H a: ‘ MICHIGAN STATE UNIVERSITY LIBRARIES Hlllll lllllll ll 3 1293 10666 5460 ll HI f LIBRARY ; Michigan State l, 3 University This is to certify that the thesis entitled TITANACYCLOBUTANE—TITANIUM METHYLIDENE REACTIONS: STUDY OF A DEGENERATE OLEFIN METATHESIS SYSTEM presented by JohnBosco Chi—Bun Lee has been accepted towards fulfillment of the requirements for M. degree in LhemisLLy Major professor . WM“ 0-7639 MSU LIBRARIES m x. RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. WWW TITANACYCLOBUTANE—TITANIUM METHYLIDENE REACTIONS: STUDY OF A DEGENERATE OLEFIN METATHESIS SYSTEM BY JohnBosco Chi—Bun Lee A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1985 To My Family ii ACKNOWLEDGMENTS I would like to express my appreciation for the patience extended to me by my research preceptor, Dr. R. H. Grubbs. I am grateful to Drs. T. J. Pinnavaia, M. W. Rathke, W. H. Reusch, G. E. Leroi (substituting for C. H. Brubaker) for serving on my oral examination committee. I would like to thank the ”Group” at Caltech for providing much friendship and memorable moments and making my stay in Pasadena enjoyable. To my parents, special thanks for the love and encouragement they have given me through the years. I hope someday I might fulfill their hopes and expectations. Also, I would like to mention that my brothers have provided a great deal of support during the difficult times. Finally, I would like to thank my wife, Helen, for the care and understanding which she has shown me, especially during the preparation of this Dissertation. Her sacrifices and encouragement are appreciated more than words can expressed. TABLE OF CONTENTS Page LIST OF TABLES . . . . . . . . . . . v LIST OF FIGURES . . . . . . . . . . . . . . . . . . . vi ABBREVIATIONS . . . . . . . . . . . . . . . . . . . . viii INTRODUCTION . . . . . . . . . . . . . . . . . . . . . 1 RESULTS AND DISCUSSION . . . . . . . . . . . . . . . . 11 1. Preparation and Characterization of Q . . . . 11 2. Description of the Structure of .Q and Related Complexes . . . . . . . . . l6 3. Reaction of Q with AlMeZCl— Stepwise Metathesis . . . . . . . . . . . . 25 4. Reactions of Q with 0— Olefins and Diphenylacetylene . . . . . . . . . . . . . 33 5. Reactivity of the Parent Titanacyclobutane Q. 55 EXPERIMENTAL . . . . . . . . . . . . . . . . . . . . . 72 LIST OF REFERENCES . . . . . . . . . . . . . . . . . . 92 iv LIST OF TABLES 1 13 H and C NMR data for Q in benzene-d 6 Selected bond angles and distances for Q and Q 13 Isotopic perturbation of C chemical shifts of TiC C C m 1320313) 13C Chemical shifts for alkylidene and olefin complexes . . . . . . . . . . . . . . . Observed rate constants for the reaction of Q with PhCECPh at various temperatures . Observed rate constants for the reaction of Q with PhCECPh in the presence of added Olefins at 40°C . . . . . . . . . . . . . . . . . . . Page 13 17 21 23 42 43 10. ll. 12. 14. LIST OF FIGURES Proposed mechanism for the photolysis of the tungstenacyclobutane . . . . . . . Reaction pathways for metallacyclobutanes 90 MHz 1H NMR spectrum of Q in benezene-d6(*). 1. H3 and Hb regions of trans-Q-d12. Ha and Hb regions in selective homodecoupling of HC, + indicates the position of the decoupling power . . . . . . . 125. 7 MHz 13C NMR spectrum of a 4:1 mixture of Q— d2 and Q at 25° C in benzene— —d6 . Mechanism for Lewis acid cocatalyzed degenerate olefin metathesis . . . . . . . . . . . . 76.8 MHz 2H NMR spectra of the reaction of trans—Q—d1 with AlMeZCl in toluene (*) Proposed mechanism for AlMeZCl induced isomerization of trans—Q—dl Alternate mechanism for the isomerization of trans-Q-dl 76. 8 MHz 2H NMR spectra of the exchange reaction between Q— d2 and Q at 40° C Catalysis cycle for the degenerate metathesis of Q— dO and Q— d3 . . . . . . . Proposed mechanism for the reaction of Q with PhC: CPh . . . . . . . . . . . . . . . . Natural log plot of the percentage of Q as a function of time . . . Eyring plot for the reaction of Q with PhCECPh Plot of the ratio [4¢><]/[PhCECPh] vs. 1/kobsd for the reaction of Q and PhCECPh with added Q in benzene-d6 at 40°C . vi Page 14 22 27 28 30 31 34 35 37 4O 41 44 15. 16. 17. 18. 19. 20. 21. Metathesis of trans-Q—dl with Q—do Olefin exchange of trans—Q-d1 with 3,3-dimethy1—l-pentene Bimolecular mechanism for olefin exchange 90 MHz 1H NMR spectrum of a. Expansion of Ho and H B in toluene-d8(*) region Experimental (3. and b.) and calculated (c.) spectra of Hg (a. and c.) and H8 (See Experimental Section for simulation parameters) Formation and reactions of titanacyclopentane Q (b.) of Q-d Proposed mechanism for the formation of the enediolate complex QQ vii 2 Page 47 51 S3 56 S8 63 68 Cp DMAP Ph >0 t-Bu PY Me Et iPr R or R' Cp* ABBREVIATIONS nS-CSHS N,N-dimethylaminopyridine Phenyl angstrom t-butyl pyridine methyl ethyl isopropyl an alkyl group nS-CSMeS metal center Ligand viii INTRODUCTION The mechanism of the transition metal catalyzed olefin metathesis reaction1 has been intensively studied for many years, and a consensus has emerged that this reaction proceeds via a non-pairwise chain mechanism2 involving the interconversion between metal alkylidene and metallacyclo— butane intermediates (equation 1). H R LnM=CHR RHC=CHR' H > Q + T'——‘ LnM R' '\— + (1) R‘HC=CHR' LnM=CHR' R' H The syntheses and isolation of metal alkylidene complexes and studies of their reactions with Olefins have convincingly demonstrated the intermediacy of such species in metathesis. In contrast, the evidence for the involve— ment of the metallacyclobutane intermediates has not been as well—documented and is largely indirect. To date, examples of well-characterized complexes in this class are rare, being limited to only a few transition metals. A brief review of published results that are pertinent to this work will be presented here. Z The earliest known metallacyclobutane complexes were prepared by the reaction of cycloprOpanes with Zeise's . . . . l ' 3 dimer as illustrated 1n equation 2 for CHZCHZCHR (R=H,Ph). [ClZPt(C2H4)]2 c1 R pyridine FYI", i d + %> "Pt (2) Py”l [:>> R c1 These Pt(IV) complexes have been characterized by X-ray diffraction techniques.4 The results indicated that the complexes possess a puckered metallacyclobutane ring. The degree of puckering is least in the unsubstituted deriva- tive. It is generally assumed that the structures of the intermediates involved in metathesis are similar to these platinacyclobutanes. However, there has been no report of metathesis activity for any of these complexes. The products in the thermal degradation of the parent compound are a mixture of cyclopropane and propylene (equation 3). Products arising from fragmentation of the metallacycle were not observed. PyZClZPt<::::> _——Ji——) 1Z2: -+ 4é?\\ (3) Puddephatt5 found that the o-phenyl substituted platinacycle rearranges to a 1:2 equilibrium mixture of o- and B-phenylplatinacycles on heating (equation 4)., 50°C \ PyZClZPt<> \ PyZClZPt Ph (4) Ph Initially, it seemed possible that this rearrangement proceeds by a mechanism closely related to that proposed for metathesis. Labeling and crossover experiments indicated that it is an intramolecular process and proceeds with retention of stereochemistry. Puddephatt concluded that the data is most consistent with a concerted mechanism without the formation of a Pt=CH2 complex. The fact that phenylcyclopropane was formed in the thermal decomposition of these complexes has led Casey6 to propose an alternate scheme for this skeletal rearrangement. It involves the formation of an edge-metallated cyclopropane, followed by an edge—to-edge isomerization and subsequent ring opening. A rather different approach to stable metallacyclo- butane was discovered by Green and co-workers.7 The synthesis arises by regiospecific addition of a nucleophile to the central carbon of n-allyl derivatives of bis-n-cyclo- pentadienyl-tungsten and molybdenum. Thus, treatment of a cationic n-allyl tungstenocene complex with either sodium borohydride or methyl lithium results in the formation of tungstenacyclobutanes (equation 5). Green has examined the thermal decomposition of these complexes. In all cases studied, Olefins arising from hydrogen migration reaction were found as major products. NaBH 4 CpZWO Cp2W Me MeLi Cyclopropanes were formed in only trace quantities. szw+ —) PF6 Photolysis of the methyl—substituted tungstenacycle yields substantial amount of propene, a metathesis—like product, along with isobutene. To account for these results, Green postulated a scheme (Figure 1) whereby the initial photo— chemical reaction causes a nS—CSHS—to—n3—C5H5 ring shift. Rearrangement of the resulting l6—electron intermediate leads to the observed product. Q, pig. 1% %& 9 ———————%> //S§> + [Cp2W=CH2] ”t Figure 1. Proposed mechanism for the photolysis of the tungstenacyclobutane Recently, metallacyclobutanes have been isolated from the thermolysis of neopentyl—metal complexes. Whitesides8 reported that heating the dineopentyl—bis(triethylphosphine) platinum(II) complex results in the formation of neopentane and the cyclometallated complex as shown in equation 6. / l 157°C (EtSP)2Pt ? (Et3p)zpt<>< + CMe4 (6) This reaction involves a rather unusual C-H scission on a saturated hydrocarbon substrate. Mechanistic studies have shown that it proceeds via an initial phosphine dissocia- tion, followed by oxidative addition of the y-CH bond of the neopentyl group to Pt and subsequent elimination of neopentane. This reaction appears to be general for low valent Group VIII metals, since similar intramolecular cyclometallation reactions leading to isolable metalla- cyclobutanes have also been observed for Ir, Rh9 and RulO. The platinacyclobutane decomposed thermally to give 1,l—dimethylcyclopropane. Isobutene was not observed from a fragmentation process. The studies of isolable metallacyclobutanes have shown that there are several mechanisms for decomposition for these structural types. The three main reaction pathways are: a) carbon-carbon bond cleavage b) reductive elimina- tion c) B-hydrogen transfer, as illustrated in Figure 2. Reductive elimination to cyclopropanes occurred in all cases but to widely varying degree. With the exception of the B,B-dimethyl-platinacycle, the B-hydrogen transfer reaction is competitive with cyclopropanation. However, in a metathesis chain reaction, these two pathways most likely represent forms of termination steps.1 M:::C a M———C a l C E: i l —‘ H H C::: C———C S C ell C fi ll——lh-)C '_—> + [M] 6/ <—- c—c FigUre 2. Reaction pathways for metallacyclobutanes Considerable amount of work has been conducted in the study of metal alkylidene and carbene complexes, particu- larly their reactions with Olefins. Although there has been no report of isolable metallacyclobutanes from such reactions, these complexes have often been invoked to account for the observed products. Casey11 has studied the reaction of a diphenylcarbene complex of tungsten with isobutylene which gives a new olefin and a cyclopropane (equation 7). Ph / (c0) w=c + l A (c0) w 5 \ph A 40°C “ Ph Ph J / i (7) [w:<] + R (76%) Ph (10%) Ph Ph Ph ‘————E“—‘———_———‘_—_T" 7 A metallacyclobutane was proposed as the key intermediate for both products with the formation of the olefin via path a (Figure 2). Unfortunately, this reaction is irreversible and the fate of the new carbene fragment is not known. The fact that no other olefin was formed in the reaction was rationalized by invoking an electrophilic character of the carbene carbon. Schrock and his co—workers12 have isolated many tantalum and niobium alkylidene complexes. The chemistry of these complexes closely resembles that of the phosphorus ylides.13 Propylene was found to react stoichiometrically with a tantalum neopentylidene complex to afford two new Olefins (equation 8). Ta CpClZTa=CHC(CH3)3 +/\ —9 l ‘ (8) ____> W (3%) +W (86%) On the basis of labeling studies, Schrock proposed that the Olefins are most plausibly derived from the rearrangement of an intermediate tantalacyclobutane via path c (Figure 2). The reaction course is consistent with the nucleophilic character of the alkylidene carbon despite potential steric hindrance from the bulky tert-butyl group. More recently, Schrockl4 has synthesized a number of tantalum and tungsten alkylidenes which metathesize Olefins 8 catalytically. For example, (PEt3)C12(O)W=CHCMe reacts 3 slowly with ethylene, propylene or styrene to give t-butyl- ethylene and the analogous alkylidene complexes (equation 9). L Cl 0 1 da u>iv< + CH2=CHR y > l\CHm463 25-50°c (9) L L L - p CH -CHCM 1>01i¢ — EtS 2— 83 + C1 /|§§C CHR R = H, Et, Ph L In the presence of traces of AlCls, these same reactions were reported to occur considerably faster (1 hour). Furthermore, Schrock noted that the resulting solutions are active for the metathesis of l—butene and cis-2~pentene. No intermediates were detected during any of these exchange reactions. A bridging methylidene complex which also exhibits catalytic activity was isolated by TebbelS from the reac- tion between titanocene dichloride and trimethylaluminum (equation 10). CH Me Cp TiCl + 2A1Me3 ——> CpT i/ 2\A1/ + CH 2 2 21\. \x 4 Cl Me (10) & + ClAlMeZ Specifically, complex Q was reported to catalyze a metathesis process in which the methylene groups of isobutene and methylenecyclohexane exchange (equation 11). 13 CH CH 13CMZ 2 % CH2 2 A A + fix + 51°C Metallacyclobutanes were proposed as important but unobserved intermediates in this degenerate metathesis reaction. Interestingly, ethylene and propylene were found to react with Q irreversibly to give the homologated products propylene and isobutylene, respectively. The difference in reactivity of Q with these Olefins was attributed to the structures of the presumed intermediate metallacyclobutanes. As part of our study of the stereochemistry of the metathesis reaction, we have reinvestigated the homolo- 15 In the course of gation reaction described by Tebbe. this study, we unexpectedly discovered that metallacyclo— butanes could be isolated from the reaction of complex Q 16 It was found with monosubstituted terminal Olefins. that when Q was treated with 3,3-dimethy1—l-butene Q in the presence of 1 equiv of pyridine at room temperature, the 3—t—butyl-titanacyclobutane was formed quantitatively (equation 12).17 This reaction has proven to be a general route to numerous other titanacyclobutanes. 10 pyridine . : __________+ I + CH2 CHCMe3 t Cp2T1 oluene 2 t (12) t + py-AlMeZCl In this work, details of the preparation and charac— terization of Q, its reactivity with ClAlMeZ, terminal Olefins, and diphenylacetylene are reported. The solution structure of Q as determined by 13C NMR spectroscopy is also described. In addition, some of the features of the reactivity of the parent compound CpZTTCHECHECHZ, prepared from I and C2H4, are presented and compared with the t-butyl complex Q. These results provide important insights into the basic steps of the alkylidene transfer mechanism. RESULTS AND DISCUSSION 1- PESEE‘EE‘EE‘BE-§B§-§§§E§SEEEEE§EESB-9§-3’ Although the titanium methylidene complex Q was isolated as an alkyl aluminum adduct, its behavior was consistent with that expected of a free methylene species. In the paper describing the reactions of Q with Olefins, Tebbe15 has reported that the reactivity of the complex is greatly affected by the presence of bases such as THF and EtZO. He found that with THF in the reaction mixture, the rate was faster by a factor of at least 103. It was thought that the bases promote the dissociation of ClAlMe2 from the reagent, thereby freeing the methylidene for reactions. The reaction shown in equation 12 is a case of Lewis base- assisted conversion of Q to a titanacyclobutane. Pyridine (or its derivatives, vide infra) removes ClAlMe2 from Q, and the reactive ”Cp2T1=CH2" fragment is trapped immediately by the olefin to form the metallacyclobutane. The titanacyclobutane complex Q was initially prepared by treatment of a toluene solution of Q and 3,3-dimethyl-1— butene with 1 equiv of pyridine.17 However, this proved to be inconvenient on a preparative scale as the pyridine— aluminum byproduct and complex Q could only be separated with great difficulty by repeated recrystallization. We 11 12 have since found that this lengthy workup procedure could be circumvented by employing N,N—dimethylaminopyridine (DMAP) as the Lewis base. We also have some success using a 1:1 4—vinylpyridine—styrene copolymer in place of pyridine, but the yield is erratic. The titanacyclobutane complex Q was isolated in 54% yield as a red crystalline solid. It is readily soluble in hydrocarbons, dichloromethane, and ether. When pure, it is quite stable thermally, and the solid may be handled in air for moderate periods. The complex has been characterized by elemental analysis, molecular weight measurement, NMR spectroscopy as well as chemical reactions. Acidolysis with dry DCl liberated 2,2,3—trimethylbutane—d2 quantitatively. The 1H and 13C NMR spectra were invaluable in characterizing compound Q. A summary of the spectral data is presented in Table 1. Examination of the 1H NMR spectrum (Figure 3) reveals a high field signal at 6 -0.02 ppm, assignable to the B-hydrogen, He. The resonances of the a—hydrogens appear as two unresolved AB quartets at 6 2.24 and 1.94 ppm. For the chemically distinct cyclopentadienyl rings, the resonances are the expected two singlets separated by 0.18 ppm. The assignment of the resonances for the a—hydrogens was accomplished with the aid of the 1H NMR spectrum of the specifically deuterated metallacycle, 1 and (E)-3,3—dimethy1—1—butene— '\: trans—Q—d1,prepared from l—dl. The spectrum showed the u—methylenic hydrogen resonances at 6 2.24 and 1.94 in an intensity ratio of 1:2 13 Apofimfipunu .uofinsovnvv mpcwpmcou mcwfimsoo oo>pompo as wozoafiow .mowpfloflfimflpfise one momonucoemm :fl mofiuflucmson mzz UMH you wmaa ecu mzz :H toe mfi.ka .66-6:6N=6n on 6>Hpmfioa momnam Hm6a56eum N: «NH “zoe Nm wNH "mom N: mmH "mom m.OHH flew m.mN mmv m.em flee k.mH and m.ko a.OHH II II .C Mmmzuvu- mmmzueu- mu 0 mmmu own N: 6.x ”664 N: e.w Home .m m N: m.m u pa N: m.w up a me.m mo.o flu no he No.0- fie no we NQ.H me mo 6V 4N.N Ho.m U m mflmzovu- : a: : mzmu :H o e IQCGNfiQ CH .HO m m CW 6 Q m n.a e n . m m o e mzz UMH e :H H H6 e l4 T1 Ha H6 2. AM . ' Iia lib * UL Mei] He 1 J 8 7 6 5 4(ppm)3 2 1 O -1 Figure 3. 90 MHz 1H NMR spectrum of Q in benzene d6 (*). 1. H3 and Hb regions of trans- Q- d 2.6 H and H regions in selective homodecoupling of HC + indicates the position of the decoupling power 15 (l. in Figure 3). If we make the reasonable assumption that the formation of the titanacycle proceeds with retention of configuration, then we may assign the signal at 6 2.24 ppm to the hydrogens trans to the t-butyl group, H3, and the signal at 6 1.94 ppm to the hydrogens cis to the t-butyl 1 group, Hb. The H NMR spectrum of the titanacycle prepared from (Z)-3,3-dimethyl-l-butene-l-d1 displayed signals simi- lar to those described above, but the intensity ratio for the a-hydrogens was 2:1 (trans cis). The location of the B—hydrogen resonance is unusual. A plausible explanation for this high-field shift is that HC is in the shielding environment of the cyclopentadienyl ligands. This is apparently also true for the B-carbon whose resonance 13 appears upfield at 6 18.7 ppm in the C NMR spectrum. The coupling constant J for Cd is slightly higher than that C-H expected for an sp3carbon(d' =125 Hz), but compares well C-H =134 Hz), while J for C with that of cyclobutane (J C-H B C-H falls in between these two values (Table 1). Finally, it should be mentioned that at no time did we observe the a-substituted titanacycle in the reaction of l with 3,3-dimethyl-l-butene. Although this could be attri— buted to the polarization of the methylidene moiety (CpZTi=CH +—+ CpZTE-CHZ), steric crowding is probably the 2 main reason why the B-substituted derivative is favored. 16 2. Description of the Structure of 3 and Related Complexes ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The molecular structure of compound g has been firmly established by single-crystal X-ray diffraction techniques. Unfortunately, difficulty was encountered in the refine- ment process due to disorder, and consequently the preci- 18 sion of the diffraction data was limited. A recent structural determination of a related titanacycle 19 I CpZTiCHZCHPhEH2 é revealed a structure in which the basic features are analogous to those of compound 3.18 For comparision, Table 2 lists selected bond distances and angles for both complexes, and in the following discussion reference is made to the parameters of compound 3. It is interesting to note that despite the asymmetry at the C(S) positions, both metallacyclic rings are nearly planar. This is evident in the displacement of C(3) from the plane defined by C(l)-Ti-C(2), 0.09 X and 0.05 X for g and g, respectively. The displacement in 3 corresponds to a dihedral angle of 3.25°. It appears that in both struc- tures, the R(tert—butyl or phenyl)-Cp interaction is being relieved by a rocking motion of the C(3) fragment in the plane instead of puckering. The factors governing this choice is not apparent, however. The near planarity of the ring is in sharp contrast to the common assumption that metallacyclobutanes involved in metathesis are highly puckered. This assumption, which has been used to explain the moderate stereospecificity observed in the metathesis 20 of acyclic Olefins, is largely based on the structural 17 Table 2. Selected Bond Angles and Distances for 3 and 4 /C2\ 5 . C1 (2 4. Bond Angles (deg) Cl-Ti-C2 75 75.3(1) Ti-Cl—C3 84 86.0(2) C2-C3-C1 116 112.0(3) C1-C3-C4 118 115.9(3) C2-C3-C4 115 109.1(3) Bond Distances (A) Ti-C1 2.16 2.127(3) Ti-C2 2.14 2.113(4) C1-C3 1.55 1.546(5) C2-C3 1.53 1.579(5) C3-C4 1.52 1.521(5) Ti-C(n5) 2.07-2.51 2.36-2.40 Displacement (A) of C(3) from C(1)-Ti-C(2) 0.09 0.05 18 data from a series of stable platinacyclobutanes where puckering of the metallacyclic ring have been reported.4 Unfortunately, these systems do not engage in metathesis. The question of the symmetry of the metallacyclic plane is more subtle. The difference between the two independent titanium—carbon distances in 4 is 0.14 A and the C(1)-C(3) and C(2)-C(3) distances differ by 0.033 A, but the internal angles Ti-C(l)-C(3) and Ti—C(2)-C(3) are nearly identical. The bond length differences appear to be outside experimental error and may suggest a slight distor- tion. Hoffmann and his co-workers have recently conducted a theoretical study on the parent unsubstituted titanacyclo- butane CpZTiCHZCHZCH2 5.21 Interestingly, they found that the optimum geometry for 5 is not a symmetrical metalla- cycle, but an intermediate structure in which the metalla- cyclic plane is highly distorted.22 If indeed the contributions from the valence isomers, Ta and $2 (equation 13), are important as Hoffmann's calcu- lation21 has suggested, a highly distorted metallacycle would be expected. 2 C \\ ’5; 2 ______2s .,;>\C§__R (3132“ —-R CP T1 (13) c1 C1 Ia Ib 19 However, this does not seem to be the case. It is note- worthy that complex 4 probably represents a limiting case since conjugation is increased in structures Ia and Ib. ’b’b ’b’b Since the data of g are not particularly helpful, we probe this question further by 13 C NMR spectroscopy. Deuterium isotope effectscnicarbon chemical shifts have been used as a tool for the determination of solution structures.23 Saunders and co-workers24 have recently presented a new criterion for differentiating symmetrical molecules from systems that are asymmetrical but appear symmetrical on the NMR time scale due to rapid degenerate rearrangement. They have studied deuterium isotope induced splitting in the carbon chemical shifts of some deuterium substituted carbocations and found that the relative iso- topic splitting, 6/A (where 6 is the observed isotope splitting and A is the chemical shift difference estimated for the "frozen" equilibrium), was much greater for equi- librium processes than for resonance phenomena. In the equilibrium case, the energy surface has two minima sepa- rated by a barrier and one observes a splitting as a result of isotopic influence on the relative stability of the two minima. In the resonance case, there is only one minimum and one observes a splitting due to the change in a single structure perturbed over vibration upon isotOpic substitution. If the titanacycles presented here were in fact rapidly equilibrating alkylidene-olefin complexes, Ia and lb, 20 deuteration of one of the d-carbons should lift this degen- eracy and results in a significant shift of the other a- carbon. A significant shift is anticipated because of the large frequency difference between an olefinic carbon and an alkylidene carbon (vide infra). Consequently, labeled 25 26 13 (0,6-d2) 3, 4, and 5 were prepared and their C NMR ’b ’\J ’b spectra were measured. The pertinent results are collected in Table 3. In the 13C NMR spectrum (Figure 4) of a 4:1 mixture of g-dz and 3 (the non-deuterated complex was added as an internal reference), the deuterium—labeled carbon (C1) appeared as a 1:2:3:2:l quintet (JC_D = 22.9 Hz), 99.4 Hz (0.791 ppm) upfield from the corresponding resonance of the nondeuterated complex. An important finding was the reso- nance for C2 of é-dz, which appeared 7.6 Hz (0.060 ppm) downfield from the corresponding resonance of 3. Similar downfield shifts were also observed for C2 of 4—d2 and i-dZ' The other carbon shifts showed the normal intrinsic isotope effect.27 In é-dz, for example, C3 was shifted 30.5 Hz (0.243 ppm) upfield (a B-deuterium shift) from the corre- sponding resonance of 3. The observed splitting, 6 (per deuterium), in these systems ranged from 0.045 ppm for E-dz to 0.121 ppm for i-d2 (the structure which showed some distortion). Significantly, the observed splitting for i-dz remained unchanged when the temperature was lowered to -70°C. The expected (or static) shift, A, is estimated from a series of known metal alkylidene complexes (Group 4 and 5) and the only known Group 4 ethylene complex to be ca. 100 .zpwfifinsfiom 30H Op 636 wo>somno “oz n .N: m.on paw Ema noo.on Hospo Hmucoeflhmmxm .m can w 90% 2:5 93: $6538 28 m .86 2:5 cam: eeueceucee 3 2,322 362... $355 m N 3 can.on omw.-- mam.mn w-m mvo.o e Hmo.cn mmc.HH- Hmo.om on- .o m 1 N 8 Nwm.mn mov.HH n-- 6-? HNH.O 8 How.mn won.HH How.ms ca- v N a «Ho.wo wa.wfi mofi.no w-m ooo.o e wmm.no mmo.wH 6mm.no mm m @ .wcfiuufifiom wo>homno Nu mo _ Ho Uo.QEoH vasomEou AEQQV NomuHuHH mo mumwzm HmoHEogo u wo :oflpmnhsuhom oflQOpomH .m manmh m FIIIIL ma 22 e N a ou-o:o~:on en 60mm on m eee e-m me annexes Hue e we sseeeeem mzz onH em: e.m- .e enemHe e 3 .22 8. L 5 p b b b b b b b b h b b b \ f .2 NUQOUQAMU { N: 3N ... 3.. Emu 36-8.6 Q m...» N 2 m PJ\; 3-9m. :3 z 55 [H n e 4 ENG eeee :1 men 2. e; m m N w 23 Table 4. 13C chemical shifts for alkylidene and olefin complexes Compounds 6 (ppm) Reference CpZTa(CHZ)(CH3) 228 28 CpZTaCHZAlMe3 177 28 Cp2T1CH2AlClMe2 188 15 CpZTigHZAlMe3 204 15 (szTiEH2)2 236 29 * . Cp2T1(CZH4) 105 30 ppm (Table 4). Thus we obtain 6/A to be of the order 10’3- 10‘4 for these titanacycles. It is important to note that rather large changes in A will have a small effect on this relative splitting. Previously reported symmetrical systems such as the cyclohexyl H and cyclopentyl H cations have values of 3.5 x 10‘3 and 8.5 x 10-2, respectively, while unsymmetrical system such as the 1,2-dimethylcyclopentyl cation 9 has a value of 0.18.24 H D 82 82 O m 24 The allyl cations H and H serve as good models for the deuterium shifts in our cases. _The above observations are consistent with a symmetrical but easily distorted metalla- cycle structure. These structural studies suggest that metallacyclo- butanes involved in metathesis may be planar and symme- trical. If this is the case, the source of stereoselection in olefin metathesis20 may result from factors other than those arising from the conformational effects due to puckering of the intermediate. In this titanium system, the symmetrical metallacyclobutane appears to rest at the mini- mum of a broad potential surface, such that distortion toward the required transition state for metathesis is facile (vide infra). 25 3- BS29319029f-§-riEb,9151992:-§39203§§-¥§E§Eh€§i§ The successful isolation of metallacyclobutanes from a metal alkylidene-olefin reaction clearly represents a major advance in the study of the mechanism of the olefin meta- thesis reaction. However, it remains to be demonstrated that such complexes are indeed intermediates in a metathesis reaction. Treatment of a toluene-d solution of complex 3 with N 8 an equimolar amount of ClAlMe at room temperature resulted 2 in the gradual disappearance of the titanacycle as observed by 1H NMR spectroscopy. The resonances of 3 were replaced ’\J by those of the methylidene complex 1 and 3,3-dimethy1-l- butene. The conversion follows second-order kinetics, 4 -ls-l first-order in g and ClAlMeZ, with k = 9.8 x 10_ M 2 at 20°C. An identical experiment employing 3-d2, prepared from-l and 3,3-dimethyl-1—butene-1,1-d instead of 3 '\J 2, afforded a mixture of 1 (-d0 and -d2) and 2 (-d0and -d2) on treatment with ClAlMe2 as shown in equation 14. (D)H\ /H(D) Me CpZTi + c1A1Me2——> CpZTi~ CpZTi //Al -—e> CpZTi R + "" \\c1 \xMe““’ T *CH2=CHR 01-A1::E: * CpZTi<::>>——R + ClAlMeZ Cp Ti/ l//Al/ /Me Cl-Al/IM 2 \:H2 \Me 1 *\Me C Me ___1 // + ‘—-—- szTi/ Al —'_“ CPZTI<>‘R \ x\\\/// Me” CH2=CHR * CH2=CHR Figure 5. Mechanism for Lewis acid cocatalyzed degenerate olefin metathesis 28 Mimwewwwwwww ' ' ‘4 ' _ _ a C13 3 d1 / (E)‘ : (Z) - -d a: 2-d I WM ”$101011:th)‘XWWWWW‘MKJWM)WW,W.]¢Niyfix‘/aniMW (WWW? 2 . Figure 6. 76.8 MHz H NMR spectra of the reaction of trans-g-dl with AlMeZCl in toluene (*). 29 AlMe7C1 f j . t H \ c T B '7 m... - - (15) U NH! trans-[\3-d1 r2 (E)-,(Z)-r2J-d1 completely scrambled upon mixing with ClAlMeZ, giving equal amounts of trans- and cis-é-d1 before cleavage to 1 and l-dl and the Olefins. Another significant finding of this study was that the isomerization of trans-é—d1 requires only catalytic amounts of ClAlMeZ. These observations may be rationalized in terms of two possible mechanisms. In the first scheme (Figure 7), a fast, reversible transmetallation reaction occurs initially between trans—é-d1 and ClAlMeZ, yielding an intermediate A. Following transmetallation, A can undergo reversible inversion at the methylene group a to aluminum, and subse- quent ring closure accomplishes the observed scrambling. There are ample precedents for transmetallation reactions in Ti-Al and Zr—Al chemistry.31 It has been shown that in the case of Zr, transmetallation proceeds with retention of configuration at the exchange site, suggesting a four- centered transition state.32 Main group metal alkyls have been shown to undergo rapid inversion at the a-carbon by NMR spectroscopy.33 The inversion rate in the case of aluminum alkyls has been found to follow first-order kinetics. This has led to the proposal of a mechanism ( CpZTi:::>P‘FBU+- AlMeZCl :;:::::i CpZTi tBu % llgiAIH ‘ \\Me trans-é-d1 Me I 1 r inversion ______s <______. \ Me Me ) k A ’b r w ’ tB ' tBu Cp2T1 u CpZT + AlMeZCl \‘ H“ D C1-—Al D !\ i . * MeNW ClS-é-dl Figure 7. Proposed mechanism for AlMeZCl induced isomerization of trans-é-d1 31 involving the ionization of the aluminum alkyl to a carbanion and a metal cation.34 As an alternate to the above scheme, an n-allyl inter- mediate may be involved as depicted in Figure 8. This species can be formed via a reversible abstraction of the B-hydrogen of trans—g-d1 by ClAlMez. An nS-nl-n3 transfor- mation of the allylic-Ti bond with rapid rotation of the carbon-carbon bond in the n1 state would interconvert the syn and the anti groups. Subsequent ring closure would lead to the observed isomerization. This scheme is remini- scent of the synthetic routes to stable W and Mo metalla- cyclobutanes reported by Green7 (equation 5). C) 2 t + HClAlMe CpZTi Bu + ClAlMe ———=~ CpZTi 2 V‘— H trans-3-d ql/ m 1 + HClAlMe 29 C3 "an <3 U \ ’1 CpZTi tBu + ClAlMeZ D Cis-é-d1 Figure 8. Alternate mechanism for the isomerization of trans—3-d m 1 While both proposals may be speculative, we favor the first scheme on the basis of the following crossover experi- ment. A toluene solution containing equimolar amounts of 32 3-d and 3-d was treated with ClAlMe and the olefins m 0 '0 S 2 released were collected and analyzed by GC-MS. The results indicated that the conversion produces only g—d and é-d 0 3 and no crossover products (equation 16). D D D [Al] 7: D CpZTi + CpZTi -—————€> -—- + -—' D D D 2 2 d m m- 3 3 3-dS m N -d (16) The absence of crossover products in the above experi- ment seemed inconsistent with the second scheme but entirely consistent with the first. Thus, although by no means proven to be correct, the first scheme is the more appealing mechanistic sequence for the isomerization of trans—é-dl. As mentioned earlier, Lewis acids are important com- ponents in many active metathesis catalysts. It has been noted that the stereospecificities are often low in these cocatalyzed metathesis systems. Katz35 attributed this to the isomerization of the metallacyclic intermediates by the cocatalysts. He suggested that the Lewis acids facilitate cleavage of the metal-carbon bond in the metallacyclobutane, forming a 3-metallapropyl cation in which the C—C bond rotate, and consequently leading to scrambling of stereo- chemistry. Our results and mechanistic interpretation are compatible with this hypothesis. 33 4- Besstiees-9f-§-wi§b-9:91sfiss-eséiPinserlessEylses The reaction of g with aluminum alkyl discussed earlier provided direct evidence for the intermediacy of metalla— cyclobutanes in metathesis reactions. We have discovered that compound 3 is also an effective catalyst for the methylene exchange of monosubstituted terminal olefins. Thus, 3 catalyzes the degenerate metathesis of 3,3—dimethyl— l—butene—d (2—d ) and 3,3—dimethyl-l—butene—l,l,2—d3 (2-d3) O O as shown in equation 17. D —_4;%L 7: 3 _ :f + >:g< ‘43—: 1> + (17) D D 40°C D D 2—d g—d g—d 2—d O 3 2 1 All the intermediate labeled metallacyclobutanes and olefins are observable by NMR spectroscopy. For example, when an equimolar solution of 8 and 2—d2 was heated to 40°C and examined periodically by 2H NMR spectroscopy, the spec— tra (Figure 9) showed depletion of deuterated olefin and incorporation of deuterium into the cis- and trans—a—posi— tions of the metallacyclobutane (6 1.86 and 2.17 ppm, re- spectively) in equal amounts. Thus, the stepwise exchange of labeled and unlabeled olefins (equation 17) can be repre— sented by the catalysis cycle depicted in Figure 10. It is significant to note that the catalysis is accomplished without the addition of a cocatalyst. To our knowledge, this is the first example of a metathesis reaction catalyzed 34 t = 180 min t = 30 min A l M— 'l I l' ‘I I I I ‘_'*T t = O I i ' 'é' "as " ' 1 0 (cm) Figure 9. 76. 8 MHz 2H NMR spectra of the exchange reaction between 2- <12 and 3 at 40°C -|— ,. - van-o we: ‘0. or»- -— — . - .-.a ~ ‘\—--a .v-g...“"’" -—- - -~ . ”>26 DH D D D D CpZTi CpZTi D D D D __ :7< Figure 10. Catalysis cycle for the degenerate metathesis ' of 2-d0 and 2-d3 l————fi. « _ . 74.“... £2,513.41 36 by an isolable metallacyclobutane complex. The reaction of 3 with ethylene at room temperature resulted in the formation of the parent titanacyclobutane complex 2 (equation 18). szTi<>——*— + CH2=CH2 __\____,(__.A szTi<> +:/<(18) 5 '\/ Preliminary results indicated that the reverse reaction does not occur. In contrast, isoamylethylene reacted with 3 reversibly to yield an equilibrium mixture containing the complex 6 as shown in equation 19. szTi<>_+ + J—g-f‘ CpZTiO_‘_/+ J (19) Q In the presence of 1 equiv of PhCECPh, 3 afforded CpZTiCHZCPhCPh Z, a compound previously noted by Tebbe36 as F—-—_—_—_7 the product of the reaction between szTiCHZAlMe2C1 l and PhCECPh in THF (equation 20). 3 + PthCPh ——————€> Cp2T1<:;;>—Ph + ;==/;7< (20) Ph 7 ’\4 A proposed mechanism for the exchange reactions de- scribed above is shown in Figure 11. It involves an initial ring—opening of the titanacyclobutane 3 to a methylidene 37 h j CpZTi CpZTi=CH2 + ——— k—1 3 . B 2 m ’\J ’\.: k _ 2 . 2 ——__> E + PhC CPh szrl // Ph Ph 7 "\1 Figure 11. Proposed mechanism for the reaction of 3 with PhCECPh ” intermediate 8. This highly reactive and unstable species is then rapidly trapped by an incoming olefin or PhCECPh to yield titanacyclobutane or -butene, respectively. The scheme in Figure 11 is illustrated with PhCECPh as the substrate. The presence of E in this reaction system seems plau— sible in view of the thermal behavior of 3. Heating a toluene solution of 3 at 100°C for 1 h under vacuum resulted in the decomposition of the titanacycle and gave a solution containing products shown in equation 21. ‘ A J J CpZTl ————%> CH4 + CZH6 + —— + + >%——+- (21) 3 39% trace 57% 15% 6.6%/Ti The proposed mechanism provides a route to the observed metathesis products. In the absence of substrates, the methylene species 8 decomposed to CH4 presumably via 38 abstraction of hydrogens from the cyclopentadienyl rings.37 Also, when é—ds was mixed with an equimolar amount of the titanacycle 2 in benzene-d6, heated to 50°C, and examined periodically by 1H NMR spectroscopy, scrambling of protio— (or deuterio—) methylene units was observed (equation 22). The spectra showed resonances due to incorporation of protons into the a—positions of Q. The observation of 3—d3 is consistent with an intermolecular exchange of olefins between é—ds and Q proceeding through the intermediate 8. D D D . 4;______ CpZTl + CpZTf:::>>-—+—// D D 6 ’\I _ (22) Q d5 H D D D CpZTi + CpZTi D D Q'dz é'ds The proposed mechanism depicted in Figure 11 should follow the rate law shown in equation 23, derived by applying the steady-state approximation to the concentra— tion of the reactive intermediate 8. d[3] k1k2[PhCECPh][g] dt — k_1[ 4¢><1 + k2[PhCECPh] (23) 39 In the limiting case where k2[PhCECPh] >> k_l[ 4¢§>(], the equation simplifies to that shown in equation 24, which predicts that under these conditions the observed rate should exhibit a first-order dependence on the concentra— tion of 3 and no dependence on the acetylene concentration. (24) g] = kobsd [3] Thus, the process in Figure 11 reduces to a two step reac- tion in which the first is rate-limiting. Experimentally, this condition is satisfied by employing a large excess of PhCECPh. With the use of 1H NMR spectroscopy to monitor the reaction,kdnetics data were obtained at 40°C in benzene-d6. The reactions were all followed to at least three half- lives and displayed excellent first—order kinetics in agree- ment with the predicted rate law. A representative plot is shown in Figure 12. The measured rate constant kObsd at 40°C was (9.2 i 0.3) x 10"5 5-1.38 The rate remained constant upon varying the concentration of PhCECPh from 0.056 to 0.79 M. Rate data were also obtained at 30, 50 and 60°C (Table 5). The activation parameters calculated from a least—squares fit to the Eyring plot of-ln(k/T) versus l/T (Figure 13) are AH# = 26.9 i 0.7 kcal mol-1 and AS# = 8.9 i 2.1 eu. The large excess of PhCECPh employed in the above experiments was designed to trapped the presumed "szTi=CHZ intermediate as it formed, thereby preventing any back 40 mEHu wo coHuoch 6 mm w mo mwmuCQUme osu mo uoHQ on Housumz .NH muswfim owe com wmchHE ooN . _ OOH o . a (E %)uI . - emeo m X I H- mice N e x 6.0e n .8569 8 amomuzm :ufiz m we :ofiuomos 6:» new poHQ wcfiszm .mH ensued MOH x AH\HV mnm N.m H.m o.m m.N o.na- 41 _ _ _ _ 1.0.eH- 1,0.mH- I.o.va- lio.mH- llo.NH- _ _ o.HH- (i/X)UI 42 Table 5. Observed rate constants for the reaction of 3 with PhCECPh at various temperatures a o -1 Temperatures ( C) kobsd (s ) 30 1.8 x 10'5 40 9.2 x 10'5 50 5.9 x 10’4 60 1.1 x 10'3 a [g] = 0.19 M, [PhCECPh] = 0.79 M reaction (k-l) from occurring. To verify that the back reaction pathway was accessible, we next carried out a set ofkinetics runs using added excess olefin 2. Under these conditions, k_1[4¢>< ] is competitive with k2[PhCECPh], and the full rate law (equation 23) holds. At high [.49><] and [PhCECPh], these concentrations remain effectively constant during each run and allow determination of the psuedo-first-order rate constant kobsd. The reciprocal of kObsd is given by equation 25. k 1 _1_ + '1 [4¢><] (25) obsd 1 klkz [PthCPh] W This predicts that a plot of l/kObsd versus [49><]/[PhCECPh] should be linear having a slope equal to k_l/k1k2 and an intercept value of 1/k1. Experimentally, the presence of added olefin retarded the reaction rate. Still, the dis- appearance of 3 obeyed first-order kinetics (Table 6). The Observation that added olefin 2 inhibits the reaction ’\1 43 Table 6. Observed rate constants for the reaction of 3 with PhCECPh in the presence of added olefin“ 2 at 40°C, [3] = 0.015 M : S _1 [PhC_CPh],M [ALE 10 kobsd (S ) 0_59 0.17 8-8 0.59 0.38 8-2 0.59 0.77 7-6 provides strong evidence that this olefin is binding to the intermediate. This is consistent with Figure 11 if we assume the first step is reversible. Another possibility is that 2 could be reversibly forming an adduct with 3. However, such an adduct must form in appreciable amounts to account for the observed rate inhibition. This pathway can be discounted since no other products or intermediates were detected by 1 H NMR spectroscopy at any time during the reactions. A plot of the data according to equation 25 is displayed in Figure 14. From the intercept of this linear plot follows k1 = 9.1 x 10_5 5‘1, the rate of ring-opening. The ratio of the rates of trapping of E by olefin and -acetylene k_1/k2 = 0.15 is obtained from the slope and k1. It is possible to augment thekineticsevidence for the preequilibrium mechanism in Figure 11 by direct detection of the equilibrium itself. Thus, a toluene solution of 3 was treated with a four—fold excess of both g-dz and PhCECPh and heated to 40°C. After 140 min, the reaction mixture was examined by ‘H NMR spectroscopy. The spectrum exhibited 44 #64 6.04 seememe-eeeeeee as w eeeee gen: geomuee eee m we eeneeeez eee new x\H .ms Heeumueec\fiqst_ efleee exp we seem _;e6mueen\_oxu\1 o.H m.o .eH ensued GA 1701 X [quOX/I] «A 4S signals at 6 1.86 and 2.17 ppm, indicative of é'dZ' The approximate extent of D incorporation was determined to be 4% of the total [D]. Thus, both the kinetics and labeling studies are consistent with the proposed mechanism. Our mechanistic interpretation was further supported by the observation of a large secondary isotope effect when 3-d2 was allowed to react with excess PhCECPh. The ratio of z to 2 was 2.2:1.0 (:kHZ/kD7) as determined by integra- tion of the corresponding signals in the 1H NMR spectrum (equation 26). 7 "\1 Ph C k T 2 2 D HZ > CpZTi=CH2 + “‘ D 2 D2 B-d2 2 "J (\J lphzc, z’dz 'Fhe magnitude of the isotope effect is consistent with a net decrease in p character of the deuterium—substituted carbon - . . 39 111 the trans1tion state (sp3 methylene + sp2 carbene). Sintilar reactions with trans-3—d , or trans-3-h (d ) m 1 m 1 4 Ylelxied isotope effects of kH /kHD = 1.4 and kHD/kD = 1.9, 2 2 TBS}? ectively.40’4l’ 46 H D D CPZTiN—‘i— CpZTi D H H D trans—g—d1 trans—é—h1 In these exchange reactions, the complexes trans-é—d1 and trans—i—hl stereospecifically released the olefins from which they were synthesized, (E)—3,3—dimethyl—l—butene—l—d1 ( E i_dl) and (Z)-3,3—dimethyl-l-butene—l,Z-dZ, respec— tively. The kinetics and labeling experiments have presented evidence for the existence of'theequilibrium step in accord with the proposed mechanism. We werethen able to test this conclusion by examining the stereochemistry of a metathesis reaction. Our mechanism proposes that olefin 2 dissociates from the CpZTi=CH fragment, and this species is then 2 rapidly trapped by a substrate. We wanted to determine the stereochemical outcome of a metathesis reaction between trans—é—dl and excess 2 under conditions identical with the PhCECPh exchange. Since the methylidene complex CpZTi=CH2 is symmetrical and incapable of retaining any stereochem- ical information, the exchange is expected to afford the olefin (E)-2-d1 and the titanacycle cis—é—d1 in the ratio of 2:1 as outlined in Figure 15. The reactions between trans—é—dl and excess 2 were run in toluene in NMR tubes sealed under vacuum. The reactions were monitored by 2H NMR spectroscopy and the relative 47 k HD H2 \p_-j;%: 7L D ——— + CpZTi=CHD CpZTi=CH + ‘—’ 2 (”é-<11 Figure 15. Metathesis of trans-é—d1 with 2-d0 48 amounts of products were determined by integration of the corresponding signals. After 1 h at 40°C, the 2H{1H} NMR spectrum showed the signals due to trans-é—dl, cis-é-dl, (E)-g-d and (Z)-2-d1 at 6 2.17, 1.86, 4.84, 4.91 ppm, 1, respectively. The ratio of (E)—2—dl to cis-3-d1 was 3.4:1. ’\1 Since the presence of (Z)—2-d1 could only have arisen from 8 exchange of cis—é-d1 with 2-d0, we must therefore also allow for this secondary reaction in figuring the concentra- tion of cis—é-dl. In addition, it is necessary to account for the loss of deuterium due to the formation of trans—g-d1 (Figure 15, path A), albeit this correction is minimal. Thus, we obtain the following expression for the total concentration of cis-é-d1 (equation 27). [cis-g-dlitot = [cis-g—dli + Hug-<11] + emu—2,91] (27) Using this equation, we calculate a new ratio of 2.6:l.0 for [(E)-%-dl]/[cis-é-dl]tot. However, this value is based on the assumption that no deuterium isotope effect was operating. It should be recalled that earlier it has been determined from the PhCECPh exchange with trans-(é—d1 that kHz/kHD = 1.4. Hence it is necessary to include this ratio in the calculations. For this purpose, the following rate expressions (equations 28 and 29) which have been derived from Figure 15 were employed. 49 MW) gal] = k [trans—3—d ] (28) dt H2 q. 1 d[cis—3-dl] kC ——~——¥3——~— = kHDItrans-3—dl][———————) (29) dt ” kC+ kt Dividing equation 28 by equation 29 gives an equation representing the relative amounts of (E)-%—dl and cis—[é—dl (equation 30). d[(E) 1 1 = Z ( C t ) d (30) d -2- ,\.I d[C1S_§' 1]tot After inserting the experimental value of 2.6 into equation 30, it can be rewritten as and from this follows kt ——— = 0.9. kc This ratio indicates that the olefin exchange proceeds with little, if any selectivity in agreement with the proposed scheme. 50 Additional and perhaps even more convincing evidence concerning this aspect of olefin exchange was provided by the results of a similar metathesis experiment carried out with trans-3b-d1 and 3,3-dimethy1—l-pentene. This reaction, as outlined in Figure 16, is expected to give a 1:1 mixture of cis- and trans-Q-dl. Our plan was to quench the reaction with PhCECPh which we know displaces the olefins with retention of stereochemistry. The products can then be analyzed for the stereochemical fate of the label since the resonances of the methylenic deuterium are distinguishable by 2H NMR spectroscopy. Another advantage of this experi- ment is that no correction for the deuterium isotope effect is needed. The metathesis experiment was carried out by allowing trans-é-dl to react with a ten-fold excess of the olefin in toluene at 40°C for l h, after which the volatiles were removed and excess PhCECPh was then added, and the mixture was heated for 12 h at 40°C. The olefins evolved contained a mixture of 2 (dOand d1) and 3,3-dimethyl-l-pentene (d0 and d1) as determined by GC/MS. The C6 and C7 olefins were separated by preparative gas chromatography. The 2H{1H} NMR spectrum of the 3,3-dimethyl-1-pentene displayed two signals of approximately equal intensity at 6 4.89 and 4.92 ppm. In the proton-coupled 2H NMR spectrum, two overlapping doublets were observed. The peak centered at 6 4.89 had J(H-D) = 2.56 Hz, and the peak at 6 4.92 had J(H-D) = 1.59 Hz. From the magnitude of the coupling constants, the ('3 T5 N v—i Q (‘f CU Cl 366 H *‘S 9) :3 m I 8N l D... H / _ _ \ trans 3 d1 kHD D 8'd1 g’do ktrans kcis V Cp2T{:::>—-i-amyl CpZTi ivamyl 13 D trans-Q-dl Cis-Q-d1 Figure 16. Olefin exchange reaction of trans-é-d1 with 3,3-dimethyl-l-pentene 52 6 4.89 and 4.92 peaks were assigned to (Z)- and (E)-3,3- dimethyl—l—pentene-l—dl, respectively. Examination of the C6 olefin 2 by 2H NMR spectroscopy revealed the presence of only a trace of cis—g-dl, indicating that only minimal amount of isomerization of the starting trans-g-dl had occurred during the exchange reactions. These results, coupled with previous observations, are most consistent with the olefin exchange occurring via the intermediacy of the methylidene complex. In summary, the titanacyclobutane complex 3 provides the first example of an isolable metallacyclobutane which exhibits metathesis activity, albeit its scope is limited to the exchange of methylene groups of41—olefins due to steric constraint. The studies detailed above have shown that g reacts with a-olefins and PhCECPh by a dissociative mechanism involving a high energy intermediate methylidene species "CpZTiCHZ". This titanium metathesis system may be contrasted with the well-defined Ta and W catalyst systems 14 In these systems, it is the metal reported by Schrock. alkylidene that is the lower energy intermediate. Finally, there is an alternate mechanistic pathway for olefin metathesis for systems in which the dominant intermediate la,c The key feature in is the metallacyclobutane complex. this scheme, as depicted in Figure 17, is that the reactive alkylidene species is present as a transient intermediate and is formed only when a suitable trapping substrate is available. The first step involves the coordination of an S3 R R R' T__/ ..___A M R' + RHC=CHR ‘—— 1"4—\R. R' RHC=CHR RHC=CHR' W M=CHR' I RHC=CHR K / 8' RHC=CHR' M ’/R , __;. M R + RHC—CHR V... I R \R R Figure 17. Bimolecular mechanism for olefin exchange 54 olefin to the metallacyclobutane, followed by ring—opening to a bis(olefin)-metal-alkylidene intermediate. For coor- dinatively unsaturated systems, this associative mechanism is an attractive pathway for olefin metathesis. This scheme was initially considered as a viable alternative to the free alkylidene mechanism (Figure 11) for the bis(cyclo- pentadienyl)titanium system discussed in this work. There is perhaps one unfavorable situation which arises in the case of a bis(olefin)—titanium—methylidene intermediate, such a complex would be formally a 20-electron species. In principle, however, an nS-to-n3 slippage of a cyclopenta- dienyl ligand could alleviate this problem. But, kinetics evidence points to a dissociative mechanism rather than an associative scheme, and stereochemical studies also supports the involvement of an olefin-free alkylidene species. 55 5. Reactivity of the Parent Titanacyclobutane 5 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The parent titanacyclobutane complex CpZTiCHZCHZCH2 5 was prepared according to the procedure described earlier for the t-butyl derivative 3. This complex proved to be much more sensitive thermally than compound 3. Whereas 3 can be handled in the solid state at room temperature for long periods without any noticeatde decomposition, complex 5 decomposes gradually at the same temperature over a period of ca. 1 hour, and is best handled at or near 0°C. In solution, however, the complex is moderately stable for extended periods at room temperature in saturated or aromat- ic solvents. Treatment of 5 with BrZ in diethyl ether at -65°C affords 1,3-dibromopropane and CpZTiBrZ. Similarly, treatment of 5 with 2 equiv of anhydrous DCl results in the immediate liberation of propane-1,3-d2 and generation of the dichloride CpZTiCl2 (equation 31). 2 g A + CpZTiBrZ Br Br Br CpZTi (31) 5 DCl g//A\\¥ + Cp2T1C12 ’b l The H NMR spectrum of i at either 90 or 500 MHz displayed only one signal at 6 5.22 ppm for the two cyclo- pentadienyl rings, indicating that they are equivalent, (Figure 18shows the 90 MHz spectrum of 5 in toluene-d8). This would arise if the metallacyclobutane ring is planar. Another possibility is that a dynamic puckering motion is 56 L 1‘ ’2 1 L 1.1. I I 1 I l 1 L I 1 7 6 5 2 1 O -1 (mm) Figure 18. 90 MHz 1H NMR spectrum of 5 in toluene-d8(*) a. Expansion of Ha and HB region S7 . . . . . 42 occurring in solution as shown in equation 32. < /\ /\ Q /1V \____\/Tl© _ T1 (32) \ \/ This ring—flip process, which also equilibrates the four respective a—protons (6 2.95 ppm) and the two respective B-protons (6 —O.l6 ppm) in the temperature range -90 to 250C, must be fast on the NMR time scale. Close examination of the resonances for the trimethylene unit (a. of Figure 18) reveals the average cis and trans Ha/HB coupling constants to be very different. Hence, instead of a simple A4X2 spectrum,43 rather complex non-first-order splitting is observed. The spectrum of such cyclic —CHZCH2CH2- system should in principle be assignable in terms ofeniAA'XX'AA' spin system (oran1AA'XX'A"A”‘ system if there is cross—ring 44 No attempts were made to analyze this six-spin coupling). system. We did, however, examine a simpler case, the 0,0— deuterated titanacycle szTiCHZCHZCDZ 2-d2° The methylene resonances consist ofeniAA'XX' pattern as shown in a. and b. of Figure 19. In the AA' half, ten of the twelve theore— tical lines were observed (a. of Figure 19).45 The XX' half was unfortunately broadened by the quadruple moment of the a—deuterons, and appeared as a broad triplet. The AA' half was analyzed using a simulation program to give the Coupling constants to generate the spectrum shown in c. 0f Figure 19. Within the limits of visual fitting, the 58 D D H Cp Ti 8 2 H B H H (1 (X a . b b . WU I I I I I I I I | '7 340 320 60 40 Hz I 340 I 310 I Hz Figure 19. Experimental (a. and b.) and calculated (c.) spectra of Ha (a. and c.) and H8 (b.) of é—dz (See Experimental Section for Simulation parameters, (*)impuritY) r-_.v 59 simulated spectrum matches well with the experimental one. The thermolysis reaction of 5 was next examined to determine if this complex would follow paths similar to those shown by the t-butyl derivative 3. Recall that 3 yielded predominantly carbon-carbon bond cleavage products ‘on thermolysis, this despite the presence of a B-hydrogen. 15 Note, also, that Tebbe has reported that ethylene reacts fl with CpZTiCHZAlMeZCl l to give propylene, presumably via the intermediacy of complex 5. Thermolysis of a toluene solution of 5 at 60°C under vacuum for l h resulted in the complete decomposition of the complex and yielded the products shown in equation 33. +CH+/\+/\ (33) CpZTi . -——--9 CH4 2 6 5 7.8% 12.3% 50.4% 29.3% 'b Similarly, heating a solid sample of 5 under vacuum at 50°C yielded methane (2.2%), propane (78.3%), propene (19.4%), but no CZ-hydrocarbons (ethane or ethylene). On the basis of the product distribution in these experiments, it is apparent that decomposition via B-H transfer and olefin elimination represents a lower energy pathway for 5 than for 3. The production of the saturated hydrocarbons may be rationalized in terms of secondary reactions involving ”Ti-H" species resulting from the decomposition.46 Although the factors influencing the partitioning of the trimethylene unit of 5 between B-H elimination and C—C bond cleavage on thermal decomposition are not immediately 6O obvious, puckering of the metallacyclobutane ring is almost certainly necessary in order for a B-hydrogen to transfer to the metal as illustrated in equation 34. CpZTiAH—é (min-V flCpZTi—IK (34) H H K/ Earlier, we have suggested that complex 5 might be a floppy molecule in solution. Recently, Rappe and Goddard22 con- ducted an ab initio calculation on the hypothetical compound ClzTiCHZCHZCH2 and found that the energy to pucker the ring by 10° is only 0.3 kcal/mol. If ring puckering is rela- tively easy in these systems, the puzzling question then is why the t-butyl complex undergoes C-C bond cleavage predomi— nantly. A plausible explanation is that steric interaction between the bulky t—butyl group and the cyclopentadienyl ligand in the intermediate n3—a11y1 complex may have ren— dered the B—H transfer pathway energetically unfavorable. If indeed puckering of the metallacycle ring favors the B-H transfer reaction path, then it is not unreasonable for one to speculate that a planar conformation favors C¢C bond cleavage. This conformation provides the best geometry for reaching the required transition state for the cleavage of the metallacyclic ring as the n-bonding orbitals of both the olefin and the methylene fragment are c0planar.47’21 61 That the reaction chemistry of complex 5 is dominated by B-—H elimination is best illustrated by its ”reactions" with ethylene-d4 and-d0. When a toluene-d8 solution of 5 was sealed under C2D4 (~4 equiv), allowed to stand at room temperature, and observed periodically by 1H NMR spectro- scopy, the signals due to 5 decreased in intensity. This was accompanied by the gradual appearance of a new signal at 6 5.88 ppm along with the resonances due to propylene and traces of propane. No ethylene-d0 was detected. The general features of the spectrum remained unchanged over the course of the reaction, but slight broadening of the signals was noted, suggesting the presence of small amounts of paramagnetic species, presumably derived from the decompo- sition of 5. When a similar experiment was carried out with C2 4, the 1H NMR spectrum displayed two broad multiplets at 6 1.78 and 1.22 ppm in addition to the signals described above. Traces of C2H6 was also detected. Integration of the final spectrum after all the starting complex had dis— appeared showed the ratio of the resonance at 6 5.88 ppm to the two new signalstolxaca.10:4:4. On the basis of these spectral observations, the species responsible for the NMR resonances was assigned a structure containing the titanacyclopentane moiety. CpZTi 800 62 The titanacyclopentane complex 8 has been previously 48 as the product of the noted by Whitesides and co-workers reaction between 1,4—dilithiobutane and CpZTiCl2 at low temperature, and also between ethylene and functional equiv- 46 Unfortunately, no NMR data for alent of "titanocene". this complex was available. To assure ourselves that 8 was indeed the species we observed in the NMR tube reactions, the following experiments were performed. A toluene solu- tion of 5 was stirred under 1 atm of ethylene at room temp- erature for 48 h, after which the mixture was cooled to -30°C and the excess ethylene was removed, and anhydrous HCl was then introduced. The volatiles thus evolved was analyzed by GC-MS. The major product was found to be n-butane. Similarly, carbonylation of a toluene solution containing the presumed titanacyclopentane 8 yielded a yellow compound which exhibited an IR absorption at 1614 cm-1. This species decomposed rapidly in CHZClZ at room temperature to give cyclopentanone (GC-MS). Thus, both of these reactions afforded products characteristic of the 1,4-tetramethylene moiety. A sequence which summarizes these results and the formation of 8 from 5 and ethylene is depicted in Figure 20. We believe the formation of 8 began with the decomposition of 5 via B-H elimination, generating a propylene complex. In the presence of C2H4’ the propylene was displaced, yielding initially a mono- (ethylene)-titanium species and, ultimately, the titana- cyclopentane 8, presumably through the intermediacy of a _ 2 CpZTi ————€> Cp2T1——‘ -——__+> Cp2T1_il . ./\ CpZTi <5~ €> Cp2T1\‘ + y 80 FigureZO . Formation and reactions of titanacyclopentane 8 bis(ethylene) complex. Complex 8 has been reported to be extremely sensitive thermally and decomposes to give mainly ethylene. Our ability to observe 8 spectroscopically may be attributed to the presence of the excess ethylene which stablizes 8 toward decomposition. In the carbonylation reaction, we obtained a yellow compound which we formulated as the 2-titanacyclohexanone complex 8 on the basis of the IR data. The observed vCO (1614 cm_l) is consistent with a 1.49 48 has reported a tran- sient species which has an IR band at 1720 cm—1, attributed bihapto acy Earlier, Whitesides to a complex believed to be 8. This result is somewhat puzzling in view of the much lower C=O stretching frequen- cies observed for other Ti and Zr acyl complexes.SO It is interesting to note that the VCO for cyclopentanone is at 1745 cm—1. 64 A number of other reactions for 8 have been investi- gated briefly. Treatment of a toluene solution of 8 at 40°C with excess PhCECPh yielded a dark green solution containing the known metallacyclopentadiene complex 88.51 Ph Ph \ CpZTi / Ph Ph 10 ”VD _GC analysis of the gas phase above the solution showed: propylene (89%), propane (3.4%), methane (5.6%) and traces of ethylene. Addition of ClAlMe2 to a toluene solution of 8 at room temperature resulted in the decomposition of the complex and formation of organic products similar to those obtained in the PhCECPh reaction. The organometallic products were 1H NMR spectrum of the not easily characterizable; the reaction mixture showed only severely broadened peaks. At no time was the methylidene complex 8 observed. The differences in the reactivity of 8 with ClAlMe2 and PhCECPh (vis-a—vis 8) are attributable to the facile B-H elimination reaction discussed earlier. Interestingly, however, it was discovered that 8 reacts with acetone in a similar fashion as the t-butyl derivative 8.38 Heating a toluene solution of 8 at 40°C in the presence of 1 equiv of acetone afforded ethylene (64%) and isobutylene (27%) as the major products. Some propylene and propane were also 65 produced (~6%). In the case of g, the reaction with acetone proceeded cleanly according to equation 35.38 O epzriO—k + A—az/ix + [CpZTi(O)]n (35) Hence, both titanacycles g and 3 underwent fragmentation reaction to give neohexene and ethylene, respectively. In both cases, methylenation of the organic carbonyl functions also occurred, yielding Wittig-type products. The organo- metallic product was the polymeric titanocene oxide.52 Formation of this stable oxide must have contributed greatly to the driving force of these reactions. The mechanism for the methylenation of acetone involving é probably consists of at least three distinct steps.53 The first step may simply be the formation of an adduct in which the carbonyl oxygen is bonded to titanium. This is probably followed by the formation of an intermediate oxytitanacyclobutane complex %%.54 Subsequent metathesis then yields the olefin, leaving the oxygen strongly bound to Ti. 0 66 The small amount of C3-hydrocarbons present in the organic products indicated that thermal decomposition of g via B-H elimination could not be completely prevented during the reaction. During our study of the carbonylation of the titana— cyclopentane 8, discussed earlier, the reaction of complex g with CO was examined briefly. Complex é reacted quickly with CO. The carbonylation was complete within a few minutes at room temperature, giving a dark purple solution from which a purple microcrystalline product could be iso— lated. Monitoring the reaction in sealed tubes by 1H-NMR spectroscopy revealed the decay of the signals attributed to g concmitant with the appearance of resonances at 5 5.53(s), 2.31(t), 1.87(q) ppm, in the ratio of ca. 10:4:2. A small amount (~5%) of szTi(CO)2 was also detected ( szT1 \0/ ‘ J ”I Cp Ti lll" 2 \o 12 mm Figure 21. Proposed mechanism for the formation of the enediolate complex 12 ’b’b 69 is expected to involve an intramolecular dimerization, yielding the enediolate complex lg. Earlier Bercaw and co—workers57 have reported that bis(alkyl)zirconium derivatives undergo facile reactions with CO to give final products in which the carbonyl oxygen is bonded to zirconium. For example, Cp*ZZr(CH3)2 reacts with CO to give initially Cp*ZZr(CH3)(CH3CO), which then undergoes further carbonylation to afford only the purple enediolate (equation 36). CO Cp*ZZr(CH3)2 .rEE:E__ Cp*2Zr(CH3)(CH3CO) (36) CO * /0 CH3 _ o GC 22r\ | /O C O The observed patterns for the reactions of alkyl deriva- tives of zirconium with CO have been attributed to the carbenoid character of the carbonyl carbon resulting from the unusual ”side-on“ coordination of the acyl group. The reactivity of g with CO may be similarly rationalized (vide supra). The fact that g reacts with CO to yield the enediolate complex 1% as the final product while the metallacyclo- pentane 8 reacts with CO to give the six-membered titana- cyclohexanone suggests that the intermediate five-membered titanacyclopentanone may be sufficiently strained to 70 encourage a second insertion of CO into the other Ti—C bond. In summary, significant differences, but also some similarities have been observed between the behavior of the t—butyl substituted titanacyclobutane complex 3, discussed 9m? earlier in this work, and the parent compound Whereas 3 undergoes carbon-carbon bond cleavage reaction, giving the reactive methylidene complex 8 and neohexene, é undergoes s—H elimination yielding propylene, followed by reductive reactions involving titanium hydrides. As noted before, the s—H elimination reaction most likely representserform of termination step in a metathesis reaction. In the presence of ethylene and PthCPh, the metal fragment "CpZTi” was trapped to form the titanacyclopentane 8 and the titana- cyclopentadiene 19, respectively. Presumably driven by the high oxygen affinity of titanium, both 3 and g reacted with acetone to yield the Wittig-type product, isobutylene. The reactivity of these titanacycles with organic carbonyl functions may find impor- tant applications in organic reactions. In this context, it is interesting to note that if functional groups such as organic carbonyls are present as impurities in metathesis reactions, they may lead to the destruction of the catalysts via formation of stable metal oxide as observed here.58 Finally, while not related to olefin metathesis, but' interesting nevertheless, is the reaction of g with CO. The final product of this reaction is an enediolate complex, 71 derived from the coupling of carbonyl carbon atoms. The formation of 12 may best be explained by invoking an uncon— ventional n2 bonding of the acyl group to the " CpZTi" moiety. Perhaps most intriguing is the contrast between the behavior of titanacyclohexanone 2, obtained from the reaction of the titanacyclopentane 8 and CO, and the criti— cal intermediate "titanacyclopentanone" outlined in the mechanism (Figure 12) for the carbonylation of §. The latter apparently undergoes a second insertion extremely rapidly even below room temperature. It was suggested that perhaps the ring strain in the five—membered complex is enough such that a second insertion is facile. EXPERIMENTAL General Considerations ~~~~~~~~~~~~~~~~~~~~~~ All manipulations of oxygen and/or moisture sensitive materials were conducted either under argon (purified by passage through BASF RSvll catalyst and Linde 4A molecular sieves) using standard Schlenk or vacuum line techniques, or in a nitrogen filled Vacuum Atmosphere Dri-Lab equipped with a M0-40-1 Dri-Train and a DK-3E Dri-Kool. Routine 1H nuclear magentic resonance (NMR) spectra were recorded on either a Varian EM-39O or a JEOL FX9OQ 13 2 (Fourier Transform mode) spectrometer. C and H NMR spectra were recorded on either the JEOL instrument (22.53 MHz, 13c; 13.76 MHz, 2H) or a Bruker WM-SOO (500.13 MHz, 1 13 1 H; 125.7 MHz, c; 76.76 MHz, 2H) spectrometer. H NMR spec- tra were reported in units of 6 (ppm downfield from tetra- methylsilane) but were most often measured relative to the residual lH absorption in the deuterated solvent: benzene- d6 (7.15), toluene-d8 (2.09), THF-d8 (1.73), methylene chloride—d2 (5.32). 13 C NMR spectra were also measured relative to a solvent absorption: benzene-d6 (128 ppm), toluene-d8 (137.5 ppm) and reported in units of 6. Infrared spectra were recorded on either a Beckman 4240 or a Perkin Elmer 257 spectrophotometer. GC-MS 72 73 analyses were obtained using a Kratos M825 GC~mass spec- trometer. Analytical gas—liquid chromatography was conducted on a Varian Aerograph Series 1400 GC equipped with a Flame Ionization Detector and interfaced with a Spectra Physics Autolab System 1 computing integrator. The columns used were (1) 8' x 1/8” Porapak Q, 80/100 mesh; (2) 17' x 1/8" Durapak, 100/120 mesh; (3) 10' x 1/8” FFAP, 19% on Chromo- sorb G. Preparative gas-liquid chromatography was per- formed on a Varian Aerography Model 920 GC. The columns used were (1) 8-1/4' x 1/4" Durapak, 100/120 mesh; (2) 10' x 1/4” 5% SE-30 on Chromosorb W, AW, 60/80 mesh. Tetrahydrofuran (THF), diethyl ether, toluene, hexane and pentane were distilled from dark purple solutions of sodium/benzophenone ketyl under vacuum. Before distilla- tion, both hexane and pentane were stirred for 24 h each over two portions of concentrated 112804 and two portions of saturated KMnO4 in 15% H2804, washed with two portions of two addi- distilled water, one portion of saturated Na C0 2 3’ tional portions of distilled water, and dried over anhy- drous MgSO4. Dichloromethane was dried over P205 and deaer- ated on a vacuum line. Pyridine was dried over KOH pellets and deaerated by freeze-pump-thaw cycles prior to use. Deu- terated solvents such as benzene-d6, toluene-d8 were dis— tilled from sodium/benzophenone ketyl solutions and stored under vacuum in vessels equipped with Teflon stopcocks. Titanocene dichloride was obtained from Strem Chemicals 74 and purified by Soxhlet extraction with dichloromethane under N2. AlMe3 was purchased either as a 2 M solution in toluene from Aldrich or as a neat liquid from Alfa and used as received. Zirconocene dichloride, D20 (99.7 atom %D), n-butyl lithium, lithium aluminum deuteride (LAD) from Aldrich; ethylene, propylene and C0 (CP grade, Matheson Gas Products); and ethylene—d4 (99.6 atom %D) from KOR IsotOpe; AlMeZCl (Texas Alkyl) were used as received. Diphenylacetylene (Aldrich) was recrystallized from heptane. Dimethylaminopyridine (DMAP) from Aldrich was recrystal- lized from benzene. 3,3-dimethy1—1-butene (Chem SampCo); 3-methyl-1—butene (Aldrich); 3,3-dimethy1-1-pentene (Chemi- cals Procurement) were stored over Linde 4A molecular sieves and deaerated on a vacuum line prior to use. Methyl iodide-d3 (Merck or Aldrich) was dried over 4A molecular sieves and stored over a drop of Hg. Preparation~pf cpZTECHZAineZCH 1 This procedure represents a modification of Tebbe's 15 original preparation. To 20 g (80.4 mmol) of recrystal- lized CpZTiCl in a 250 ml Schlenk flask was added 80.4 ml 2 (160.8 mmol) of a 2 M solution of AlMe3 in toluene. The CpZTiC12 dissolved immediately to give a homogeneous dark red solution and evolution of methane began. The reaction mixture was allowed to stand at room temperature for 30 h, after which all the volatiles were removed in vacuo. The resulting dark red solid was washed with pentane (2 x 35 m1) and dried under vacuum briefly. This crude material 75 2A1C1MeCl (the -CH2- resonance of this complex appears at 67.66 as an unresolved was assayed for the presence of CpZTiCH AB quartet). This complex could be converted to l by addi- tion of an equivalent amount of AlMe3 to the crude product, which was first redissolved in 80 m1 of toluene. The resulting solution was then passed through a short pad of Celite supported on a coarse frit, and the filtrate was concentrated in vacuo until crystallization occured. At this time, 80 ml of n-pentane was carefully layered on top of the mixture, and allowed to stand at -20°C for a day. The supernatant was then removed via cannula. The red ij 2C1 was washed with several portions of n-pentane at -10°C. The solids were dried under high crystalline CpZTiCHZAlMe vacuum overnight (Yield: 14.5g, 63%). A second crOp of product (2g) was recovered by concentrating the mother liquor as described above. 1H NMR (C6D6), 68.29 (s, 2H, -CHZ-), 5.66 (s, 10H, Cp's), -0.29 (s, 6H, A1(CH3)2). CpZTi-CDZAlMeZCl was prepared as described above except A1(CD3)3 was used instead of A1(CH3)3. Iizsrzeétémf-eezfietzemaeztez..é 1.5g (5.27 mmol) of 1 was mixed with 0.8 ml (6.22 mmol) of 3,3-dimethy1—l—butene in 10 m1 of toluene in a Schlenk tube at room temperature. 645 mg (5.27 mmol) of DMAP was then added, and the resulting mixture was stirred for 10 minutes. The solvent was removed in vacuo, and the red residue was extracted with 40 m1 of n-pentane. The. 76 resulting red solution was passed through a short pad of Celite supported on a coarse frit to remove the DMAP-AlMe C1 2 adduct. The filtrate was concentrated by removal of the solvent in vacuo, and cooled to -50°C. The crude product thus obtained was further recrystallized from diethyl ether to give red needles of 3 (785 mg, 54%). 1H and 13C NMR data are compiled in Table 1. Anal. Calcd. for C17H24Ti: C, 73.91; H, 8.76; Ti, 17.34. Found: C, 73.88; H, 8.81; Ti, 17.44. Molecular weight determined cryoscopically (benzene): 292 (calculated 276).16 To a frozen solution of 43 mg (0.156 mmol) of 3 in 0.5 m1 toluene was added 0.312 mmol of anhydrous DCl (generated from NaCl and D2S04) at 77 K. The mixture was warmed to 25°C. After 3 h, the volatiles were removed by vacuum transfer. 0.148 mmol (0.95 mmol/mmol 3) of 2,2,3—trimethyl- butane—d2 was collected and identified by GC-MS analysis. Trans-Q-dl, 3—d2 and trans—Q—h1 were prepared similarly as described above with use of 3,3—dimethy1—1—butene—1—d1, —1,1-d2 and —1,2—d2, respectively. 3—d5 was prepared from l-d and 3,3—dimethyl—1—butene—1,1,2—d 2 3° FE§B§E§EESE-9§-£§Z:§zEZQETEEEZ}:}:PEE§E§:Z:§1 The procedure of Schwartz and Carr32 was followed to prepare the deuterated olefins. To 8.0 g (31.0 mmol) of CpZZr(H)Cl suspended in 20 ml of toluene was added 3.8 m1 (31.0 mmol) of 3,3-dimethy1butyne. After stirring for 8 h, the reaction was quenched by addition of 0.6 m1 of D20. All the volatiles were vacuum transferred into another vessel. 77 The deuterated olefin was purified by preparative glc 1 (column 2, 100°C). H NMR (C6D6), 65.78 (1H, d of t, JHH= 17.6 Hz, JHD = 1.8 Hz), 4.87 (1H, d, J 0.95 (9H, 5). HH = 17.6Hz), 3,3-dimethyl-l—butene-l,l-dz, g-dz, was prepared simi- larly with use of 3,3-dimethylbutyne-1-d1 and CpZZr(H)C1. 3,3-dimethyl-l-butene-l,1,2-d3, 2-d3, was prepared from 3,3-dimethylbutyne-l-d1 and CpZZr(D)Cl as described above. Isotopic perturbation of C NMR chemical shifts ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ A 120 mg (0.432 mmol) sample of é-dz and 30 mg (0.109 mmol) of 3-d0 were weighed into a 10-mm NMR tube, and 13 dissolved in 1.8 m1 of benzene-d6. A C {1H} NMR spectrum (127.8 MHz) was recorded. The observed splitting 6 (per deuterium) was determined from the difference of the chemi- cal shifts of C of 3-d and 3—d ’b ’b 13 2 0 Compounds 4 and 5 were thermally labile, and the 2‘ C NMR spectra were recorded in toluene—d8 at -10 and 0°C, respectively. An NMR spectrum of 5 and é-dz was also recorded at -70°C. No apparent change in the chemical shifts was observed. The results are tabulated in Table 3. Reaction of 3 with AlMeZCl 20 mg (0.072 mmol) of 3 was weighed into a 5-mm NMR tube. 6.7 p1 (0.072 mmol) of AlMeZCl was added by syringe 1H NMR apectra recorded along with 0.4 ml of toluene-d8. over the next 2 h showed the disappearance of g and the appearance of l and 2. Measurement of the rate of this 78 reaction with 1.0, 1.5, and 2.0 equiv of AlMeZCl indicated that the conversion proceeds with second-order kinetics, first-order in 3 and AlMe Cl. Second—order plots of the 2 disappearance of 3 versus time gave a second-order rate constant k2 = (9.8 i .2) x 10.4 M.1 s.1 at 20°C. An identical reaction employing é-dz instead of g afforded a mixture of 1 (<10 and d2) and 2 (d0 and d2) as determined by 2H NMR spectroscopy. A small secondary deute- rium isotope effect was observed in this reaction; the ratio cxf 1-d0 to g-do was 1.4:1.0. A similar reaction was conducted with trans-[é-d1 and AllAeZCl. 1H and 2H NMR analysis revealed that the stereo- cfliendstry of the q-carbon was completely scrambled before clxeavage to l-do and -d1 (Figure 6). Similar results were obsserved when trans-é-h1 was treated with AlMeZCl (see disscussion in text). §I£A§§9ysr-seBEIEP§93-P§EE§§P-§-§99-§:§5 A 15 mg (0.054 mmol) sample of é-do and 15 mg (0.055 nmual) of é-d5 were weighed into a small Schlenk tube. 10.2 01 (0.11 mmol) of AlMeZCl was added via syringe along with (L 35 m1 of toluene. After 2 h at 25°C, all the volatiles Vfirre vacuum transferred into another vessel for GC-MS ENIalysis. No crossover products, g-dl or g-dz, were observed in the mass spectrum. Prepere}399-9§.§31§9323 Trimethylaluminum-d9 was prepared from aluminum and 79 Hg(CD3)2. The method of Gilman and Brown59 was used to prepare Hg(CD 20 g (0.084 mmol) of Hg(CD3)2 was added 3)2' to a toluene (80 ml) suspension of 8.5 g (0.31 mmol) of aluminum powder. The mixture was heated to reflux for 24 h. A 1.0 ml aliquot of this solution was titrated by pyridine in xylene, using methyl violet as the indicator.60 The concentration was determined to be 0.7 E. Theseelz§5§-ef-§ A 108 mg (0.391 mmol) sample of 3 was weighed into a small Schlenk tube, and dissolved in 1.0 m1 of toluene. A Teflon needle valve was attached. The solution was cooled to 77K and evacuated on a vacuum line. The tube was immersed in an oil bath maintained at 100°C. After 1 h, the noncondensable gas was collected by a Toepler pump, and identified as CH4 by GC analysis. Methane amounted to 0.152 mmol (0.39 mmol/mmol 3). The remaining volatiles were vacuum transferred into another vessel and analyzed by GC-MS: 2,2-dimethy1butane, 0.059 mmol (0.15 mmol/Ti); 3,3-dimethy1-l-butene, 0.223 mmol (0.57 mmol/Ti); 2,2,3- trimethylbutane, 0.026 mmol (0.066 mmol/Ti); 2,3,3-trimethyl- l-butene, 0.003 mmol (0.007 mmol/Ti) (yields were determined by GC using n-hexane as internal standard). A similar thermolysis was conducted in toluene-d8. The decomposition products were analyzed by GC—MS, and showed no incorporation of deuterium. 80 88E8588§E§-95-?:8s-8¥8-§:8o A 30 mg(0.108 mmol) sample of 3 was weighed into a 5-mm NMR tube attached to a ground glass joint. A Teflon needle valve adaptor was then attached and the apparatus evacuated on a vacuum line. 44 01 (0.334 mmol) of 2-d and 42 ul 3, (0.326 mmol) of E-dO’ and 0.6 ml of toluene were transferred in at 77 K. The tube was sealed under vacuum and thawed, and the contents were mixed. The sample was then immersed in an oil bath maintained at 40°C. After 13 h, the tube was opened under an inert atmosphere, and the volatiles were vacuum distilled into another vessel for GC-MS analysis. Mass spectrum (70 ev) m/e (relative intensity), 87 (26), g-ds; 86 (27), 2-d 85 (31), z-dl; 84 (49), g-do. 2; Protonolysis (HCl) of the recovered metallacycle yielded a mixture of 2,2,3-trimethylbutane (do-d5). 885§558§%§-95-§:§o-5255~%:§2 A 25 mg (0.091 mmol) sample of 3 was weighed into a S—mm NMR tube attached to a ground glass joint. A Teflon needle valve adaptor was then attached and the apparatus evacuated on a vacuum line. 24 ul (0.091 mmol) of 2-d2 along with 0.5 m1 of toluene were transferred in at 77 K. The tube was sealed under vacuum and thawed, and the con- tents were mixed. The tube was immersed in an oil bath maintained at 40°C. The reaction was monitored periodi- cally by 2H NMR spectroscopy (76.8 MHz). The spectra showed depletion of the olefin signals at 64.91 and 4.84, 81 and growth of resonances due to 3-d2 at 61.86 and 2.17. Preparation of CpZTiCHZCH(i-amyl)CHz 6 The procedure is the same as that for 3 except that 3,3-dimethy1-1-pentene was used in place of olefin 2. 1H NMR (C6D 500.1 MHz) 65.62 (5H, s, Cp), 5.45 (5H, s, Cp), 6, 2.19 (2H, d of d, Ha, Jab = 8.3 Hz, JaC = 8.1 Hz), 1.94 (2H, d of d, H Jba = 8.3 Hz, ch = 8.1 Hz), 1.37 (2H, q, b, -cHZCH3, J = 5.9 Hz), 0.92 (3H, t, -CH2C§3, 0.85 (6H, s, -(CH3)2), 0.11 (1H, t of t, H 13 J = 5.9 Hz), J = 8.1 Hz, C’ C8 J cb 8.1 Hz); 0 NMR (0606, 125.7 MHz) 6110.6 (Cp, d, J 171 Hz), 110.2 (Cp, d, J = 171 Hz), 66.7 (cu, t, CH 135 Hz), 36.5 (-g(CH3)2Et, s), 35.1 (-9H CH Jc CH t, H 2 3’ 124 Hz), 25.1 (-C(§H3)2-, q, JCH - 124 Hz), 16.2 (C8, JCH d, JCH = 126 Hz), 9.0 (-CHZCH3, q, JCH To a frozen solution of 53 mg (0.183 mmol) of 6 in 0.5 124 Hz). m1 of toluene was added 0.366 mmol of anhydrous DCl (gene- rated from NaCl and D2804) at 77 K. The mixture was warmed to 25°C. After 2 h, the volatiles were removed by vacuum transfer. 0.169 mmol of 2,3,3-trimethylpentane-d2 (0.92 mmol/mmol 6) was collected and identified by GC-MS analysis. Reaction of 3-d5 with 6 A 14 mg (0.049 mmol) sample of é-ds and 14 mg (0.048 mmol) of g were sealed under vacuum in a 5—mm NMR tube along with 0.5 m1 of benzene—d6. When the solution was heated at 50°C for 1 h, incorporation of protons into the 82 a-positions of 3 was observed. ’b Reaction of 3 with 3-methyl-l-butene A 30 mg (0.109 mmol) sample of 3 and 0.5 ml of benzene- d6 were placed in a septum-capped NMR tube in the dry box. 55 ul (0.491 mmol) of 3-methyl-l-butene was introduced through the septum with a syringe. NMR spectra were recorded over the next 3 days, showing the disappearance of I 16 3 and growth of signals due to CpZTiCHZCH(i-Pr)CH2. Due to the long reaction time, extensive decomposition of the newly formed metallacycle also occurred. Identical reactions carried out with ethylene, propy- lene, and isoamylethylene instead of 3-methyl-l-butene led to the formation of the analogous titanacycles. Reaction of 3 with diphenylacetylene A 25 mg (0.091 mmol) sample of 3 and 16 mg (0.091 mmol) of diphenylacetylene were dissolved in 0.5 ml of benzene-d6 and sealed in an NMR tube under vacuum. The solution was then heated to 40°C. 1H NMR spectra recorded periodically over the next 7 h showed the disappearance of 3 and appea- rance of signals due to CpZTiCH2C(Ph)C(Ph) 3 (67.0 (10H, m), 5.6 (10H, s), 3.4 (2H, 5)) and 3. The titanacyclobutene had been prepared earlier by Tebbe,37 and our sample of 3 was characterized by comparison of spectral data with those reported for Tebbe's complex. An identical reaction employing 3-dZ instead of 3 afforded a mixture of 3 (<10 and d2) and 3 (<10 and d2).as 83 . . l 2 indicated by H and H NMR spectroscopy. A large secondary deuterium isotope effect was observed in this reaction. The ratio of 7-d ’b to 2-d was 2.2:1.0 k k . A similar m ( H2/ D2) 0 0 reaction with trans-3-d1 yielded an isotope effect kHz/kHD= 1.37. In this reaction, only a trace of cis-E-d1 was produced indicating no significant isomerization of the starting titanacycle had occurred during the exchange. 51883???-9?-E¥?-f??€319¥-95-§-YEE§-?§9§§?§ Stock solutions of 3 and PhCECPh were prepared in benzene-d6. Appropriate aliquots were transferred via syringe into S-mm NMR tubes attached to ground-glass joints. The total volume in each tube was brought up to 400 pl with benzene-d6. The samples were frozen in liquid N2, evacuated and sealed off. After thawing, the sample was mixed and inserted into the probe maintained at 40°C. Sufficient time was allowed for the sample to come to thermal equili- brium in the probe, and spectra were recorded periodically by using the JOEL STACK* WAIT program. The data were obtained by monitoring the decrease in integrated intensity of the cyclopentadienyl signal of 3 (the phenyl resonances of PhCECPh and 3 were used as internal standard). The decay of the cyclopentadienyl integration as function of time showed first-order behavior for more than three half- lives (see Figure 12 forua typical plot). The slope (kobsd) was determined from a least-squares fit. This procedure was followed for several concentrations of 3 (0.036-0.21 M) 84 and PhCECPh (0.056-0.79 M). The activation parameters were derived frmnthe kinetics data obtained at 30-60°C (Table 5). The dependence of the rate of disappearance Of 3 on the concentration of olefin 3 was obtained similarly, and the data summarized in Table 6. R88€EE99-95-%-5}Eb-€99:C59-%9-Eb€-85€§8955-9f-§:82 A 20 mg (0.073 mmol) sample of 3 and 52 mg (0.289 mmol) of PhCECPh were weighed into a 5—mm NMR tube attached to a ground-glass joint. 38 pl (0.289 mmol) of E-dz was trans- ferred in at 77 K along with 400 pl of toluene. The tube was sealed under vacuum, and thawed and the contents were mixed. The tube was immersed in an oil bath maintained at 40°C. The progress of the reaction was monitored periodi- cally by 2H NMR spectroscopy (76.8 MHz). After 140 minutes (ca. 1 half—life), NMR analysis indicated that 4% of the total deuterium had been incorporated into the metallacycle. After 7 h, Z-dz (1.7% of total deuterium) was observed. An identical reaction employing trans-3-d1 and 3-d0 instead of 3 and 3-d2 showed that the deuterated 3,3-di- methyl-l-butene contained 7.5% of the cis isomer after 7 h. Metathesis of trans- 3- d1 with 3,3-dimethy1-l-pentene ~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 120 mg (0.433 mmol) of trans-3-d1 was weighed into a Srnall Schlenk tube. 610 01 (4.33 mmol) of 3,3—dimethy1-1- Pelltene and 1.5 m1 of toluene were added via syringe. The mitxture was cooled to 77 K and evacuated on a vacuum line. Tllen it was heated to 40°C in an oil bath for 65 minutes, 85 after which all volatiles were removed in vacuo. The red solids were dried at room temperature an additional 1 h under vacuum. 232 mg (1.3 mmol) of diphenylacetylene was added to the Schlenk tube along with 1.0 m1 of toluene. The mixture was again heated to 40°C. After 12 h, all volatiles were vacuum transferred into another vessel. The volatiles contained a mixture of 3 (d0 and d1) and 3,3-di- methyl-l-pentene (<10 and d1) as determined by GC—MS. The C6 and C7 olefins were separated by preparative glc (column 1, 120°C). The 2H {1H} NMR spectrum (76.8 MHz) of 3,3—di— methyl—l-pentene showed two signals of approximately equal intensity at 64.89 and 4.92. In the proton-coupled 2H NMR spectrum, two overlapping doublets were observed, The peak centered at 64.89 had J(H-D) = 2.56 Hz, and the peak at 64.92 had J(H-D) = 1.59 Hz. From the magnitude of the coupling constants, the 64.89 and 4.92 peaks were assigned to (Z)- and (E)-3,3-dimethyl-1—pentene—l—dl, respectively. The C6 olefins were also analyzed by 2H NMR spectroscopy. The spectrum revealed that only a trace cis-3-d1 had been formed during the reaction. 8959599§1§-9§ trans-3-d1 with Z’do This reaction was run with molar ratios of trans-3—d1 'b to 3—d0 of 1:2 and 1:10. Trans—3—d1 (21 mg, 0.076 mmol) was weighed into a 5—mm NMR tube attached to a ground glass joint. A 500 pl amount of toluene and 3-d0 (20 ul, 0.155 mmol) were condensed into the tube. The sample was then 86 2H {1H} sealed under vacuum, and heated to 40°C for 1 h. NMR spectroscopy indicated that the ratio of trans—3d1 (64.84) to cis—3-d1 (61.86) was 3.421. A ratio of 3.5:1 was obtained when the experiment was run with a 10 molar excess of 3-d0. 3:888:85199-95-99215982992992-§ To a cold (—20°C) suspension of 1.29 g (10.54 mmol) of DMAP in 10 ml of toluene saturated with C2H4 was slowly added a solution of 3 g (10.54 mmol) of 3 in 6 ml of toluene. After completion of addition, the red solution was stirred for 15 min. Then it was slowly transferred via cannula into 150 ml of vigorously stirred cold n—pentane (-20°C). The DMAP-AlMeZCl adduct was removed by rapid filtration through a short pad of Celite supported on a medium frit. The filtrate was then concentrated by removal of solvent in vacuo until crystallization began to occur. After cooling at —50°C overnight, the supernatant was removed via cannula and the red solid was washed once with 5 ml of cold n—pentane. The product so obtained is of suitable purity for further studies. Further purification can be accomplished by redissolving the red solid in diethyl ether at 0°C and cooling slowly to —50°C: yield 0.85 g (37%). 1 H NMR (90 MHz, toluene-d8):65.22 (s, 10H, Cp's), 2.95 (m, AA' of AA'XX'AA' spectrum, 4H, Ha)’ —0.16 (m, XX‘ of AA'XX‘AA' spectrum, 2H, H8). Gated decoupled 13C NMR (125.7 MHz, toluene-d8):6107.6 (d, J = 170 Hz), 76.6 (t, J = 140 Hz), -ll.7 (t, J = 139 Hz). Anhydrous DCl, 87 generated from D2504 and NaCl, reacted with 3 to produce a quantitative yield of propane-1,3—d2 (GC—MS) and CpZTiCl2 (1H NMR). Addition of 0.1 m1 of Br2 to a diethyl ether solution of 3 (177.6 mg, 0.808 mmol) at —65°C and allowing the mixture to warm up slowly to room temperature resulted in the formation of 1,3-dibromopropane which was identified by comparison to a GC/mass spectrum of an authentic sample, and CpZTiBr2 (1H NMR, 6 6.69, CDCl The yield of the 3)" organic product was determined by gas chromatography (column 3, 200°C) using 1,4-dibromobutane as an internal standard. The relative response of the product vs. standard was assumed to be the ratio of their molecular weight. Compound 3—d2 was prepared similarly, except 3—d2 was 1 used instead of 3. H NMR (90 MHz, toluene-d8): 6 5.22 (s, 10H, Cp'S), 2-95 (m, 2H, Ha, AA' of AA'XX‘ spectrum), —0.16 (t, 2H, HB’ XX' of AA'XX' spectrum). Simulation of 1H NMR spectrum was conducted on the Bruker spectrometer equipped with an Aspect 2000 computer. Simulation parameters: v(A)=v(A‘)=329.9 Hz; v(X)=v(X')=50 Hz; J(A,X)=J(A',X')= 4.9 Hz; J(A,X')=J(A',X)=12.9 Hz; J(A,A')=i7.9 Hz; J(X,X')= $10.35 Hz. Thermal decomposition of 5 A solution of 3 (40 mg, 0.182 mmol) in 3 m1 of toluene was heated under vacuum, at 60°C for 1 h. The volatile components of the thermolysis mixture were collected with a Toepler pump. GC analysis of the gases (column 1, 2) 88 indicated the presence of methane, ethane, propane and propene. A solid sample of 3 (35 mg, 0.159 mmol) was heated at 40°C for l h in a septum-capped 10 m1 Schlenk flask. The gas mixture generated in this way contained methane, propane and propene. No CZ—hydrocarbons were detected by GC analysis. 899953995-95-§-8339-93921999:94r-:go A solution of 3 (29 mg, 0.132 mmol) in 0.6 m1 of toluene-d8 was sealed under ethylene—d4 (0.5 mmol), allowed to stand at room temperature, and monitored periodically by 1H NMR spectroscopy. After 23 h, the spectral changes noted in the text were observed. After 90 h, all the starting material had disappeared. Due to the long reaction time, decomposition of the product occurred also. An identical reaction employing CZH4 instead of C2D4 showed the spectral changes noted in the text. In a separate experiment, a solution of 3 (115 mg, 0.52 mmol) in 4 m1 of toluene was stirred under an atmosphere of ethylene at room temperature for 48 h, after which the reaction mixture was cooled to —50°C, and the excess ethylene was removed. CO was then introduced and the reaction was allowed to warm up slowly to room temperature in ca. 2 h. A yellow precipitate appeared between —10° and 0°C from the deep red solution and was isolated by filtration. The toluene was removed from the filtrate in vacuo giving a 89 dark red solid. On the basis of 1H NMR (6 4.56) and IR (VCO=1975, 1897 cm 1) data, this dark red solid was iden— tified as CpZTi(C0)2. The IR spectrum of the yellow solid displayed a band at 1614 cm.1 (Nujol) assignable to an nZ—acyl. The complex decomposed in CHZCl2 at room tempera— ture to give cyclopentanone identified by IR, GC/MS (VCO= l). A reaction of 3 (52.6 mg, 0.239 mmol) with C 1745 cm- 2H4 was quenched after 48 h at room temperature with anhydrous HCl affording CpZTiCI2 and n-butane as the major product (GC/MS). Some propane was also detected due to unreacted 5. r1. 89555399-9f-§-5259-éiebsszlsssEzlsBs A solution of 3 (20 mg, 0.091 mmol) and diphenyl- acetylene (48 mg, 0.27 mmol) in 0.5 m1 of toluene—d8 was sealed in an NMR tube in vacuum. 1H NMR spectra were recorded periodically and the conversion of 3 into 39 (6 6.24, Cp's) was observed. Propylene was produced as the major organic product. GC analysis (column 1) of the gas phase above the reaction mixture confirmed the identity of the major products noted in the text. Reaction of 5 with acetone A solution of 5 (57.6 mg, 0.262 mmol) and acetone ’b (22 p1, 0.262 mmol) in 2 m1 of toluene was heated to 40°C for 7 h. The organic products in the gas phase and in solution were analyzed by GC (see text). The toluene was 90 then removed in vacuo, giving a yellow solid which matched the literature description of polymeric titanocene oxide.52 Reaction of~§~with AlMeZCl A solution of 12.5 ul (0.134 mmol) of AlMe C1 in 0.3 2 ml 0f toluene—d8 was added via syringe to 29.5 mg (0.134 mmol) of 3 in 0.25 ml of toluene—d8 in a septum-capped NMR tube. An immediate reaction occurred, and an NMR spectrum did not show any characterizable organometallic product. GC analysis of the organic products showed CH4 (8.1%), C2H6 (trace), propane (4.7%), propene (86.8%). Reaction of 5 with C0 ~~~~~~~~~~~~~~~~~~~~~ Excess CO (0.5 mmol) was partially condensed at -l96°C into an NMR tube containing a frozen solution of 3 (41.5mg, 0.189 mmol) in 0.6 m1 of toluene—d8. The tube was sealed 1 13 and the reaction monitored by H and C NMR spectroscopy. The reaction was complete after 30 min at room temperature giving a dark purple solution. The spectral changes noted 13 in the text were observed. In a separate experiment, C0 (Stohler, 12CO 9.9%, 13CO 90.1%) was used instead of 12C0. The 13C NMR spectrum showed the 1:SC-label only at the signal at 6 143.6 ppm. On a preparative scale, a solution of 3 (146 mg, 0.664 mmol) in 5 ml of toluene was stirred under a C0 atmosphere for 90 min at room temperature. The toluene was removed in vacuo and the dark solid was recrystallized from toluene/ pentane at —50°C to give 125 mg of dark purple crystals 91 (68%). Anal. Calcd. for C15H16Ti02: C, 65.23; H, 5.84; Ti, 17.34. Found (Schwarzkopf): C, 64.43; H, 6.23; Ti, 17.00. 1H NMR (500 MHz): 6 5.53 (s, 10H, CSHS), 2.31 (t, JHH=7.63 13C NMR (benzene-d6, Hz, 4H), 1.85 (q, JHH=7.63 Hz, 2H). 1H decoupled): 6 110 (CSHS), 143.6 (9:9), 28.8 (~CH2CHZCH2-L 18.9 (—CHZCH2CHZ—). IR (KBr): 3122(w), 3095(m), 2955(w), 2900(m), 2840(m), 1495(vs), 1480(sh), 1442(m), 1365(m), 1305(w), 1285(w), 1242(m), 1190(w), 1125(w), 1082(m), 1065(sh), 1025(m), 1015(m), 945(vw), 925(vw), 915(vw), 890(w), 837(m), 805(vs), 650(m). LIST OF REFERENCES 10. LIST OF REFERENCES For recent reviews of the olefin metathesis reaction, see: a) R. H. Grubbs, Prog. Inorg. Chem. 24,1 (1979); b) T. J. Katz, Adv. Organomet. Chem. 16, 283 (1977); c) N. Calderon, J. P. Lawrence, E. A. Ofstead, ibid. 12, 449 (1979); d) J. J. Rooney and A. Stewart, Catalysis, 1, 277 (1977). L Herisson and Y. Chauvin, Makromol. Chem., 161 (1970). (H 34> 214—“ u a) D. M. Adams, J. Chatt, R. G. Guy, N. Sheppard, 3; Chem. Soc., 738 (1961); b) F. J. McQuillin and K. C. Powell, J. Chem. Soc., Chem. Commun., 45 (1972). a) R. D. Gillard, M. Keefon, R. Mason, M. F. Pilbrow, R. D. Russel, J. Organomet. Chem., 33, 247 (1971); b) D. J. Yarrow, J. A. Ibers, M. Lenarda, M. Grazini, ibid., 20, 133 (1974). a) R. J. Puddephatt, M. A. Guyser, C. F. H. Tipper, J. Chem. Soc., Chem. Commun., 626 (1976); b) R. J. Puddephatt, Coord. Chem. Rev.,33, 149 (1980); c) R. J. Al—Essa, R. J. Puddephatt, M. A. Quyser, C. . H. Tipper, J. Am. Chem. Soc., 101, 364 (1979). C. P. Casey, D. M. Scheck, A. J. Schusterman, ibid., 101, 4233 (1979). a) M. Ephritikhine, B. R. Francis, M. L. H. Green, R E. Mackenzie, M. J. Smith, J. Chem Soc., Dalton Trans., 1131 (1977); b) G. J. A. Adam, S. G. Davies, K. A. Ford, M. Ephritikhine, P. F. Todd, M. L. H. Green, J. Mol. Cata1., 0, 15 (1980). P. Foley and G. W. Whiteside, J. Am. Chem. Soc., 101, 2732 (1979). T. H. Tulip and D. L. Thorn, J. Am. Chem. Soc., 103, 2448 (1981). R. A. Anderson, R. A. Jones, G. Wilkinson, J. Chem. Soc., Dalton Trans., 446_(1978). 92 12. 13. 14. 15. l6. l7. l8. 19. 21. 22. 23. 93 C. P. Casey and T. J. Burkhardt, J. Am. Chem. Soc., 96, 7808 (1974). a) R. R. Schrock, ibid., 96, 6796 (1974): b) R. R. Schrock, Accounts Chem. Rest, 12, 98 (1979). R. R. Schrock, J. Am. Chem. Soc., 00, 5399 (1976). a) J. H. Wengrovius, R. R. Schrock, M. R. Churchill, R. J. Wissert, W. J. Youngs, J. Am. Chem. Soc., 103, 4515 (1980); b) R. R. Schrock, S. Rocklage, J. H. Wengrovius, G. Rupprecht, J. Fellman, J. Mol. Catal., §, 73 (1980). a) F. N. Tebbe, G. W. Parshall, G. S. Reddy, J. Am. Chem. Soc., 100, 3611 (1978); b) N. Tebbe, G. W. Parshall, D. W. Overnall, ibid., 1 5074 (1979); c) V. Klabunde, F. N. Tebbe, G. W. arshall, R. L. Harlow, J. Mol. Catal., 0, 37 (1980). 3CD (H'Ti v - Part of the studies reported in this work have been published in preliminary form, T. R. Howard, J. B. Lee, R. H. Grubbs, J. Am. Chem. Soc., 102, 6876 (1980). T. R. Howard, M.S. thesis, California Institute of Technology, 1980. J. B. Lee, G. J. Gajda, W. P. Schaefer, T. R. Howard, T. Ikariya, D. A. Straus, R. H. Grubbs, J. Am. Chem. Soc., 103, 7358 (1981). Compound 4 was prepared from styrene and 1 by Dr. T. Ikariya from this laboratory and x-ray analysis was performed by G. J. Gajda. a) C. P. Casey and H. E. Tuinstra, J. Am. Chem. Soc., 100, 2271 (1978); b) J. M. Basset, J. L. Bilhou, P. Martin, A. Theider, ibid., 92, 7376 (1975); c) T. J. Katz and J. McGinnis, 151d., 9], 1572 (1975); d) J. Wang and H. R. Menapace, J. Or . Chem., 33, 3794 (1968fi e) W. B. Hughes, J. Chem. Soc., Chem. Commun., 431 (1969). O. Eisenstein, R. Hoffmann, A. R. Rossi, J. Am. Chem. Soc., 103, 5582 (1981). However, a second calculation shows a symmetrical metallacycle which may easily pucker, T. Rappe and W. A. Goddard, ibid., 104, 297 (1982). F. W. Wehrli and T. Wirthlin, ”Interpretation of Carbon—13 NMR spectra”, Heyden: London, 1978. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 94 a) M. Saunders, L. Telkowsi, M. R. Kates, J. Am. Chem. Soc., 99, 8070 (1977); b) M. Saunders and M. R.Kates, 1 id., 99, 8071 (1977); c) M. Saunders, M. R. Kates, K. B Wiberg, W. Pratt, ibid., 99, 8072 (1977). The author would like to acknowledge the gift of a sample of 4—d2 from S. Ho for this study. Compound -d was prepared from 1—d2 and CZH4 (see experimental section). a) H. Batiz—Hernandez and R. H. Bernheim, Prog. Nucl. Ma n. Reson., 3, 63 (1970); b) T. W. Marshall, Mol. Phys., 4, 61 (1961); c) H. S. Gutowsky, J. Chem. 31, 1683 (1959). ”O \< (I) R. R. Schrock and P. R. Sharp, J. Am. Chem. Soc., 100, 2389 (1978). K. C. Ott and R. H. Grubbs, J. Am. Chem. Soc., 103, 5922 (1981). S. Cohen and J. E. Bercaw, private communication, (Cp* = n5-(CH3)SCS). a) P. C. Wailes, R. S. P. Coutts, H. Weigold, "Organo— metallic Chemistry of Titanium, Zirconium, and Hafnium”, Academic Press: New York, 1974; b) Reaction of CpZTi(CH )Cl with AlMe3—d9 results in rapid equili— bration with CpZTi(CD3)Cl and AlMe3—d6, J. B. Lee and R. H. Grubbs, unpublished results. D. B. Carr and J. Schwarz, J. Am. Chem. Soc., 101, 3521 (1979). a) G. M. Whitesides, M. Witanowski, J. D. Roberts, J. Am. Chem. Soc., 82, 2860 (1965); b) D. S.Matteson, "Organometallic Reaction Mechanisms of the Non- transition Elements”, Academic Press: New York 1974. a) G Fraenkel, D. T. Dix, M. Carlson, Tetrahedron Lett., 579 (1968); b) J. J. Eisch, J. Organomet. Chem. Rev. Ann. Surv., 4B, 314 (1967); c) L.Lardicci, L. Lucarini, P. Palagi, P. Pino, J. Organomet. Chem., 4, 341 (1965). T. J. Katz and W. H. Hersh, Tetrahedron Lett., 585 (1977). 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 95 F. N. Tebbe and R. L. Harlow, J. Am. Chem. Soc., 102, 6151 (1980). ~~~ Ring—hydrogen abstractions find precedents in titanocene chemistry, see for example: a) G. J. Erskine, D. A. Wilson, J. D. McGowan, J. Organomet. Chem., 114, 119 (1976); b) M. D. Rausch, W. H. Boon, H. G. Alt, ibid., 141, 299 (1977). ~~~ In a parallel study, the reaction of 3 with trimethyl— acetylaldehyde to yield 2 and [CpZTi(O)]n was found to proceed with kobsd=9-3XlO—5 s‘1 at 40°C. K. Brown— Wensley, Ph.D. Thesis, California Institute of Technology. R. W. Alder, R. Baker, J. M. Brown, ”Mechanism in Organic Chemistry"; Wiley—Interscience: New York, 1971. Although the large secondary isotope effect is also consistent with the ring—opening of 3 to an metal- olefin—alkylidene complex, ourresultscertainly do not require such an intermediate. However, we cannot rule out the intervention of this species in the conversion of 3 to 3 and g (Figure 11). The actual secondary isotope effects should be larger than the ones observed experimentally due to the con- current formation of the olefins in these reactions. R. J. Klinger, J. C. Huffman, J. K. Kochi, J. Am. Chem. Soc., 104, 2147 (1982). See for example, some indane system: W. R. Jackson, C. H. McMullen, R. Spratt, P. Bladen, J. Organomet. Chem., 4, 392 (1965). R. M. Silverstein, G. C. Bassler, T. C. Morril, ”Spectrometric Identification of Ogranic Compounds”, 3rd ed., John Wiley and Sons, Inc., New York, 1974. E. D. Becker, ”High Resolution NMR: Theory and Chemical Application,” 2nd ed., Academic Press, New York, 1980. . P. Pez and J. N. Armor, Adv. Organomet. Chem., 19, l (1981). In szTi(CH2)(C2H4), the metal orbital is the lal orbital characteristic of CpZMLZ complexes: J. W. Lauher and R. Hoffman, J. Am. Chem. Soc., 98, 1729 (1976). 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 96 J. X. McDermott, M. E. Wilson, G. M. Whitesides, J. Am. Chem. Soc., 98, 6529(1976). Typical Erequencies for bihapto or n2 acyls are 1530— 1620 cm‘ See: G. Fachanetti, C. Floriani, G. Fochi, J. Chem. Soc., Dalton Trans., 1946(1977); b) C. Fachenetti, C. Floriani, ibid., 2297(1977). Perhaps even more perplexing is the fact that typical frequencies for non—conjugated transition metal mono- hapto(n1) acyls are lower than the value reported by Whitesides. The range is 1630-1680 cm'l. See: a) E. Maslowsky, ” Vibrational Spectra of Organometallic Compounds", Wiley-Interscience, N.Y., 1977, p.155; b) M. L. H. Green, "Organometallic Compounds", Vol. 2, Methuen, London, 1968, p. 257-261. K. Sonogashira and N. Hagihara, Bull. Chem. Soc. Japan, 39, 1178(1969). S. A. Giddings, Inorg. Chem., 3, 684(1964). Kinetics studies of the reaction of 3 with (CH3) CO indicated a first— order reaction (ref. 38) in which the rate was independent of [(CH ) CO]. However, the proposed mechanism did not involve a free methylidene. A zwitterion or a betaine—like species analogous to that invoked as intermediate in the Wittig reaction are all possible intermediates and cannot be excluded. See: G. Wittig, J. Organomet. Chem., 100,279(1975). ~~~ Similar reactivity with CO was also displayed by the t- -butyl complex 3. The resultant enediolate from 3 and CO yielded the cyclic acyloin on hydrolysis with HCl. D. Strauss. unpublished results. A M.W. determined cryoscopically in benzene by Schwarz- kopf Microanalytical Laboratory suggested that this complex may be a dimer in solution. Calcd for C H 6T1 02: 276; Found: 555. But it was indicated that e solution was cloudy suggesting possible decomposition of the complex. An x-ray study of an enediolate is in progress, preliminary data show it to be a monomer, G. J. Gajda, D. Strauss, unpublished results. a) P. T. Wolczanski and J. E. Bercaw, Acc. Chem. Res., 13, 121(1980); b) J. M. Manriquez, D. R. McAlister, R. D. Sanner, J. E. Bercaw, J. Am. Chem. Soc., 100, 2716(1978). ~~~ 58. 59. 60. 97 The metathesis of functionalized olefins is one of the most promising synthetic applications of the metathesis reaction. However, with only a few exceptions, many of the active metathesis catalysts are easily poisoned by polar groups. For a review, see: J. C. Mol, J. Mol. Catal., 13, 35 (1982). H. Gilman and R. E. Brown, J. Am. Chem. Soc., 32, 3314 (1930). D. F. Hagen and W. D. Leslie, Anal. Chem., 814 (1943). "I11111111111“