.; Marsal, C.; Lempereur, F.; Tarieu Ma 1 e i s s ye, J. ; I nL. .J.,_ .C. he .ID..t... .Ki.net. , 1 9 7 9 , il, 5 9 5. 8. Lew is, K. E. ; Mc Mi 11 en, D. F. ; Go 1 den, D. M. ; .iL.. .EhYh .c..hg.ID..t..., 1980, H, 226; Berman, M. R.; Comita, P. B.; Moore, C. B.; Bergman, R. G.; .iL.. .8.ID..t... .C,he.ID..t... ~ , 1980, in press. 9. Golden, D. M.; Spokes, G. N.; .I.n..t..,_ Ed. Eng_,_, 1973, 12, 534. 10. Golden, 11. Glatt, I.; Yogev, A.; .iL.. .8.ID..t... ~.ID..t... ~ , 1976, 2-..a, 7087. 12. Yogev, A.; Lowensein-Benmair, R. M. J.; .J.,_ .8.ID..t... .c..hg.ID..t... ..S.Q.£...., 1973, 95, 8487; Buechele, J. L.; Weitz, E.; Lewis, F. D.; .iL.. .8.ID..t... Che .ID..t... ..S. o c. , 19 7 9, 1 0 1 , 7 0 8 7 • 13. Danen, D. M.; Benson, s. de W.; ,bn.ggli.L .C..hg.m..... unpublished results. w. C.; Munslow, W. D.; Setser, D. W.; .J.i. A.ID..t... .c..h~.m..... ..S. o c. , 19 7 7 , ti , 6 9 6 1 ; Re 1 s e r , C. ; st e i n f e 1 d , J. I. ; .J.,_ P hY.§..L .C.he.ID..t..., 1980, 84, 680. 14. 15. Walsh, R.; Bailey, I. M.; .iL.. h l i , 1146. Walsh, R.; lQ.2., 1210. Hopf, H.; Priehe, ~ Faraday Trans, 1., 1978, H.; .iL.. .8.ID..t... .C..h~.ID..t... ~ , 1980, 116 16. Ideally, one would like to generate cyclopropene upon irradiation of both propyne and allene. The weak IR absorption of allene in the output region of the co 2 laser, however, may make the allene photolysis difficult. 17. Spielmann, w.; Kaufman, D.; de Meijere, A.; ~ l i i . ~he.ID...t.. .I.n..ti. ~_g_,_ ~119. u I 9 7 a , ll, 4 4 o. 117 CIDNP AS A TOOL FOR THE STUDY OF THE PHOTOCHEMISTRY OF ORGANOSILICON COMPOUNDS Introduction While the field of silicon chemistry has been actively investigated for a number of years, many of the mechanistic aspects remain to be characterized. This is especially true of the photochemistry of silicon compounds~ while the photochemical generation of silyl radicals and silylenes is of growing importance as a preparative tool, 1 numerous mechanistic questions await study. In general, the events immediately following electronic excitation of the precursors of these reactive species are not well elucidated and, with the exception of some very recent studies of silylenes, 2 the electronic excited states (singlet or triplet) involved in the formation of these reactive intermediates has not been examined. Proposal The technique of chemically induced dynamic nuclear polarization (CIDNP) 3 has been applied with great success to the study of organic reactions which involve radical intermediates. It is proposed that this method be used to elucidate numerous mechanistic details of the photochemical generation of silyl radicals from several different precursor molecules. Application of the CIDNP technique to photochemical silyl radical-generating reactions is likely to provide insight into the mechanism of decomposition of various silyl radical precursors. Furthermore, 118 it should provid e a method of determining the spin multiplicity of the precursor excited states leading to the formation of the radicals. This will also help to establish the role of various photochemical sensitizers, which currently are used without detailed understanding of the spin multiplicity transferred nor the mechanistic role played in the photochemical generation of silyl radicals. It seems likely that 13 c- and 29 si-CIDNP will provide the greatest insight into the mechanism of decomposition of photochemical silyl radical precursors. 13 c-CIDNP (recorded with an FT-NMR spectrometer) has been employed on numerous occasions in the study of organic reaction mechanisms 4 and is conceptually identical to 1 H-CIDNP. Typically, the enhancements observed with carbon nuclei are even larger than those observed with protons. 29si-CIDNP apparently has not been recorded before. This situation may be due to the fact that silicon NMR is only lately becoming widely recorded. 5 The natural abundance of 29 s i 1/2) is 4.7%, appreciably greater than that of 13 c. the recording of silicon NMR spectra, (spin Complicating however, are long relaxation times and the presence of a negative nuclear overhauser effect. Cr(acac) 2 can be added to the NMR solution to reduce these problems. Four compounds that generate silyl radicals on photolysis are shown in Schemes I through IV. mercury compounds and The photolysis of silyl silanes are widely producing silyl radicals. used methods of Determination of the excited states 119 responsible for the direct and sensitized photochemical decomposition of these compounds and other mechanistic details are the intended goal of the proposed CIDNP investigations; the application of CIDNP to the study of each of these reactions is briefly outlined below. In Scheme I are outlined the two major reaction pathways which have been postulated to explain the photochemical reactivity of acylsilanes Cl}. Pathway (A} is observed in a variety of organic solvents 6 and appears to involve the cleavage of the silicon acyl bond by an excited state of 1. Products of trapping with cc1 4 strongly imply the presence of the silyl and acyl radicals. The second pathway (B} is the exclusive reaction mode of cyclic acylsilanes (2) 7 and is also followed when the with acyclic acylsilanes when the photochemical reaction is carried out in alcohol solvent in the presence of a trace of base. 8 In pathway (B} silicon migration to the carbonyl oxygen gives a carbene which inserts into the hydroxyl OH bond cf the sol vent. The application of CIDNP to the photochemistry of acylsilanes might further elucidate several mechanistic details: (l} Evidence for the intermediacy of radical pairs in the migration of silicon to the acyl oxygen would be obtained if CIDNP c13 c or 29 si} were observed in the acetal reaction product. (2) The possibility exists that the photochemical cleavage of 1 to give the radical pair is reversible; the observation of CIDNP in the starting material would provide evidence on this point. (3) The importance of the spin state on the photochemical 120 (A) /1<"0µ, k( '3S~ DR'' + l<~'S~OH ( B) 1 hv A., 1<.''oH, bo.s~ "r) ::;,,- K. .. S~ o c... ~' 3 121 reaction may be determined by CIDNP studies. The direction (absorption or emission) of the polarized signals observed in the trapping products will depend on the spin state of the radical pair for med upon photolysis. 3 Direct and sensitized photolyses may produce different excited states of 1. It is possible that the reaction pathway followed depends on the spin multiplicity of the excited state. The use of sensitizers to effect the reaction of 1 does not appear to have been examined. A last point that might be addressed by a CIDNP study is related to the observation that silyl radicals undergo inversion much more slowly than carbon radicals; 9 in fact, free silyl radicals in solution (generated from optically active precursors) undergo solvent configuration. trapping in competition with loss of Such an observation has been made in the photolysis of an optically active acyl silane (where the substituents on silicon were methyl, phenyl and naphthyl).10 About 60% net retention of configuration was observed in the products of trapping with cc1 4 ; the product of trapping with alcohol showed about 80% net retention and the products derived from silicon migration to oxygen showed about 90% net retention. It would be interesting to examine these reactions in an NMR probe in the presence of a chiral shift reagent which could separate the signals of enantiomeric products. 11 The observation of retention of substantial CIDNP in the products with configuration 12 would dramatically confirm the configurational stability of silyl radicals. It would also allow an estimate to 122 be made of the rate constant for inversion based on the time scale of the CIDNP phenomenon cio- 8 to 10- 9 seconds). 3 In Scheme II the photochemistry of (bis)silyl mercury and silyl alkyl mercury compounds (3) is outlined. Photolysis of these compounds is an effective method of generating silyl radicals in solution. 13 Several mechanistic details which may be elucidated by the application of CIDNP to these reactions are the following: First, the mechanism of formation of the silyl radicals is not understood at present. Two possibilities may be enumerated: one bond cleavage (pathway C) or simultaneous two bond cleavage (pathway D). The lifetime of alkyl mercury and silyl mercury radicals in solution are believed to be very short; this has prevented their detection by ESR spectroscopy. CIDNP observed in compound 3 would indicate that decomposition must occur as shown in equation (C) and that cage combination of the radical pair is possible. This observation would establish the finite lifetime of the mercury radical and would even allow a rough estimate of its lifetime. The direction polarizations observed in reaction products of (such as the silyl chlorides for reactions performed in the presence of halogenated solvents) will depend on whether the radical pair of pathway (C) or (D) is present. The difference in g factors for silyl and mercury radicals 14 should be large enough to produce a net effect in the reaction products; the direction of polarization is determined by the g factor difference. The g factor difference for paired trimethylsilyl radicals (pathway CD)) will be zero for R = R'; thus polarization will be weak and should show a 12 3 Sc.he_ me. II h\) 1) ~ "3 s· L • • + v' ",3 s L • + ~\_s + •~d s~ "R~ t ~ld + •<;~ ~; ("D) I hv =>1<3'>~-1--\j SK~ cc.1'l,,,.7<3S~-C.I 1( I sL' 3 'R"5 ('_ \ s~ - s~ ~; 124 multiplet effect. The g factor of silyl radicals depends strongly on the groups bonded to silicon; 14 thus photolysis of asymmetric mercury compounds may lead to CIDNP signals even if decomposition occurs by pathway (D). A second piece of information that may be obtained concerning the photochemical decomposition of these mercurials is the effect of photochemical sensitisers on the spin state of the radical pairs generated. There could conceivably be a change in decomposition mechanism ((C) vs (D)) depending on the spin state of the precursor. Direct photolysis of phenyl substituted silanes (4) leads to cleavage of the silicon hydrogen bona. 15 In the absence of the phenyl chromophore silicon-hydrogen bond cleavage may be induced by irradiation of elemental benzophenone 18 sensitizers. mercury, 16 acetone 17 and Two mechanisms which have been proposed to explain these reactions are illustrated in Scheme III. The mercury sensitized photolysis has been explained by both mechanisms; the ketone "sensitized" reactions are usually described by pathway (F). Pathway (E) is intriguing because analogous hemolytic cleavage of carbon hydrogen bonds by UV photolysis is unknown. It would therefore be unusual to observe CIDNP arising from the hydrogen-silyl radical caged pair shown in pathway (E). CIDNP studies could confirm the presence of the radical pair and also provide information as to the occurrence of cage recombination to regenerate 4. The spin state of the radical pair generated upon direct and sensitized photolysis may 125 hv (d ~rte:.+) ---::;:;,, 4------- e>r '2...~erj~ "t<'tlM. s ~~ \ fr-~ ~s:+:~e.r Sc.heme__ N 'R 3 S~-N==-N- SL' 'DI t'\ 3 5 CC.\4 r Hl. +("R~ ~~ J~ + l<~ S~e\ 126 be determined by the CIDNP observed in the products of the reaction. Evidence for pathway (F), in which reaction is induced by abstraction of hydrogen by the sensitizer, would be found if CIDNP were observed in the 1 H- or 13 c-NMR spectra of the sensitizers. The last photochemical reaction to be discussed is the reaction of (bis)silyl azo (5) compounds (6, Compounds 5, compounds, undergo bimolecular reactions on thermolysis. 19 Photolysis, however, Scheme IV). and phenyl trialkylsilyl azo unlike organic azo produces silyl radicals. 2 ° Compound 6 generates free silyl radicals either upon photolysis or heating 21 at 115°c. The photochemical reactions of these compounds are formally similar to that of organic azo compounds and raises the question as to whether or not the same general rules of one bond versus simultaneous two bond cleavage apply. CIDNP has been very effectively employed to elucidate organic azo decomposition mechanisms; 22 similarly, monitoring the photolysis of 15 N labelled 5 and 6 by 15 N-NMR for polarization of the nitrogen atoms would provide evidence for cleavage (G). the presence of stepwise The simultaneous bond cleavage pathway CH) may produce CIDNP in the reaction products if R is different from R' ( in 5. Again, it would be interesting to examine the effect of sensitizers on the homolysis reaction. 127 References 1. For reviews see (a) West, R.; Barton, T. J.; .iL.. ~Ilh. Ed., 1980, 57, 334; (b) Sakurai, J., in "Free Radicals,; Vol. 2; Kochi, J. K., Ed.; Wiley-Interscience, 1973; (c) Jackson, R. A., in "Free Radical Reactions," Int. Rev. of Science, Organic Chemistry Series Two, Vol. 10, Waters, w. A., Ed., Butterworths, Boston, 1975; (d) Gaspar, P. P.; in "Reactive Intermediates,: Vol. l; Moss, R. A.; Jones, M., Eds.; WileyIntersciene, 197 8. 2. (a) Ishikawa, M.; Nakagawa, K.-I.; Kumada, M.; .iL.. Organomet, ,C.he.Ilh., 1979, 178, 105; (b) Tortorelli, V. J.; Jones, M.; ,ih. .Anh. .c.hg.ffi.L ~ - , 19 8 0 , lQ2_ , 1 4 2 5. 3. For reviews see (a) Kaptein, R., in "Advances in Free Radical Chemistry,", Vol. 5; 1975; (b) Muus, L. T.; Atkins, P. W.; McLauchlan, ,K. A.; Pedersen, J. B.; "Chemically Induced Magnetic Polarization," D. Reidel, Boston, 1977. 4. See reference 3(a) and Branski, C. F.; Moniz, w. B.; Sojka, S. A.; .iL.. .Anh. ~Ilh. ~ , 1975, il, 4275. 5. Harris, R. K.; Mann, B. E.; "NMR and the Periodic Table," Academic Press, N. Y., 1978. 6. (a) Brook, A. J.; Dillon, P. J.; Pearce, R.; ~ .ih. ~Ilh., 1971, ti, 133; (b) Brook, A. G.; Duff, J. M.; ,C.an, .iL.. ~ill.Lr 1973, 51, 1622. 7. (a) Brook, A.G.; Kucera, H. w.; Pearce, R.; ~ .iL.. .C..h..e.Ilh., 1971, 49, 1618; Cb) Brook, A. G.; Pearce, R.; Pierce, J. B.; ~ .iL.. .c.hg.ffi.L, 19 7 1 , t i 1 6 2 2. 8. Brook, A. G.; Duff, J. M.; .iL.. .Anh. .c.hg.ffi.L ~ , 1967, ll, 454. 9. See reference l(b), 10. (a) Brook, A.G.; Duff, J. M.; .iL.. .Anh. .c.hg.ffi.L ~ , 1969, ll, 2118; (b) see reference (8). 11. Shift reagents have been used with success in the study of transition metal chlorides; this suggests that silicon chlorides will also be susceptible to shift reagents: Marks, T. J.; Porter, R.; Kristoff, J. S.; Shriver, D. F.; in "Nuclear Magnetic Resonance," Sievers, R. E., Ed.; Academic Press, N. Y., 1973. 12. Shift reagents have been used to separate polarized signals in CIDNP experiments. Low concentrations of the shift reagent must be employed in order to avoid quenching of the pp. 763-767. 128 polarization: Bargen, J., in "Nuclear Magnetic Resonance," Sievers, R. E., Ed.; Academic Press, N. Y., 1973. 13. (a) Neumann, w. P.; Reuter, K., in "Organometallic Chemistry Reviews," Vol.7; Seyferth, D., tl tl., Eds.; Elsevier, N. Y., 1979; (b) see reference l(b). 14. Silicon radicals typically have values of g greater than 2.003; Kochi, J. K., in "Advances in Free Radical Chemistry," Vol. 5, N. Y., 1975. Mercury radicals have only rarely been observed. When substituted by Hor CN low g factors are found (<2.000); (a) Knight, L. B.; Weltner, w.; .J.i,. .c..h..§..m..,_ Phys., 1971, 55, 2061; (b) Knight, L. B.; Lin, K. C.; .J.i.. ~.m..,_ Phys., 197 2, a, 6 O44. 15. Nelson, L. E.; Angelotti, N. C.; Weyenberg, D. R.; .J.i,. _Am..,_ .c..hg.m..,_ ~ , 1963, .8....2., 2 6 6 2. 16. (a) Jung, I. N.; Weber, w. P.; ~ OrganomgL_ .c.h.g.m..,_, 1976, l.li, 257; (b) Childs, M. E.; Weber, W. P.; .J...t.. Organom.!tl..t._ .C.h.f. .ID..1..• 1 9 7 5 , .a.Q. , 1 6 9 ; ( c ) Nay , M • A. ; Wo d d a 11 , G. N. ; Strausz, O. P.; Gunning, H. E.; .J...t.. Am.... .c.h.g.m..,_ ~ , 1965, ll, 17 9. ~ 17. Calas, R.; Duffaut, N.; 906. 18• ( B k H. - D. , ~ Q.I..9-L .C.h..e. .m..,_, 1 9 6 9 , J.i , 2 4 6 9 ; ( b ) McIntosh, A. R.; Wan, J. K. s.; ~ .J...t.. .c..h..§..m..,_, 1971, il, 812. 19. a ) e c e r 1h. Acad, ££.L. Paris, 1957, 245, , Wiberg, N.; Joo, w. C.; Uhlenbrock, W.; Ange\tLt.. .c.h.g.m..,_, 1968, .a.Q, 661. 20. Watanabe, H.; Inoue, K.; Furutachi, M.; Nagai, Y.; quoted in reference 1 (b), p. 7 53. 21. Watanabe, H.; Cho, Y.; Ide, Y.; Matsumoto, M.; Nagai, Y.; b OrganomgL_ .c..h..§..m..,_, 1974, ll, C4. 22. Porter, N. A.; Marnett, L. J.; Lochmuller, c. H.; Closs, G. L.; Shoba tak i, M.; .J...t.. Am.... Che.ll..L ~ , 1972, ll, 3 6 6 4. 129 INVESTIGATIONS IN ORGANOMETALLIC PHOTOCHEMISTRY Introduction The field of organic photochemistry commands the attention of organic chemists precisely because the photochemical reactivity displayed by organic compounds frequently differs substantially from that observed in thermal reactions. photochemical reactions have proven to be The novel fascinating mechanistically and, most importantly, permit a number of useful synthetic transformations to be carried out which do not occur upon simple heating. In comparison to organic photochemistry, the field of organometallic photochemistry is still very much in its infancy. Certainly, the possibility that unique photochemical reactions may occur is a compelling reason for intensive investigation of this field. Only a few general classes of organometallic photochemical reactions have been characterized thus far (Scheme I). The most widely investigated is the labilization of CO ligands upon irradiation with UV light. 1 This appears to be the primary photochemical process in almost all CO containing complexes studied thus far. Typically, loss of CO is al so the 1 owe st energy unimolecular process and is observed thermally. Because the photochemical reaction can be performed at, or below, ambient temperature, thermally unstable products sometimes can be isolated and bimolecular processes which unimolecular thermal reaction can be avoided. may obscure the Similar to the CO labilization reaction, ethylene appears to be photolabile in at 130 Sc.hen\e. I ~ hv (2.) (co'\ M-M (co')r\ ~v M + co 'l ~ (c 61Y1 M • t ('C,R 4 (co )V\ M-C\ (5) cp To._ (~1\ Cf\.,_ 'S~ Me_ 3 ~ C(fa. C13 + c~ 2 s~ Me 3 131 least some complexes. 2 Another well-studied photochemical process is the hemolytic cleavage of metal-metal bonds in binuclear (and polynuclear) organometallic complexes. 3 While the reactions described above have proven to be both general in nature and of some utility in preparative chemistry, it is likely that they represent only a fraction of the photochemical reactivity of organometallic complexes. Because CO labilization dominates the photochemistry of metal carbonyl complexes, new modes of photochemical reactivity are most likely to be found in CO-free organometallic complexes. Surprisingly, there has been very little investigation of the photochemistry of complexes which do not contain CO as a ligand. Several examples which have been observed are shown in Scheme I. Photolysis of n5-cp 2 TiMe2 (equation (3)) appears to generate methyl radicals in the primary photochemical event, 4 although the reaction is not clean and remains somewhat poorly characterized. equation (4), As outlined in elimination of H2 from the iridium dihydride complex is a facile photochemical process; similar thermal reactivity is not observed even after prolonged heating at 150° c. 5 Photochemical cleavage of the tantalum silylmethyl bond has also been observed (equation (5) ). 6 Thus, the few photochemical investigations of CO-free organometallic complexes that have been reported demonstrate that interesting reactivity may take place. These observations strongly encourage additional investigation of non-CO containing complexes. Again, investigations in organic chemistry, in analogy with it is hoped that these studies will result in the discovery of novel and useful 132 molecular transformations. Proposal As discussed in the preceding section, the study of the photochemistry of organometallic complexes promises to reveal unusual and, hopefully, useful modes of reactivity. Outlined in this section is a general program for the study of organometallic complexes (metallacyclopentanes), compounds which are likely to display unusual reactivity. 7 It is proposed that photochemistry of complexes 1 through 5 be determined. the These compounds have been characterized previously 8 - 12 and in most cases their thermal chemistry has been examined. Some of the molecules undergo thermal decomposition at ambient temperature; irradiation at low temperature, however, complications due to thermal reaction. should eliminate These complexes have been selected because,~ priori, several different modes of reactivity seem possible for each. It would be interesting to look for reactivity dependence on the wavelength of the incident radiation 13 (reflecting the nature of the electronic transition). If the primary photochemical process in complexes 1 through 5 is metal-carbon bond homolysis (in analogy with the photochemistry of Cp 2TiMe 2 ) unusual reactivity may result because the alkyl radical will still be bound to the metal. In addition, acyl complex 4 has several possible points of bond homolysis. Examination of the photochemistry of complexes 1 through 5 may not only provide novel examples of the excited state chemistry of organometallic complexes, but may also provide a unique 133 1 ,..._, 2 -- 134 opportunity to examine the dynamics of reaction of unusual metal alkyl "biradicals". Possible reaction modes of several of the complexes are outlined in Schemes II through V and are discussed below. The speculative; photoreactions postulated are necessarily 1 ikely primary photochemical processes and some of the interesting rearrangements which may follow are outlined in the schemes. The thermal reactions of (bis)triphenylphosphine and (tris)triphenylphosphine metallocyclopentane complexes 1 have been characterized in elegant experiments by Grubbs and coworkers 8 (Scheme II). Thermal reactivity proved to be dependent on the number of phosphine ligands attached to nickel. Electronically excited 1 may simply repeat the thermal chemistry or it may explore new reaction pathways. Several possible photochemical reaction channels are outlined in Scheme II; most interesting are those in equation (9) which involve initial nickel-carbon bond cleavage to give 6 and may be followed by any of several conceivable reaction pathways. Cobaltacyclopentanone 4 has recently been prepared in these laboratories 9 and produces a mixture of cyclopropane and propylene on thermolysis (Scheme II I). Five conceivable primary photochemical processes are outlined in Scheme III. Especially interesting would be light-induced elimination of CO (equation (11)) to give cobaltacyclobutane 7, a molecule which has thus far eluded synthesis. Equations (13), (14) and (15) outline several possible homolysis pathways; (13) and (14) are analogous to the 135 j_ ,...,. D A~ ( ¢.a P)t\. N;_) ~~ <"" f rod.l,\~t 1t22 ( (o) ~ c. Hi== c. W2. ~o.jor -p<-oclu.c..+ "l ... 3 hV / ~O. )' Con<?e,ted (l\~3 N~ Jt\ + □ (1) r~du.c.t~ve. e_ \: ~ :f\l). ~ Or\ t i I""\,., ~v ./~ ('Pef3 \N°y "'-...:..:....:hv~~► (V~)~ N~0 ~ "-- (8) 136 A./ c? ('P~) Co:) + co ( 11) 1 ;"./ 0 \\ c~ CPi3) Co-\\ (12) + C. \.\2. == Cl-h_ Lf "-' ~v ,o c r (N3 ) Co·0 ( 13) 8 I'\.,; /co ( ) Cp W3 Co\__). (\I{) 9 0 Cp (N3 )~0)) • /0 ;"\..., ( 15) 137 Norrish photochemical cleavage observed in ketones. As outlined for the biradicals in equation (9) and, again, in analogy to the reactivity of biradicals generated by Norrish photocleavage, intramolecular hydrogen transfer and cyclization reactions may be possible. Upon heating, tantalum metallacycle 5 undergoes reversible exchange with ethylene and more slowly produces l-butene. 1 electronically excited 5, (Scheme IV). ° For several pathways may be possible It is intriguing to consider the possibility that Ta-carbon bond homolysis (equation (17)) could be followed by intramolecular chlorine transfer to the alkyl-centered radical. Cp'TaC1 2 is believed to form metallacycle complex 5 under an ethylene atmosphere and it is conceivable that performing the photolysis of 5 under ethylene would lead to a catalytic conversion of ethylene to the photochemical products. require that the final step of This would the photochemical reaction generate a free organic molecule and Cp'TaC1 2 or Cp'TaC1 2 (olefin) which could react with more ethylene to reenter the cycle. Mechanistic investigations of the photochemistry displayed by these compounds will be aided by deuterium labelling of the metallacycles. Further details may be elucidated by trapping unsaturated organometallic intermediates (with added 2-electron donor ligands such as phosphines and also with solvents such as cc1 4 which may intermediates 14 ). trap odd-electron organometallic It may be possible to trap some of the intermediates containing alkyl radical fragments by adding radical scavengers such as hydrogen donating molecules. 138 S'c he.me. \V ~ er' To..:) C/ Ta. Cl-i. +- q.h_::::: c... \-\ '2.. I I\ a. et ( I lo) "'( (c y,' = Cs Mes ) "' s I Cp Ta. / H )\ id ~ ~/\/ Cl I ~ (11) 13 ...,__ c,/To.~ I\ C.l C..I 139 References 1. For an outstanding recent review of the field of organometallic photochemistry see Geoffroy, G. L.; Wrighton, M. S.; "Organometallic Photochemistry," Academic Press, N. Y., 1979. 2. Hafa, G.; Kondo, H.; Miyake, A.; .J:...,_ AI!h. ~I!h. ~ , 1968, ~ ' 2278; Sostero, S.; Traverso, O.; Lenarda, M.; Graziani, M.; .J:...,_ Organorn~ ~I!h., 1977, l.ll, 259. 3. See reference Cl) Chapter 2. 4. Alt, H.; Rausch, M. D.; .J:...,_ AI!h. ~I!h. ~ , 1974, 96, 5936; Rausch, M. D.; Boon, w. H.; Alt, H. G.; .J.i. Organorn~ .c.hg.I!h., 1977, 141, 299. For a similar example involving a platinum complex see reference (1), p. 321. 5. Geoffroy, G. L.; Pierantozzi, R.; .J:...,_ AI!h. ,C_hg.I!h. ~ , 1976, .9....8., 8054. 6. McLain, S. J.; Schrock, R. R.; private communication. 7. The photochemistry of one metallacyclopentane has been reported. The hexacoordinate complex r 2 (PMe 2 Ph) 2 PtCH 2 cH 2 cH 2 CH2 gives I-butene on therrnolysis, apparently via loss of phosphine ligand followed by /3 hydrogen elimination to produce the coordinated I-butene. Ethylene is the major product observed on photolysis, but 1butene is also observed. I-Butene appears to be formed in the photolysis by a mechanism similar to the thermal reaction since excess phosphine prevents its formation. Perkins, D. C. L.; Puddephatt, R. J.; Tipper, C. F. H.; .iL.. Organornet, Chern,, 1978, 154, Cl6. 8. Grubbs, R. H.; Miyashita, A.; Lui, M-I. M.; Burke, P. L.; .J:...... AI!h. ,C_hg.I!h. ~ , 1977, 9 9, 3 86 3. 9. Theopold, press. K. H.; Bergman, R. G.; J...,._ Alli ~hgmi ~Q.Q_.__, 10. McLain, s. J.; Wood, C. ~ ; 1979, .lfil, 4558. 11. Young, G. B.; Whitesides, G. M.; .J:...,_ AI!h. .c.hg.I!h. ~ , 1978, 100 5808. 12. Diversi, P.; Ingrosso, G.; Lucherni, A.; Porzio, W.; zocchi, M.; ~g.!Jh. .!:QIDID.Y..Il.L, 1977, 811; Evitt, E. R.; Bergman, R. G.; unpublished results. 13. (a) Strohmeier, W.; von Hobe, D.; .c.hg.I!h. ~gL.. 1961, ~A., 2031; D.; Schrock, R. in R.; .J:...,_ Am ~MI!h. 140 ( b ) St r ohm e i e r , W. ; v on Hob e , D. ; .L. Phys • ~nh. ( F r an k f u r t £ID Mfil.Ill., 1962, ll, 393. These results are considered in detail in reference Cl), pp. 68 to 74. Cc) Rest, A. J.; Sodeau, J. R.; ~nh. ,C,Qmmun., 1975, 696. 14. See reference Cl) pp. 81-87. 141 A NOVEL SYNTHETIC ROUTE TO BINUCLEAR ORGANOMETALLIC COMPOUNDS CONTAINING A BRIDGING METHYLENE FUNCTION Introduction Although the first binuclear organometallic compound containing a bridging methylene group was prepared in 1969, 1 only in the past five years has interest in these molecules become widespread. 2 Interest in these compounds is due to their structural novelty and the possibility that they may display unique reactivity. One unusual characteristic is that, although formally related to metallocarbenes, bridging methylene complexes show much greater kinetic (and perhaps thermodynamic) stability. 3 Another intriguing feature of these molecules is that they belong to a small group of metallacycles containing two metals in the ring. 4 It has been noted that methylene bridged binuclear complexes may be analogues of intermediates formed on the surface of Fischer-Tropsch catalysts. 5 In spite of the activity to prepare and study various representatives of this class of compounds, the synthetic routes discovered to date remain limited in applicability. Several metal complexes react with diazo compounds to give bridging methylene complexes (Scheme I), but the number prepared by this route remains sma11. 1 , 2 a,e,f discovered in two A synthetic route recently laboratories employs the di iodomethane with anionic binuclear complexes reaction of (Scheme I I). 2 h, k The generality of this route is likely to be limited by the number of appropriate anionic binuclear complexes which may be 14 2 Sc.he~e.. I Sc.he.me. 1I. Sc.he Me. ID G?> I M-MI y\ G) e Y-CR1 © L CR"2. Y-C~2. ~ Co \ ~ M-M I e ~ \ c...o 1 0_0 2 -- ""-" /CR1. 2 "'-- I 0 fv\-=-=-=M ~ M~M 't:/ II 0 3 ~ + y 143 prepared. In this proposal a novel synthetic approach to methylene bridged binuclear complexes is outlined; the method could prove to be more general than those previously employed. Proposal The proposed synthetic route to bridging methylene complexes involves the displacement of an appropriate substituent in a binuclear complex by a phosphorus or sulfur ylid. 6 In Scheme III the general reaction sequence is outlined for the attack of an ylid (Y+cR 2 -) on a binuclear complex (1) containing a leaving group, L, and a CO ligand which may bridge the two metals in the final product. The betaine (2) may eliminate the neutral phosphine or sulfur compound to give a methylene complex (3). Similar phosphine eliminations have been observed in both organophosphorus and sulfur betaines (Scheme IV). 7 The reaction of ylides with numerous organometallic complexes has been studied. displacement of co, 8 weakly bound tetrahydrofuran 10 all have been reported. mononuclear Examples of the olefins 9110 and Such reactions have been carried out on a number of carbonyl and cyclopentadienyl complexes of the group VI, VII and VIII transition metals (Cr, Mo, W, Mn, Re, Fe, Co, Rh, Ni, Pd) with phosphorus ylides;S,lO a similar displacement using a sulfur ylid has also been reported. 9 Alkyl, olefin and phenyl substituted phosphorus ylides all have been observed to undergo ligand displacement reactions; 8 c thus it may be possible to prepare a variety of bridging methylene groups 144 E= e/edroY\ w/f/..drawin9 3rouf J 145 (gg., cR 1 R2 ) by this synthetic method. In Scheme V are given several examples of the application of this method to the preparation of binuclear bridging methylene complexes. Reaction (1) would produce a bridging methylene complex (4) previously prepared by Hermann and coworkers (R = H) , 2 a and would provide a good initial test of the method. The nonacarbonyl dimanganese complex (5) has not been reported. The remaining examples (reactions potential of this method for previously (3) to (6)) illustrate the the synthesis of numerous, unavailable methylene complexes. The binuclear complexes chosen represent each of the group VI, VII and VIII triads. In all cases, the starting binuclear complexes are well characterized and readily available. Bridging methylene-substituted containing two synthetically binuclear complexes different transition metals may accessible by this route. also Several be well characterized mixed binuclear complexes which may react with ylides to form bridging methylene complexes are also shown in Scheme V (reactions (7), (8), (9), (10)). The latter two reactions would involve anionic complexes; this could hinder substitution of ylid for co. Elimination of triphenylphosphine from the intermediate betaine, however, may be especially facile in these cases. Because the res ul tan t bridging methylene complexes (12, 13) would be negatively charged, they may be particularly conducive to subsequent functionalization. 146 (2) [ M" (to\ l CK1.. '' => ( c_o \ M:~ MV\ ( s "c_t' ( 3) [ Co (CO\ l ( l-\) [ Cr Mo (co)3 cg, ,, l /"'-.. Ct>)~ - Co (CO /3 "et' ;;:> (co )3 Co - (o "'--' Cl?2 ~ (co ')Cp M o / " MoCr (co) I\ 2 '- / CD 2. -:,"- ( 5) [ Cp i:-e_lco).,_ l_ II 1 ;;:, (Co)Cf re :f'c('(Co) 'co 8 "'"'- 147 I\./ 1l G> (q) ( Co )5 M~- W(Co\s17. 13 148 References 1. Cooke, J; Cullen, w. R.; Green, Che.!!h. ~ J..Al., 1969, 1872. 2. (a) Hermann, W. A.; Reiter, B.; Beirsack, H.; J_._ Qrganom~ .C.hsl .Ilh., 1 9 7 5 , !J..2 , 2 4 5 ; ( b ) H o 1 t o n , J • ; L a p p e r t , M • F . ; Ballard, D. G. H.; Pearce, R.; Atwood, J. L.; Hunter, W. E.; .C.h.g.!!h..C.Q.m.mun., 1976, 480; (c) Green, M.; Laguna, A; Spencer, J. L.; Stone, F. G. A.; J_._ .C.h.g.!!h. ~ Dalton, 1977, 1010; (d) Calvert, R. B.; Shapley, J. R.; .J:..,__ A.!!h. .C.h.g.!!h. ~ , 1977, !J..-2. 5 2 2 5 ; ( e ) H e r m an n , w. A. ; K r u g e r , c. ; G o d d a r d , R. ; Be r n a 1 , I. ~lis.. .C.h.g.!!h. .I.nL.. Ed, Eng 1 , 1 9 7 7 , 16 , 3 3 4 ; ( f ) Brown, M. P.; Fischer, J. R.; Franklin, s. J.; Puddephat, R. J.; Seddon, D. R.; .C.h.g.!!h. .C.Q.m.mun., 1978, 748; (g) Hursthouse, M. B.; Jones, R. A.; Malick, K. M. A.; Wilkinson, G.; J_._ A.!!h. .C.h.g.!!h. .£2£.,_, 1 9 7 9 , 1 0 1 , 41 2 8 ; ( h) S um m e r , C. E. ; R i 1 e y , P. E.; Davis, R. E.; Pettit, R.; h .A.!!h. .C.h.g.!!h. ~ , 1980, 102, 1752; Ci) Berke, H.; AL Naturforsche, 1980, J2, 86; (j) Jeffreys, J. A. D.; J..... .C.hsl!!h. ~ Dalton, 1980, 435; (k) Theopold, K. H.; Bergman, R. G.; .J...... AID~ .C.h~.Ilh. EQg_,_, in press. 3. This statement is based on the observation that metallocarbenes, in particular the methylene complexes, are very reactive and frequently difficult to isolate. Bridging methylene complexes, on the other hand, appear to be generally amenable to isolation and are stable at room temperature. 4. See for example: Knox, S. A. R.; Stansfield, R. F. D.; Stone, F. G. A.; Winter, M. J.; Woodward, P.; .C.hsl.!!h..C.2.ID.ID.illlt.., 1978, 221; (b) Guthrie, D. J. S.; Khand, I. U.; Knox, G. R.; Kollmeier, J.; Paulson, P. L.; Watts, w. E.; J_._ Qrganom~ .Che.!!h., 1975, .2.Q, 93; (c) Garnier, F.; Krausz, P; Dubois, J.E.; .i:L.. Qrganom~ .C.h.g.!!h., 1979, l1Q, 195; (d) Pertici, P.; Vitulli, G.; Tetrahedron Lett,, 1979, 1897. 5. Pettit, 6. The ylid Ph 3 PC(SnCH ) has been allowed to react with 3 Fe 3 Cco> 12 and gave a binuclear bridging methylidene complex in whicli extensive and complex bond reorganization had taken place; Churchill, M. R.; Rotella, F. J.; Abel, E. W.; Muckel john, s. A.; J..... A!!h. .C.hsl.!!h. ~ , 1977, li, 5820. 7. For an example observed with a phosphorous ylide see Johnson, A. w., in "Ylid Chemistry", Academic Press, N. Y., 1966, p. 107. For examples involving sulfur ylides see Trost, B. M.; Melvin, L. S.; in "Sulfur Ylides", Academic Press, N. Y .. 1975, pp. 38-48. 8. (a) R.; M.; Stone, F. G. A.; J_._ unpublished results. Heydenreich, F.; Wilke, A. M. u. G.; Dreeskamp, H.; 149 Hoffman, E. G.; Schroth, G.; Stempfle, K. S. u. W.; Uh JL,. .c.h..g_I!h., 1972, 10, 293; (b) Kaska, W. C.; Mitchell, D. K.; Reichelderfer, R. F.; Kate, W. D.; JL,. AI!h. .c.h..g_I!h. ~ , 1974, l l , 2847; (c) Starzewski, K. A. O.; tom Dieck, H.; Franz, K. D. ; Hohm a n n , F. ; JL. 0 r g an o me t, .c.h..g__m.u 19 7 2 , i l , C 3 5 . 9. Brazo, P.; Frouza, 1974, l i , 143. 10. Reger, 297. D. L.; G.; Ticozzi, C.; iL.. Organom~ .cfilI!h., Culbertson, Ji Organom~ ~I!h., 1977, Ul..,