Formation of Bonds to Carbon at Transition Metal Centers Citation Brown-Wensley, Katherine Ann (1981) Formation of Bonds to Carbon at Transition Metal Centers. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/sz44-eg97. https://resolver.caltech.edu/CaltechTHESIS:03022018-142501266 Abstract The reactions of Cp 2 TiCH 2 Al(Me) 2 Cl 1, Cp 2 TiCH 2 C(R) = CR 2, and Cp 2 TiCH 2 CHRCH 2 3 with organic carbonyl compounds are explored. 1 with carbonyl, in a Wittig-type reaction, exchanges methylene for oxygen and yields alkenes; no intermediates can be observed. When treated with carbonyl, 2a (R = Ø) inserts O = C into the Ti-CH 2 bond, generating oxytitanacyclohexenes. Other metallacyclobutenes similarly insert carbonyl; or, if RC ≡ CR is labile, exchange methylene for oxygen, producing alkenes. 3a (R = t-butyl) with carbonyl compounds also exchanges methylene for oxygen, and the mechanism of this reaction is explored. 3a is in equilibrium with a titanocene-carbene-olefin complex 11, which is trapped by ketones. 11 is also in equilibrium with titanocene carbene 12 and free olefin; 12 is trapped by esters (as well as ketones}. The formation of 11 is rate-determining. Formation of 12 from 11, trapping of 11 by ketone, and the reverse reaction, formation of 3a from 11 are all competitive. Trapping of 12 by esters is competitive with the reverse reaction, formation of 11 from 12 and olefin. Relative rates of reaction of several ketones and esters are determined. A new platinum vinyl complex, L 2 Pt(CH = CH 2 ) Me 3 (L = PMe 2 Ø) is prepared by treatment of the known L 2 Pt(CH = CH 2 )Cl with MeLi at low temperatures. Oxidative addition of MeI to the platinum (II) complex 3 generates L 2 Pt(CH = CH 2 )Me 2 I 4· The thermal decomposition of 4 is examined, and the relative amounts of ethane and propene formed lead to the conclusion that a vinyl group reductively eliminates twenty times faster than methyl. CpCr(NO) 2 Me is treated with various ligands in an attempt to see insertion of NO into the Cr-Me bond. A clean reaction is observed with PMe 3 . Instead of insertion, deoxygenation of NO by PMe 3 generates O = PMe 3 and, ultimately, CpCr(NO) (:NMe) - (PMe 3 ) 5a· Similar reactions occur with CpCr(NO) 2 R, R = n-hexyl, isobutyl, Ø, Thermal decomposition of 5a or treatment with H + generates CH 4 . 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): Dervan, Peter B. (advisor) Grubbs, Robert H. (co-advisor) Thesis Committee: Dervan, Peter B. (chair) Grubbs, Robert H. Bercaw, John E. Gagne, Robert R. Defense Date: 10 October 1980 Funders: Funding Agency Grant Number Earle C. Anthony Fellowship UNSPECIFIED NSF UNSPECIFIED NIH UNSPECIFIED Record Number: CaltechTHESIS:03022018-142501266 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:03022018-142501266 DOI: 10.7907/sz44-eg97 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 10749 Collection: CaltechTHESIS Deposited By: Mel Ray Deposited On: 06 Mar 2018 21:39 Last Modified: 23 Jan 2025 21:06 Thesis Files Preview PDF - Final Version See Usage Policy. 74MB Repository Staff Only: item control page

FORMATION OF BONDS TO CARBON AT TRANSITION METAL CENTERS thesis by Katherine A. Brown-Wensley In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1981 (submitted October 10, 1980) ii To Boogh iii Acknowledgements I would like to thank Bob Bergman for introducing me to organometallic chemistry and for his excellent work as a teacher. Bob Grubbs kindly adopted an orphaned graduate student, and I am grateful for his cheerful help as advisor and for his friendship. I am grateful to much of the Caltech community, but I especially thank the various members of the Bergman and Grubbs research groups for their scientific input, emotional support and wonderful friendships. I sincerely appreciate the encouragement of my family and friends, and wish to give a special thanks to my husband Wayne, without whose love and support I would never have succeeded. I acknowledge financial support from an Earl c. Anthony Fellowship, 1976-1977, and the National Science Foundation and National Institutes of Health. iv Abstract The reactions of Cp2TiCH2Al{Me)2Cl ~, Cp2ficH2C(R)=tR 2, and cp 2ficH CHRCH ~ with organic carbonyl compounds are 2 2 explored. 1 with carbonyl, in a Wittig-type reaction, exchanges methylene for oxygen and yields alkenes; no intermediates can be observed. When treated with carbonyl, 2a (R=¢) inserts O=C -into the Ti-CH bond, generating oxytitanacyclohexenes. Other 2 metallacyclobutenes similarly insert carbonyl; or, if RC:CR is labile, exchange methylene for oxygen, producing alkenes. 3a (R=t-butyl} with carbonyl compounds also exchanges methylene for oxygen, and the mechanism of this reaction is explored. 3a is in equilibrium with a titanocene-carbene-olefin complex 11, which is trapped by ketones. 11 is also in equilibrium with titanocene carbene 12 and free olefin; 12 is trapped by esters (as well as ketones}. determining. The formation of 11 is rate- Formation of 12 from 11, trapping of~~ by ketone, and the reverse reaction, formation of 3a from 11 are all competitive. Trapping of !~ by esters is competitive with the reverse reaction, formation of 11 from 12 and olefin. Relative rates of reaction of several ketones and esters are determined. A new platinum vinyl complex, L Pt(CH=CH }Me 2 2 ~ (L=PMe 2 ¢) is prepared by treatment of the known L Pt(CH=CH 2 )Cl with MeLi 2 at low temperatures. (II) complex ~ Oxidative addition of Me! to the platinum generates L Pt(CH=CH 2 )Me 2 I 2 ~· The thermal v decomposition of ~ is examined, and the relative amounts of ethane and propene formed lead to the conclusion that a vinyl group reductively eliminates twenty times faster than methyl. CpCr(NOl Me ·is treated with various ligands in an attempt 2 to see insertion of NO into the Cr-Me bond. A clean reaction is observed with PMe . Instead of insertion, deoxygenation of 3 NO by PMe generates O=PMe and, ultimately, CpCr(NO) (:NMe)3 3 (PMe3} ~~· Similar reactions occur with CpCr(N0) 2 R, R=£-hexyl, isobutyl, H+ ¢, Thermal decomposition of Sa ar treatment with generates CH . 4 vi Abbreviations Bu= butyl, . -c H 4 9 5 Cp = Cn -c H ), cyclopentadienyl 5 5 5 . Cp* = Cn -c 5 CcH 3 ) 5 ), pentamethylcyclopentadienyl diphos =· 'bis-1,2-(dipheriylphosphino)ethane Et = ethyl, -c 2H5 L = ligand, generally a phosphine or CO M = transition metal ¢ = _p henyl, -c 6H5 R = alkyl THF = tetrahydrofuran vii T'a ble ·o·f Contents Chapter I: Reactions of Titanacyclobutanes and -Butenes With Heteroatom-Carbon Double Bonds 1 Introduction 2 Results and Discussion 6 Experimental 49 References and Notes 67 Appendix I 71 Chapter II: Synthesis of and Reductive Elimination from Platinum-Vinyl Compounds 80 Introduction 81 Results and Discussion 83 Experimental 96 References and Notes 99 Chapter III: Formation of Carbon-Nitrogen Bonds via Transition Metal Nitrosyl Complexes 102 Introduction 103 Results and Discussion 105 Experimental 116 References and Notes 122 Propositions Abstracts 124 Proposition I 126 viii Table of ·contents, Continued Proposition II 136 Proposition III 147 Proposition IV 156 Proposition V 162 -1- CHAPTER I REACTIONS OF TITANACYCLOBUTANES AND -BUTENES WITH HETEROATOM-CARBON DOUBLE BONDS -2- Introduction Interest in metallacyclobutanes, particularly as intermediates in olefin metathesis (Scheme I) has a long . 1 h istory. The currently accepted mechanism requires a transition metal carbene, the origin of which is not Scheme I RH==CHR M==CH 2 + RCH==CHR • . ~ M==CH 2 R M=CHR + <>-R' R'CH=CHR' - - - . M==CHR' + R' HC=CHR R' M = transition metal always clear. Many metathesis catalyst systems contain a low-valent transition metal (frequently Mo or W) and a good Lewis acid (such as A1Me 3), and may be homogeneous or heterogeneous. An important clue to the role of the co-catalyst is found in the methylene-bridged titanium -3- aluminum complex !' isolated by Tebbe and Parshall from the reaction between Cp 2TiC1 2 (Cp = n5 -cyclopentadienyl) and A1Me 3 , 2 the same product can be obtained from Cp TiMe 2 2 and Me 2A1Cl. 1...., Not only does 1 metathesize olefins, 3 but it does so more rapidly in the presence of a good Lewis base (e.g., THF). 4 Tebbe succeeded in isolating metallacyclobutenes 2 from the reaction of 1 and an acetylene in THF solution (Reaction 1) . 4 1 + RC=CR THF .. Reaction 1 ~~' R = ¢ ~' R = SiMe 3 Recently, metallacyclobutanes 3 have been isolated from this system, using pyridine as Lewis base (Reaction 2). 5 Work in our laboratories is continuing to elucidate the nature of the interaction between Me 2A1Cl and ~· -4- R 1 :=/ + pyridine . Reaction 2 ~~' 12, Schrock 6 R = !_-butyl R = i-propyl . has isolated a tungsten carbene 4 which, in the presence of A1Cl 3 , metathesizes olefins for many turnovers (Reaction 3). PEt3 c1"- L~o w~ c~l~cHCMe 3 PEt 3 A1Cl 3 ¢Cl Reaction 3 4 + ~CH + 3 Tebbe has also reported that !' with organic carbonyl compounds, in a Wittig-type reaction, will exchange methylene for oxygen, 2 and Pine et al 7 demonstrated that this is a general reaction, again promoted by the presence of a Lewis base (Reaction 4). It is tempting to consider that a mechanism similar to that suggested for olefin - 5- THF 1 OT pyridine ~R' Reaction 4 metathesis is operating (Scheme II), and this led us to investigate the reactions between !' compounds. Scheme II R + o==(R' ~' and ~ and carbonyl -6- Results and Discussion I. Reactions of 1 With Heteroatom-Carbon Double It has been reported 2 ' 7 that the reaction of 1 with ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Bonds. organic carbonyl compounds produces Wittig-type alkene products in good yields. The organometallic product, which is formed more slowly than the :alkene, is Cp Ti(Me)Cl. 8 2 We have found that a similar reaction is observed when a C=N (instead of C=O) bond is present (Reaction 5), but in poorer (up to 38%) yields. 1 + ¢CH==NMe ¢CH==CH 2 Reaction 5 Our attempts to observe intermediate organometallic products (such as ~) were unsuccessful, even using pyridine as base (the best base for synthesis of metalla- ~~R cp 2 Ti~~R• 5 cyclobutanes ~ s) and trimethylacetaldehyde as the carbonyl substrate (3a is the most stable of the metallacyclobutanes synthesized to date 5 ). Even at low temperatures, one -- - 7 .. could see no evidence for ~ in the NMR, although eventually some 3a is produced, presumably from ~ and 3,3-dimethyl-lbutene, both of which can be observed at intermediate stages of the reaction. Either 5 is not an intermediate in Scheme II, or it reacts to give alkene and titanocene oxide very rapidly. The discussion of these possibilities will be delayed until the entire mechanism of the reaction between 1 and 3 and organic carbonyl compounds is considered (vide infra). Reaction of 1 with the simplest "carbonyl" compound, CO, has previously been found to yield neopentane. 9 In our hands, the organic products varied with reaction time, but included methane, ethylene, ethane, propene, isobutylene, and other c4 and higher hydrocarbons. 10 No ketene, allene or acetylene was observed. The only organometallic product observed is Cp 2Ti(Me)Cl, 11 and this is apparently unstable to the reaction mixture as it eventually disappears (although a poor yield can be isolated at early reaction times). It is too early to speculate on the mechanism of these reactions, and while work clearly remains to be done on this system, we have instead focused on the more tractable systems described below. -8- II. Reactions of Titanacyclobutenes With Carbonyl ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~:· Tebbe reports that titanacyclobutene 2b has a labile bis-trimethylsilylacetylene group which can be displaced by diphenylacetylene or Me 2A1Cl to yield 2a or ~' respectively. In the hope that an acetylenic group could be displaced by a carbonyl, several titanacyclobutenes were treated with a variety of carb6nyl compounds. 2a reacted with a variety of carbonyl compounds to give the insertion products ~' shown in Reaction 6 and identified on the basis of their NMR and IR spectra, color, and HCl decomposition products. Reactions required approximately two days to go to completion at ambient temperatures in benzene solution; yields were quantitative by NMR. R 0 R~' Za R' 6a, R=R'=Me b, R=cp , R' =H <:_, R=¢ , R' =Me d, R=!_-Bu, R'=H -- Reaction 6 -9- 1. NMR data for the various products are given in Table The IR spectra showed Ti-0 bands at 715 cm- 1 , C-0 stretches at 1160 cm- 1 , and bands due to the Cp rings at 3050, 1435, 1120, 1010 and 815 cm- 1 . C=C ring modes due to the phenyl groups appeared at 1590, 1480, and 1435. No bands appeared between 3050 and 1590 cm- 1 . Previously unreported 2c 12 showed similar reactivity with acetone (Reaction 7); data for this complex are also included in Table 1. Reaction times for 2c were somewhat less than for 2a. Reaction 7 Zc The products 6c ~~ and ~~ were isolated as yellow-orange crystals. They are quite robust: they are air-stable 14 an d 6 a h as b een h eate d at 170°C f or . so 1 ut1on, . even 1n weeks (in benzene-d 6 in a sealed NMR tube) without significant decomposition. The yellow-orange color and air stability are typical of other titanium alkoxide compounds. 15 Decomposition by HCl in dry ether occurs over a period of -10- Table 1. Oxytitanacyclohexenes Obtained from the Reaction of 2 and a Carbonyl Compound. Compound (Prepared from) 6a (2a + acetone) R R' R" R" I NMR in Benzene-d 6 Me Me ¢ ¢ 1. 20 (s, 6H), R and R' 2.55 (s' 2H), methylene~ 5.75 (s, 1 OH), Cp rings£ 6.7-7~15 (m, lOH), R" and R" I H 6b 2. 85 (d, J = 4 Hz, lH ) , Ha .£ c 2.92 (s, lH)' Hb5.40 (m' J = 4 Hz, lH ) ' R'~ 5.56 (s, SH), Cp~ e 6.03 ( s, SH), Cp6.9-7.3 (m, lSH), ¢ groups (~~ + benzaldehyde) 6c acetophenone) c3~ + Me 1. SS ( s' 3H), R' 2.99 (s' lH), H.£ a ¢ 3.03 ( s, lH), H c b; S.61 (s' SH), Cpe S.97 (s' SH), Cp6.7-7.2 (rn' 1 SH) , ¢ groups 6d t - Bu H Me Me 0.85 (s, 9H), R 2.61 (d, J = 5 Hz, lH), Ha .£ c 2.69 (s' lH), Hb4.07 (m' J = 5 Hz, lH), R'~ S.S8 (s, SH), Cp-e 6.02 (s' SH), Cp~ 6.7-7.2 (m' lOH), R" and R"' (2a + trimethyl- acetaldehyde) (2c + 6e acetone) Me <P 1.17 (s, 3H), R" 1. S4 (s' 6H), R and R' 2.30 (s, 2H), methylene! 5.80 (s' lOH), Cp rings£ 6.7-7.3 (rn' SH), R''' a) Ha and Hb, which are equivalent. b) Equivalent Cp rings. c) Ha or Hb, which are inequivalent. No specific stereochemical assignment with respect to R and R' is intended. d) Coupled to Ha and Hb. Actually a non-first order 4-line pattern. e) Cp rings are inequivalent. No specific assignment with respect to R and R' is intended. -11- approximately 15 min at room temperature, and yields the expected products, shown in Reaction 8 and identified on the basis of their NMR, IR and mass spectral data. These data is included in Table 2. R PR' HCl .Et 20 Cp 2T ¢ ¢ (80-88%) CpzTiClz ~~' R=R'=Me b, R=¢, R'=H Reaction 8 + R H;:<R' (98-100%) ¢ z~, ¢ ~' R=R'=Me R=¢, R'=H Reaction of ~~ with benzaldehyde yielded a minor product (approximately 20%) as well as the major product ~~ (80%). The minor product is an isomer which can occasionally be obtained nearly pure as large, dark red crystals from a toluene-pentane recrystallization of the mixture obtained from the reaction of 2a and benzaldehyde. It has been identified as isomer 8.... on the basis of its NMR spectrum, color, and HCl decomposition products. These data are contained in Table 3. deep red color is In particular, the atypical of titanium alkoxide compounds, and the HCl decomposition product contains no alcoholic group, so a Ti-0 bond is unlikely. Furthermore, the -12- Spectral Data for the HCl Decomposition Products 7a and 7b. Table 2. o 1.10 (s, 6H), Me groups 1. 13 ( s , 1H) , - OH 2.68 (s, 2H), methylene 6.46 (s, lH), vinyl proton! 6.8-7.3 (m, lOH), phenyl groups 3590 cm -1 , OH IR (KBr): 7a MS: ~/~ (rel. int.) 252 (2), parent ion 234 (2), parent-H 2 o~ 219 (2), 234-CH 3 194 (100), parent-acetone 179 (43), 194-CH 3 NMR (CC1 4 ): 02.04 (s, lH), -OH 2.80 (d, J = 6.8 Hz, 2H), methylene 4.48 (t, J = 6.8 Hz, lH), Ha 6.33 (s, lH), vinyl proton! 6.8-7.3 (m, 15H), phenyl groups ¢ ¢ IR (KBr): 3590 cm- 1 , OH 7b MS: ~/~ (rel. int.) 300 (1), parent ion 282 (3), parent-H o~ 2 194 (100), parent-benzaldehyde a) Compare to values of 66.39 for cis-methylstilbene and 6.75 for trans-methylstilbene: M. Michman and H. H. Zeiss, J. Organomet. Chem., 15, 139 (1968). HCrefecomposition of 2a also yields cis-methylst1lbene. b) this is-a major fragment at low eV. -13- !~~!~-~· cf> Data for Oxytitanacyclohexene Isomer 8. He )f-o <S NMR (CC1 ): 04.95 (s, 2H), Ha 5 • Z5 ( s , 2H) , Hb 6.4Z (s, lH), He 6.S-7.3 (m, lSH), phenyl groups IR (CC1 4 ): no bands above 3ZOO cm-l Hb CpzT>=«Ha cf> cf> (s' lH)' H ~ aa (s' lH)' Hb( s' lH), Hc a (s' SH), ·c pa 6~19 (s' SH), Cp6.7-7.Z (m, lSH), phenyl groups 1.49 1.60 2.12 S.6Z NMR (C 6D6 ): 8 4 MS: Product of HCl Decomposition ~/~ (rel. int.) 300 (0.1), parent Z98 ( z) ' parent-Hz Z96 ( z) ' parent-Z Hz Z8Z (100)' Z98-0 19Z (100)' 298-benzaldehyde a) No specific stereochemical assignment with respect to He and the phenyl group on the same carbon as He is intended . -14- the vinylic proton . on the HCl decomposition product is clearly seen so that the Ti-C(~) bond must have remained intact and insertion must have occurred into the Ti-CH 2 bond. Compound 8 is therefore the structural assignment. 8 Compounds ~~' and the previously unreported Zd 13 and ~~, 12 reacted very slowly with carbonyl compounds. No tractable organometallic products were formed, and small amounts of alkene products from methylene for oxygen ~~' ~, R'=R"'=Me R"=SiMe , R '" =¢ 3 exchange (e.g., styrene from benzaldehyde) were observed. If, as Tebbe suggests, 4 ~~ gives rise to the intermediate responsible for olefin metathesis (by virtue of the labile -15- bis-trimethylsilyl~cetylene) and if this intermediate is also responsible for the methylene for oxygen exchange reaction described for ~ in the presence of a Lewis base, then the alkene products are not surprising. The acetylene- carbene complex may also be responsible for some or all of the methylene for oxygen exchange. There may be, however, ~~' alternate decomposition pathways available to ~~'accounting ~~' and for the poor yields of alkene (Scheme III). Scheme III 1 decomposition products i -R'"c=cR' R"' ~~' R"=R"'=SiMe 3 ~' R"=R"'=Me e, R"=SiMe 3 , R"'=¢ O C>< l Cp 2Ti R R' -16- No reaction at all was observed between 2a and Nbenzylidenemethylamine, up to temperatures sufficient to decompose 2a. A brief study of the reactivity of 6a toward CO was conducted. Although the reaction was slow, in two weeks at 70°C in THF-d 8 in a sealed NMR tube under 1 atmosphere of CO, all 6a was observed to disappear, and a new product 9 was formed in 78% yield (by NMR) (Reaction 9). -product was identified on the basis of NMR, IR and Themass 0 co ¢ OH Reaction 9 THF ¢ ¢ ¢ spectral data; 9 these are included in Table 4. While these few reactions do not establish wide generality, this does suggest a route to a,8 -unsaturated acids or lactones (if an intermediate titanacycle can be made to reductively eliminate or if the product is condensed to the lactone)(Scheme IV). While the first step, formation of the titanacyclo13 butene 2, has been used only once on a preparative scale, it has also succeeded on a NMR-tube scale 12 and may prove to - be applicable even to terminal acetylenes. 16 With the 9 0 &1.10 (s, 6H), Me groups 1.20 (s, lH)~ -OH 2.77 (s, 2H), methylene 6.9-7.3 (m, lOH), phenyl groups (COOH not located) MS: ~/~ a 279 (30), parent-OH178 (100), diphenylacetylene (rel. int.) IR (CC1 4 ): 3450 cm - 1 (broad, intense), -OH 1660 cm -1 (strong), CO stretch NMR (THF-d 8 ): Spectral Data for CO Insertion Product 9. a) Parent ion not visible, due to poor signal to noise ratio. !~~!~-~o I 1 -......] ~ .-18- Scheme IV 1 pyridine R '" C:::CR" 0 R' co R'" OH R R'' R"' R' R" exception of particularly labile acetylenes, the formation of 6 from 2 appears to be well established, and perhaps CO insertion and removal from titanium will be equally general. Work remains to be done to determine which functional groups (on R, R', R", and R"') will not interfere with the course of the reaction, whether or not a one-pot synthesis can be effected, optimum reaction conditions, and how widely applicable the scheme is, but the approach is an attractive one. The reaction of titanacyclobutenes with carbonyl compounds is, on the basis of a few preliminary experiments, -19- second order (first order in titanacyclobutene, first order in carbonyl). Carbonyl compounds which typically show sharp resonances in the NMR show instead broadened signals in the presence of titanacyclobutene; when the reaction has gone to completion, any excess carbonyl compound again shows a sharp resonance, suggesting that the carbonyl is reversibly binding to 2 before reaction . ..;. The reaction shows a solvent dependence (the reaction is approximately ten times faster in THF than in benzene). III. Reaction of Titanacyclobutanes and Heteroatom- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Carbon Double Bonds. Grubbs 5 has reported that the alkene (3,3-dimethyl-l-butene) can be displaced from titanacyclobutane ~~by diphenylacetylene or Me 2A1Cl. We investigated the possibility that alkene might also be displaced by carbonyl to give oxytitanacyclobutane 5 or that insertion might occur to give an oxytitanacyclohexane 10. All investigations along these lines were performed with ~~' the most stable titanacyclobutane to date. R R' R" 10 -20- Compound ~~' with all organic carbonyl compounds tested, gives the alkene product of methylene for oxygen exchange in good yield (Reaction 10). in Table ·s. Yields are listed Ketone and ester substrates react with one0 ~R' ~2 + [Cp 2Ti==O] + CH R R' \j ~ Reaction 10 to-one stoichiometry, while two equivalents of aldehyde substrate are required for complete reaction of 3a. also is shown in Table ·s. This The organometallic product is titanocene oxide polymer, an intractable yellow solid described in the literature. 17 It is presumably formed from mononuclear titanocene oxide, although the inter- mediacy of this is not directly observed. 3,3-Dimethyl- l-butene (100% per Ti) is also produced. Compound 3a also reacts with C=N double bonds, but -- the yield is much poorer than reactions with C=O double bonds (Reaction 11). The mechanism of the reaction between 3a and carbonyl is of interest, especially since one might expect some common intermediates in this and other reactions -df 3a and perhaps even of 1. This will be discussed below. -21- I~~!~-~· Reactions of 3a With Carbonyl Substrates. 3a Substrate Product (Yield per Ti) Equivalents Required (n) acetone isobutylene (90%)a 1 benzaldehyde styrene (92%)b 2 acetophenone a-methylstyrene (94%)a 1 trimethylacetaldehyde 3,3-dimethyl-1-butenea (100%) 2 methyl benzoate a-methoxystyrene (90%)b 1 a) NMR yield. b) Yield by gas chromatography. Reaction 11 <PCH=CH 2 (9 %) -22- IV. Mechanism of the Reaction Between 3a and Carbonyl Compounds. One can envision a number of ~~~~~~~~~~~~~~~~~~ mechanisms for the reaction between 3a .and a carbonyl substrate. -- Several reactive organometallic intermediates, including a titanocene carbene ~~' are suggested by analogy to the proposed ·olefin metathesis mechanism (References 2, 3, and 4, and see also Scheme I), and they are certainly in equilibrium, either with or without the intermediacy of carbene-olefin complex ~~ as a discrete intermediate (Reactions 12 and 13). YCp Tic:>--<- ----z i 3a Compound 3a and/or .....--- Cp2Ti=CH2 11 + Y- 12 13 Reaction 12 Reaction 13 + 3a 12 Cp2Ti=CH2 13 any of these intermediates might react with carbonyl substrate to yield alkene. In the presence of added 3,3-dimethyl-l-butene ~~' carbonyl, solvent, or another potential ligand, several other species are also possible (14, 15, and 16, for instance), and these may also be -23- species which are important in the mechanism. Steric 18 requirements for!~ may be severe, and 15 is formally a 20-electron species (although one could argue that a 5 3 Cp ring has backed off from n to n to alleviate this situation). 12. When L = !~' ~~is, of course, 11 in Reaction Depending on the rate-determining step, one might see first or second order kinetics, or one might observe L Cp 2 T~ L L Cp2Ti==CH2 CpzTi==CH2 I ~ 15 14 some intermediate species. I 16 Derivation of kinetic expressions to describe various mechanisms is possible, but before further discussion we shall narrow the field somewhat by considering some of the data. Kinetic data have been obtained by following the reaction between Compound ~~ ~~ and a carbonyl substrate by NMR. reacts at approximately (± 20%) the same rate with all carbonyl substrates (Table 6). The reaction rate is neatly independent of carbonyl concentration (from one to fourteen equivalents) and shows first order kinetics -24- Table 6. Rates of Reaction of 3a With Various Substrates. Substrate (Concentration, M) Rate (Temperature, in sec""l oc)~ tri~ethylacetaldehyd~ (0.44)b 1.6 x 10-s (25) acetophenone (0.22)b 1.5 x 10-s (25) 1.3 x 10 ... 5 (ZS) methyl benzoate (0.22)b acetone (0.72)c acetone (13.7)c 2.43 x 10-s (27) 2.67 x 10- 5 (27) methyl benzoate (0.88)b 4.89 x 10-s (37) 4.94 x 10- 5 (37) methyl benzoate (l.76)b 5.42 x 10-s (37) trimethylacetaldehyde (0.36)£ 9.3 x 10-s (40) trimethylacetaldehyde (0.72)c 9.8 x 10-s (40) trimethylacetaldehyde (1.54)c 10. 7 x 10-s (40) methyl benzoate (0.44)b a) Plotted as first order in 3a, zero order in substrate. See also Footnote 19. b) [~~f"= 0.22 M. c) [~~) = 0.36 M. -25- for the disappearance of 3a for greater than one half-life. 19 f':'- The reaction is independent of solvent, with approximately the same rates being observed in benzene and THF. Other workers have observed similar rates and kinetics (first order in ~~' zero order in substrate) with acetylenes. 12 No new species are observed to grow in and then disappear (except, of course, that one observes the final products, 13 and the alkene from methylene for oxygen exchange, . . ) . 20 growing 1n The NMR spectra of 3a are the same in the presence or absence of carbonyl substrate, added alkene 13, or pyridine. The NMR resonances for the carbonyl substrates remain sharp and are not shifted in the presence of 3a. This suggests that one is not rapidly forming a complex such as !~ (L = carbonyl compound) and then actually observing the disappearance of the complex. Furthermore, one might expect an observed rate of disappearance of 14 to vary greatly as carbonyl ligand L varies 21 (from ketones to esters, and small to bulky substrates), and this is not observed. 16. Similar arguments apply to 15 and The first order kinetic data rule out a reaction between 3a and carbonyl (or between any species and carbonyl) in the rate-determining step. This leaves us with the following possibilities: either 11 or 12 or both are trapped by carbonyl, and the -26- rate-determining step is formation of !! from 3a or of 12 from !!' as long as it occurs before trapping by carbonyl (Scheme V, Cases 1 through 4). Trapping of!! by carbonyl will involve 15 as either a transition state or an inter· mediate further along the reaction pathway; trapping of !~ leads to 16. similarly, Little can be learned about these from the experimental data, · since only the starting materials and final products can be observed, but such intermediates will be discussed later (vide infra). Scheme V ....,,,,,,,.,l'V..._,..-.,,-.,,"W,..., Ti(>-+ kl k_l ~ Ti=CH 2 3a Ykz + k~z Ti=CH 2 11 k 3 [carbonyl] products k 4 [carbonyl] products 13 12 -27- Notes: Cp rings omitted. Depending on relative rates and the rate-determining step, the following cases are envisioned: Case 1: k 3 [carbonyl] << k_ 1 or k 2 , so that no trapping of !! by carbonyl occurs. k 1 or k 2 can be rate- determining. Tie>-+ k1 k-1 3a Y- ~ kz Ti=CH 2 k-2 I 13 + Ti=CH 2 12 11 k 4 {carbonyl] j products Case 2: kt~< k 3 [carbonylJ or k_ 1 , so that no free carbene 12 is formed or trapped. determining. k 1 is rate- Actually, one obtains identical kinetic expressions with or without including the equilibrium between ~! and 12, if trapping of 12 does not occur (although it seems unlikely that 12 would not be trapped by carbonyl), and -28- both of these possibilities are derived in Appendix I, to prov.e this. To-+ 3a k 3 [carbonyl] I 11 products Case 3: Trapping of both 11 and ~~ occurs, just as shown in Scheme V. k 1 or k 2 rate-determining, but to see first order kinetics, k 3 Icarbonyl] must be negligible if k 2 is rate-determining. Mixed kinetics are possible. Case 4: Compound ~~might be directly in equilibrium with !~' which is trapped by carbonyl. We must now assign different rate constants, since we are talking about different reactions. determining. k 5 is rate- -29- Ti~ Ti=CH 2 12 l + Y13 k 4 [carbonyl] products For the present, we will just consider trapping of 11 and 12 to lead to products with a rate k 3 or k 4 , respectively, which represents the overall rate of the reaction (i.e., the rate-determining step) and which is assumed to be second order (first order in 11 or 12, and *"V#'V first order in carbonyl). _,.,,_..., The various mechanisms are given in Scheme V, Cases 1 through 4, and kinetic expressions have been derived for each of these in Appendix I. While the reactions with rate constants k 3 and k 4 in these schemes cannot be rate determining (or second order kinetics would be observed), they may be either comparable -30- to or much faster than competing reactions. If 12 (titanium carbene from which alkene 13 has dissociated) is (one of) the intermediates which is trapped (Cases 1, 3 or 4) and if petitive with k 4 Icarbonyl] k_ 2 [!~] or k_ 5 [!~J is com- then addition of alKene 13 to the reaction mixture should slow the formation of products and the disappearance of 3a. Indeed, when methyl benzoate is the carbonyl substrate, we do observe that the rate of disappearance of ~~' while still best fitting first order kinetics, shows a slight decrease (Table 7). Therefore, methyl benzoate must be trapping 12 and k 4 [carbonyl] and k~z[!~] or k_ 5 [!~J mnst be competitive. Methyl benzoate may also be trapping !!~ However, one would expect trapping of 11 (which is formally 18-electron) to be considerably slower than trapping of !~ (which is 16-electron) for a given substrate. Let us make the approximation that methyl benzoate (=MB) traps (Cases 1 or 4). only~~ We can then do the following exp_eriment: using large amounts of 13 and MB relative to 3a (so that the concentrations of 13 and MB cart be assumed to remain approximately constant during the first half-life of the reaction), we can vary from sample to sample the ratio of !~ to MB, and obtain for each sample an observed rate constant kobs by plotting the first order disappearance of 3a. For Case 1 (12 is formed from 11,·and 12 is -31- trapped by carbonyl) or [13] k ={l+k-1} obs We can plot k 1 /k 0 .b s vs + [MB] .k z I!~]/ .[MB] and obtain a plot with intercept (1 + k_ 1 /k 2) and slope (k_ 1k_ 2)/k 2k 4 ). To do this, of course, we need a value for k 1 . For Case 4 (where 12 is formed directly from ~~, and 12 is trapped) [13] + [MB] and a plot of l/kobs vs I~~]/ [MB] yields a plot with intercept l/k 5 and slope k_ 5 /k 5k 4 . When 3a is allowed to react with a mixture of excess acetone and excess methyl benzoate, the sole product observed (by NMR or gas chromatographic analysis) is isobutylene, indicating that acetone traps the reactive -32- organometallic intermediate(s) at least 100 times faster than methyl benzoate. This is consistent with the observations made by Pine for != 1 reacts much more rapidly with acetone than methyl benzoate does. 21 We - have also generated !~ 22 in the presence of acetone and methyl benzoate (by adding a mixture of excess pyridine, acetone and methyl benzoate to 1), and again observed acetone to react at least 100 times faster. Since we are seeing what appear to be first order kinetics for the disappearance of ~~ even in the presence of only low concentrations of methyl benzoate, we can assume that in high concentrations of acetone, trapping is very efficient, and we can use the approximation that k 1 ~ kobs when the reaction is followed in neat acetone-d 6 . The observed rates at varying ratios of [13]/[MB] are contained in Table 7. -- Following Case 1 (trapping only 12, where 12 is formed from 11) and using the. approximate value ~~ ~~ of k 1 obtained in neat acetone-d 6 , the plot of k 1 /kobs versus f!~J/[MB] yields an intercept of 1.57 and a slope of 0.284, which leads to the following: = 0.57 = 0.284 0.57 = a.so - 33- Table 7. Observed Rates of Disappearance of ~~ in the Presence of Varying Amounts of Methyl Benzoate and Alkene 13.a,~,c [!-~JI [MB] -1 kobs' sec 1.21 2.00 2.50 x 10- 5 2.42 2.42 1.00 2.85 x 10-s 3 1. 21 1.21 1.00 4 2.42 3.63 0.67 2.82 x 10-s 2.98 x 10- 5 5 1.21 2.42 0.50 3.17 x 10-s [13] [MBJ M M 1 2.42 2 Sample # a) Plotted as first order for the disappearance of ~g. b) At 31.5°C. [~~] = 0.242 M. c) kobs in neat acetoneds = 4.98 x 10- 5 sec- 1 at 31.5°C. This is consistent with our assumption that k_ 2 and k 4 are competitive. If we follow Case 4 C!~ is trapped after being formed directly from~~), we obtain an intercept of 2o97 x 10 4 sec, and a slope of 5.37 x 10 3 sec, which gives us k 5 = 3.37 x 10- 5 sec -1 = 0.181 -34- This also gives a result consistent with our assumption of competitive k 4 and k_ 5 • However, it is impossible for ks to be 3.37 x 10-S sec-l if we have already observed faster reactions when acetone is the substrate, as there is no way the reaction can proceed faster than k 5 (the rate . . determining step), no matter how efficiently and rapidly we trap 12. This suggests that Case 4 · is incorrect, and further experimentation will support this contention (vide infra). If acetone reacts much more rapidly with 12 than methyl benzoate does (k 4 1 for acetone>> k 4 for methyl . reasona bl e to suppose t h at acetone a 1 so b e.nzoate ) , 2 3 1. t 1s reacts much more rapidly with 11 than methyl benzoate (k 3 ' for acetone >> ~ 3 for methyl benzoate). Case 1, in which carbonyl substrate traps only ~~ is now (possibly) insufficient to describe the reaction of 3a with acetone, so we must consider Cases 2 (in which carbonyl traps 11) and 3 (in which carbonyl traps 11 and 12). Case 2 will apply only if acetone traps 11 so rapidly that 12 never has a chance to form (k 2 -- [~~] ~~ k 3 I~~] [acetone]). We can do the following experiment: a large excess of acetone (such that the concentration remains approximately constant over the course of the first half-life) can be used in each sample, with the concentration of acetone varying from sample to sample. The rate constants -35- for this experiment are given in Table 8. If we assume for a moment that acetone traps 11 so rapidly that 12 never has a chance to form, then we can use Case 2 to describe our system, and the rate constant observed for disappearance of ~~' kobs for each sample, will be = kl kobs k 3 ,J acetone] k_l + k 3 ,Iacetone] or 1 1 = - + kl kobs k . -1 k •k [acetone] 3 1 and a plot of 1/k 0 b S . versus l/[acetone] gives a plot with intercept l/k 1 and slope k_ 1 /k 3 •k 1 . From the intercept of 1.94 x 10 4 sec, we obtain a value for k 1 24 of 5.15 x 10-S sec-l = and the slope (1.68 x 10 4 sec·M) yields a value for k_ 1 /k 3 ' of = 0.87 M -36- Table 8. Observed Rates of Disappearance of 3a in the Presence of Varying Amounts of Acetone-d 6 .a,b Sample # [acetone] , M k obs' sec ·1 1 13.7 4.98 x 10- 5 2 6.85 3 3.65 4.76 x 10-s 4.36 x 10- 5 4 2.42 4.21 x 10- 5 5C 6.85 6c 2.42 4.9 x 10-s 4.3 x 10- 5 a) Plotted as first order for the disappearance of 2e· b) At 31 . 5 ° C • L~ ~] = 0 • 24 2 M. c) These s amp 1 es a 1 so contain added alkene, I!~] = 3.87 M. Due to the difficulty in integrating somewhat poorly resolved peaks in the NMR in the presence of large amounts of 13, these rate constants are somewhat less accurate than those determined for Samples 1 through 4, and so k 0 b is only expressed to two significant figures. s Since we know from the methyl benzoate work that k_ 1 /k 2 = 0.57, we can determine that = 0.66 M-l In neat acetone-d 6 (13.7 M), k 3 ,Iacetone]/k 2 is, therefore, -37- 9.04/1, so our use of Case 2 to describe the reaction in neat acetone-d 6 is certainly reasonable. If we are to be rigorous, however, we should instead consider .Case 3 in which both 11 and 12 are trapped by acetone [i.e., ·k , [acetoneJ is not so much greater than 3 k 2 that we can neglect k 2 ]. We know that k_ 2 Jk 4 for methyl benzoate is 0.50, and that least 100. 23 k4,{acetone]/k 4 [MB] is at In deriving kinetic expressions for Case 3, therefore, we can use the fact that k 4 ' some simplifications. k >> k 2 , and make We then obtain: obs or + k 3 ,{acetone] k ... 1 Using the value of k 1 observed above (since this is our best number for this rate constant) we can plot (k 1 /kobs 1)-l versus (acetone] to obtain values for k_ 1 /k 2 (the intercept) and k 3 /k_ (the slope). Since the difference 1 between k 1 and kobs is very small, there is a large amount of error in this determination, but it is encouraging to note that the numbers so obtained: -38- = 0.26 = 1.74 are qualitatively in agreement with those values which we had determined previously. Since the observed rate in neat acetone-d 6 is, within experimental error, k , we can now determin 1 for Reaction 14 by determining k 1 Cp 2T0-f 3a Ea and !:J. S =f at various temperatures. Reaction 14 11 The rates are :listed in Table 9, and a plot of log 10 Ck 1 ) versus l/T (temperature in °K) gives a slope of -4.73 x 10 3 °K and an intercept of 11.2. From this it can be calculated that: Ea = 21.6 kcal/mole 6S r = -9. 31 e. u. While both of these values are quite reasonable for k 1 in Case 2, 3 or 4, a negative 6S =f is inconsistent with a dissociative rate-determining step (such as k 5 in Case 4 -39- Table 9. Observed Rate Constants for the Disappearance of 3a at Various Temperatures.~ Sample # a) [~~] Temperature, oc k obs' sec -1 1 51. 0 4.43 x 10- 4 2 31.5 5 4.98 x 10- 3 27.0 4 4.8 2.67 x 10- 5 1.67 x 10- 6 = 0.242 M, in neat acetone-ds. or k 2 in Case 1). In addition to the competition between acetone and methyl benzoate described above, several other competition experiments were performed, between substrates expected to trap the same intermediate. Relative rates of reaction of two substrates were determined by measuring the relative amounts of products formed · from each. It was assumed that reaction of 1 with pyridine would generate ~~ (by removing Me 2A1Cl from the coordination sphere of titanium), which could then be trapped with ketone or ester substrates. Two competing esters, when allowed to react with ~~' would also trap ~~' so the same relative rates should be observed for competing -40- esters in either case. alkene !~' Even in the presence of added little or no trapping at 11 is to be expected of esters. Ketones, however, when allowed to react with ~~' will trap !!' and different ratios of rates might be expected for trapping 11 and 12 (from! and pyridine). Added alkene, however, should have no effect on the trapping of 11 (since 11 will never dissociate alkene 13 ·before being trapped) or !~ (since k_ 2 [!~] is so small compared to k 4 ,k 4 , [ketone]). The results of the competition experiments are shown in Table 10. Observations are as expected. Although steric effects must certainly enter into considerations of the relative rates of reactivity of different substrates, one also sees that rates increase as electron density on oxygen (as a result of more electron-donating groups on the carbonyl group) increases. It appears then that Scheme V describes our system (Scheme V and known ~elative rates are reproduced below). k 3 [MB] being too slow to compete effectively with k_ 1 or k 4 (Case 1). k_ Methyl benzoate traps only at!~' 2 [!~] k 4 [MB] is the same order of magnitude as in our system, and so a dependence of kobs on [13) is observed. Acetone, on the other hand, can trap at !! and, at very high concentrations of acetone, neither k_ 1 nor k 2 can compete effectively with applies). .:~3, {acetone) (Case 2 Acetone will also trap 12 (if 12 has a chance to -41- I~2!~_!Q. Competition Experiments. Substrate A Substrate B Reactive Ti Species [13] ! -M Rate A Rate B Acetone Acetophenone 1.1!?_ 0 2.3 Acetone Acetophenone llb 3.9 2. 3 Acetone Acetophenone 12c 0 2.0 Acetone Acetophenone 12£ 3.9 2.1 Ethyl Acetate Methyl Benzoate 12d 0 13 Ethyl Acetate Methyl Benzoate 12d 3.9 14 Ethyl Acetate Methyl Benzoate 12c 0 15 Ethyl Acetate Methyl Benzoate 12£ 3.9 15 a) Concentration of ·added alkene. This does not include the concentration of !~ produced by reaction of ~e· b) Using Case 2 to approximately describe the reaction of ~~with ketone, the reaction of ketone with 3a will result in the trapping of !!· Small amounts of !~-will also be trapped (Case 3). c) From! and pyridine. d) Using Case 1 (or Case 3, with a very small k3), the reaction of 3a with esters will trap only 12. form from 11: Case 3), and k 4 ' is undoubtedly so fast that k_ 2 has no chance of competing with it; and, therefore, no dependence of kobs on alkene concentration is observed. It would, of course, be very interesting to know k 1 /k_ 1 or k /k_ 2 . 2 Unfortunately, the present system does not allow -42- Scheme V 3a 11 12 products products 13 k 3 ,[acetone] k 1 = 5.5 x 10 - 5 sec -1 at 31.5°C ~s 9= =-9 e.u. Ea = 22 kcal/mole k -1 k2 products k = 0.57 -2 = 0.50 = 0.87 M k4 k -1 k3' k3' = 0.66 M-l k2 for the determination of these or of any absolute rates (except k 1 ) because 11 and 12 cannot be directly observed. (Examination of the kinetic expressions will reveal that all one can determine from such a system is the relative rates -43- of any two reactions which compete for a given intermediate.) Alternate approaches might, however, yield such information. If one could generate 12 in known concentrations and watch its disappearance (or the formation of products from ~~ and some other substrate) one could learn k 4 . have attempted to generate !~ from We ! and pyridine at low temperatures and then follow the subsequent reactions by NMR, but the spectral resolution is too poor to allow one to succeed in this approach. 25 Another means to 12 might be to synthesize a diazo compound such as 17, which could perhaps easily loose N2 and generate !~· Work along these lines is proceeding in our laboratories. _,/CH 2 Cp Ti '-....N . 2 "N~ 17 Photochemical approaches may also hold some promise: disappearance of ~~ in neat acetone-d 6 at -20°C is two to three orders of magnitude faster when the reaction is irradiated with unfiltered ultraviolet light. Work by several workers 26 on the photochemistry of Cp 2TiMe 2 and related compounds suggests that the initial photoproducts are radicals (Reaction 15), so that (if we are seeing -44- hr Reaction 15 photochemistry from an excited state of 3a rather than attack by an n + ~* state of acetone) we might initially expect to form a 1,4-diradical which might then lead to 11 or 12 (Reaction 16). The · formation of double bonds from an organic 1,4-diradical is well established for cyclobutanones (Reaction 17), 27 and ethylene has been observed in the photodecomposition of platinocyclobutanes. Cp 2 Ti~ 28 / ·~ 3a 11 Reaction 16 -4 5- hr • ==0 = + Reaction 17 One could perhaps distinguish between the photoproducts (i.e., between !! and 12) on the basis of competition experiments: since the relative rates of reaction of two ketone substrates with 11 and 12 are known and are different, the product ratio from irradiation ought to reflect the photochemically formed intermediate. Ketones are ideal for this study since we know they react rapidly with 11 or 12 (this prevents 11, if it is formed photochemically, from going to!~ before being trapped). One might also be able to directly observe the photoproduct (by generating it in high concentrations via flash photolysis) and watch its rate of disappearance in the presence of a reactive substrate. A great deal might be learned from such a study. Another approach is to tackle the system with theoretical studies. Rappe and Goddard 29 have reported calculations on -46- molybdenum and chromium systems. The M=O bonds in these systems strongly direct the course of the reactions (Reaction 18), and therefore a good analogy to our titanium '-/ \/ Cl". #0 M~ Cl/ ~X c cII + / Cl" Ill0 /C" / M" c, Cl/ X/ ~Cl"-.M~O Cl/ 'CH + 2 x II c /\ c o/ I /\ ' system cannot be made. c1, I /c M Cl/ ~X Reaction 18 Work is in progress to calculate the corresponding titanium reactions (Reaction 19), but it is not yet completed. 30 However, rough initial estimates of . d"1cate ' . 30 1n t h at t h e overa 11 sequence o f 12 b on d energies going to products ought to be exothermic (by 28-30 kcal/mole, if one uses bond energies of O=CH 2 = 175 kcal/mole and 31 H2 C=CH 2 = 163 kcal/mole). While one might predict that + 70-80 kcal/mole O=C - c1 2T() ' 'o / Cl 2Ti=O + I 120 kcal/mole Reaction 19 -47- Reaction 20 ought to be downhill every step of the way (and also rapid), it is premature to draw any conclusions, since so little is known about the metallacyclobutanes or Cp 2Ti=O Reaction 20 oxymetallacyclobutanes. Interestingly, Caulton 32 suggests that the alkoxy ligand in Cp 2Ti(OEt)Cl functions as a 3electron donor; but, it is too early to judge how that fact will be reflected in an oxytitanacyclobutane. Metallocyclobutenes may form oxymetallacyclohexenes if the formation of ·a carbene-acetylene complex !~ is much slower than the insertion of the carbonyl compound (k 1 << k 2 {carbonyl] in Scheme VI) . In cases where k 1 is fast, however, some oxygen for methylene exchange may occur, either via trapping of !~ or !~ or both, along with other decomposition reactions. A thorough investigation of these possibilities has not been made. R "' 2 6 R "' k2 kl R' R [o=<J Cp2TY-R" Scheme VI ,...,,...,,...., ,....,,....,,...,,..., """' t R -- 18 R~"_C==CR"~R' Pz C Ti=CHz / ==<R' R R 12 o=(R' CpzTi=CHz Decomposition ==< I I 00 .i::. -49- Experimental Section ~~~~~~~~~~~~~~~~~~~~ General. All manipulations (except work . . up of organic compounds) were carried out under purified nitrogen in a Vacuum Atmospheres Recirculating Dry Box, under argon using standard Schlenk techniques, or on a vacuum line. Benzene, toluene, THF, diethyl ether, dimethoxyethane and their deuterated analogs were vacuum transferred from sodium benzophenone ketyl prior to use. Pentane was scrubbed twice with concentrated H2so 4 , twice with KMn0 4 in 10% H2so 4 , dried, and vacuum transferred from sodium benzophenone ketyl. Acetone was deoxygenated by three 0 freeze-pump-thaw cycles, then dried over 3 A molecular sieves. Other volatile liquids were deoxygenated by freeze- pump-thawing; high .boiling liquids were deoxygenated by pumping directly on the material for 30 min. Proton NMR spectra were obtained on an EM-390 Varian Associates NMR spectrometer, and IR spectra on a Beckman 4240. Gas chromatographic analyses were performed on a Varian Aerograph Series 1400 GC, using Porapak Q (8' x 1/8" column, 80/100 mesh commercial Porapak Q, 70°C) for c1 -c 4 for hydrocarbon species, Duropak (17' x 1/8", 100/120 mesh, 80°C) for c2 -c 6 species, and Carbowax (6 or 32' x 1/8", Carbowax 20M, 10% on Chromosorb W, AW-DMSC, 100/120 mesh, 80 or 100°C) for larger hydrocarbons or heteroatom-contain- -so- ing species. Mass spectra were obtained by the Caltech Analytical Laboratory; elemental analyses were performed by the same facility or by Schwarzkopf Microanalytical Laboratory, Woodside, New York. Preparation of Methylene-Bridged Titanocene Complex 1. 2 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Cp 2TiC1 2 (AlfaDivision, Ventron Corporation) was recrystallized from toltiene, using a soxhlet extractor, in air. To the clean Cp 2TiC1 2 so obtained (24.9 g, 0.1 mol) was added Me 3Al (110 mL of 2 N Me 3Al in toluene, 0.11 mmol) under an inert atmosphere. The reaction mixture was allowed to stir at ambient temperature for 68 h; as CH 4 is evolved, it was necessary to vent the reaction flask through a mercury bubbler. In the dry box, the reaction solution was filtered through a medium frit (this could also be accomplished using Schlenk techniques) and cooled to -40°C. A solid mass of crystals formed; these were broken into small fragments and collected by filtration. The 1 so obtained was not entirely pure, containing traces of 2 2 Cp Ti-CH ~Al(Me)(Cl)-Cl and was recrystallized from toluene/ pentane containing a few percent of Me 3Al by cooling to -40°C. The red crystals were collected by filtration, washed with pentane, and dried under vacuum. (25%) of !· Yield: lOoO g The mother liquors from the procedures above contained significant amounts- of 1, which could be recovered by removal of solvent and recrystallization. NMR ( c6n6): -51- o-0.25 (s, 6H, A1Me 2), 5.66 (s, lOH, . Cp's), 8.35 (s, 2H, methylene). Preparation of Titanacyclobutene 2a from 1 and Diphenyl- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~!l!~~~: 4 Diphenylacetylene (1.3 g, 7.3 mmol) was dissolved in 10 mL THF, and cooled to -40°C in the dry box freezer. The cold solution was added to 1 (2.0 g, 7.1 mmol), which had also been cooled, and the reaction mixture was stored at -40°C for 36 h. The mixture was warmed to room temperature, the solvent was removed under vacuum, and the resulting solids recrystallized from toluene/ pentane at -40°C. Yield: 2.2 g (84%) of dark red needles of 2a . NMR ( C6D6 ) : cS 3 . 4 9 ( s , 2H, met h y 1 en e) , 5 . 74 (s, lOH, Cp's), 6.8-7.4 (m, lOH, phenyl groups). 3~, 4 3~, 12 and 2e 12 were prepared from! and bis- trimethylsilylacetyl~ne, phenylmethylacetylene, and trimethylsilylphenylacetylene, respectively. 2b: NMR (C 6D6 ) o 0.15 (s, 9H, SiMe 3), 0.29 [s, 9H, the other SiMe 3 (no specific assignment for a. or 5, position for either SiMe 3 group); 4.6 (s, 2H, methylene), 5.17 (s, lOH, NMR (C 6D6 ) o 1.40 (s, 3H, methyl), 3.09 (s, 2H, methylene), 5.42 (s, lOH, Cp's), 6.7-7.2 (m, SH, phenyl). Cp's). 3~: NMR (C 6D6 ) o 0. 02 (s, 9H, SiMe 3), 4. 21 (s, 2H, methylene), 5.35 (s, lOH, Cp's), 6.8-7.3 (m, SH, phenyl). 2e: Preparation of Titanacyclobutene 2d from 1 and 2~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ !3!:1.!Y~~· ! (1 gm), 2-vinylpyridine ... styrene copolymer -52- (Aldrich Chemical Company, 1 gm), and 20 mL c6H6 were placed in a Schlenk flask. Under argon flow, 1 mL 2-butyne was added, and the reaction was allowed to stir for 24 h at ambient temperature. (The flask was kept under a slight positive pressure of argon and was open to a mercury bubbler, as the initially exothermic reaction made the use of a closed vessel impossible.) remaining 2-butyne were The solvent and any removed under vacuum, and the reaction was worked up in the dry box. The mixture was extracted with benzene (25 mL, 2 portions), the extracts were filtered, and the filtrates were taken to dryness under vacuum. The resulting red-black solids were recrystallized twice from pentane to give a poor yield of NMR (C 6D6 ): c 5.54 (lOH, s, Cp rings), 3.24 (2H, broad singlet, methylene), 2.16 (3H, ~~(approximately 10%). broad singlet, Me), 1.55 (3H, s, Me). Compound 1 - (2.2 g, 7.8 mmol) and 2-vinyl pyridine-styrene copolymer (from Aldrich Chemical Company, 2.2 g) were placed in a flask in the dry box. 3,3-Dimethyl-l-butene (2.2 mL, 18 mmol) and then benzene (25 mL) were added, and the mixture allowed to stir for 20 h at ambient temperature. The mixture was filtered through a coarse frit, and the solvent removed from the filtrate. The resulting red- black solids were taken up in pentane (25 mL), the solution -53- was filtered through a fine frit, and the filtrate cooled to -40°C to produce dark red needles of ~~' which were collected by filtration and dried under vacuum. g ( 5 7 %) of 2~ . Yield 1.2 ·NMR ( C6D6 ) : o 0 , 0 2 (m , J = 10 Hz , 1H, Hin S-position), 0.97 (s, 9H, !-butyl), 1.94 and 2.26 (non-first order d.of d, J = 10 Hz, total of 4 H~ H's in a-position), 5.43 and 5.64 (s, SH each, CP's). For all kinetic work, 3a was recrystallized -0ne more time, from pentane. Reaction of 1 With N-Benzylidenemethylamine. Compound ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ! (28 mg, 0.1 mmol), N-benzylidenemethylamine (12.3 µl, 0.1 mmol), and anisole (10.8 µl, 0.1 mmol), as internal standard) were placed in 330 µl of c6n6 in an NMR tube. The tube was sealed, and the reaction followed by NMR. All ~ had disappeared in 5 h, by which time the NMR resolution was very poor. could be observed. No new organometallic products The yield of methylene for _NMe exchange product, styrene, was determined by gas chromatography on a 6' Carbowax column (25% per Ti). 37% of the starting N-benzylidenementhylamine remained. An identical NMR tube to which pyridine was added (8.1 µl, 0.1 mmol, syringed in immediately after all other materials had been mixed) showed immediate disappearance of~' no new organometallic products, styrene (38% per Ti), and remaining N-benzylidenemethylamine (51%). -54- Reaction of 1 With Carbonyl in the Presence of ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~r:~~~~~· Compound ~ (56 mg, 0.2 mmol) was placed in toluene-d 8 (300 µl) in an NMR tube, and cooled to -78°C. An equimolar mixture of pyridine and trimethylacetaldehyde (35 µ1, 0.2 mmol of each) was added, and the tube was then sealed. Following the reaction at -60°C in the NMR probe, one observed 3,3-dimethyl-l-butene growing in, and then some ~~ forming. No other organometallic products were observed. Similar results were obtained when the reaction was allowed to proceed at higher temperatures (up to 0°C), except that slightly less ~~ was observed. In another experiment, pyridine (70 µl, 0.88 mmol) was added over 15 min to a stirred solution of ~ (125 mg, 0.44 mmol) in 1 mL of toluene at -78°C. The initially deep red solution had changed to an orange-red solution by the end of the addition, and a precipitate had formed. After stirring at -78°C for 15 min more, acetaldehyde (25 µl in 1 mL toluene, 0.44 mmol) was added slowly, and stirred for 60 min. Gas chromatographic analysis of the volatile components of the reaction mixture showed the presence of propene, indicating that a reactive titanocene carbene of some kind remained after addition of pyridine to 1, for at least 15 min at -78°C, No oxymetallacyclo~ -55- butane could be observed in or isolated from the reaction mixture, however. Reaction of 1 With CO. Compound ! (20 mg, 0.070 mmol) was placed in an NMR tube with 500 µl · C6n6 . The solution was freeze-pump-thawed to remove all gases dissolved in solution, then placed under approximately 1 atm of CO gas and sealed (approximately 0.15 mmol.CO). The reaction was followed by NMR, and Cp 2Ti(Me)Cl could be observed to grow in several hours; within several days, no organometallic products could be observed in the NMR (including Cp 2Ti(Me)Cl, which apparently had decomposed). The organic products in the gas phase and in solution were observed by gas chromatography on Porapak Q and Duropak, and included CH 4 , c 2H4 , c 2H6 , propene, isobutylene, and other c4 and higher hydrocarbons, but not ketene, allene, acetylene or neopentane. In other runs, the composition of the hydrocarbon mixtures formed in the reaction varied. From a larger scale reaction (1 g 1 in 25 mL c6H6 under 1 atm co in 250 mL flask), Cp 2Ti(Me)Cl was isolated after 20 h of reaction time by removing all solvent under vacuum. The remaining red solids were nearly pure CpzTi(Me)Cl, identified by comparison of NMR spectrum and m.p. to literature values. 11 -56- ~:~~!~~~-~~-~-~~!~-~~:~~~r!_~~~E~~~~:· Reactions of the various titanacyclobutenes with carbonyl compounds were all carried out in the same manner; the reaction of 2a with acetone on NMR tube and preparative scales are given as examples. 6a. ~~ (19 mg, 0.05 nunol) and acetone (4 µl, 0.05 mmol) were placed in 330 µl c6 n6 in an NMR tube. was capped, and the reaction followed by NMR. The tube Initially, the acetone resonance was very broad and the solution was deep red; within 48 h, however, the solution had turned yellow-orange, a small amount of remaining acetone showed a sharp NMR signal, and all 2a had disappeared. A quantitative yield of ~~ was observed. On a preparative scale, ~~ (1 gm, 2.7 mmol), and acetone (211 µl, 2.7 mmol) were dissolved in 25 mL c6H6 , and allowed to stir for two days. The solvent was removed under vacuum, and the orange solids so obtained recrystallized from toluene/pentane at 40°C. Yield: 0.87 gm (75%) of yellow-orange crystals of 6a. Calcd. C, 78.50%; Ti , 11 . 18 . Found : C , 7 8 . 91 ; H, 6 . 6 9 ; MW: Found: 436. 428; ion), 178 (titanocene). Mass spectrum: Ti , 11 . 18 . H, 6.59; Ca 1 c d . m/e 428 (parent NMR data is included in Table 1 (see text). 6b and 7. - A preparative scale reaction between 2a -57- (1 gm, 2. 7 mmol) and benzaldehyde (287 µl, 2. 7 mmol) yielded, upon recrystallization from toluene/pentane, small orange crystals and several large dark red crystals. These were separated physically. The orange crystals were recrystallized again from toluene/pentane to yield 0.70 gm (55%) of yellow-orange 6b. Calcd. Ti , 1 0 . 0 5 • H, 6 . 0 4 ; Found : C, 8 0 . 76 ; C, 80.67; H, 5.92; Ti , 1 0 . 0 0 . The dark red crystals were scraped free of small orange crystals of ~~' although the NMR spectrum of them still showed small amounts of 6a as contaminant. These were identified as isomer 8 on the basis of their NMR spectrum, color and HCl decomposition products, which have been included in the text in Table 3. Compound 6a (115 mg, 0.24 -- mmol) was placed in 3 mL Et 2o in a round-bottom flask containing a stir bar; cap. the flask was capped with a serum 25 mL of HCl gas (1 mmol) was injected via syringe, and the mixture was allowed to stir for 15 min at ambient temperature. The color changed from orange to purple to pale red, with the formation of large amounts of red precipitate. The mixture was worked up in air: the red solid was collected by filtration through a fine frit, and the filtrate taken tb dryness on a rotary evaporator. The red solid is cp TiC1 2 (m.p. 2 280~285°C), 102 mg (88%). -58- The white solid obtained from the filtrate is the expected organic product of HCl decomposition (see text for NMR, IR and mass spectral data, Table 2), 116 mg (98%). Calcd.: C, 85.67; H, 7.99. Found: C, 85.12; H, 8.12. Compound 6b was decomposed with HCl in an entirely analogous manner. Cp 2TiC1 2 , 80%; Yield: 100%. Calcd. for organic product: Found: C, 87.04; organic product, C, 87.96; H, 6.71. H, 7.24. Isomer 8 was decomposed in an analogous manner, except that the Et 2o solution of ~ was cooled to -78°C, treated with two equivalents of HCl, and allowed to warm to room temperature. After work-up in air, there was obtained Cp 2TiCI 2 (80%) and the expected organic product (80%). (See Table 3 for data on the organic product.) Reaction of 6a with CO. Compound 6a (20 mg, 0.047 ~~~~~~~~~~~~~~~~~~~~~~ mmol) was placed in 300 µl THF-d 8 in an NMR tube. The solution was freeze-pump-thawed to remove dissolved gases, placed under 1 atm of CO, and sealed. The reaction was followed by NMR, and was complete in two weeks at 70°C. The orange solution had become pale yellow, and a white precipitate had formed. No evidence for organometallic products could be observed, but an organic product was identified as the acid formed from CO insertion based on its NMR, IR and mass spectra (see Table 4 in text). The -59- source of H2o (or its equivalent) is unknown. Yield: 78% per Ti. The reaction was repeated in c6 n6 , but was so slow that no products could be observed in several weeks at 70°C. Reactions of 3a With Carbonyl Compounds. The reactions ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ of ~~ with various carbonyl compounds (see Table 5 in text) were all run in the same manner. Compound 3a (20 mg, 0.072 mmol) was placed in an NMR tube with 300 µl c6 n6 and anisole or dimethoxyethane as internal standard. 6ne equivalent of carbonyl compound was added via syringe, and the tube was immediately cooled in liquid nitrogen and sealed. Reactions were allowed to proceed at room temperature, with monitoring by NMR. When all 3a had disappeared, the yield of product was measured by NMR or by gas chromatography (Duropak or Carbowax columns). Products of methylene for oxygen exchange were identified by comparison to authentic samples. 3,3-Dimethyl-l-butene was also produced in quantitative yields (by NMR). No organometallic products were visible in the NMR, but a yellow precipitate formed, which matched the literature description of titanocene oxide polymer. 17 Reactions with aldehydes did not proceed to completion with the addition of only one equivalent of carbonyl substrate; two equivalents were required. One equivalent -60- of aldehyde presumably is taken up by the titanocene oxide, as it can be accounted for in no other way. The rate of disappearance of 3a in the presence of aldehyde is, however, first order (vide infra). An entirely analogous approach was used in the experiments with N-benzylidenemethylamine. Kinetic Data. General. obtained in the following way: All kinetic data were 3a was placed in an NMR -tube, which had been marked to indicate a volume of 300 µl. One of two methods was used: either 3a was weighed directly into the tube in the dry box (judged to be accurate to± 1 mg), or a stock solution of known concentration of 3a in solvent (C 6D6 ) was syringed into the tube. There was no particular advantage to either method, except that a stock solution of 3a in c6n6 could not, of course, be employed in cases where the reaction was to be run in neat acetone-d 6 . (Stock solutions of acetone-d 6 were not employed because the solvent reacts with 3a.) Anisole or dimethoxyethane was used as internal standard (chosen so that their NMR resonances would not interfere with the observation of any other peaks), and was either syringed into each individual tube or included in the stock solution. Any other "inert" materials, such as added alkene 13, were added via syringe to the tube. All tubes -61- were capped with septa when not being manipulated. Finally, carbonyl substrate was added to each tube (measured via syringe), the total vol1.lllle was brought up to 300 µl with c6n6 , and the tube was closed, removed from the dry box, cooled in liquid nitrogen, and sealed. Care was taken to work quickly once any reactive agent had been added to the tube, and all tubes were frozen in liquid nitrogen within 5 min of any such addition. Tubes were stored at or below -55°C until kinetics could be measured on them. For each set of kinetic data, a t=O spectrum was measured upon first thawing the tubes, and tubes were then piaced in a constant temperature bath, sealed end down. The amount of 3a present. at various points in the time was -- measured by NMR, integrating one of the Cp resonances of 3a and the internal standard. The NMR probe was maintained at 34°C, and samp1es were centrifuged at ambient temperature prior to "insertion in the NMR probe, but the total amount of time necessary to accomplish the centrifugation and NMR measurement was less than 4 min, and this was judged to be a negligible amount of time for a sample to be out of the temperature bath. Since solid titanocene oxide polymer precipitates out of the reaction in most cases, it was necessary to allow the reactions to proceed in the ·62- sealed end of the NMR tube and to centrifuge each sample prior to NMR measurement, so that solids would not interfere with the resolution of the NMR spectrum. The data were worked up as follows: Let A = the area of the Cp peak of 3a at time t A 0 = the area at time t=O s = the area of the internal standard at time t so = the area of the internal standard in the t=O spectrum A plot was made of ~ln(AS 0 /A 0 S) versus time t. The slope of this line is the first-order rate constant kobs· Slopes were determined using a linear regression program on a Hewlett-Packard 25 Pocket Calculator, and were checked by graphing the results to see that the calculated results were reasonable and free of obvious errors (i.e., obviously spurious points). Particular procedures for each kinetics experiment are discussed below. Determination of Reaction Order. The first kinetics experiments were plotted using both first and second order treatments, for a variety of substrates. The first order calculation has been described above, and the second -63 .. order calculation is: Let A, A0 , S, and 5 0 be as previously defined. B = area of substrate at time t. B0 = area of substrate at time t=O The area of the substrate B was measured directly if possible; otherwise it was calculated, using the amount of ~~ remaining and the known stoichiometry of the reaction. Then a plot of -ln(AB 0 /A 0 B) versus time yields a line with slope CI3a] - [substrate])k where k is the second order rate constant. (When an aldehyde is the substrate, one must use the value for equivalents of aldehyde rather than the concentration. When equal amounts of 3a and substrate are present, one must plot (A 0 -A)/(A A) versus time to obtain a slope equal to the 0 rate constant k.) While the first order plots yielded good straight lines and reasonably consistent values for the rate constant, the second order plots gave slightly curving lines and rate "constants" that varied by a factor of 4 or more. The reaction order was concluded to be first order in ~~' zero order in substrate, ~:!:~~~~~!~~~-~~-~~~:~:-~:~=~~=~::· The dependence of the rate of disappearance of 3a on the concentration -64- of alkene 13 was determined as follows. At least 5 equivalents of 13 and methyl benzoate as carbonyl substrate were placed in each NMR tube with 3a; the high concentrations assured that the concentration of 13 and MB would remain approximately constant over the first half-life of the reaction. The amounts and relative ratios of 13 and MB varied from sample to sample, however (from 1.21 to 3.63 M, and ratios of 2.0 to 0.5). The disappearance of 3a was ·followed, and plotted using first order kinetics. -- The data so obtained, and the treatment of them, are described in Table 7 and in the text. Determination of k by varying [acetone]. ~~~~~~~~~~~~~~~~~~!~~~~~~~~~~~~~~~~~~~~~ At least 10 equivalents of acetone were placed in NMR tubes with 3a; the high conce~trations assured an approximately constant concentration of acetone over the first half-life of the reaction. The acetone concentration varied from 2.42 to 13.7 Min different tubes. The disappearance of 3a was monitored, and a first order rate constant kobs o.btained for each sample. The data are presented in Table 8 and a treatment of them discussed in the text. A similar experiment was done using large amounts of acetone and added alkene 13. Within experimental error, no effect of added alkene could be observed with acetone as the carbonyl substrate. -65- Dependence of k on Temperature. ~~~~~~~~~~~~~~~obs~~~~~~~~~~~~~~~ It had been determined by the acetone concentration dependence experiments that kobs in neat acetone-d 6 is within 3% of k 1 (see text). It was therefore a simple matter to observe the rate of disappearance of ~~ in neat acetone-d 6 at various temperatures (from 4.8 to 51.0°C) and thereby determine Ea and ~sf for the reaction. The results of these experiments are contained in Table 9, and discussed in the text. Competition Experiments. Mixtures were prepared ~~~~~~~~~~~~~~~~~~~~~~~ which were equimolar in each of two competing substrates (acetone and methyl benzoate, acetone and acetophenone, methyl benzoate and ethyl acetate), and these were added to 3a in large enough amounts that the concentration of the two substrates would not vary over the course of the reaction. For example, acetone (580 mg) and aceto- phenone (1200 mg) were mixed together. 300 µl of this solution (approximately 1.5 mmol of each substrate) was added to 3a (20 mg, 0.072 mmol) in an NMR tube. The tube was sealed, and when the reaction was complete (3 days) the tube was broken open and the product ratios determined by gas chromatography. To determine the effect of added alkene, 300 µl of the acetone/acetophenone mixture was added to a tube containing 3a (20 mg) and 13 (150 µl, -66- 1.2 mmol), and the products again analyzed by gas chromatography. Competition experiments with ~were performed by adding an equimolar mixture of pyridine and two competing substrates to 1. For example, pyridine (316 mg) and the acetone/acetophenone mixture described above (710 mg) were mixed, cooled to -40°C, and the mixture then added to an NMR tube at -40°C containing 1 (20 mg, 0.07 mmol). - The reaction tube was immediately cooled in liquid nitrogen, and sealed. Upon warming to room temperature, reaction was complete. Product ratios were determined by gas chromatography. Photochemical Reaction of 3a with Acetone. 3a (20 mg) was placed in 300 µl acetone-d 6 in a Pyrex NMR tube, with dimethoxyethane as internal standard. The tube was irradiated with an unmodified UV light (Hanovia 579A36 High Pressure Hg Light Source) at -20°C, and the first order rate of disappearance of 3a was followed by NMR and determined to be 3.1 x 10-S sec-l The calculated rate of the thermal reaction at this temperature is 3.5 x 10-S sec-l At higher temperatures, the thermal reaction began to be competitive with the photochemical one, so that the overall rate of disappearance of 3a with or -- without irradiation at ambient temperatures was nearly the same. -67- ·Pro.gr. in Inorg. Chem. 1979, _..,,,..,....., ....... (1) (a) Grubbs, R. H. (b) Calderon, N• ; Lawrence, J. p. ; ~' 1. Of stead, E. A. Adv. Organomet. Chem. ~Ez~, _, 17 449. (2) Tebbe, F. N. ; Parshall, G. w. ; Reddy, G. s. J. Am. Chem. Soc. 1978, 100, 3611. (3) Tebbe, F. N.; Parshall, G. W.; Ovenall, D. W.. J. Am. Chem. Soc. 1979, 101, 5074. __,,,,,.,,,,,...,""'111 -- (4) Tebbe, F. N.; Harlow, R. L. Submitted for publication. (5) Howard, T. R.; Lee, J.B.; Grubbs, R. H. Submitted for publication. (6) Wengrovius, J. H.; Missert, J. R.; Schrock, R.R.; Youngs, W. J. Churchill, M. R.; J. Am. Chem. Soc. 1980, 102, 4515. (7) Pine, S. H.; Zahler, R.; Evans, D. A.; Grubbs, R.H. J. Am. Chem. Soc. 1980, 102, 3270 . .............. """"' ....... (8) Pine, S. H.; -- Howard, T. R. Unpublished results. This material is probably formed from titanocene oxide or some other titanocene product and Me 2A1Cl. Aluminum oxides are also formed. (9) Tebbe, F. N. Private communication. (10) No neopentane was observed in gas chromatographic analyses, but this is undoubtedly due to variations in reaction conditions, length of reaction, etc., to -68- which the system is quite sensitive. (11) Identified by comparison of NMR spectrum and melting point to published data: G. A. Waters, J. A.; J. Organomet, Chem. 1970, 22, 417. ...,"""'"""' - Mortimer, The material also showed a positive Cl test (green flame). (12) Howard, T. R. Unpublished results. (13) Compound 3d was prepared from ~ and 2-butyne in the presence of 2-vinylpyridine-styrene copolymer, as described for the preparation of the metallacyclobutanes in Reference 5. Complete data are included in the Experimental Section. (14) They can even be chromatographed on alumina on a bench-top. They appear to be slightly unstable in benzene solution in air, but are not at all sensitive in diethyl ether or CC1 4 solutions. (15) Nesmeyanov, A.H.; A.; Nogina, O.; Shatalov, G. V. Berlin, A.; Girshovich, Bull. Acad. Sci. USSR, Div. Chem. Sci. 1961, 2008. """""""" (16) Brown-Wensley, K.; Howard, T. R. Unpublished results. With 3,3-dimethyl-l-butyne or trimethylsilylacetylene, 1 and pyridine, NMR evidence suggests the formation of metallacyclobutenes. Inorg. Chem. 1964, ,,....,,,..,,...,,,..,, -3, 684. (18) A crystal structure of 3a is in progress: Grubbs, (17) Giddings, S. A. -69- R.H.; Schaeffer, W. A,; Howard, T. R. (19) The data for several different substrates at several different concentrations were plotted using both first and second order treatments. A much better fit was obtained for the first order plots; in addition, the second order rate constant so obtained varied by huge amounts. (20) Titanocene oxide polymer precipitates out of most solvents and appears as a broad lump at o 6.0 to 7.0 in the NMR when it does remain in solution. When aldehyde is the substrate, one equivalent of aldehyde appears as alkene product and the other is evidently incorporated into the insoluble organometallic polymer, as it appears nowhere else. greatly, ketones being several orders of magnitude (21) Rates of reaction of ketones and esters with 1 vary faster: Pine, S. H. Unpublished results. (22) This assumes (a) that pyridine will react with 1 faster than acetone or methyl benzoate will, and (b) that the pyridine·A1Me 2Cl complex is entirely removed from the coordination sphere of the titanium, thus generating 12. (23) Since k 3 and k 4 will be different for different substrates, we have designated k 3 and k 4 to represent -70- the rates of reaction of !! and ~~' respectively, with methyl benzoate, and k 3 i and k 4 , to refer to the analogous reactions with acetone. (24) Note that this value is only 3% different than the kobs obtained in neat acetone-d 6 . (25) Lee, J. B.; Brown-Wensley, K. A. (26) (a) Harrigan, R. W.; Unpublished results. Hammond, G. S.; J. Organomet. Chem. 1974, 81, 79. Gray, H. G. (b) Rausch, M. D.; ~~.....,....., Boon, W. H.; Tsai, Z.; Alt, H. G. Ibid. 1977, 141, - 299. (c) Brubaker, Jr., c. H. Ibid. !~z~, 166, 199. (d) Samuel, E.; ...,,,._...,,~ Maillard, P.; Giannotti, C. Ibid. 1977' 142' 289. ,....,......,......,#ftJ -(27) Morton, D. R.; Turro, N. J. Adv. Photochem. 1974, 2_, 197. (28) Perkins, D. C. ·1.; Tipper, C. F. H. (29) Rappe, A. K.; Puddephatt, R. J.; Rendle, M. C.; J. Organomet. Chem. 1980, 195, 105. Goddard, III, W. A. J. Am. Chem. Soc. 1980' -102' 5114 . .....,....,,,.....,~ (30) Rappe, A. K. Priv~te (31) Gordon, A. J. D.; panion"; communication. Ford, R. A. Wiley-Interscience: "The Chemist '·s ComNew York, 1972, pp. 113, and references therein. (32) Huff, J. C.; K. G. Moloy, K. E.; Marsella, J. A.; J. Am. Chem. Soc. 1980, 102, 3009. Caulton, -71- APPENDIX I Derivation of Kinetic Expressions -72· A. For Scheme V, Case 1 (see text): kl 3a k2 11 k_l 12 k_2 + 13 i k4 [CJ products C = carbonyl substrate -d{3a] -- dt = kl[~~] - k_1[!~) (1) Using a steady-state approximation, d [11] -- dt = o = k 1 [~~J - k .. 1 [!!] - k 2 I~!J + (2) k_ 2 I~~J I~~J [11] -- d[l2] -- dt = k 1 [~~] + k_ 2 I!~JI!~] k_1 = + (3) kz o = k 2 [~!] - k_ 2 I!~l I!. ~J - (4) k 4 [!~][C] [12] = Substituting for one obtains: (5) I!~] in Equation (3) and simplifying, -73- = Substituting for (6) [~!] in ·Equation (1) and simplifying gives: kzk4[C] = } { k:_ (k_ [~~J + k [CJ) + k k [cJ 2 4 2 4 1 (7) If [!~] and [C] are large enough relative to [~~] that they do not vary with time, then: kobs = kl k2k4(C] ~ k,.1 (k~z [~~J + k 4 [C]) + kzk4(CU (8) or kl k . obs B. = (1 k;21) + + {k-lk-2} -(13] -[C] kzk4 For Scheme V, Case 2, including the equilibrium between 12 and 13 (see text): (9) -74- kl 3a kz 11 k_l 12 + 13 k_z 1 k 3 IC] products C = carbonyl substrate -d [ 3a] -- dt = (1) Using a steady-state approximation, d[ll] -- dt = 0 = kl[~~] · k. 1 I~~J - k 2 I!!J + (2) k. 2 r~~JI!~J - k 3 I!!J [11] -- d[l2] = -- = (12] = dt -- kl [ ~~] k_ 1 + + k k_z [~~J [~~] 2 + k 3 [C] 0 = kz[~!J - k_2[!~J [!~J kz[~~] k_z[~~J Substituting for [ 12] in Equation (3) and simplifying, one obtains: (3) (4) (5) -75- [11] = (6) Note that this is the same expression one would obtain if one ignored the equilibriwn between !! and !~, and derived an expression for 11. Substituting for [!!] based on a steady-state of [!!] in Equation (1) and simplifying gives: -d [ 3a] = dt kl[~~] k 3 [CJ ( ) ( 7) k .. 1 + k 3 [C] If [CJ is large enough relative to f 3a] so that it does not vary with time, then k obs = 1 = k k 3 [C] 1 k_ 1 + k 3 [C] (8) or kobs C. 1 kl + 1 ( kkl_k 3) ( For Scheme V, Case 3 (see text): [~]) (9) -76- kl 3a 11 k_l kz 12 + 13 k_z J k 3 [C] J k 4 fC] products products C = carbonyl substrate -d [ 3a] -- dt = kl [ ~~] - k_l[~!] (1) Using a steady-state approximation for 11 and simplifying gives: [11] -- (2) = Using a steady-state approximation for 12 and simplifying gives: = (3) Substituting this value back into Equation (2) gives: Ill] = (k_z[~~] + k4[C])(k_l + k3[C]) +kzk4[C] (4) -77- Substituting for [~~] in Equation (10 and simplifying gives: = (5) If [13] and {C] are large enough relative to 3a that they -- do not vary with time, then: k obs (6) OT 1) -1 = + (7) For substrates such as acetone, for which k 4 ICJ>>k_ 2 [~~], this simplifies to: 1) -1 D. = + For Scheme V, Case 4 (see text): [CJ (8) -78- 3a 12 ! + 13 k 4 [C] products -d I 3aJ -- dt = (1) Using a steady-state approximation for 12 and simplifying gives: = (2) Substituting this expression back -into Equation (1) and simplifying yields: -d [ ~~] (3) dt If [13] and [C] are large enough relative to 3a so that they do not vary with time, then: = or (4) -79- 1 k obs = + (5) -80- CHAPTER II: SYNTHESIS OF AND REDUCTIVE ELIMINATION FROM PLATINUM-VINYL COMPOUNDS -81- INTRODUCTION The examination of reductive elimination from transition metal complexes (Reaction 1) has received a great deal ....R M --+ M + (1) R--R' 'R' of attention. Studies have been done with a variety of organic frag- ments R and R'; for example, the systems described in Table I are representative of the work on platinum systems alone (which are admittedly studied more extensively than many other metal systems, although work Table I Reductive Elimination from Pt Systems R' R Me Me n-Bu n-Bu Me H aryl aryl Me CH 3C 0 metallacycles Reference # (metal) l(Pt IV), 2(Pt IV) 3(Pt I I) 4(Pt I I) 5( Pt II) 6(Pt IV) 7(Pt II), 8(Pt II), 9(Pt IV) has a1so been done with Zr, lO ' 11 Ta, 12 Mo, 13 Os, 14 Co, 15 Ni 16 - 18 and Au 19 , 20 ). In spite of this tremendous amount of research, reductive elimination involving vinyl ligands has not been examined, and, in fact, relatively few vinyl compounds are known. Some of the best-characterized vinyl compounds are those of platinum, L2Pt(Cl+==CH 2 )X, L = PMe 20 or PEt 20, X =Cl, Br, I. 21 There is also evidence of reductive elimination -82- of propene from (PEt 2 ~) 2 PtMe(CH=CH 2 )I 2 . 22 Furthermore, as is obvious from Table I, reductive elimination from a great many platinum II and IV systems is already well studied. Therefore, it seemed reasonable to undertake a study of the reductive elimination reactions of platinum II and IV vinyl compounds. -83- RESULTS AND DISCUSSION Syntheses of Platinum-Vinyl Compounds Syntheses of L2Pt(CH=CH 2)x ~ have been reported, and one can make · L2Pt(CH:.CH 2)Me(I)X ~ via the oxidative addition of Mel to the Pt(II) compounds. The synthesis of 2 takes approximately four weeks at room temperature, however, and cannot be greatly accelerated by heating without the occurrence of reductive elimination to fonn L2Pt(I)x. 22 Platinum chlorides are substituted easily with Meli or MeMgBr in a number of. cases, so that one might expect to be able to make L2Pt(CH=CH 2 )Me. 23 , 24 Oxidative addition to Pt(.II) dialkyls appears to proceed much more rapidly than to Pt(II) alkyl halides. 24 , 25 With these facts in mind, the following synthetic scheme was planned: Cl L 'Pt/ L/ 'CH=CH 2 la, Meli L=PMe 20 L2Pt(CH=CH 2)Me Mel j L2Pt(CH=CH 2)Me 2I Scheme I -3 4 - -84- Treatment of !~ with Meli at -78°C does, indeed,afford ~ in 87% isolated yield. IR and 1H NMR spectra of 3 are shown in Figure I, and the 1H and 31 P NMR data are listed in Table II. The fact that two different phosphine resonances (Jp_p = 14 Hz) are observed by 31 P NMR identifies ~ as the cis isomer, a structure which is also consistent with the coupling constants JPt-P and JPt-H" The NMR assignments are based on the assumption that the chemical shift and coupling constants of the 2 ligand trans to -CH 3 will be more similar to the ligands in cis~ (PMe 2 ~) 2 PtMe 2 ~ than the ligand trans to -CH=CH 2. The 31 P NMR spectral PMe ~ data for 5 are included in Table II for reference. The IR spectrum of 3 shows a C=C stretch at 1557 cm- 1 . This may be compared to a C=C ..., stretch at 1564 cm- 1 in la. 22 Care must be taken in the synthesis of 3 to avoid the fonnation of 5. 26 An excess of Meli, fast addition rates or a higher reaction temperature will result in a final product containing up to 20% 5. Use of MeMgI (instead of Meli) ·or excess ~~ will not prevent the formation of ~; the alternate synthetic approach in Reaction (2) is unsuccessful as no reaction at all occurs. Furthermore, ? 3 (2) separation of 3 and 5 is difficult by recrystallization or column chromatography. The presence of 5 is not evident on the basis of melting point (~ containing 20% ~melts at 70.5 - 71.5°C, and the expected lowered broadened melting point is only observed in nearly 50:50 mixtures) nor is it easily distinguishable by 1H NMR. 31 p NMR, however, readily distinguishes the two, and the IR spectra between 1150 and 1250 cm -1 also ·-85- Figure I IR and 1 H NMR Spectra for 3 IR (KBr) a 4 3.5 s 6 7 8 6 s 9 1s 10 2 0 ( c) 8 c. 7 Expansion (5x) of region between 61.0 and 1.5. 3 2 l 0 -86- Table II NMR Data 1H Spectrum in c o 6 6 assignment: a: 3 b: c: d: e: f: g: o 1.30(3H),multa,b,c 8.40(1H),mult 5.82(1H),mult 6. 61 ( 1H) , mu 1t d 1. 33(6H) ,do.ubb 1.17(6H) ,doubb 7. 03(6H) , mu 1t 7.41(4H) ,rnult ~P-H(Hz) ~Pt-H(Hz) 9.5 69 7.1 7.8 7.7 21.0 18.6 a,e,f: 1.15 to 1.8 ( 15H) ,mul te b: 8.06(1H),mult, Jbc = 12 Hz c: 5.84(1H),mult, Jed= 2 to 3 Hz d: 7.36(1H),rnult, Jbd = 9 Hz, JP-Hd = 18 Hz 4 g: h: 7.15 (6H) ,rnul t 7.21(4H),rnult 0.84(3H),tripb 7.5 67 31 P Spectrum in c o , Proton Decoupled 66 Compound 2 L=PMe ~ 5 -12.1(1824) -87- Table II (continued) a L, Lb/ 31 p Spectrum in c o , Proton Decoupled 6 6 Compound Pt /CH=CH 2 'cH -13 . 6 (1821). JP- P = 14 Hz -14. 4 ( 1724) 3 a: b: 6 -47. 0(1164) 4 a: 3 b: -47.4(1174) -47.4(1152) a. Non-first-order four line pattern. b. Multiplet due to phosphorous couplings is further split into a doublet due to coupling to 195 Pt, which is present in approximately 30% abundance. · c. Only the resonances coupled to Pt are clearly visible; the central, non-Pt-coupled resonance is obscured by other peaks. The position of the central resonance is inferred from the positions of the Ptcoupled peaks. d. Tentative assignment. e. The spectrum is too complex to permit definite assignments for each methyl group. -88- allow one to see relatively small amounts of 5 in 3 (Figure II). Dissolution of 3 in Mel followed by standing for approximately one week at room temperature or 20 hours at 400 C results in oxidative addition of Mel to the Pt(II) compound. IR and 1H NMR spectra for 4 are shown in Figure III, and the 1H and 31 P NMR data are listed in Table II. The 31 P NMR spectrum shows coincidentally identical chemical shifts for the two PMe 20 ligands, but different JPt-P's, indicating inequivalent phosphorus atoms. Both coupling constants (JPt-P = 1152, 1174 Hz) are too small to indicate an iodine trans to a phosphorus (such coupling constants are generally in the range of 3500 to 4500 Hz 22 , 27 , 28 ). Thus, the structure of 4 .is as shown in Table II. The 1H NMR spectrum is too com- plex to permit the assignment of all resonances and coupling constants, but JPt-H for -Ctf~ is. consistent . with the assignment. The IR spectrum shows a C=C stretch at 1578 cm-1 . Attempts to generate 4 or another isomer, possibly 7a, or a CH=CH 2 L,. I ~CH 3 ~''Pt''' L~I ~R x L = PMe 0 2 7a: 7b: R = Me R = CH=CH 2 Pt(lV)divinyl complex (such as ~~) via oxidative addition of vinyl halide to Pt(II) (~ or ~) were unsuccessful. Reaction of ~ or ~ with BrCH=CH 2 is much slower than the reaction with Mel (in reaction times comparable to those used to make 4, large amounts of unchanged starting material However, even in low percentage conversion reactions, no Pt(IV) product is observed (by 31 p NMR), and the species observed remained). -89l 4 0 0 Figure II a. L2PtMe2 5 L=PMe 2 ¢ NMR(expanded scale) b. 80% 3 IR l 2 0 0 l 0 0 0 -90Figure III: 3 a. 9 IR and s . 35 1 H NMR Spectra for 4 7 6 9 8 15 1 0 20 IR (KBr) 8 7 6 5 4 3 2 l 0 c. ):xpansion of region between ol.O and 2.0 Ln -91- generally correspond to the expected Pt(II) products of reductive elimination from ...,7. IR spectra show no evidence of new vinyl C=C stretches . Gas chromatographic analysis of the organic products of thermal decomposition of what was expected to be 7 (after work-up) 29 showed small ..., amounts of the products expected from reductive elimination of 7, but much larger amounts of the products expected from thennal decomposition of Pt(II) species (see Table IV)~ 9 The failure to observe any Pt(IV) species with a vinyl group trans to halogen, in addition to the observation of the expected products of reductive elimination from such a species, suggests that the rate of decomposition (reductive elimination or other decomposition pathways) of 7 (or similar molecules) is equal to or greater than the rate of its ..., formation. Other possibilities (secondary reactions or variations in the mechanism of oxidative addition of vfoyl halides) have not, however, been ruled out. 30 Thermal Decomposition of Platinum Compounds - Thennal decomposition of 3 and 4 was accomplished by heating solid ..., samples and/or benzene solutions of the compounds. Gases evolved were analyzed by gas chromatography, and some organometallic products were identified f>y NMR. Results of the studies are summarized in Table III. Thennal decompositions of 5 and 6 were also examined briefly for comparative purposes. Decomposition of 3 produces CH 4 and CH 2=CH2 in a ratio of about 1:2, as well as small amounts of the reductive elimination product propene. The major products may arise from a Pt(lV) hydride formed by oxidative 0 52.7 45.7 ·2. 4 97.2 32. 4 61. 2 1. 3 0.3 0.4 0 Melt (140°C) Melt (155°C) Soln (90°C) Melt (100°C) Melt (160°C) So 1n (160° C) e 3 4 5 6d 0 0 2 88 97 1 0 0 0 5.5 trans-L 2PtMel 31p NMRb trans-L 2PtMel 1H NMRc Puddephatt et al. report c2H6 as the sole .organic product. e. Some CH 4 may be forming from decomposition of L2PtMel. d. See Ref. 1. c. 1H NMR can be used to identify relatively pure materials, but spectra of complex mixtures cannot be used to reliably identify products. · a. No evfdence- of reaction of Pt(.11) species with pro-pene- wa-s oota-ined. b. 31 P NMR spectra were taken when appropriate and when instrumentation was available. ·12 3 96 -- CH 4 C2H4 C2H6 propenea - Products Gas Chromatographic Analysis (%) Decomposition Method Comeound Thermal Decomposition of Platinum Compounds Table III I I \.Q I\) Products 33 39 + BrCH=CH 2 ~ 0 16 4.5 9.5 24 0 2 trans-L 2PtMeBr (19%) trans-L 2Pt(CH=CH 2 )Br (14%) (28%) -§3 (40%)a cis-L 2PtMe2.(92%) trans-L 2PtMeBr (5%) trans-L 2Pt(CH=CH 2 )Br (3%) 31 P NMR, L = PMe ~ a. The source of this product, which is not present in the starting material-l, -is unclear. It suggests that simple oxi'dative addition/reductive elimination""'is not· sufficient to explain these reactions. 1.4 73 + BrCH=CH 2 CH 4 c2H4 c2H6 CH 2=CHCH 3 butadiene - -- -- Gas Charomatographic Analysis after work-up (see Note 29) as %of total organic products. ~ Reactants Products from the Reaction of Pt(II) and Vinyl Halide Table IV I I w \0 -94- addition of Pt(II) into a C-H bond (i.e., ortho-metallation of a phenyl group on PMe 2 ~ or perhaps insertion into a P~H 2 --H bond), which then eliminates reductively. 7 No acetylene, which might have been produced via a-abstraction from the vinyl group, was observed. 31 The solution decomposition of 4 is first order for disappearance of 4 (followed by 1H NMR), and indicates that the rate of reductive elimination.of vinyl is 20 times faster than that of methyl (after correction for the statistical consideration that there are two ways to make propene and only one way to make ethane). t 1 for the reaction at 83.5°C is 14 ~ min. Melt decomposition of 4 products approximately equal amounts of ethane and propene. Puddephatt1 also obtained different results from solution and melt decompositions, although his results would have led on~ to expect even more propene than is produced in the solution decomposition. He suggests that the two groups trans to PMe 2 ~ leave, and that incomplete scrambling of these two positions and the .position trans to halogen prior to reductive elimination occurs in melt decomposition. (Complete scrambling occurs in solution decompositions.) If scrambling is incom- plete in melt decompositions, then the observation of even greater amounts of ethane suggests that one of the leaving groups is trans to halogen. The lability of a vinyl group trans to halogen may also explain the failure to observe any products of the type 7a or b in additions of vinyl halide to Pt(II). Small amounts of CH 4 are produced in all decompositions of Pt(IV) compounds (the large amount of CH 4 in the solution decomposition of~ may be due to a reaction of L2PtMel). Whether the source of the CH 4 is Pt(II) -95- or Pt(IV) in all cases is unclear, but it certainly constitutes a minor pathway. The materials produced in attempted syntheses of 7a and b showed no evidence by 31 P NMR for Pt(IV) species, and yet subsequent decomposition -- of these materials showed what appears on the surface to be incorporation of a vinyl group . Considering that small amounts of Pt(IV) species would not have been observed by NMR and that studies of the decomposition modes of L2Pt(R)X (R = Me, CH=CH 2 ) have not been done, this should not be conObservation of 3 or 4 in neat vinyl halide at varying temperatures by 31 p NMR (.and subsequent observation of the orsidered inconsistent. ganics produced by gas chromatography, before work-up) would be certainly . f . 29 1n onnat1ve. -96- EXPERIMENTAL General Spectro-grade benzene, and reagent-~rade toluene, THF and diethyl ether were vacuum-transferred from sodium benzophenone ketyl prior to use. Petroleum ether (bp. 35 - 60°C) was washed twice with H2so4 , twice with saturated KMn0 4 in 10% H2so 4 , and dried and distilled from sodium K2PtC1 4 , PMe 20, (n-Bu) 3Sn(CH=CH 2 ) and Meli were used Me! and BrCH::CH 2 were degassed (3 freeze-pump-thaw cycles) benzophenone ketyl. as supplied. prior to use. Proton NMR spectra were recorded ·on a Varion Associates EM-309 NMR spectrometer. 31 P NMR spectra were recorded on a Bruker WH-90, IR spectra on a Beckman 4240, and UV spectra on a Cary 14. Gas chromatographic analyses were performed on a Varian Aerograph Series 1400 GC, using Porapak Q (8' x l/8 column, 80/100 mesh corranercial Porapak Q, 70°C) for 11 c1 - c4 species and Duropak (20' x 1/8 11 , 100/120 mesh, 80°C) for c2 - c6 species. Manipulations requiring an inert atmosphere were perfonned in a Vacuum Atmospheres Recirculating Dry Box or under purified argon using standard Schlenk Techniques. Syntheses The synthesis of (PMe 20) 2PtC1 2 has been described. 32 The synthesis of la from this has also been described, 22 and we have discovered that the use of (n-Bu) 3Sn(CH=CH 2 ) instead of the described Me 3Sn(CH=CH 2 ) -97- results in improved yields (up to 52%). (PMe 20) 2Pt(CH=CH 2 )Me, ~: ~ was made by adding Meli (1,56 M in diethylether, 1.1 ml, 1.7 JTJTlol) dropwise over a period of 45 min. to a solution of la in THF (0.92 g, 1.7 tmlOl in 150 ml) in a 250 ml roundbottom fl ask stirred magnetically, at -78° C under an inert atmosphere. The reaction mixture was stirred for an additional 2 hrs. while it warmed to -50°C. Meli. At -SOOC, 3 ml MeOH was added to destroy any remaining Work-up was continued at ambient temperature, in air: the sol- vents were removed with the use of a rotary evaporator, water was added to the solids, and the suspension that fanned was extracted with three portions of ether. The combined ether extracts were dried over MgS0 4 (anhydrous), the solvent was removed, and the resulting solids recrystallized from toluene/petroleum ether at -40°C in the dry box. 0.77 g 3 (87% yield), mp. 70.5 - 71.5°C. Cale: Obtained: C 44.44, H 5.5%, found: c 44.35, H 5.44%. UV(MeOH): shoulder at 310 nm on a peak which continues to increase until the solvent cut-off. (PMe 0) Pt(CH=CH 2)Me 2I, ~: ~was prepared by dissolving ~ (300 Mg, 2 2 0.59 ITITiol) in Mel (3 ml) under an inert atmosphere, in a vial which was then sealed with a Teflon-lined screw-on cap. The reaction mixture was then allowed to stand in the dark at room temperature (outside of the dry box) or at 400C. The course of the reaction was monitored by TLC, and was judged to be complete in one week at room temperatur~ or 20 hrs. at· 40°C. Mel was then removed with the use of a rotary evaporator, and the solid recrystallized from toluene/petroleum ether under an inert at~ mosphere at -40°C. Pale yellow material, mp. with decomposition and -98- bubbling 127.5 -129.5°C, was obtained (363 mg, 95% yield). 36.65, H 4.77%, found: 7a: Cale: C C 36.25, H 4.56%. Synthesis of 7a was attempted in a manner analogous to that used to prepare ~' using BrCH=CH 2 instead of Mel. Samples were examined by 31 P NMR after removal of BrCH=CH 2 , with and without recrystallization. The reaction was also allowed to proceed in one attempt at -40°C for one week, after which the BrCH=CH 2 was removed at -40°C and the resulting solids examined by IR at ambient temperature. No incorporation of a vinyl group {judging by the absence of any C==C stretch) could be observed. ~~: The attempted syntheses of 7b proceeded in a manner analogous to that used for synthesis of 4. - Decompositions Melt decompositions were accomplished in sealed tubes under N2. Solutions for decomposition were prepared under an inert atmosphere in c6o6 (0.05 to 0.1 M), in sealed tubes (which were broken before sampling for GC analysis) or in flasks having a straight-bore stopcock capped with a serum cap through which gases and liquids could be sampled via syringe. GC analyses were calibrated with absolute response factors for the gases and/or propane as internal standard. In the solution decompositions, gas and liquid phase were sampled and the total amount of gases fanned is reported. In all cases except those in Table IV), gases fanned accounted for a11 the Pt decomposed (± 10%). -99- References and Notes 1. M. P. Brown, R. J. Puddephatt, C. E. E. Upton, J. C. S. Dalton, (1974), 2457. 2. H. C. Clark, L. E. Manzer, Inorg. Chem.,~~ (1973) 362. 3. G. M. Whitesides, J. F. Gaasch, E. R. Stedronsky, J. Amer. Chem. Soc.,~~ (1972) 5258. 4. L. Abis, A. Sen, J. Halpern, J. Amer. Chem. Soc., 100 (1978) 2915. 5. P. S. Bratennan, R. J. Cross, G. B. Young, J. C. S. Dalton, (1977), 1892, and previous work. 6. M. P. Brown, R. J. Puddephatt, c. E. E. Upton, S. W. Lavington, J. c. s. Dalton, (19 74 ) ' 1613 . G. B. Young, G. M. Whitesi'des, J. Amer. Chem. Soc., ~99 (1978 ), 5808. M. P. Brown, A. Hollings, K. J. Houston, R. J. Puddephatt, M. Rashidi, J. C. S. Dalton, 0976) 1 786. 9. D. C. L. Perkins, R. J. Puddephatt, C. F. H. Tipper, J. Organomet. Chem.,~~~ 10. (1978}> C16. D. R. McAlister, D. K. Erwin, J. E. Bercaw, J. Amer. Chem. Soc.JlOO (1978), 5966. 11. M. Yoshifuji, K. I. Gell, J. Schwartz, J. Organomet. Chem., 153 (1978). C15. 12. S. J. McLain, R. R. Schrock, J. Amer. Chem. Soc., ~9~ (1978), 1315. 13. M. H. Chisholm, D. A. Haitko, C. A. Murrilo, J. Amer. Chem. Soc., 100 ......... (1978), 6262 . 14. J. Evans, S. J. Okrasinski, A. J. Pribula, J. R. Norton, J. Amer. Chem. Soc.,~~ (1977),5835. -100- 15. E. L. Muetterties, P. L. Watson, J. Amer. Chem. Soc., 98 (1976),4665. 16. D. G. Morrell, J. K. Kechi, J. Amer. Chem. Soc.,~~ (1975),7262. 17. K. Maruyama, T. Ito, A. Yamamoto, J. Organomet. Chem.,~~~ (1978), 359. 18. R. H. Grubbs, A. Miyashita, J. Amer. Chem. Soc.,100 (1978),1300, and previous work. 19. A. Tamaki, S. A. Magennis, J. K. Kechi, J. Amer. Chem. Soc.,96 (1974), -- 6140. 20. A. Johnson, R. J. Puddephatt, J .. C. S. Dalton, (1978),980. 21. C. J. Cardin, K. W. Muir, J. C. S. Dalton, (1977),1593. 22. C. J. Cardin, D. J. Cardin, M. F. Lappert, J. C. S. Dalton, (1977), 767. 23. J. D. Ruddick, B. L. Shaw, J. C. S. (A), (1969), 2801. 24. J. D. Ruddick, B. L. Shaw, J. C. S. Chem. Comm., (1967),1135. 25. T. G. Appleton, H. C. Clark, L. E. Manzer, J. Organomet. Chem.)~~ (1975),C39. 26. While the exact mechanisms of formation of§ is uncertain, it should be noted that L2PtMeCl cannot be made directly from L2PtC1 2 ; L2PtMe 2 forms instead. Also, exchange of groups has been noted frequently in Pt(II) and (IV) systems; see for example, R. J. Puddephatt, P. J. Thompson, J. C. S. Dalton~(1977})1219. 27. J. F. Nixon, A. Pidcock, Ann. Rev. of NMR Spect., ~ (1969),345. 28. D. A. Slack, M. C. Baird, Inorg. Chim. Acta,~~ (1977)>277. 31 P NMR spectra could only be obtained occasionally on a few samples. 29. Consequently, earliest attempts at the syntheses of Z~ and Z~ were examined by 31 p NMR only after removal of excess vinyl halide (so -101- preliminary 1H NMR could be obtained) and frequently after recrystallization (so that spectra of pure samples would be obtained). Decomposition studies were carried out on several samples prior to 31 P NMR, as a means of preliminary characterization. 30. (a) B. E. Mann, B. L. Shaw, N. I. Tucker, J.C. S. (A), (1971),2667,and (b) M. Yamamura, I. Moritani, S. I. Murahashi, J. Organomet. Chem., ~! (1975),C39, both report oxidative addition of vinyl halides to proceed with retention of stereochemistry about the C=C bond. 31. J. Schwartz, D. W. Hart, B. McGiffert, J. Amer. Chem. Soc., 96 (1974); 5613. 32. F. Ro Hartley, Organomet. Chem. Rev. A,§ (1970),119. -102- CHAPTER III: FORMATION OF CARBON-NITROGEN BONDS VIA TRANSITION METAL NITROSYL COMPLEXES -103- INTRODUCTION The formation of carbon-carbon bonds at transition metal centers is well established for a variety of systems. In particular, formation of carbon-carbon bonds via insertion of CO into a metal-alkyl bond has been observed in many instances (Reacti'on 1). 1 The presence of two-electron donor ligands frequently promotes the reaction 2 co 0 L II Reaction 1 ~-R A large number of transition metal nitrosyl compounds are known., 3 including some nitrosyl-alkyl complexes. Several of the complexes appar- ently form carbon-nitrogen bonds via insertion of NO into a metal-alkyl bond in a reaction analogous to the CO insertion shown tn Reaction 1. In Reactions 2 and 3, for example, the intennediacy of a metal nitrosyl alkyl compound seems certain. Reaction 4 seems to involve 0 Ni (.NO)L 2Br II Li0 Ni(NO)L 20 ~ L2Ni +- N0 Reaction 2 Reference 4 0 Ru(N0) 2L2 .0CH 2Br co I Ru (NO) (.NCH 2i) L2 Ru (.CO) (NO) L2Br + 0CH=NOH Reaction 3 Reference 5 Reaction 4 Reference 6 coordinated O=NCH 2 ~, formed via NO i.nsertion. Other cases (Reactions 5 and 6) involve free NO, but it seems that NO is again inserting into metal-carbon bonds, presumably after coordination to the transition metal. -104- Reaction 5 Reference 7 Reaction 6 Reference 8 A great many transition metal compounds also contain carbonyl ligands, and it is often observed that reactions proceed at the carbonyl ligands with almost total disregard for the nitrosyls. An example is given in Reaction 7, and one should be aware that this is a widespread Na+ RX [Fe(C0) 3(NO)J- ~~-- Fe(C0) 3(COR)(NO) Reaction 7 . Reference 9 phenomenon. It seemed to us th~t the most logical system in which to look for insertion of NO into a metal-carbon bond would be one where both NO and an alkyl group were attached to a transition metal, with no coordinated CO ligands. Such a system has been reported by Legzdins et al. for the Group VI metals: CpM(N0) 2R (M =Cr, Mo, W) (.Cp = n5-cyclopentadienyl).10 We decided to explore the chemistry of these to see if we could effect insertion of NO into the metal-carbon bond. -105- RESULTS AND DISCUSSION The chromium nitrosyl compounds ~ can be prepared from CpCr(N0) 2c1 ~ as described by Legzdins et al. , 10 and CpCr(N0) 2Cl can be prepared from CrC1 3 in a procedure reported by Wilkinson 11 (Reaction 8). There is some indication in the literature that the molybdenum and tungsten compounds analogous to 1 and 2 are somewhat less stable than the chromium 11 compounds, and therefore they were not prepared or investigated. 12 - Na +Cp CrC1 3 - - - - - CpCr(N0) 2Cl 2. NO 1. Reaction 8 2 la, R = Me b, R = n-hexyl c, R = isobutyl d, R = 0 - We prepared 2 as described by Wilkinson, but with the use of a large excess of NO bubbled through the reaction solution (instead of the approximately one equivalent used by Wilkinson), with the result that the yield improved somewhat (up to 59% from 30-50%). 2 is air stable as a solid and in solution for periods of up to an hour, but apparently decomposes upon storage at 18°C under rigorously air-free conditions over a period of several months. ...1 was also prepared as described. It sim- ilarly is air stable as a solid or in solution, but decomposes over a period of weeks as a solid, stored carefully (18° C, air free). -106- Treatment of la with Two-electron Donors In experiments designed to see insertion of NO into the chromiumalkyl bond, la was treated with a variety of ·two-electron donors. experiments are described below. of la during this discussion: These One should keep in mind the stability except for the aforementioned sporadic tendency of la to decompose upon standing, it is quite stable thermally, -- remaining unchanged after heating in benzene solution at 100°C for 24 hr. Photochemically, tt is considerably less stable: a solution of la will oe entirely decomposed in less than an h.our of irradiati'on with an unmodified mercury lamp. Treatment of la with CO caused no change in the starting materi a1. Treatment of ~~ with P~ 3 resulted in decomposition of ~~ at a rate faster than that observed in the absence of P9J 3 but with formation of no tractable products. When ~~ was irradiated in the presence of P9J 3, how- ever, a new organometallic product could be observed by NMR to be growing in. Attempts to drive the reaction to completion only resulted in de- struction of both starting material and product, until a modified light 0 source (< 4750 A) was used. We could then obtain reaction mixtures con- taining more product than !~' but long irradiation times would still destroy both materials. The ultraviolet spectrum of la exhibits a maxi- 0 mum at 4610 A, while the spectrum of a mixture of la and product shows a 0 shoulder at approximately 5750 A, the apparent maximum for the product. Evidently, the product is photoreactive, and is destroyed at rates comparable to those of its formation. It was subsequently discovered that other phosphines reacted more quickly and cleanly with la. -107- The thermal reaction of ~~ and PMe 20 was complete in 5 hrs. at ambient temperatures; the new product appeared, on the basis of its NMR spectrum, to be quite similar to the product of the P~ 3 reaction. PMe 3 reacted even more rapidly, so that all ~~ had reacted within two hours at ambient tempe·r atures. The reaction with PMe 3 was also quite clean (there seemed to be a correlation with the faster reactions being cleaner), and it was decided to concentrate our efforts on this reaction. Reaction of ~~ with PMe 3 Two equivalents of PMe 3 are required to convert all of ~~ to product; if only one equivalent is used, half of the ~~ is rapidly converted to product and the other half remains unchanged. There are actually two products, one of which is O=PMe 3 , identified by comparison of NMR and IR spectra and melting point to that of an authentic sample. McPhai1 13 has reported that CpMo(C0) 2 (~0) in the presence of P0 3 is converted upon ir~ radiation to CpMo(NCO)(CO)L 2 , and speculates that this is accomplished via the mechanism shown in Scheme I. A nitrosyl ligand is deoxygenated by the P0 3 to fonn a nitride species; the nitride then inserts into a Mo-CO bond. Scheme I r ro CpMb-NO 'co Phosphines are also known to deoxygenate organic nitroso hv P03 / CO CpM~O 'P¢3 +pm ""3 CO I-... CpM6-N: + 0 P=O 'P~3 3 +2L NCO CpMoL 3 I compounds, generating organic nitrenes, 14a and to deoxygenate nitroso compounds datively bond to transition metals. 14 b This suggests that in our reaction, PMe 3 is ,deoxygenating a nitrosyl or nitroso species, perhaps -108- via one of the two mechanisms shown in Schemes II and III. S.cheme I I CPC~ \ CH 3 NO + PMe ~ CpCr-N: + 0==PMe 3 3 \ J' CH 3 I .. la 3 Sa 4 Scheme III NO (+PMe3) --~ I CpCr=NCH 3 + O=PMe 3 l PMe; la 6 Sa In Scheme II, deoxygenation of a nitrosyl ligand generates a nitride species 3, which then inserts into the Cr-CH 3 bond. The 16-electron nitrene 4 which is so formed then adds another molecule of PMe 3 to achieve an 18-electron configuration. All of the intermediates in this mechanism are expected to be very reactive, which is consistent with the fact that no intermediates are observed by NMR during the course of the reaction. In a variation on this Scheme, PMe 3 might initially coordinate. to la and the coordinated phosphine deoxygenate a nitrosyl ligand to l"'J- -109- generate nitride species 3. (This is the variation favored by McPhai1 13 for his system.) In Scheme III, a datively-bound nitroso ligand is fanned first, in one or more steps; PMe 3 may or may not be involved in any of the steps required to form 6, and may or may not be coordinated when deoxygenation occurs to fonn the nitrene 4. One might expect a nitroso species such as ~ (L = PMe 3, so that an 18-electron configuration is achieved) to be stable and observable; for this reason, the mechanism shown in Scheme 1·1 is favored. Neither mechanism has been proven, and another mechanism may be operating. However, note that the same product, ~~, is predicted by either of the Schemes. Isolation and Assignment of Structure of Sa Isolation of the product of the reaction between ~~ and PMe 3 proved to be more difficult than anticipated. After removal of solvent under vacuum from the completed reaction mixture, O=PMe 3 could be separated from the solid residues by sublimation (room temperature, 0.1 µ) onto a -78°C cold finger. 15 In this way, crude ~~ containing only a few percent of O=PMe 3 and other impurities (visible in the NMR spectrum) was obtained in 80% yield. A variety of techniques were explored to further purify the material. During attempts to recrystallize ~~' it was discovered that all.owing Sa to stand in benzene, THF, acetonitrile or diethyl ether for periods of #~ hours resulting in apparent partial decomposition as evidenced by formation of black, gummy, intractable solids. The problem was less pronounced -110- with methylene chloride or hydrocarbon solvents, but recrystallization from or precipitation from any combination of these solvents resulted in material which showed only slight enrichment of ~~ to impurities. Chromatography of Sa on Florisil, neutral alumina I, neutral alumina V, neutral alumina I deactivated with PMe 3 , or basic alumina V resulted in apparent decomposition. One could obtain mixtures of Sa and a new material by rapid (less than one minute) washing of Sa through basic alumina I or chromatography at -50°C on the same support, but no pure Sa was obtained with any of these efforts. Sa does not sublime. The structure of 5a was assigned on the basis of its NMR, IR and mass spectra. These are contained in Table I. The NMR spectrum of Sa clearly shows a cyclopentadienyl resonance at oS.50; a doublet at ol.21 which, based on its chemical shift and 31 P-1 H coupling constant can be reliably assigned to a coordinated PMe 3 group; and one other CH 3 group, . most probably derived from the methyl group in la. The sensitivity of the chemical shift of this group to concentration and solvent is reminiscent of the solvent shifts of NCH 3 in MeC1 2M(ON(Me)N0) 2 , M=Nb, Ta. 16 The IR spectrum exhibits a strong band at 1630 cm- 1 , which can be reasonably assigned to a linear nitrosyl ligand (although this is somewhat ambiguous, as both bent and bridging nitrosyls can also display bands in this region). 3b IR data on many of the k~own metal nitrenes is unfortunately lacking, but Chatt has assigned bands in the region of 1090 cm- 1 to a metal-nitrogen stretch in some rhenium nitrenes, and has noted a band at approximately 1315 cm- 1 which seems to be chara~teristic of the NMe group in these compounds. 17 The latter band is often obscured by bands due to phosphine ligands in molybdenum nitrenes. 18 The band we have observed at 1040 cm- 1 may then be assigned to a Cr=N stretch, and -111- Table I Spectral Data for 5a 1.08 (d, J = 1 Hz, 3H), :N--{v'lea 1.21 (d, J = 13 Hz, 9H), PMe 3 5.50 (s, 5H), Cp 1630, 1585 (sh), NO stretch 1040 (broad), Cr=N stretch MS, m/e (rel. int.)b ae b. 252(4) parent ion 234(1) 222(31) parent ion - NO 220(28) 207(100) 222 - CH 3 205(.74) 192(13) 186(11) parent - c5H6 177(19) 162(19) 117(31) CpCr This value 1s very solvent and concentration dependent. Peaks which can be accounted for by chromium isotopes have been omitted for clarity. Some of the peaks which are not assigned may be due to loss of ethylene (from Cp) from various fragments, which has been observed for CpCr(NO)zCl: J. Muller, F. Ludemann, S. Schmitt, J. Organomet. Chem.,!~~ (1979), 25. -112- the NMe band obscured by phosphine bands. The mass spectrum of the organometallic product shows a peak at m/e 252, which corresponds to the parent ion, and the most abundant fragment at m/e 207 corresponds to a loss of Me and NO. Tentative assignments have been made for most of the major fragments. (A mass spectrum of Sa, when it had been briefly exposed to air, showed a peak at m/e 253 and none at m/e 2S2, indicating that it perhaps acquired a proton from the air; this is not inconsistent with the stability and chemistry of the compound described below.) The structural assignment is therefore as shown in Scheme II. Since Sa could be obtained only in 9S% purity, elemental analyses, molecular weight determination or an .X-ray crystallographic structure have not been performed. Reactions .of 5a The brief studies described in this section have been performed with approximately 9S% pure Sa, obtained as described above. Heating Sa for four hours at 70°C in benzene solution (0.3 M} resulted in nearly quantitative conversion to one new organometallic product (which has not been identified} and methane. Methane is also instantly obtained upon dissolution of Sa in methanol, or upon treatment of 5a with H+. Treatment of ~~ with H- resulted in formation of free PMe 3 as the only isolable product. No reaction of 5a with ethylene occurred, up to temperatures sufficient to decompose 5a. -113- Other Chromium Nitrenes Sb, Sc and 5d CpCr(N0) 2R, R = ]-hexyl,isobutyl,and ¢ (~~,~=and ld, respectively} were prepared in a manner analogous to that described for Sa. 10 They were similarly allowed to react with two equivalents of PMe 3 , and reactions were followed by NMR. The reaction of CpCr(N0) 2Cn-hexyl) ~~with PMe 3 proceeds at roughly the same rate as observed for la. O=PMe 3 and the ni_trene ~~ are fonned. This nitrene is apparently more reactive than the methyl nitrene, for it began to disappear even before all of the starting material had reacted. At least six new resonances appeared in the cyclopentadienyl region of the NMR. No attempt was made to isolate these compounds. NMR data for the reaction are listed in Table II. It was thought that the wide variety of products from Sa was perhaps due to insertion into C-+l bonds along the ~-hexyl chain. We hoped· to simplify the system by changing from an n-hexyl to an isobutyl group. 19 . CpCr(N0) 2(isobutyl) ~~ forms nitrene ~= and O=PMe 3 , but unfortunately this nitrene, too, further reacts to give a mixture which shows at least seven different peaks in the cyclopentadienyl region of the NMR spectrum, at rates comparable to those of the ~-hexyl nitrene. NMR data are again listed in Table II. The reaction between CpCr(N0) 2 ~ ~Sand PMe 3 is complete within 15 minutes; this is at least twenty times faster than when R=Me. ni trene does not undergo further reactions. The phenyl ld is deoxygenated so read.ily that we have been able to generate a P~ 3 -substituted nitrene, 3 CrCr(NO)(:N~)(P~ ) ~!· Data for 5d and 5e are included in Table II. -114- Table II - NMR Data for Chromium Nitrenes 5 . ~ I N.::;;Cr••• •NO R.-- ' L NMR(C 6o ) ,o Compound R L Sa Me PMe · Sb ~-hexyl 0.81 (m, 2H), -C~ 2 -c 5 H 11 1.3 to 1.8 (m, 20H), -CH 2 -c 5 ~ll + PMe~ 5.49 (s, 5H), Cp Sc isobutyl 1.0 to 1.8 (~, 18H), 5. 47 (s, SH) , Cp a. b. 6 3 1.08 (d, J = 1 Hz, 3H), R = Mea 1.21 (d, J = 13 Hz, 9H), PMe 3 5.50 (s, SH), Cp R = isobutyl + PMe~ 5d 1.14 (d, J = 12 Hz, 9H), PMe 3 S.68 (s, SH), Cp 7.2 (m, SH), R=0 Se 5. 65 (s, SH) , Cp 7.2 (m, 20H), phenyl groups Very solvent and concentration dependent. Spectrum too poorly resolved to allow a more specific assignment. -115- Neither of the phenyl nitrenes was found to be significantly more stable than Sa, so that purification was again difficult. -- 21 While organometallic nitrenes are relatively novel and deserving of investigation, the difficulty in obtaining pure materials in this system makes it a poor choice for such a study, and further work is not planned. -116- EXPERIMENTAL General THF, diethyl ether, Benzene and benzene-d 6 were vacuum transferred from sodium benzophenone ketyl prior to use. Acetonitrile and CH 2c1 2 were distilled from CaH 2 and degassed (three freeze-pump-thaw cycles) prior to use. Toluene and petroleum ether (35 - 60°C) were stirred over two portions of H2so 4 , then over two portions of ~no 4 in 10% H2so 4 , then dried before vacuum transfer from sodium benzophenone ketyl. Proton NMR spectra were obtained on an EM-390 Varian Associates NMR Spectrometer; IR spectra were obtained on a Perkin-Elmer 257 grating infrared spectrometer. Mass spectra were perfonned by Jet Propulsion Laboratory, Pasadena, California. Gas chromatographic analyses were performed on a Hewlett-Packard 5750 Research Chromatograph with Porapak Q (9' x l/8 11 , 80°C) for hydrocarbon gases and a Varian Aerograph 90-P on Penwalt (28% on Chromosorb R 80/100, 5' x 1/4 11 , 50°C) for amines. Manipulations of air sensitive materials were accomplished in a Vacuum Atmospheres Recirculating Dry Box, on a vacuum line, or under argon using standard Schlenk techniques. CpCr(N0) 2Cl was prepared according to the method of Wilkinson11 -117- except that NO gas was bubbled through the reaction solution, which improved the yi e1d (from 30 - 50% to 59%). The compounds 1 we re prep a red following the procedure described by Legzdins~ 10 Irradiations were perfonned with an Oriel medium-high pressure Hg lamp; the samples were placed tn Pyrex tubes and water-cooled. Stabili'ty of la la was heated in c6o6 (0.5 Ml at 100°C in a sealed NMR tube for 24 hrs. All of the la remained unchanged. Irradiation of la for 1 hr with an unmodified light source resulted in its complete decomposition. 0 Use of 3900 or 4750 A filters greatly increased the photochemical stability of ~~' which is then largely unchanged after 1 hr of irradiation. la absorbs in the visible with a 0 maximum at 4610 A. Reactions of l with Phosphines Reactions with phosphines were all perfonned similarly. 1 (.0.5 M) tn c6o6 and 5 equivalents of phosphine (either wei·ghed into an NMR tube or measured on a vacuum line) were allowed to react in an NMR tube, and the reaction followed by periodically recording the NMR spectrum. In- dividual results are described below. !~ and P0 3 gave no reaction at room temperature and, when heated at 100°C for 24 hrs., ·resulted only in the destruction of la with no -118- appearance of new products. When~~ and P~ 3 were irradiated with an unmodified light (10 min.), a new product could be observed to grow in (NMR: 1.21 (s, 3H), Me) . 55.53 (s, SH), Cp; It was impossible to drive the reaction to com- pletion, however, as both la and the product disappeared upon further irradiation (20 min. more). 0 Use of a 3900 A filter improved the situation somewhat, but the 0 best results were obtained with a 4750 A filter. in 1 hr., a 2-to-l mixture of~~ to product. One could then obtain, Further irradiation caused partial disappearance of both species, and a new peak in the NMR at cS5.47, in small amounts. For the product CpCr(NO) (:NMe)(P0 3 ): NMR (C 6D6 ) cSl.21 (s, 3H, Me), 5.53 (s, SH, Cp), 7.2 (m, 15H, phenyl); IR(c 6o6 ): 1640 cm-l (NO stretch); UV 4610 A (residual la and perhaps 0 also product), 5750 A (shoulder). and PMe 20 reacted completely in 5 hrs. at ambient temperature to !~ give CpCriNO)(:NMe) (PMe 20): NMR(C 6D6 ) ol.25 (partially obscured by PMe20, 3H, Me), 1.30 (d, J = 9Hz, PMe20), 5.50 (s, phenyl). SH, ·cp), 7.2 (m, 5H, The reaction was not very clean and the NMR resolution was poore la + CO No reaction at all was observed between la and CO (1 atmosphere) over a period of days. -119- la reacted completely with PMe 3 in 2 hrs. at ambient temperature. The spectral data for the product are included in Table I. Preparative scale reactions between !~ and PMe 3 were run in roundbottom flasks, usi'ng 0.5 M la and 2.2 equivalents of PMe 3. Solvent was removed under vacuum, and the resulting red-brown residues purified by various techniques. O=PMe 3 was best removed vi a sublimation (0.1 µ, _ room temperature), occasionally cleaning the cold finger and stirring the solids. This afforded an 80% yield of 95% pure Sa. Recrystallization, precipitation and chromatography were attempted in the dry box using a variety of solvents (benzene, THF, acetonitrile, diethyl ether, CH 2c1 2 and petroleum ether, either alone or in combination) and solid supports (Florisil, neutral alumina I, neutral alumina V, neutral alumina I deactivated with PMe 3 , basic alumina V). None of these efforts produce ~~ that was any purer than the material simply obtained after removal of O=PMe 3. Chromatography at -50°C on basic alumina I under argon was similarly unsuccessful. O=PMe 3 Authentic O==PMe 3 was prepared by treating PMe 3 (0.1 M~ with· NO gas in methanol solution. 20 NMR (c 6o6 ) 80.97 (d, J = 12.6Hz); IR 1295, 1290, 1195, 1975, 930, 847, 730 cm-l; mp 140 - 141° C. No reaction occurred between PMe 3 and NO in benzene solution. -120- lb, le and ld with Phosphines These reactions were run in an entirely analogous manner. Spectral data for the complexes can be found in Table II. Reactions of Sa Using 95% pure Sat a few preliminary studies of its reactivity were made. The thermal decomposition of ?~ was accomplished by heating ~~ in benzene c6o6 (0.3 M) in a sealed NMR tube at 70°C for 4 hrs. A peak in the NMR spectrum at 55.19 lIR: 1595 cm- 1 } was evidently due to a new organometallic product, which, along with methane, was formed in nearly quantitative yield. The new product was not identified. Dissolution of 5a in co 3oo _(0.3 M) or co 3oo containing one equivalent of H2so 4 resulted in an immediate color change (from red-brown to green) and formation of methane. Addition of more acid or NaOH resulted in no further changes. Treatment of a solution of~~ in THF-d 8 (0.2 M) with 2.5 equivalents of LiAlH 4 at 0°C, followed by warming to room temperature, resulted in a color change to dark brown. After two hours, the reaction mixture was treated with H20 (50 µl/ml THF), 15% NaOH (50 µl/ml THF) and additional H2o (200 µ1/ml THF). All volatile materials were then vacuum transferred away from the reaction mixture and examined by gas chromatography. Only P(CH 3 ) 3 could be detected. §~ in c6o6 (0.3 M) was treated with an equivalent of ethylene in a sealed NMR tube . The reaction was followed by NMR. No reaction occurred -121- in 10 hours at room temperature; heating the tube resulted in decomposition of 5a identical to that observed in the absence of ethylene. -122- References and Notes 1. c. Kotz, Inorganic Chemistry, W. B. K. F. Purcell, J. Saunders, Philadelphia (1977), 933, and references therein. 2. ibid., 937, and references therein. 3. (a) J. H. Enemark, R. D. Feltham, Coard. Chem. Rev. 13 (1974) 339. (b) R. Eisenburg, C. D. Meyer, Accts. Chem. Res. 8 (1975) 26. (c) N. G. Connelly, lnorg. Chem. Acta, Reviews 6 (1972) 47. (d) W. P. Griffith, Adv. Organomet. Chem. 7 (1968) 211, and referenc es therein. (e) s. K. Satija, B. I. Swanson, Inorg. Syn. ~~ (1976) 1. ( f) K. G. Caulton, Coard. Chem. Rev. 14 (1975) 317. 4. W. Seidel and D. Geinitz, Z. Chem. 15 (1975) 71. 5. J. A. McCleverty, C. W. Ninnes, I. Wolochowicz, J. Chem. Soc., Chem. Cormn. (1976) 1061. 6. M. F6a, L. Cassar, J. Organomet. Chem. 30 (1971) 123. 7. H. F. 8. A. J. Shortland, 9. F. M. -- Klein, H. H. Karsch, Chem. Ber. 109 (1976) 1453. G. Wilkinson, J. Chem. Soc. Dalton, (1973) 872. Chaudhari, G. R. Knox, P. L. Paulson, J. Chem. Soc. (C),(1967) 2255. 10. J. K. Hoyano, P. Legzdins, J. T. Malito, J. Chem. Soc. Dalton, (1975) 1022 . 0 11. T. S. Piper, G. Wilkinson, J. Inorg. Nucl. Chem. ~ (1956) 38. 12. The observation that the Mo and Wcompounds are less stable has, however, been questioned. P. Legzdins and J. T. Malito, Inora. Chem. 14 (1975) 1875. 13. A. T. McPhail, G. R. Knox, C. G. Robertson, G. A. Sims, J. -Chem. Soc. (A), (1971) 205. -123- 14. (a) T. L. Gilchrist, C. W. Rees, Carbenes, Nitrenes and Arynes, Appleton-Century-Crofts, New York (1969), 37, and references therein. (b) 15. G. L. Merica, S. Cenini, J. Chem. Soc. Dalton, (1980) 1145. To achieve good separation, the re~idues had first to be finely crumbled. During the course of the sublimation, the cold finger had to be cleaned off several times and the solids stirred. The residues could not be heated. 16. J. D. Wilkins and M. G. B. Drew, J. Organomet. Chem. 69 (1974) 111. 17. J. Chatt, R. J. Dosser, F. King, G. J. Leigh, J. Chem. Soc. Dalton, (1976) 2435. 18. J. Chatt and J. R. Dilworth, J. Less-Common Met. 36 (1974) 513. 19. There are only three types of C-H bonds into which insertion could occur in an isobutyl group, versus six in an ~-hexyl group. Additionally, insertion is probably not occurring with the C-H bonds in position 1, since the methyl nitrene does not undergo the reaction, so that the isobutyl group presents two C-H bonds versus five in ~-hexyl. 20. G. M. Kosalapoff. L. Maier, Organic Phosphorus Compounds~ WileyInterscience, New York (1972), 429. 21. -- w. A. Nugent, B. L. Haymore, Coord. Chem. Rev.,31 (1980) 123. -124- Proposition I: Intermediates in the Photochemical Decomposi- tion of Cp TiMe 2 2 cp 2Ti=CH2 has been suggested as an intermediate in the photodecomposition of cp TiMe • To confirm this, a cross2 2 over experiment (to rule out intermolecular processes) and trapping of cp Ti=CH 2 with alkenes, alkynes and/or carbonyl 2 compounds is proposed. It is also suggested that polymerization of alkenes by the photoproducts of Cp TiMe be more 2 2 carefully examined, and methods. are suggested for seeing informative intermediates. Proposition II: New Mixed-Metal Clusters Containing Small Organic Fragments Instead of treating preformed transition metal clusters with organic molecules, it is proposed that clusters be built by careful addition of a monomeric organometallic species to a dimeric one. In this way, one may be able to achieve clusters containing organic fragments in specific positions. Mixed-metal clusters containing organic fragments are of particular interest, and they are discussed. Proposition III: A Two-Dimensional Organometallic Polymer The formation of transition metal polymers containing metal-metal bonds is discussed. It is proposed that the -125- monome.r ic metal species be preformed into some particular array before polymerization {_formation of metal-metal bonds), and a monolayer, anchored to glass· via Si-O-Si bonds, is suggested as a means of generating a two-dimensional array. Proposition IV: Studies of . Metallabicyclobutanes It is proposed that transition metallabicyclobutanes be synthesized by the action of carbenes on metal alkyne complexes (which can also be viewed as metallacyclopropenes). The chemistry of metallabic-1clobutanes is discussed; in particular, one might hope to see formation of metallacyclobutenes. 3 Some Aspects of Organometallic (n -allyl) Proposition V: Fluxionality The endo and exo isomers of [CpMo(CO) (NO) (n 3 -c 3 H5 )] + ~ interconvert, but at a much slower rate than that observed for CpMo(C0) 2 (n 3 -c 3 H5 ) ~· The latter isomerizes by an apparent rotation about a Mo-allyl bond, without interconversion of the syn and anti protons. NMR and deuterium labelling experiments are proposed, which will determine whether isomerization in 2 occurs via a rotation or a n-cr-n mechanism. The deuterium labelled compound will also yield information about the mechanism of formation of 2 from 1 and NO+. -126Proposition I: Intermediates in the Photochemical Decomposition of cp TiMe 2 2 The photochemical decomposition of a number of titanocene species has been studied by various groups, and in many (although not all) cases parallels the thermal decompositions. The cyclopentadienyl groups can be photolabilized 5 in cp Ticl 2 (Cp = n -cyclopentadienyl) 112 or Cp2TiC1 2 2 5 2 (Cp* = n -c Me ), and irradiation of these species in chlor5 5 inated hydrocarbon solvents results in the formation of CpTiC1 3 or Cp*TiC1 , respectively (Reaction 1). 3 L TiC1 _ _h_.v_...) L • 2 2 CH Cl + LTiC1 2 3 ) LTiC1 L = Cp or Cp* species CpTiC1 2 The radical 3 Reaction 1 and cp· have been observed by ESR spectros- copy {cp· as its spin adduct with nitrosodurene). Irradiation of Cp TiMe results in the formation of 2 2 3 methane and titanocene (Reaction 2). Deuterium labelling studies indicate that other methyl groups and cyclopentadienyl rings serve as sources of hydrogen in the formation of 1.... Reaction 2 methane; during irradiation in benzene or toluene, no hydrogen is abstracted from solvent. The initial reaction is homol- -127- ysis of .the Ti-CH 3 bond, and both ep TiMe and Me· can be 2 observed by ESR (the latter as its spin adduct with 5,5- dirnethyl-l-pyrroline-l~oxide) ~ Cp 2 TiMe has been observed in the reaction of Cp 2 TiCl with MeLi at -78°C, but it is unstable at temperatures above -50°C. 5 The presence of free Me· (outside of the coordination sphere or titanium or a solvent cage) has been questioned~ and this may not be inconsistent with the observed spin adducts, which may be the result of trapping a species other than free Me·. Rausch suggests the possibility of intermediates such as a titanocene carbene 2 or a dimeric - species ~' formed from abstraction of hydrogen by Me· from a methyl group.3 Homolysis of M-C bonds has been observed in /CH2 Cp 2 T i ' 'TiCp 2 CH/ 2 2 3 other systems, including Cp 2 VMe,7 cp 2 Ti(aryl) 2 ~ (bipyridyl) PtMe 2 r~ .and Photolysis of cp 2TiMe 2 ~ in the presence of diphenylacetylene produces methylstilbenes as well as metallacyclo- pentadiene 4 and titanium methyl alkenyl complex ~ (Reaction · 3)~,lO Note the insertion of a carbon-carbon multiple bond into a metal-carbon bond. hv 1 ) + rac=c¢ ( cis and trans) + /CH 3 Cp2Ti CH ~ F<~ 3 Reaction 3 -128- Thermal decomposition of Cp 2TiMe also yields methane, 2 and the source of hydrogen is again both the methyl groups and the cyclopentadienyl rings~ 1 Small amounts of ethane are also produced. It seems likely that there are some common intermediates in the thermal and photochemical decompositions of Cp 2TiMe 2 2 and the reactions of cp T~CH Al(Me )Cl ~ and titanocene metallacyclobutanes: 3 ' 14 Indeed, a titanocene carbene £ 2 2 12 has been postulated as an intermediate in the reactions of titanacyclobutanes with esters 14 and is probably formed when 6 is treated with pyridine~ 5 It seems worthwhile, therefore, to propose several experiments to learn more about the mechanism of photochemical decomposition of 1, and to bring together the results of the several lines of research. Let us first address ourselves to the possibility of a dimeric species such as ~· Since one can obtain two molecules of methane for each molecule of 1 irradiated~ a species such as 3 (formed after the loss of one molecule of methane from each of two molecules of ~) must decompose further to give two more molecules of methane. If a dimeric species is formed, or if other intermolecular processep are taking place, it should be possible to detect this in the following cross-over experiment: a mixture of Cn 5-c 5H5 ) 2Ti(CH 3 > 2 and (n 5 -c 5o5 ) 2Ti(CD 3 ) 2 is irradiated. If no methane is formed via an intermolecular mechanism, then only CH 4 and CD 4 will be observed. If a dimeric species such as 3 contributes, or -129- if some other intermolecular process occurs (such as attack of cp 2TiMe on !),one will see CD 3H and CH 3 D and perhaps CD H • 2 2 The same type of cross-over experiment could be applied to the thermal decomposition of ! or decomposition of methylene· · d.iates b ri·d ged species _6 wi·th pyri·d·ine,15 an d i·f common interme are involved, the same results should be obtained. If a carbene species 3 is formed, it should be possible to trap it. For example, carbene 3 should react rapidly with a ketone to give the alkene product of methylene for oxygen exchange (Reaction 4)~ 6 wi th an acetylene to give a metallacyclobutene (Reaction S)t 2 or with an alkene to give a metallacyclobutane. (Reaction 6>~ 3 Thus, one could irradiate 0 h\> ' ep Ti=CH . 7 2 2 JlR' ) f.H2 RA_R' Reaction 4 2 1 -1 ~ h\> h\> ' - ) - 2 2 RC:.CR~ ep 2Ti<>-R Reaction 5 R /R ) in the presence of acetone and, if Cp 2Ti<>--R Reaction 6 t actually is formed from the irradiation of~' obtain isobutylene. Irradiation of cp Ticl in the presence of acetone gives cp TiC1 COCHMe 2 ), 2 2 2 2 but our system is different: since free Me· is not formed, neither is free cp TiMe (Me· always extracts H before "escape" 2 as CH 4 , and leaves behind ~or Cp(C 5H4 )TiMe or some similar -130species), and a reaction analogous to that of CpTiC1 2 is not necessarily to be expected. The fact that no titanacyclobutene is observed upon irradiation of 1 in the presence of diphenylacetylene (Reaction 5, R=¢) does not argue strongly against the intermediacy of ~ in photodecompositions in the absence of diphenylacetylene. The initial excited state formed by the absorption of light by ~ is a Cp-+Ti charge transfer state~ tron density on titanium; with increased elec- diphenylacetylene is a good elec- tron acceptor~ 7 It may be that an exciplex is formed between the excited state l* and diphenylacetylene c1·+ DPA._) and - - that the products one observes (Reaction 2) reflect the chemistry of this exciplex rather than a generalized reaction of intermediates from photolysis of ! with alkynes. The experi- ment should be performed with poorer electron acceptors such as 2-butyne (Reaction 6, R=Me) to give the known 14 metallacyclobutene Cp TiCH C(Me)=C(Me). 2 2 Irradiation of ! in the presence of 3,3-dimethyl-l-butene should afford the metallacyclobutane Cp 2TiCH 2CH(t-butyl)CH (Reaction 6)~ 2 3 Isolation or detection of titanacyclobutenes or titanacyclo~utanes might be complicated by their photoreactivity. Such problems might, however, be avoided by careful irradiation at wavelengths at which ! absorbs but the metallacycles do not. It is also conceivable that the ability to abstract a hydrogen is much different for Me. (in a solvent cage, -131perhaps) and di radicals 7 and 8, which cannot maneuver as easily. Additionally, the two radical centers are perhaps -8 held in closer proximity in 7 and 8, so that recombination to give ground state metallacycle is more likely. The quantum yield, then, of photoreactions of metallacycles may be much lower than the quantum yield of formation of methane from 1, and this might allow one to build up significant concentrations of metallacycle in photolysis mixtures containing 1 and alkene or alkyne. Puddephatt has observed that photolysis of ~ in the presence of methyl methacrylate results in polymerization of the alkene, presumably by a free radical mechanism~ If, as Rausch suggests~ the insertion of diphenylacetylene into the Ti-cH bond is an example of the type of insertion which has 3 been suggested for Ziegler-Natta polymerization catalysts (Reaction 7), then it is perhaps worthwhile to determine which mode or modes of polymerization can occur in this system. ) M-CH -CH -CH ·2 2 3 Reaction 7 One should irradiate 1 in the presence of ethylene or propene -132- (alkenes typically polymerized with Ziegler-Natta catalysts) and look for insertion products (irradiate ! in the presence of one equivalent of alkene) or polymers (irradiate 1 in the presence of a large excess of alkene) • The insertion products may be quite unstable, but low temperature photolyses are possible. While it is unlikely that the methyl methacrylate polymerization proceeds in an analogous manner, one might want to explore this further, perhaps again looking for insertion products by irradiating ~ in the presence of one equivalent of methyl methacrylate (instead of the excess used by Puddephatt) . Depending on the results obtained above, one might be able to contribute information to the resolution of the controversy over whether Ziegler-Natta polymerization occurs 18 . . . via a lk ene insertion 19 or meta 11 acyc l o b utanes. . h t, One nu.g at low temperature, be able to observe (in the irradiation of 1 in the presence of ethylene) either 2 or 10; if the system polymerized ethylene, one might see 11 or 12. One /Me Cp Ti 2 'cH cH R 2 2 2, R=CH 3 ±E, R=H ]..].. , R=CH CH CH 2 3 2 }_3, R=CH 2 CH.J might not be so lucky, however, but a careful labelling experiment might still yield some information. If intermediates -133- 9 or 10 decomposed via e-hydride abstraction, then irradiation of : in the presence of ethylene-d 4 would yield hydrocarbon mixtures indicative of the intermediates (Scheme I): Scheme I Insertion Mechanism Metallocyclobutane Mechanism h\) D2 C:- S-H cp T< )cn _ _..,.) 2 c 2 D C=CDCH D 2 2 + H C=CDCD 3 2 H2 Other modes of decomposition might similarly yield characteristic mixtures of hydrocarbons. It seems, then, that a great deal of information is potentially available from the photochemistry of cp 2TiMe 2 , especially as it relates to the chemistry of Cp 2ticH 2Al(Me 2 )tl ~ and polymerization schemes, and further investigation is suggested. -134- References and Notes 1. z. Tsai, C. H. Brubaker, Jr., J. Organomet. Chem., 166 2. R. w. Harrigan, G. s. Hammond, H. B. Gray, J. Organomet. Chem., ~! (1974), 79. 3. M. D. Rausch, w. H. Boon, H. G. Alt, J. Organomet. Chem., !~! (1977) t 299. 4. E. Samuel, P. Maillard, Chem.,~~~ (1977), 289. So E. Klei, J. A. Teuben, J. Organomet. Chem., 188 (1980), 97. 6. c. H. Bamford, R. J. Puddephatt, D. M. Slater, J. Organomet. Chem., 159 (1978), C31. -- 7. D. F. Foust, M. D. Rausch, E. Samuel, J. Org:anomet. Chem., 193 (1980), 209. (1979), 199. c. Giannotti, J. Organomet. --- 8. D. c. L. Perkins, R. J. Puddephatt, Org:anomet. Chem., 166 (1979) I 261. 9. M. D. Rausch, c. F. H. Tipper, J. w. H. Boon, E. A. Mintz, J. Organomet. Chem., 160 (1978), 81. --- 10. w. H. Boon, M. D. Rausch, J. c. S. Chem. Comm., (1977), 397. 11. G. J. Erskine, p. A. Wilson. J. D. Mccowan, J. Organomet. Chem.,}~~ (1976), 119. 12. F. N. Tebbe, G. W. Parshall, G. S. Reddy, J. Amer. Chem. Soc., ~99 (1978), 3611. 13. T. R. Howard, J. B. Lee, R. H. Grubbs, submitted for publication. 14. K. Brown-Wensley, Ph.D. Thesis (see Chapter 1). 15. It is assumed that reaction of 6 with pyridine produces 2, by removing Me A1Cl entirely- from the coordination sphere of Ti. F. 2N. Tebbe, R. L. Harlow, submitted for publication. 16. s. H. Pine, R. Zahler, D. A. Evans, R. H. Grubbs, J. Amer. Chem. Soc., ~2! (1980), 3270. 17. Diphenylacetylene has been observed to form an excimer: -135- B. Brocklehurst, D. C. Bull, M. Evans, P. M. Scott, G. Stanney, J. Amer. Chem. Soc., 97 (1975), 2977. It may reasonably be expected to behave in a manner similar to many other aromatic compounds which have been observed to form exciplexes: J. B. Birks, Photophysics of Aromatic Molecules, Wiley-Interscience, New York (1970). Stilbene is an-example: F. D. Lewis, Accts. Chem. Res., 12 ( 19 79) I 152 • 18. E. R. Evitt, R. G. Bergman, J. Amer. Chem. Soc., (1979) I 19. ~~~ 3973. K. J. Ivin, J. J. Rooney, c. D. Stewart, M. L. H. Green, R. Mahtab, J. c. s. Chem. Comm., (1978), 604. -136- Proposition II: New Mixed-Metal Clusters Containing Small Organic Fragments Because of their own potentially interesting chemistry and because they serve as models for understanding catalytic metal surfaces~ transition metal cluster compounds have received a great deal of attention~' 3 Both homometallic and mixed-metal clusters are known, although more work (particularly concerning reactions of clusters with small organic molecules) has been done with homometallic clusters. The synthetic methods applied to the formation of clusters have relied largely on luck, although rational schemes have been used, mostly consisting of building a cluster by adding one metal at a time (Reaction 1). The reaction conditions used in this approach are frequently not gentle. M + M Step 1) M--M Step 2 ) M' M--M \ I M' Reaction 1 Interesting work with clusters and small molecules has demonstrated chemistry which has also been observed at single metal centers and on heterogeneous catalytic surfaces. For example, formation of a carbene group from an alkyl group (by a-hydrogen abstraction) has been observed in osmium clusters 4 (Reaction 2) as well as tantalum mononuclear species 5 (Reaction 3). Furthermore, Pettit 6 has suggested that insertion of carbene units (on one metal center or bridging two metal -137- centers) into metal-hydride bonds gives methyl groups on heterogeneous Fischer-Tropsch catalysts (Reaction 4). > \ Reaction 2 LiR LiR) R=neopentyl Reaction 3 Reaction 4 or MMMMMMM Other examples of analogous mononuclear and cluster chemistry are shown below (Reaction 5 and 6). Insertion of NO into M-C bonds has been observed for rhenium clusters 7 and mono- s nuclear tungsten. NO NO WMe6 ~ Me4W Reaction 6 2 -138- of a $-hydrogen from a vinyl group has been observed in osmium clusters 9 and mononuclear iridiurn1.0 ~bstraction (Reactions 7 and 8}. S-H ) Reaction 7 L ---) r lfl t ' Hir(CO)L 3 Reaction 8 The synthesis of cluster compounds containing small organic fragments often consists of allowing a preformed cluster to react with, for example, alkene or N2 cH , and taking 2 as the organometallic product whatever Nature chooses to give. Wilkinson has preformed a Re c1 9 cluster and treated it with 3 MeMgX to place methyl groups on the cluster~ 1 It seems that a new synthetic approach, whereby one could more carefully design the organometallic cluster, might be valuable. 1: The .approach is similar to that described in Reaction put together mono- or dinuclear organometallic fragments. Either the mono- or dinuclear species involved in Step 2, Reaction 1, might contain a particular organic fragment of interest. For example: -139- + Reaction 9 While particular fragments used to construct the cluster might not retain their integrity (for example, when R=H, a bridging hydride, as shown in Path B, might form), one can probably find systems in which the organic fragments of interest remain (approximately) where one wants them. To do this successfully, one must employ fairly mild synthetic conditions, so as not to destroy the starting organometallic complexes. Several techniques seem promising: CO ligands can be easiiy labilized photochemically 12 or with Me 3No 13 a or (£-butyl)Po; 3b generating coordinatively unsaturated fragments which can combine to form the clusters; the photochemical approach has been used in some instances~ 4 A similar approach has been demonstrated to be general for metal carbide clusters and mononuclear compounds containing readily displacable ligands~ Re 2 (CO) lO + 5 h 'V Fe (CO) S - - - - . . Et o 2 [Re Fe (CO) 2 12 ] Reaction 10 -140- Metal anions can be allowed to react with metal halides (Reaction 12). 16 A coordinatively unsaturated metal species can be allowed to react with another metal (Reactions 13 and l4) • 17I18 (diphos)Ptcl (C0) 3 /Co (diphos)Pt I 'X=o 2 "-cd" (C0) 3 Reaction 12 L H-0~ (R P) 3 + I./"' ' ~QsL 2 L~t"os-H 3 L3 Reaction 13 + CpW (CO) ~ 2 ~ (R3 P) 2 Pt"i!R R '-wcp(coi 2 Reaction 14 It is certainly reasonable to suppose that similar methods will succeed with metal complexes containing small organic fragments. A wide variety of mononuclear species are avail- able for consideration; they might contain one or more alkyl or hydride groups, or a metallacycle. A number of interesting -141dimeric species are also known, including a diiron methylene ~; 9 dicobalt species. ~; 0 ana dialkyl dimolybdenum ~~l One (CH ) 2 ,,, CH.2.__ (CO) 4Fe ?e (CO) 4 CpCoHCoCp 0 - 1 0 L=NR n=l,3 2 3 - 2 could hope to construct clusters from such fragments. For example, photochemically labilizing CO might lead to the formation of cluster 4a (Reaction 15). (In this and all reactions listed below, the product is a tentative formulation only, and carbonyl, methyl, or other groups may prove to be bridging rather than bonded to one metal, or vice versa.) It may prove to be necessary to choose irradiation wavelengths such that only one starting material is irradiated • 1 + CpRu(C0) R 2 R=Me or H hv ) .,...........cH 2 ......__. (C0) Fe 3 Fe(C0) ~~o Ru I ep "-R 3 4a Reaction 15 Metal alkyl halide and a metal anion could lead to compound 5 (Reaction 16). An unsaturated metal-metal bond and an electron-rich metal might lead to cluster 6. (Reaction 17). A synthesis analogous to Reaction 11 ought to yield yet another cluster 7 (Reaction 18). - -142- Me L '~t [Fe (CO) 4H]- -t A'e (CO) 3 H I'a Mel'& ............. '\ / L "Fe (CO) 5 3 Reaction 16 3 + L PtMe 2 2 L =COD, 2 PR , 2 3 etc. [ Fe5(C0)14C] 2- Me ~ ' / Me Pt Reaction 17 L 2 ~ '\oRL 2 -6 + Reaction 18 Hornometallic or mixed-metal clusters ought to be obtainable via these methods. The examples in Reactions 15-18 show mixed-metal clusters as products. One would want, of course, to examine the chemistry of the products 4a, 5, 6, and 7. For example, 4a might insert CH 2 into the Ru-R bond to give a elimination from 5 (to give CH 4 -cH R group. 2 Reductive or c H ) might occur, or 2 6 similar processes might be observed with 6 (giving MeR or c H ). 2 6 Formation of a carbyne (!CR) from R observed for 7. and C might be Of course, one can speculate that a number of other reactions might also occur, but no attempt is made -143to discuss all of them here. It should be noted, however, that mixed-metal species have the potential to demonstrate a greater v·ariety of reactions than homometallic compounds do. For, example, if one were able to observe insertion of CH 2 into the Ru-R bond in 4a, one might also see an analogous reaction in 4b (the synthesis of which is shown in Reaction 22 In the products, the CH R group may be found on the 19). 2 metal atom on which it is formed (which would depend on whether CH inserted or R migrated), or its position may 2 reflect the tendency for the heavier metals to prefer the higher oxidation states (Scheme I). CpFe {CO) R 2 Reaction 19 Scheme I Fe /CH 2 'Fe~ Fe ' Ru/ R Ru/ CH 4a 2' Ru ---4 "/ Fe R 4b Fe or "Ru/ R\._/u Fe I CH 2 R / " Fe Ru I CH R 2 f:2R (or a bridging CH R) 2 or 'tH2R Ru " / Ru Fe (or a bridging CH 2 R) -144- It is clearly too early to make specific predictions concerning the results of the experiments sug~ested here. However, one will certainly be able to obtain new clusters containing intriguing small organic fragments, which should display rich chemistries of their own. -145References and Notes 1. A metal cluster is defined as containing three or more metal atoms. Work with bimetallic species provides an important link, and some of it will be referred to. 2. M. Moskovits, Accts. Chem. Res., 12 (1979), 229. 3. w. L. Gladfelter, G. L. Geoffroy, Adv. Organomet. Chem., 18 (1980) I 207. 4. R. B. Calvert, J. R. Shapley, J. Amer. Chem. Soc., 99 (1977), 5225. 5. R.R. Schrock, Accts. Chem. Res., 12 (1979), 98, and references therein. 6. R. Pettit, Division of Inorganic Chemistry, Second Chemical Congress of the ·North American Continent, Las Vegas, Nevada, August 25-29, 1980. 7. P. Edwards, K. Mertis, G. Wilkinson, M. B. Hursthouse, c. S. Dalton, (1980), 334. K. M. A. Malik, J. a. 9. c. S. Dalton, (1973), 872. B. F. G. Johnson, J. w. Kelland, J. Lewis, s. K. Rehani, A. J. Shortland, G. Wilkinson, J. J. Organomet. Chem., 113 (1976), C42. 10. J. Schwartz, D. w. Hart, B. McGiffert, J. Amer. Chem.Soc., 96 (1974), 5613. 11. A. F. Masters, K.. Mertis, J. F. Gibson, G. Wilkinson, Nouveau J. Chim., ...l (1977), 389 • 12. Many reviews are available on the subject. {a) c. R. Bock, E. A. K. von Gustorf, Adv. Photochem., 10 (1977), 221. (b) M. Wrighton, Chem. Rev., Z~ (l974);-40l. 13. (al . D. J. Darensbourg, N. Walker, M. Y. Darensbourg, J. Amer. Chem. Soc., 102 (1980), 1213. (b) Y. Shvo, E. Hazum, J. c. s. Chem: comm., (1975), 829. 14. M. R. Churchill, F. J. Hollander, Inorg. Chem., 16 (1977), 2493. 15. M. Tachikawa, A. s. Sievert, E. L. Muetterties, M. R. Thompson, C. S. Day, v. W. Day, J. Amer. Chem. Soc., 102 (1980), 1725. -14616. J. Dehand I J. F. Nennig I rn·o rg .. Nucl. Chem. Lett. I 10 (1974} I 17. 18. 875. L. J. Farrugia, J. A. K. Howard, P. Mitrprachachon, J. L. Spencer, F. G. A. Stone, P. Woodward, · J. · c. s. Chem. Comm., (1978}, 260. This method was applied to H os (C0) 2 3 10 and AuMe(P¢ 3 ), but HAuOs 3 (co) (P¢ ) was formed. 3 10 T. V. A. Ashworth, J. A. K. Howard, F. G. A. Stone, J. c. S. Chem.· Comm., (1979) , 42. 19. C. E. Sunmer, Jr., P. E. Riley, R. E. Davis, R. Pettit, J. Amer. Chem. Soc., 102 (1980), 1753. 20. K. H. Theopold, R. G. Bergman, J. Amer. Chem. Ebe., 102 (1980), 5695. 21. M. H. Chisholm, D. A. Haitko, J. Amer. Chem. Soc., 101 (1979) I 22. 6784. (Co) Rtf.2!ZRu(C0) is not yet known, but it ought to be 4 4 available by a method analogous to that used to prepare the Fe analog (Reference 19). -147- Proposition III: A Two-Dimensional Organometallic Polymer Considering the importance of organic polymers in our lives, it is not surprising that interest has sprung up in organometallic polymers as well. 1 Organometallic polymers may or may not contain formal metal-metal bonds, or be designed for specific through-space metal-metal interactions; they may have organometallic groups incorporated into the polymer chain or pendant from an organic backbone; and a variety of synthetic schemes can be envisioned. However, there 2 are (of course ) problems which are rapidly encountered by the experimentalist. Organometallic polymers may display poor solubility, so that purification, handling and characterization are difficult. 3 Clusters 4or cyclic 5organometallic species may be obtained instead of linear polymer chains. There are a few examples of linear oligomers containing three transition metals. 516 Thus, while the polymerization of a solution of organic monomer is frequently easily accomplished, different techniques may be required for polymerization of organometallic compounds. It is not difficult to think of methods of making organometallic crosslinks (Reaction 1). These are not t~ue organo- metallic polymers, but organic polymers contaminated with a bit of metal, and one would expect their properties to be very similar to those of other crosslinked organic polymers. It is a simple matter conceptually to extend this idea to a -148- M ) Reaction 1 M L L LM L LM LM Reaction 2 polymer with many pendant organometallic groups, complexed to pendant ligands. only Such materials may be quite insoluble, and a small percentage of the pendant ligands might be sub- stituted before precipitation occurred. Steric or other prob- lems might also lead to a low percentage of substituted M M M M + -M-M-M- 'CJ'C -M ' \ M-M Reaction 3 ligands. Furthermore, if one wished to make metal-metal bonds crosslinking would certainly ensue (Reaction 3). It is proposed to explore a new approach to organometallic polymers containing metal-metal bonds: with some organic polymer or other material, the monomeric metal species are -149- aligned in some configuration, and then bonds between the monomers are formed. The method might allow one to construct polymers containing metal-metal bonds in an orderly fashion. It is difficult to think of a linear (one-dimensional) array which would not present the crosslinking problems described above (Reaction 3) , wiless the metals could perhaps be sterically constrained to react only with their neighbors on the chain. achieved. However, a two-dimensional array might be 7 Sagiv has described the formation of mixed mono- layers on various surfaces. The most stable of the monolayers contained n-octadecyltrichlorosilane (OTS) chemically bonded to a glass surface. Other materials (fatty acids, cyanine dyes) were physically absorbed on the surface, and good '\ s/°' ~o\ / Q.. Si Si / 0 /////)i ///~/////// glass surface quality monolayers could be achieved with dye constituting up to 50-60 % of the mixed monolayer coverage. Such mono- layers supply a method of achieving ordered arrays of organometallic species, depicted below (l and 2), either by using an organometallic species with a long hydrocarbon tail as one of the components of the monolayer, or by arranging organometallic species at the other end of the hydrocarbon chain attached to a -sicl 3 group. -150- ) ) Si Si M Si M M M M M Si Si Si Si Si I I I I l 2 1 - Such monolayers might have interesting properties of their own, or they might allow the preparation of two-dimensional polymers from the formation of metal-metal bonds in the array (~ or 4). 4 3 To form an array of type l, one would need to know more about the interaction of the metal species Mand the surface (i.e., whether it is chemically bonded or physically absorbed on the surface). Metals which are physically absorbed will quite likely be attached reversibly, just as cyanine dyes and fatty acids are reversibly attached; so that the array will not be as stable. Chemically bonded species would be an improvement, but still be not be as stable as the OTS monolayers, which derive stability from polymeric bonding; the M-0-Si bond might not be as strong as the Si-0-Si bond, either. -151- Additionally, .the choice of metals that .f orm strong M-0 bonds is rather limiting, especially if one wants to subsequently polymerize the metals via metal-metal bonds~ Clearly, further investigation into the (chemical or physical) bonding of metals with long hydrocarbon tails to form monolayers (or to modify surfaces) is warranted, but it will not be discussed further here. Taking advantage of the chemical bond formed by OTS (or other RSiC1 species), it is suggested that the following 3 materials be formed into monolayers: Cl Si-(CH 3 2 )nPR~ R = Me or ~; n = 18 3 Cl Si-(CH R =Me 2 )nPR~-M or~; M = Fe(C0) 4 or other metals; n = 18 -5 -6 c1 3 si-CH 2 CH 2 Pt¢) (CH 2 )nCH 3 3 2 2 2 Cl Si-~ cH f (¢) (CH ln~J n = 16 M 7 M = Fe(C0) n 4 or other metals; = 16 -8 c1 3 si-i(¢) (CH 2 )n~J M M = Fe(C0) 4 or other n = 18 = 18 10 metals, n 9 Cl 3 Si-M-(CH )nCH 2 3 M = Co(C0) or other metals; 3 -152- · a mono 1 ayer w1'th p h osp h'1ne compoun d 5 9 ,lO ought to give 11 groups at the upper surface. The choice of the group R may be governed by steric effects (two phenyl groups at the upper surface may be too bulky to allow the formation of a good monolayer). A transition metal complex can be attached after (M=Fe, Ru, Os), or [Mn(C0) J 5 2 (M=Mn, Re) are good choices for the metal, since they are the monolayer is formed. M(C0) 5 known to form metal-metal bonds (in clusters) readily; they 4 Other metals might, of t ria . t omic . species. . . even f orm 1inear course, be used. Monomeric or dimeric metals might be acceptable, although dimeric species might be too bulky to complex to all of the (closely packed) phosphine groups. One could make a mixed monolayer, consisting of ~ and OTS, to achieve a suitable spacing of phosphine groups at the upper surface. Alternatively, one could complex the metal to the phosphine before forming the monolayer (Compound 6). - The long hydrocarbon chains holding the transition metal complex in a "fixed" array may be too flexible. Compounds 1 12 and 8 would largely solve this problem, as they will hold the metal complex rigidly near the glass surface. Compounds 9 13 and io ~ould also hold the metal rigidly; however, their 1 ability to form stable monolayers will partly depend on the . . b on d s. 13 stability of the Si-P an d S1-M A monolayer formed from these species might be interesting in its own right. Even more fascinating, though, is the polymer that will be formed as photolysis or treatment -153- with (n-butyl} 3 Po or Me No gently removes CO and promotes 3 5 the formation of metal-metal bonds~ With the metal species constrained to be co-planar (or nearly so) by the long hydrocarbon chains z, ~,~and~~), (and attachment near the glass surface for planar polymers might form. An investigation of the properties (physical, chemical, catalytic, electrical) of these, while beyond the scope of the present proposal, would certainly be interesting. -154- 1. c. E. Carraher, J. E. Sheats, c. U. Pittman, · Organometallic Polymers, Academic Press, New York (1978). 2. Murphy' s Law. 3. G. D. Goretsas, R. H. Grubbs, unpublished results. 4. Many reviews are available which discuss the formation of clusters from monomeric organometallic species. One of the most recent is: W. L. Gladfelter, G. L. Geoffroy, Adv. Org·a nomet. ·chem., 18 (1980), 207. There are many references therein-fo earlier reviews. 5. P. A. Christian, Ph. D. Thesis, Stanford University, 1978. 6. (a) G. O. Evans, J. P. Hargaden, R. K. Sheline, J. c. S. Chem.· Comm., ( 1967) , 186. (b) M. W. Lindauer, G. O. Evans, R. K. Sheline, Inorg. Chem., 7 (1968), 1249. J. Sagiv, J. Amer. Chem. Soc., 102 (1980}, 92. 8. D. c. Bradley, Adv. Inorg. Chem. and Radiochem., 15 (1972), 259. 9. Compound 5 can be easily prepared using the standard methods for synthesis of RSiC1 3 and RP¢ 2 (Reference lOa). 10. (a) G. M. Kosolapoff, L. Maier, Organic Phosrhorus Compounds, Wiley-Interscience, New York (1974~. (b) s. N. Borisov, M. c. Voronkov, E. Ya. Lukevits, Organosilicon Derivatives of Phosphorus and Sulfur, Plenum Press, New York (1971). (c) B. J. Aylett, Organometallic Chemistry Reviews, Journal of Organometallic Chemistry Library 9, Elsevier Scientific Publishing Company, New York (1980), 327. 11. As the monolayer is depicted in 1, 2, 3, and 4, the bottom of the monolayer is attached to €he-glass surface, and the upper "surface" is that formed by the ends of the hydrocarbon tails. 12. ¢2 PcH,cH 2 SiC1 1 is known (Reference lOa and patents cited tfiere~n), as are a number of analogous compounds. 13. Fe(Co) 4P(SiMe 3 ) 3 is known: H. Schumann, o. Stelzer,~ Organomet. Chern., l l (1968), C52, although it is uncertain whether c1 3SiPR 2 will be a good ligand. Si-P bonds -155- are susceptible ·to hydrolysis, so that water on the glass surf ace might cleave this bond before hydrolyzing ·the ·si~Cl bonds (Reference lOb}. 14. 15. CpFe(C0} 2 SiMe 3 and other transition metal-Si bonded species are kfiown, although I am not aware of any metal-Sicl 3 species (Reference lOc) • (a} M. Wrighton, Chem. Rev., 74 (1974}, 401. (b} D. J. Darensbourg, N. Walker, M. Y: Darensbourg, J. Amer. Chem. Soc., 102 (19 80}, 1213. (c) Y. Shvo, . E. Hazum, J. c. s. Chem:-comm., (1975), 829. -156- Proposition IV: Studies of Metallabicyclobutanes Bicyclo{l.l.O]butane isomerizes at temperatures above 150°C to butadiene, 112 and concerted 3 and diradical anisms have been proposed for the rearrangement. 4 mech- In the pres- ence of nickel(O) catalysts, bicyclobutanes undergo an intramolecular retro-carbene reaction 5 to apparently give a nickel carbene, which is trapped by electron-poor olefins (Reaction 1). Silver cation opens bicyclobutanes rapidly at R R--4> -'z - Ni(O) ) R Ni Reaction 1 moderate temperatures; and reactions of other strained ring molecules with metals have been studied. 1 One example of a transition metallabicyclobutane is known: the reaction of platinum(O) and cyclopropene yields 1 (Reaction 2) . 6 The compound has considerable a-character R R R__A_R~ R=Me or H Reaction 2 in the Pt-C bonds. 7 Compound! releases cyclopropene when treated with carbon disulfide. 6 -157- A number of other transition metal alkyne .. complexes, which may be viewed as cycloproperies, are known. Cp~M<JC.R 5 Cp*=(n -CSMes> M=Ti, R=~ or Me R M=Zr, R=¢ (Reference 8) R I -Cp=(ns-CSHS) Cp2M+lll I 3 M=Mo, R=¢ -or Me (References 9,10) R M=W, R=Me (¢3P)2M+~ ++ (¢3P)2M~: M=Ni, R=¢ (Reference 9) (Reference 11) M=Pt, R=¢, Me or H (References 11,12) 4 The addition of carbenes to carbon-carbon multiple bonds 13 is well known, and i t is proposed that the transition metal alkyne complexes 2, 3 and 4 ·be treated with N CH 2 to see if 2 metallabicyclobutanes can be obtained in this manner (Reaction 3) • + ) Reaction 3 It seems logical to start with (¢ P) Pt(c H ), since 3 2 2 2 the resulting metallabicyclobutane is known, but it will be interesting to extend the method to other transition metal systems. It should be noted that the approach used to -158synthesize .~ (displacement of ethylene by cyclopropene) . appears not to b e .genera 1 • 14 It will then be interesting to explore the chemistry of the compounds obtained via these reactions. Some of the anticipated products are discussed below. In an analogous reaction to the isomerization of bicyclobutanes or oxybicyclobutanes 15 one might see formation of a vinyl metal carbene (Reaction 4). Reaction 4 S-hydrogen abstraction might be observed, leading to (n 3 -c 3 R3 ), a ligand which has been observed before, although not via this route. Reaction 5 Metallacyclobutenes may be formed via a diene intermediate (Reaction 6), a diradical intermediate (Reaction 7), or an n 3-c 3 R3 intermediate (Reaction 8), or via a concerted mechanism (Reaction 9). Some of these routes will be distinguishable from others by appropriate variation of groups on the metallabicyclobutane. It will be especially interesting to examine the reactions of Cp2Ti(C 2 R2 ), since titanacyclo- -159- butenes are well known. 17118 D M<><H • • 1 1 2 shift Reaction 6 Reaction 7 R H M-(9 ' Reaction 8 concerted Reaction 9 While it is too early to predict the reactions of the metallabicyclobutanes formed from reaction of carbene on transition metal alkyne complexes, they will undoubtedly be interesting, and syntheses via be attempted. this route should certainly -160- References ·and Notes 1. A. Greenburg, _J. F. Liebman, Strai·ned Organic Molecules, Academic Press, New York (1978). 2. H. M. Frey, I. D. R. Stevens, Tran:s.· F'araday Soc., 61 (1969)' 90. 3. G. L. Closs, P. E. Pfeffer, ·J. Alne·r·.· 'Chem. Soc., 90 (1968} I 2452. s. Kirschner,· ·J·.· Ame·r·.· Chem. · Soc., 97 S. Dewar, (19 7 5} I 2 9 31. 4. M. J. 5. R. Noyori, H. Kawauchi, H. Takaya, Tetrahedron Lett., 19 (1974), 1749. 6. J. P. Visser, A. J. Schipperijn, J. Lukas, J. Organomet. Chem., 47 (1973), 433. 7. J. J. de Boer, D. Bright, J. 8. R. s. Threlkel, Ph. D. Thesis, California Institute of Technology, 1980. 9. J. L. Thomas, J. Amer. Chem. Soc., 95 (1973), 1838. c. s. Dalton, (1975), 622. 10. A. Nakamura, S. Otsuka, J. Amer. Chem. Soc., 94 (1972), 1886. 11. E. O Greaves, C. J. L. Loch, P. M. Mathis, Can. J. Chem., 46 (1968), 3879 •. 12. M. W. Lister, R. B. Poyntz, Thermochem. Acta, 17 (1976), 177. c. W. Rees, Carbenes, Nitrenes, and Arynes, Appleton-Century-Crofts, New York (1969). 13. T. L. Gilchrist, 14. The following reactions have been observed: (a} Ni (0) + ~4 Ni'j M. J. Doyle, J. McMeeking, P. Binger, J. (1976), 376. 0 (b) Fe 2 {CO) g + ~ ~ c. S. Chem. Comm., f::<co> 3 G. Dettlaf, u. Behrens, E. Weiss,· Chem. Ber., 111 (1978), 3019. -161- The same reaction as in (b), with Fe (co) and 2 9 CpMn .cco) 2 (THF) • . P. Binger, B. Cetinkaya, c. Krueger, J. Organomet. Chem., 159 (1978) t 63 • - (c) 15. L. E. Friedrich, R. A. Leckonby, D. M. Stout, J. Org, Chem., 45 (1980), 3198. 3 3 16. One n - _cycloproperiyl .i .s '.k.nown·' · ._in Cp_ N.i Cn -c 3 H3 ): W. K. Olander, T. L. Brown, J. Amer. Chem. Soc., 94 {1972), 2139. 17. metallacyclobutenes are known: F. N. Tebbe, G. Parshall, D.• w. Ovena11, · J. Amer. Chem. Soc., 101 ~itanocene (19 7 9 ) t 5 Q7 4 • 18. K. A. Brown-Wensley, Ph.D. Thesis {Chapter I). -162- Proposition V: Some Aspects of Organometallic Cn 3 -allyl) Fluxionality The molecule CpMo(C0) 2 (n 3 -c 3H5 ) (Cp = n 5 -c 5 H5 ) can exist in two isomeric forms, la and lb, which interconvert rapidly at room temperature.1 la en do The isomerization occurs without exchange lb exo of syn and anti hydrogens, thus ruling out a TI-o-TI mechanism, and appears (from the simplest, though not the only viewpoint) to be a rotation about an axis between the metal and the center of gravity of the.n 3-allyl. This motion best describes 3 the fluxionality in several other n -allylic systems, while isomerization is well established for others~ 3 Fluxionality in Cn -c H5 )Mo(C0) 2 (diphos)Cl can be 3 explained by the rotation wherein the <n 3 -c 5H5 )-CO-CO system the TI-o-n remains fixed with respect to the metal and with respect to 13 each other~ This is the simplest explanation for the ~H, c, and 31 P NMR spectra observed from -100 to 30°C, although a simultaneous metal-allyl rotation is not excluded by the data (Reaction 1). A similar motion can explain the!~~~ isomeri- -163- zation, if the molecule is viewed as pseudo-octahedral, with 3 the (n -c 3H ) occupying two positions. Note that, to explain 5 the observed equivalence of the protons on C-1 and C-3, one must not draw a distinction between axial and equatorial positions in the schematic representation (Reaction 2). ico oc-;-:-x lJ. ' ' ~co x"LJ \ lb exo Reaction 1 OC-M-P' Reaction 2 la en do c Interestingly, treatment of la and lb with NO+PF - results 6 4 in the following tr~sformations: · la (endo) NO+PF 6 ) lb (exo) NO+PF 6 3 + 3 + [CpMo (CO) (NO) (n -C3H5)] (exo) 2a Reaction 3 ~ [CpMo{CO) {NO) (n -C3H5)] (endo) 2b Reaction 4 SE2 mechanism, explaining these transformations, is shown in Scheme I. 5 This explanation raises several interesting An points. Faller observes that the products 2a and 2b interconvert -164- Scheme I ::di-,,,,co -=21 ...• + oc-tl _.co NO+ ~ oc-M-NO H c-1H 21 CH lb exo ~ OC-M~ 2 o~M-NO : + ~ @ (§7 oc/r& H2 C-,CH -- l,cH~ ~M-0 oc~ + 2b en do at a rate 10 6 times slower than that observed for the endoexo isomerization of la and lb. 4 Rotation of alkenes bonded to metals is in general faster for cationic species, due to decreased back-bonding contributions from a less electronrich metal. Metal nitrosyl compounds, however, tend to rotate olefins less easily, and an explanation may lie in a strong preference for NO to be in some particular geometry (eg. trans) relative to some particular ligand. Faller's observation and this possible explanation lead one to wonder what fluxionality 3 (or lack thereof) might be observed in other n -allylic systerns where NO + has replaced CO {a facile reaction) • Systems 6 7 containing ligands such as (n 2-RN=CR) or (n 3-RN-N=NR) might be suitable, although such studies are not the intent of the present proposition. One wonders what mechanism explains the slow endo-exo isomerization of 2a~2b. A similar, but much slower, rotation .,,..,,,..., .,,..,,...,, about the metal-(n 3-c H5 ) axis is possible, but so is a n-o-n 3 -165interconversion. One can distinguish between these possibil- ities by examining whether a syn-anti interconversion occurs during the endo-exo transformation. This can be accom- plished in two ways. First, one can examine the high temperature coalescence of the proton-NMR signals of~~ and 2b. One will observe coalescence of the ABCDX spectrum of 2a and the A'B'C'D' ' spectrum of 2b into one ABCDX or one A X spectrum as one 4 approaches fast enough rates of isomerization, implying, respectively, a metal-(n 3 -c 3H 5 ) rotation o~ a TI-cr-TI conversion. This observation will only be possible at high (which may not be accessible). temperatures 8 Figure 1 x+x' a+a' I I I I c+c' I I t I I I I b+b' Hd ' I I I I I 4 I I I I d+d' 9H H 'H c' I exo a+b+c+d+ x ' ~+x' a'+b'+c'+ d' An alternative approach is the spin-saturation technique. -166- Nuclei in a particular site are saturated, and, provided they move to a new site at a rate comparable to relaxation times (but too slowly to observe coalescence of the signals of the new and original sites), some saturation of the signal of the new site will be observed. Thus one can saturate the H siga nal in 2a, for example, and observe some saturation in either the Ha' or in the Ha'' Hb'' He' and Hd' signals in~~' implying either rotation or n-o-n interconversions, respectively. One should note that it has not been established whether or not syn-anti interconversion occurs in 2a or 2b without endo-exo isomerization. Such a question can also be answered by the spin-saturation technique: saturation of Ha in 2a, for example, will lead to partial saturation of Hb' He and Hd if n-o-n conversion occurs without endo-exo isomerization. The deuterium-labelled molecules 3a and 3b or 4a and 4b can answer the same sorts of questions, as well as another which will be discussed momentarily. The deuterium will allow one to follow the fate of a particular nucleus without resorting to spin-saturation techniques, but entirely analogous observations will be made. An advantage is that one need not worry about attaining proper temperatures and rates of isornerization, as one will always be able to see the position of the deuterium. ~H ON•••• Mo'A.._ oc~ ~ ira 3a (exo) ~ ON ·~~oa ?YH OC ~ H 3b (endo) <? ~ ON• 1 11Mo~H oc" ""' H 4a (exo) ON uu MoH ~D D O~. H~ H 4b (endo) -167- Cn 3-c 3H 4D) ~ can be accomplished 2 9 Using Z-3-chloropropene-d there are two routes: 1 The synthesis of CpMo(C0) easily. Mo(C0) CH CN 3 6 ) H-~Cl D CpMo(C0) \H (endo) --Sa Sb (exo) 2 ~ [CpMo(C0)3] - Na + ~ H\~Cl D ~ CpMo (CO) 3 V~ D Reaction 5 Reference 10 hv ) 5 6 - Reaction Reference 11 The first route (Reaction 5) gives better yields, but may not work if syn-anti scrambling via a n-a-n route occurs in the 3 intermediate Mo(C0) 2 (cH 3 CN) 2 Cn -c 3H4D)Cl. No such fluxionality is reported for these compounds.12 The second route (Reaction 6) should avoid this potential problem. One will probably obtain a mixtu.re of endo and exo isomers, which may be separable, or it may be necessary to use a 2-methylallyl derivative, where the exo/endo ratio is higher. Treatment of Sa with NO+ will certainly give 3a or 4a or -- both, and this brings us to an important question: what is the transition state in the proposed SE2 mechanism, and, in particular, is the allyl n- or a-bonded? The reaction still needs to be checked kinetically, to be sure it is first order in -168- each of the reactants (consistent with an SE2 mechanism). Assuming this to be the case, one can ask whether the transition state is a twenty electron species (as drawn previously to explain the endo-exo conversion upon treatment with NO+, with a n-bonded allyl) or an eighteen electron species (with a a-bonded allyl). One might expect an eighteen electron transition state, in which case Sa will, via a a-bonded allyl (which can rotate), give a mixture of 3a and 4a. One could also obtain an eighteen electron transition 14 state in two other ways. First, it is well documented that nitrosyl ligands can function as either one or three electron donors. In this case, the nitrosyl ligand enters as NO+ (NO+ is usually used to describe a three electron nitrosyl ligand, and NO- to describe a one electron donor) and is a three electron donor in the product, so it seems unreasonable to envision a one electron NO in the transition state. Alternatively, one can argue that since· NO must be donating electrons to the metal as CO leaves, it is unlikely that NO would simultaneously be "backing off" to a one electron donor. The second 3 possibility is that the cyclopentaaienyl ring is n instead of n,5 a mechanism for achieving an eighteen electron config5 3 . uration which has been observed in <n -c 5H5 ) (n -c 5H5 )Mo{NO)Me, 15 for example. It seems reasonable that one may observe products 3a and 4a, implying a a-bound allyl in the transition state. Alter- natively, one might argue for a twenty electron transition -169- state or an 3 n -cyclopentadienyl group, a reaction which is too fast to allow rotation {or is constrained in some other way to prevent rotation), or another mechanism. In the event that one does obtain either 3a or 4a {but not the mixture, which would be difficult to separate), one can then complement the spin-saturation study described previously, to determine whether syn-anti scrambling accompanies the endo-exo isomerization {or occurs, perhaps, at a faster rate, although still a slow one on the NMR time scale). In any case, the experiments described should allow one to determine (1) whether endo-exo isomerization occurs in 2a~2b via the same rotational motion as in la~lb or via a TI-o-TI mechanism, and (2) whether the transition state for the SE2 conversion of la~2a or lb~2b contains a cr- or TI-allyl. -170Reference·s and Notes l. J. W. Faller, C. Chen, M. J. Mattina, A. Jukubowski, J. Organornet. Chem. I s_~ {197 3) I 361. 2. J. W. Faller, Adv. Organomet. Chern. , 16 {19 77) , 221, and references therein. -- 3. J. W. Faller, D. A. Hartke, R. D. Adams, D. F. Chodosh, J. Arner. Chern. Soc., 99 {1977), 1654. 4. J. W. Faller, A. M. Rosan, J. Amer. Chem. Soc., 98 {1976), 33 88. 5. It is easiest to see the endo and exo forms in a pseudooctahedral arrangement, ~ut it is easiest to see concerted substitution of CO by NO ~. from the epi;resi te side in a tetrahedral arrangement. Both representations are shown. 6. R. D. Adams, D. F. Chodosh, Inorg. Chern., 17 {1978), 41. 7. E. Pfeiffer, J. Kuyper, K. Vrieze, J. Organomet. Chern., 105 {1976) I 371. 8. Sharp signals as the result of coalescence for la and lb occur at 65°C. Since 2a and 2b isornerize on the-order-of one million times rnore-slowly;-one would need a temperature of perhaps 300°C to see a comparable spectrum, and decomposition of the compounds might occur previous to the attainment of this temperature. A typical NMR spectrometer cannot achieve these temperatures, either. One might, however, see informative spectra at lower temperatures, if not nicely resolved coalesced spectra. 9. The synthesis of Z-3-brornopropene-d 1 has been described: J. A. Katznellenbogen, A. L. Crumrine, J. Arner. Chern. Soc.,~~ {1976), 4931. An entirely analogous scheme ought to lead to the chloride. 10. R. G. Hayter, J. Organomet. Chem., J} {1968), Pl. 11. M. L. H. Green, A. N. Stear, J. Organornet. Chern., {1964), 230. J 12. H. D. Murdoch, ·R. Henzi, J. Organornet. Chern. , 2 {19 6 6) , 553. 13. No distinction is made between a transition state and a highly reactive intermediate in this discussion, and the proposed study would not distinguish between the two. -17114. See, for example, R. Eisenburg, 8 ( 19 7 5 ) , 2 6 • c. D. Meyer, Accts. Chem. Res • , -- - 15. F. A. Cotton, G. A. Rusholme, J. Amer. Chem. Soc., 94 {1972) I 402.