Synthesis and Investigations into the Reactivity of Electron Deficient Organoscandium Complexes Citation Thompson, Mark Edward (1986) Synthesis and Investigations into the Reactivity of Electron Deficient Organoscandium Complexes. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/wcaf-jd79. https://resolver.caltech.edu/CaltechTHESIS:10252019-155031892 Abstract A new class of coordinatively unsaturated, monomeric scandium complexes, (Cp* 2 ScR (Cp* = η 5 -C 5 (CH 3 ) 5 ; R = H, alkyl, aryl, halide) have been prepared. Cp* 2 ScCl is prepared by the reaction of ScCl 3 (THF) 3 with LiCp*, and Cp* 2 ScR (R = CH 3 , C 6 H 5 , C 6 H 4 CH 3 , CH 2 C 6 H 5 ) by the reaction of Cp* 2 ScCl with the appropriate organoalkali reagent. Cp* 2 ScR complexes react readily with H 2 to give RH and Cp* 2 ScH. The hydride ligand exchanges rapidly with hydrogen gas and inserts olefins to give alkyl complexes (e.g. Cp* 2 ScCH 2 CH 3 ). Cp* 2 ScH reacts with allene to give Cp* 2 Sc(η 3 -CH 2 CH=CH 2 ). Cp* 2 ScR and Cp* 2 ScH react with pyridine to give Cp* 2 Sc(C,N-η 2 -C 5 H 4 N). The crystal structure of this complex was determined and is reported herein. Spectroscopic data for Cp* 2 ScCH 3 and Cp* 2 ScCH 2 CH 3 and crystallographic data for the former indicate that the methyl ligand is bound to scandium in a conventional manner, while the ethyl ligand may participate in an agostic interaction. The reactions of scandium alkyl, aryl and hydride complexes were investigated. H/D exchange between H 2 , arenes and the 1° and 2° C-H bonds of alkanes is catalyzed by Cp* 2 ScH. In C 6 H 6 solution Cp* 2 ScH and Cp* 2 ScC 6 H 5 are in equilibrium, ΔH° = 6.7 ± 0.3 kcal/mole and ΔS° = 1.5 ± 0.1 e.u.. Thus in this system a scandium-hydride bond is 1.5 ± 0.4 kcal/mole stronger than a scandium-phenyl bond. Cp* 2 ScCH 3 reacts with a wide range of hydrocarbons (RH) by C-H bond activation to give CH 4 and Cp* 2 ScR (RH = 13 CH 4 , arenes, styrenes, propyne). From the reactions of Cp* 2 ScCH 3 with styrenes, the activation parameters (ΔH ‡ = 11.5-12.6 kcal/mole, ΔS ‡ = -34 to -38 e.u.) for these C-H activation reactions were determined. A deuterium isotope effect of 2.9 is observed for the intermolecular activation of C-H in the reaction of Cp* 2 ScCH 3 with benzene. Very small differences in the rates of vinylic C-H bond activation for CH 2 =CHC 6 H 4 X-para (X = CF 3 , OCH 3 ), and the aryl C-H bonds of C 6 H 5 X (X = CF 3 , H, CH 3 , N(CH 3 ) 2 ), as well as the positional nonselectivity for the activation of the meta and para C-H bonds of toluene indicate that the scandium center does not interact substantially with the π-system of these substrates in the transition states for these reactions. Thus for these sterically encumbered organoscandium compounds, sp 2 C-H bond activation occurs without formation of a π-complex. A general mechanism for these C-H and H-H activation reactions is proposed, and is termed "σ-bond metathesis". The reactions of Cp* 2 ScR complexes {R = hydride, alkyl, aryl) with small olefins and alkynes were examined. The hydride, methyl and benzyl complexes function as ethylene polymerization catalysts, while Cp* 2 ScC 6 H 5 does not react. Cp* 2 ScH and Cp* 2 ScCH 3 react stoichiometrically with propene by a series of insertion and vinylic C-H activation reaction. The final scandium product in both cases is trans-Cp* 2 ScCH=CHCH 3 . The scandium allyl complex, Cp* 2 Sc(η 3 -CH 2 CH=CH 2 ), is not observed and is not a reaction intermediate. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: Chemistry Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Bercaw, John E. Thesis Committee: Collins, Terrence J. (chair) Bercaw, John E. Grubbs, Robert H. Chan, Sunney I. Defense Date: 19 August 1985 Funders: Funding Agency Grant Number Atlantic Richfield Foundation UNSPECIFIED Record Number: CaltechTHESIS:10252019-155031892 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:10252019-155031892 DOI: 10.7907/wcaf-jd79 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 11866 Collection: CaltechTHESIS Deposited By: Mel Ray Deposited On: 28 Oct 2019 19:21 Last Modified: 16 Apr 2021 23:21 Thesis Files Preview PDF - Final Version See Usage Policy. 48MB Repository Staff Only: item control page

Synthesis and Investigations Into the Reactivity of Electron Deficient Organoscandium Complexes Thesis by Mark Edward Thompson In Partial Fulfillment of the Requirements for the degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1986 (Submitted August 19, 1985) ii to the best chemist I've ever known, Donald A. Thompson iii ACKNOWLEDGEMENTS I would like to thank my "boys" (Steve Baxter, Reza Mohammadi, and Mikey Nolan) and A. Ray Bulls for their help with some of the kinetics experiments described in this thesis. Of the two crystal structures described herein, Dr. Bernard Santarsiero helped me solve the structure of 2 Cp* Sc(C,N-~ -C H N> and Dr. Bill Schaefer determined the 5 4 2 structure of Cp* 2 sccH 3 • I am very grateful for their help with these structures and with the experimental "write-ups" in Chapter 1. I would also like to thank Bill Schaefer for helping me learn how to write in english. I am very grateful to all of the members of the Bercaw group, past and present, for providing an interesting and stimulating atmosphere to work in. I would like to thank John Bercaw for being a good friend, mentor and role model for the last five years. The teachings of Pete Wolczanski and Dean Roddick have proven invaluable in working with a vacuum line, and I have always obeyed Pete's teachings on lab cleanliness. I am grateful to all of the members of the "cutting edge" (213 Noyes), past and present, for giving me a nutty place to work and filling the traps when I forgot. I would like to thank the founding fathers of the A.A.S.S. for teaching me the true meaning of the word "ROADTRIP". All of the members of 1130 Lura St. (Terry Krafft, Mikey Vandyke, Allan vanAssellt and Marty St.Clair) have also been important in giving me a great place to live and always iv keeping beer and milk in the refrigerator. I would like gratefully acknowledge the Atlantic Richfield Foundation for their generous fellowship. v ABSTRACT A new class of coordinatively unsaturated, monomeric scandium complexes, Cp* 2 ScR (Cp* = ~ 5 -c 5 CcH 3 >5 1 R = H, alkyl, aryl, halide) have been prepared. Cp* 2 scCl is prepared by the reaction of ScC1 3 CTHF) 3 with LiCp*, and Cp* 2 ScR (R = CH3, C6H5' C6H4CH3, CH2C6H5) by the reaction of Cp* 2 scCl with the appropriate 'organoalkali reagent. Cp* 2 scR complexes react readily with H2 to give RH and Cp* 2 ScH. The hydride ligand exchanges rapidly with hydrogen gas and inserts olefins to give alkyl complexes (~.g. Cp* 2 sccH 2CH 3 >. Cp* 2 ScH reacts with allene to give Cp* 2 ScC~ 3 -cH 2 CH=CH 2 >. Cp* 2 scR and Cp* 2 ScH react with pyridine to give 2 Cp* Sc(C,N-~ -C H N>. The crystal structure of this complex 2 5 4 was determined and is reported herein. Spectroscopic data for Cp* 2 sccH 3 and Cp* 2 sccH 2 cH 3 and crystalographic data for the f orrner indicate that the methyl ligand is bound to scandium in a conventional manner, while the ethyl ligand may participate in an agostic interaction. The reactions of scandium alkyl, aryl and hydride cornplexes were investigated. H/D exchange between H2 , arenes and the 1° and 2° C-H bonds of alkanes is catalyzed by Cp* 2 ScH. In c 6H6 solution Cp* 2 ScH and Cp* 2 scc 6H5 are in equilibrium, ~H 0 = 6.7 ± 0.3 kcal/mole and ~s 0 = 1.5 ± 0.1 e.u •• Thus in this system a scandium-hydride bond is 1.5 + 0.4 kcal/mole stronger than a scandium-phenyl bond. Cp* 2 ScCH 3 reacts with a wide range of hydrocarbons (RH) by C-H bond activation to give CH 4 and Cp* 2 ScR (RH = 13 cH 4 , vi arenes, styrenes, propyne). From the reactions of Cp* 2 sccH with styrenes, the activation parameters (~H* = 11.5-12.6 3 kcal/mole, Js* = -34 to -38 e.u.) for these C-H activation reactions were determined. A deuterium isotope effect of 2.9 is observed for the intermolecular activation of C-H in the reaction of Cp* 2 sccH 3 with benzene. Very small differ- ences in the rates of vinylic C-H bond activation for CH 2 =CHC 6H4X-para (X = CF 3 , OCH 3 ), and the aryl C-H bonds of c 6 H5 X {X = CF 3 , H, CH 3 , N(CH 3 > 2 >, as well as the positional nonselectivity for the activation of the meta and para C-H bonds of toluene indicate that the scandium center does not interact substantially with the rr-system of these sub- strates in the transition states for these reactions. Thus for these sterically encumbered organoscandium compounds, sp 2 C-H bond activation occurs without formation of a rr-complex. A general mechanism for these C-H and H-H activation reactions is proposed, and is termed "a-bond metathesis". The reactions of Cp* 2 scR complexes {R = hydride, alkyl, aryl) with small olef ins and alkynes were examined. The hy- dride, methyl and benzyl complexes function as ethylene polymerization catalysts, while Cp* 2 scc 6H5 does not react. Cp* 2 ScH and Cp* 2 sccH 3 react stoichiometrically with propene by a series of insertion and vinylic C-H activation reaction. The final scandium product in both cases is trans- Cp* 2 scCH=CHCH 3 • The scandium allyl complex, Cp* 2 sccry 3-cH 2 CH=CH 2 >, is not observed and is not a reaction intermediate. vii TABLE OF CONTENTS ACKNOWLEDGEMENTS iii v ABSTRACT LIST OF TABLES viii LIST OF FIGURES xi ABBREVIATIONS xiv CHAPTER 1. Synthesis and Structure of Alkyl and 1 Hydride Derivatives of Permethylscandocene. CHAPTER 2. C-H Bond Activation and o-Bond Metathesis 52 by Permethylscandocene Complexes. CHAPTER 3. Insertion and Vinylic C-H Bond Activation Reactions of Unsaturated Hydrocarbons with Permethylscandocene Alkyl and Hydride Complexes. 130 viii List of Tables Chapter 1. Table 1. 1 H chemical shift differences between 18 Table 2. 1 H and 13 c NMR spectroscopic data 24 Atomic coordinates Cx 10 5 > and gaussian amplitudes cl 2 , x 10 4 > for all Table 3. .31 2 non-hydrogen atoms of Cp* 2 scCC,N-~ -C 5 H4 N>. Table 4. Cp* Hydrogen atom coordinates Cx 10 4 > for .3 2 2 2 Sc(C,N-~ -C 5H4N>. Table 5. .3 .3 Table 6. .34 Table 7. Crystal data for Cp* 2 ScCH 3 • .37 Table 8. Hydrogen atom coordinates Cx 10 4 > of .37 ix Table 9. Final atomic coordinates Cx 10 4 > and 38 gaussian amplitudes Cx 10 3 ) for Cp* 2 sccH 3 • Table 10. Bond lengths (A) and angles c0 > for 39 Cp* 2 ScCH 3 • chapter 2. Table 1. Equilibrium constants and ~Go values 64 for the equilibrium between CP"* 2 ScH, Cp* 2 scc 6H5 and H2 in benzene solution Ceq 7). Table 2. Equilibrium constants and dG 0 values 68 for the equilibrium between Cp* 2 scCTHF)H, Cp* 2 scc 6H5 , H2 and THF in benzene solution ( eq 8) • Table 3. Ratio of CH 3 o to CH 4 from the reaction 73 of Cp* 2ScCH 3 with c 6o 6 (eq 10). Table 4. 86 Rate constants and activation parameters for the reaction of Cp* 2scCH 3 with substituted styrenes CH C=CHC H 2 X = OCH 3 , CF 3 ) at 60°C Ceq 19). 6 4 X-~; x Table 5. Rate constants for the reaction of CCp*-~ 15 > 2 sccH 3 with c 6 H6 and substituted arenes Ceq 25) at Table 6. 89 ao 0 c. Isomer ratio of the product mixtures 92 from the reactions of Cp* sccH , 2 3 Cp* 2 ScCH 2 c 6 H5 , Cp* 2 ScC 6 H 4 cH 3 -~, -~ with toluene. Table 7. 1 H and 13 c NMR spectroscopic data. 111 Chapter 3. Table. 1 H and 13 c NMR spectroscopic data. 153 xi List of Figures Chapter 1. Figure 1. 16 Figure 2. 16 with selected distances (A) and angles c0 >. Figure 3. ORTEP drawing of Cp*2ScCH3· 21 van't Hoff plot for the equilibrium 65 Chapter 2. Figure 1. between Cp* 2 ScH and Cp* 2 scc 6H5 in c 6H6 solution. Figure 2. van't Hoff plot for the equilibrium between Cp* 2 Sc(THF)H and Cp* 2 ScC 6H5 in c 6H6 solution. Figure 3. Partial potential surfaces for reactions 7 and 8. 67 · xii Figure 4. 90 MHz NMR spectrum of the proposed 74 •tuck-in" complex, x. Figure 5. First order plot of the data from the 76 reaction generating the •tuck-in• complex Ceq 11). Figure 6. First order plot of the data from the 78 reaction generating the •tuck-in" complex, with added x. Figure 7. Representative second order plot for 84 the reaction of Cp* 2 sccH 3 with H2 C=CHC 6H4 CF 3 • Figure 8. Arrhenius plots of the reactions of 85 Cp* 2 ScCH 3 with H2 C=CHC 6H4X (X = CF 3 , OCH 3 >. Figure 9. Representative first order plot of 90 the data from the reaction of Cp* 2 sccH 3 with c 6a 5NCCH 3 > 2 • Figure 10. Proposed mechanism for a-bond metathesis. 101 xiii Figure 11. The predicted transition state for 107 the H/D exchange reaction between Cl 2 ScH and n 2 (from Steigerwald and Goddard C48J). Chapter 3. Figure. The IR spectrum of polyethylene from the reaction of Cp* 2 sccH 3 with H2 C=CH 2 • 137 xiv ABBREVIATIONS atm atmospheres br broad Cp cyclopentadienyl, n5 -c 5 H5 Cp* pentamethylcyclopentadienyl, d doublet i iso m multiplet CNMR), medium CIR) m meta .n normal or tho para quartet sh shoulder s singlet CNMR), strong CIR) trans w weak n 5 -c CcH > 5 3 5 1 Synthesis and Structure of Alkyl and Hydride Derivatives of Perrnethylscandocene 2 INTRODUCTION Over the past three decades the organometallic chemistry of scandium has been largely ignored. Scandium has been traditionally grouped with the lanthanides, based on its low electronegativity and lack of accessible oxidation states other than trivalent. However, scandium's ionic ra- 0 dius (0.81 A for Sc (III)) [l] is considerably smaller than 0 those seen for the lanthanides (ranging from 1.1 A for 0 Ce (III) to 0.93 A for Lu (III)) ,[l] and is comparable to some transition metals in their common oxidation states (Zr (IV) 0 0 = 0.80 A, V(III) = 0.78 A).£1] The low electroneg- ativity and ionic radius of scandium suggests that its organometallic complexes may exhibit unique reactivity patterns. ScR 3 complexes are commonly polymeric (~ [ScCp 3 Jx,[2] [ScCC 6H5 >3 Jx [3] and [Sc(C:cc 6 H5 >3 Jx) [3], and are insoluble in nonreactive solvents. Biscyclopentadienylscandium-based compounds, however, are isolated as stable monomeric and dimeric species. Ccp 2 scc11 2 and Ccp 2 scc:cc H5 J 2 6 [4] are both dimeric with bridging chloride and phenylethynyl ligands, respectively, while Cp 2 sc(THF)CH 3 [5] and cp 2 scc~ 3 -allyl) [4] are monomeric. Complexes containing bridging main group atoms have also been reported, as in Cp2SC().l-CH3>2Al(CH3>2 [6] and Cp2Sc(µ-CH2)P(CH3>2 [7]. Al- though c 5 H5 is sterically demanding enough to prevent oligo- 3 merization in Cp 2 sc-based complexes, it is not sufficiently bulky to prevent coordination of a datively bound ligand, which gives all reported scandocene-based complexes an electron count of 16. Previous work in our group has shown that monomeric, coordinatively unsaturated hydride and alkyl complexes of group IV metals, i.e. Cp*2MR2 CM = Ti, Zr, Hf: R = alkyl, hydride; Cp* = ~ 5 -c 5 Me 5 > exhibit high reactivity toward CO, H2 , olefins and C-H bonds.[81 This is attributed to the high Lewis acidity of the metal center and the coordinative unsaturation of these 16 electron, do complexes. We expect that analogous scandium complexes should be even more reactive than their group IV counterparts, because scandium is a stronger Lewis acid than the group IV elements, and Cp* 2 ScR would be a 14 electron, do complex with two vacant orbitals.[91 Herein, we report the synthesis of a new class of coordinatively unsaturated, monomeric scandium complexes, Cp* 2 ScR CR= Cl, alkyl, aryl, hydride). The alkyl, aryl and hydride complexes react readily with H2 at -so 0 c, and activation C-H bonds by these complexes is observed at room ternperature. These complexes act as ethylene polymerization catalysts, and react stoichiometrically with propene and other olef ins. In this chapter the synthesis of these com- plexes and their reactivity toward H2 and pyridine {leading 4 to Cp* 2 Sc(C,N-~ 2 -c 5 H 4 N}} are discussed. The bonding of the methyl and ethyl ligands of Cp* 2 sccH 3 and Cp* 2 sccH 2 CH 3 are also discussed. The general discussions of C-H activation and olefin reactions by these complexes chapters 2 and 3, respectively. are covered in 5 RESULTS and DISCUSSION Synthesis of Alkyl and Hydride Complexes of Permethylscandocene. We felt that a logical first step into deca- methylscandocene complexes was the synthesis of a halide compound. Cp* 2 Sc(THF)Cl is the product formed initially in the reaction of LiCp* with sccl 3 CTHF) , Ceq 1). 3 ScCl3CTHF)3 + 2LiCp* ~ Cp*2Sc(THF)Cl + 2LiCl + 2THF (1) The THF molecule of Cp* 2 scCTHF)Cl is easily removed by heating in vacuo, to give Cp* 2 scCl, (eq 2). 120°c 10- 4 torr (2) The product, Cp* 2 scCl, is generally obtained by sublimation c120°c, 10- 4 torr) directly from the residue of the initial reaction mixture after the solvent has been removed in vacuo. A solution molecular weight measurement (in C6H6 > and mass spectral analysis show that Cp* 2 sccl is monomeric in solution and in the gas phase. In contrast, Cp 2 scCl is a dimer with bridging chloride ligands. [4] Moreover, Cp 2 scCTHF)Cl does not dissociate THF on sublimation. ClOJ Decamethylscandocene chloride readily undergoes metathesis reactions. Cp* 2 scc 6H4 cH 3 Cortho, meta or para) 6 can be prepared in pure hydrocarbon solvent, by the reaction of Cp* 2 ScCl with the appropriate tolyllithium reagent, (eq 3) • (3) toluene Cp* 2 ScCH 3 and Cp* 2 scc 6 H5 are made by treating Cp* 2 ScCl with a diethylether solution of the corresponding lithium reagent, (eqs 4 and 5). (4) toluene (5) toluene If these reactions are carried out in the presence of THF, the scandium alkyl or aryl complex is not formed, but rather Cp* 2 Sc(THF)Cl is isolated, which is unreactive towards LiMe and LiC 6 H5 • One complex, Cp* 2 ScCH 2 C6 H5 , can be synthesized in the presence of THF, if KCH 2 c 6 H5 is used, (eq 6). THF toluene · Presumably, KCH 2 c 6 H5 is potent enough to react with Cp* 2 Sc(THF)Cl. Permethylscandocene complexes have a very crowded coordination sphere, which can be seen in their base coordi- 7 Both c 6 H5 and CH 2 c H5 ligands are too 6 bulky to allow coordination of THF, while Cp* 2 scMe will conating abilities. ordinate a THF molecule to give Cp* 2 Sc(THF)Me. This THF ligand is bound tightly enough to allow crystallization of an analytically pure sample of Cp* 2 Sc(THF)Me in the absence of excess THF. The 7r cloud of the phenyl portion of a ben- zyl ligand can act as an electron donor ligand, making the benzyl ~ 3 coordinate. However, the NMR spectrum of the phenyl portion of the benzyl ligand of Cp* 2 sccH 2 c 6 H5 remains a simple ABC pattern on cooling to -9o 0 c. We feel that this is an indication of ~l coordination of the the benzyl ligand, but an ~ 3 (allyl like) structure which is rapidly exchanging sides of the phenyl ring, even at -9o 0 c, can not be ruled out. Solution molecular weight determinations show that Cp* 2 ScR (R = CH 3 , CH 2 c 6 H5 > complexes are monomeric. Moreover, a crystal structure determination shows that Cp* 2 sccH 3 is also monomeric in the solid state Cvide infra). All of these scandium alkyl and aryl complexes are formally 14 electron species, with two vacant orbitals.[9] Cp*2ScR (R = CH3, C6HS' CH2C6HS' C6H4CH3> reacts cleanly with hydrogen to give RH and a single scandium species, which we presume is [Cp* 2ScHlx' Ceq 7). (7) hexane 8 This scandium hydride complex is unstable under less than one atmosphere of hydrogen. However, if a hexane solution of Cp* 2 ScH (under excess H2 > is cooled to -78°c, a solid precipitates, which can be isolated. Unfortunately, this solid decomposes immediately on addition of solvent, preventing the measurement of its molecular weight. If four atmospheres of hydrogen are put over a benzene solution of Cp* 2 ScH, it can be heated to 1S0°c for over a week with no sign of decomposition of the scandium complex (monitored by 1 H NMR). When R in (eq 7) is CH 2 c 6 H5 , the reaction consumes 0.93 equiv. of hydrogen and yields 1.0 equiv. of c 6 H5ce 3 • The kinetics of the reaction of Cp* 2 ScR CR = alkyl, aryl) with H2 were not examined in detail1 however, a half-life of 4-5 minutes was observed for the reaction of Cp* 2 sccH 3 with H2 (2 atm) at -78 0 C. A common way to demonstrate the existence of early transition metal hydrides is by their reaction with CH 3 I give CH 4 • to When Cp* 2ScH,generated in situ,is treated with CH 3 I, CH 4 is formed, (eq 8). ( 8) Solutions of Cp* 2 scH catalyze H-D exchange between H2 and n2 , giving statistical mixtures of H2 , HD and n2 , (eq 9). ( 9) 9 This exchange reaction is common for coordinatively unsaturated transition metal hydride complexes,[llJ and isoelectronic main group metal hydrides.[12) The room temperature NMR spectrum of a methylcyclohexane solution Cp* 2 ScH under four atmospheres of hydrogen shows only a single resonance for the Cp* protons and no resonances assignable to the hydride ligand or H2 • Failure to observe these signals ' could be due to broadening of the hydride resonance by the strong quadrupolar moment of the scandium nucleus C-0.22 barns, 100% spin 7/2) ,[13J or rapid exchange of the hydride with hydrogen gas, or both. Quadrupolar broadening can often be decreased by changing the temperature; this changes the quadrupolar relaxation time, giving rise to sharper lines for the observed nucleus. [14) To determine the importance of quadrupolar broadening in this compound, the NMR sample was heated to 120°c. At this temperature a new, broad peak appeared at 6.1 ppm (full width at half height Cfwhh) = 26 Hz.), which must be an average resonance for the scandium hydride and the protons of H2 , since a signal for hydrogen <6 = 4.5 ppm in c 7n14 > is not observed. This averaging process is facile even at lower temperatures. When this NMR sample was cooled to -so 0 c, a new, broad resonance again appeared, this time at 6.2 ppm Cfwhh = 32 Hz.); a similar resonance is observed in toluene at 6.65 ppm(fwhh = 33 Hz.). This exchange pro·c ess could not be frozen out in either methylcyclohexane or 10 toluene; if the sample is cooled to -95°c, a single resonance is still observed for the hydride and hydrogen. The NMR spectrum of Cp* 2 ScH in the absence of H2 has not been measured, but if the observed resonance is a simple weighted average, the chemical shift of the hydride ligand can be calculated from the observed chemical shift and the concentrations of H2 and Cp* 2 ScH.[15J We calculate a chemical shift of 7.3(3) ppm for this hydride in toluene-d 8 • If the scandium alkyl hydrogenation reaction is carried out in the presence of THF, an isolable hydride complex is formed, Ceq 10). THF {10) Cp* 2 Sc(THF)H decomposes over a period of days at room temperature, but is stable for months at -10°c. Cp* 2 sc(THF)D shows the expected shift in the V{Sc-H) IR band from 1390 cm-l to 990 cm- 1 • Like Cp* ScH, Cp* 2 Sc{THF)H re2 acts with CH 3 I to give CH 4 , {eq 11). {11) Solution molecular weight studies show that Cp* 2 Sc(THF)H is monomeric. The room temperature NMR spectrum of Cp* 2 Sc(THF)H shows peaks assignable to the Cp* ligand and the a and p positions 11 of the THF ligand, but no resonance is observed for the scandium hydride. On cooling the sample to -6o 0 c, a new peak grows out of the baseline at 4.8 ppm, and sharpens on further cooling (fwhh (-9o 0 c> = 18 Hz.). If four atmo- spheres of hydrogen is put over the solution, the resonance at 4.8 ppm remains unchanged and a sharp resonance for H2 (4.5 ppm) is observed. We therefore assign the peak at 4.8 ppm to the scandium hydride. Unlike Cp* 2 ScH, the hydride of Cp * 2 sc(THF)H is not exchanging with H2 rapidly at low temperatures; this exchange process is not evident on the NMR time scale (90 MHz) until -20°c and warmer. Even though the hydride resonance disappears above -6o 0 c, due to quadrupolar broadening,the signal for H2 remains until -20°c. In the presence of 10 equiv. of THF this signal persists until o0 c. The fact that this hydride-hydrogen exchange occurs at low temperatures for Cp* 2 ScH, and at a much higher temperature for Cp* 2 Sc(THF)H, suggests that the species actually engaged in this exchange process is Cp* 2 scH, and that Cp* Sc(THF)H 2 must dissociate the THF ligand for its hydride to exchange with H2 , as shown in Scheme 1. In support of this proposal, excess THF inhibits this exchange further , since the added THF should shift the equilibrium toward Cp* 2 Sc(THF)H. 12 + H-H + THF Cp* 2 sc-H + Scheme 1 Although Cp* 2 ScH could not be isolated pure, it has proved to be a valuable reagent when generated in situ. We have found a good route into alkyl and allyl complexes to be the insertion of olef ins into this scandium hydride bond, (eqs 12, 13 and 14). Cp* 2 ScH + H2 C=CH 2 Cp* 2 ScH + H2 C=CHCH 3 Cp* 2 ScCH 2 CH 2CH 3 (13) Cp* 2 ScH + H2 C=CH=CH 2 Cp* 2 Sc(~ 3 -cH 2 CH=CH 2 ) (14) Cp* 2 ScCH 2CH 3 <12) These insertion reactions were conducted at -78°c; at this temperature Cp* 2 ScH reacts with only a single equivalent of olefin to give the observed products. These complexes are all stable to olefin loss at room temperature. NMR coalescence of the syn and anti proton signals of the allylic CH 2 groups, of Cp* 2 scc~ 3 -cH 2 CH=CH 2 ), is observed near room temperature. At low temperature <-so 0 c) a static ~ 3 -structure is observed for the allyl ligand. The signals for the anti and syn protons are doublets at 3.93 ppm and 13 2.07 ppm, respectively, coupled to the central allylic proton by 15.6 Hz and 9.7 pling is observed. multiplet. Hz, respectively~ no geminal cou- The central proton is a non-first order If the sample is heated to 90°c the syn and anti protons become equivalent,[161 giving rise to a doublet at 2.86 ppm, which is coupled to the central proton by 12.7 Hz. The central proton is obscured by solvent resonances in toluene but is a first order quintet at 6.95 ppm in c 6o12 3 ( JH-H = 12.7 Hz). Cp*2ScR (R = CH3, C6H5' CH2C6H5' C6H4CH3), as well as Cp* 2 ScH and Cp* 2 Sc(THF)H, react cleanly with pyridine to give an ~ 2 -orthometallated pyridyl complex, Ceqs 15, 16 and 17) • (15) (16) (17) The reaction of Cp* 2 scH with pyridine presumably goes through a Cp* 2 scCpy)H intermediate, analogous to Cp* 2 Sc(THF)H. Moreover, in the reaction of Cp* 2 sccH 3 , the intermediate, Cp* 2 Sc(py)CH 3 , can be observed before it reacts to give Cp* 2 Sc(C,N-~ 2 -c 5 H 4 N> and CH 4 • The reversibility of the reaction of Cp* 2 ScH with pyridine is seen in the reaction of o2 with Cp* 2 2 Sc(C,N-~ -c 5H4N>, (eq 18). 14 02 < ' Cp~Sc~ + (18) HD D oeuterium incorporation takes place in only the ortho positions of the pyridyl ligand. In the reaction of Cp* 2 Sc(THF)H with pyridine, an 18 electron intermediate, Cp* 2 Sc(THF) Cpy)H, is possible. There is no direct evidence against such an intermediate, however, the steric crowding in it would be great. A similar inter- mediate is possible in the reaction of Cp* 2 Sc(C,N-~ 2 -c 5 H 4 N> with pyridine-~ 5 , Ceq 19). Cp* 2 2 Sc(C,N-~ -C 5 H4N) + c5 n5N -----> Cp* 2 Sc(C,N-~ 2 -c 5 n 4 N) + C5 H4 DN (18) Again, steric crowding would be large if the pyridyl ligand remained ~ 2 when the added pyridine coordinated. A 16 elec- tron structure like (A) is much more favored sterically. (A) An x-ray structural determination was performed on Cp* 2 2 scCC,N-~ -c 5 H4 N>, the ORTEP drawing and skeletal view of 15 which are shown in Figures 1 and 2, respectively. The pyridyl ligand is disordered; the carbon and nitrogen atoms bound to scandium are indistinguishable. The two atoms were modeled as 50% carbon and 50% nitrogen. Structure and Bonding of Cp* 2 ~ 3 and Cp* 2 ~ 2 .CH. 3 ~ Cp* 2 ScR CR = alkyl) complexes are very electron deficient. They are formally 14 electron species with two vacant orbitals. At least one of these orbitals is accessible for bonding, which can be seen in the formation of Cp* 2 Sc(THF)CH 3 and Cp* 2 Sc(C,N-~ 2 -C 5 H4 N). We felt that these 14 electron alkyl complexes would be excellent candidates for a three-center two-electron bond with one of the C-H bonds of the alkyl ligand, as in CB). H v \ R' M c··"' \_,/ ~R (B) Examples of this type of interaction (termed agostic) have been reviewed by Brookhart and Green.[171 By participating in an agostic interaction, the metal center increases its electron count by two; this would make a scandium alkyl cornplex a 16 electron one. Agostic interactions are only seen in electron deficient complexes and are most common in 16 Figure 1. 1.393(12) ' Figure 2. Skeletal view of Cp* sc(C,N-~ 2 -c H N) 2 5 4 with bond distances cK> and angles c0 >. 17 highly deficient ones, such as (Me 2 PCH 2 CH 2 PMe 2 >TiC1 3 R CR= CH 3 , CH 2 cH 3 > .[18] The titanium ethyl complex has an agostic interaction with a p C-H bond. The proton NMR spectrum of Cp* 2 sccH 3 shows a methyl resonance that remains a singlet on cooling to -9o 0 c; 13 Cp* 2 sc cH 3 shows a doublet with 1 JC-H = 111 Hz. The sixteen electron scandium methyl complex, Cp* 2 scCTHF> 13 cH , 3 1 also gives a doublet with JC-H = 111 Hz for the methyl resonance. The spectrum of the ethyl ligand of Cp* 2 sccH cH is 2 3 3 a simple first order pattern with a JH-H = 5.6 Hz, which does not change on cooling to -8o 0 c; 1 JC-H for the a and /J carbons are 128 and 120 Hz, respectively. Agostic interac- tions are often fluxional, such that NMR spectra of agostic ligands show an averaged resonance for the bridging (agostic) and terminal Cnon-agostic) protons, rather than discrete resonances for each type of proton. Partial deuteration of an agostic ligand can provide evidence for a f luxional agostic interaction. A large chemical shift and 1 JC-H difference is generally observed between a partially deuterated agostic ligand and a perprotonic one. This dif- f erence is attributed to an increase in metal hydride character for the protons of the partially deuterated ligand relative to those of the perprotonic one, which shifts the proton resonance and lowers the C-H coupling constant for the partially deuterated ligand relative to the perprotonic one.£17] These chemical shift differences are typically 18 0.35 ppm or greater for agostic methyl ligands. cal shift difference between Cp* 2 sccH 3 <6cH The chemi- = 0.07 ppm) and 3 Cp* 2 sccH 2 D, as well as for several other methyl compounds, are given in Table 1. TABLE 1: Chemical shift differences between CH 3 and CH 2 D ligands. ~6 C6CH --=--1.CH n> Cppm) 3 2 compound Cp* 2 ScCH 3 0.057 Cp* 2 Sc(THF)CH 3 0.047 Cp 2 WCCH 3 > 2 a 0.020 CH a 4 aNon-agostic compound. 0.019 The chemical shift difference for Cp* 2 sccH 3 vs. Cp* 2 ScCH 2 D is rather small. A partially deuterated ethyl complex, Cp* 2 sc 13 cH 2 13 cH 2D, was synthesized, and its NMR spectrum compared to that of Cp* 2 sc 13 cH 213 cH 3 • The methyl group of the ethyl ligand is shifted from 0.139 ppm to 0.049 ppm <~6 = 0.090 ppm) on deuteration. The C-H coupling constant for both the CH 3 and the CH 2n group is 120 Hz, while the CH 2 1 group has a JC-H of 129 Hz. Inf rared spectroscopy can also be used to detect an agostic interaction. Agostic alkyl ligands often have low energy C-H stretching bands in the region V(C-H) = ~ 27002400 cm- 1 .£171 Cp* 2 sccH 3 does not exhibit low energy V{C-H) 19 modes. Moreover, Cp* 2 scco 3 has two VCC-D) bands at 2195 and 2152 cm- 1 • By comparison CD 3 I has V(C-D) bands at 2160 and 2140 cm -1 • Cp* 2 sccH 2 cH 3 , however, has prominent low energy V(C-H) bands at 2593, 2503 and 2440 cm- 1 • The existence of these bands suggests strongly that and agostic interaction is present in the scandium ethyl complex. data appear contradictory. The NMR and IR One possible explanation for this is that an a-agostic interaction is present in Cp* 2 sccH 2 cH 3 • We did not partially deuterate the a posi- tion, so a large difference between its room and low temperature chemical shifts and coupling constants may not be observed. Another possibility is that a P-agostic interaction is present and both the chemical shift and coupling constant of the methyl group are not altered on partial deuteration in the same manner as other reported agostic ligands. In support of this, a crystal structure determination showed CMe 2 PcH 2 CH 2 PMe 2 >TiC1 3cH 2 CH 3 has a P-agostic interaction with the ethyl ligand,[181 but the NMR spectra of the perprotio and P-rnonodeutero complexes are identical.[191 It is also possible that an agostic interaction exists in the solid state but not in solution . The best way to ascertain if a complex has an agostic interaction is to determine its crystal structure. A crystal structure of Cp* 2 sccH 2 cH 3 is now in progress. The crystal structure of Cp* 2 sccH 3 is interesting not only as it pertains to the question of agostisism, but also 20 how it compares to the isoelectronic [Cp* 2 LuCH 3 J 2 structure. [Cp* 2LuCH 3 J 2 is an asymmetric dimer with one bridging and one terminal methyl ligand, where the bridging methyl is in a nearly linear Lu-C-Lu configuration. [20) ing of Cp* 2 sccH 3 is shown in Figure 3. The ORTEP draw- The x-ray crystal structure confirms that Cp* 2 sccH 3 is monomeric in the solid state. Unfortunately, the data set collected at 2s 0 c showed significant static and dynamic disorder in the molecule; we could not locate the methyl ligand hydrogen atoms, but information about the bonding of the methyl ligand can still be obtained from the structure. The methyl carbon has a spherical thermal ellipsoid which lies only 0.01 A out of the plane formed by the ring centroids and the scandium atom; this is consistent with simple covalent-only bonding. If there were a significant agostic interaction, the methyl carbon should have been moved out of this plane. The methyl ligand of {Me 2 PcH 2cH 2 PMe 2 >TiC1 3CH 3 is distorted appreciably from the idealized octahedral site by its agostic interaction. A low temperature data set should remove the disor- der, allowing us to locate the hydrogens, and answer the question of agostisism in this scandium methyl complex, but all of the spectroscopic data we have compiled speaks against the existence of an agostic interaction for this compound. Theoretical calculations suggest that for an agostic interaction to be the ground state, a vacant orbital of the proper symmetry and orientation, as well as a weak- 21 Figure 3. ORTEP drawing of Cp* 2ScCH 3 • 22 ened M-CH 3 bond are required.[211 These calculations pre- dict that L2 M-CH 3 complexes should not have an agostic ground state, which is consistent with our observations for Cp* 2 ScCH 3 • 23 ~XPERIMENTAL General Considerations. All manipulations were carried out using either high vacuum line or glove box techniques. Gas evolution measurements were performed using standard Toepler techniques, and the gases were identified by either mass spectroscopy or NMR. and degassed. Xylene was distilled f rorn CaH 2 THF was purified by vacuum transfer from sodium benzophenone. All other solvents were purified by vacuum transfer first from LiAlH 4 , then from titanocene, prepared as described earlier. [22] sc 2o3 , 99% (Alpha Chern. Co.), CH 2 DI, 13 CH 3 I and 13 c 2 H4 (MSD Isotopes) were used directly with no purification. NMR solvents, benzene-d , 6 toluene-g 8 (Stohler Inc.) and rnethylcyclohexane-g 14 (MSD 0 Isotopes), were vacuum transferred from activated 4 A rnolecular sieves, then from titanocene. Hydrogen was purified by 0 passage over activated 4 A molecular sieves and MnO on LiCp*,[24] ScC1 3 CTHF) 3 ,C25J and Cp 2WCCH 3 > 2 were prepared by previously published procedures. verrniculite.[23] [26] Room temperature 1 H NMR spectra were recorded in c o 6 6 with an internal Si(CH 3 > 4 reference on a Varian EM-390 (90 MHz) spectrometer. Low and high temperature NMR spectra were recorded on a JOEL FX-90Q spectrometer in either c7 o8 or c 7o14 • All NMR spectra are listed in Table 3. Infrared spectra were run as nujol mulls on KBr plates and recorded on a Beckman IR-4240 spectrometer. Mass spectra were b b a b c Cp* Sc(THF)CH 2 3 a Cp* 2 sc 13 cH 3 a Cp* 2 ScCH 2 D b (<l) c b C/J> a b b a b a 1.91 s 3.40 m 1.31 m -0.19 s 1.87 s 0.01 d 1.87 s 0.013 t 1.87 s 0.07 s a Cp* 2 ScCH a 1.83 s c 5 Ccn 3 > 5 Cp* 2 ScC1 b 6 Coom) Assignment Compound 3 1 H and 13 c NMR spectroscopic data.* Table 2. 1J 2J = 111 = 1.8 Couolino Comments f\) +=- b c a b c sc-@d f Cp* 2 Sc-C 6 H4 -CH 3 -m a f c d e b a c b b a a c c Cp* 2 Sc-C 6 H4 -CH 3 -R c b b b a or tho meta para Assignment Cp* ScC 6 H5 2 b a a Cp* ScCH 2 c H5 6 2 Compound 1.77 6.85 7.32 7.07 6.78 2.43 1.76 6.92 7.26 2.33 s d t d s s s d d s 1.73 s 7.0-7.4 m 6.78 d 6.95 t 7.15 t 1.77 s 2.23 s 6.8-7.1 m 6 (ppml 5.7 5.7 5.7 5.4 5.4 8.0 7.5 8.5 Couolin_o 500 Mhz 500 MHz 500 MHz Comments l'\) \..n b b c a b c Cp* 2 Sc 13 cH 2 13 cH 3 a Cp* 2 ScCH 2 CH 3 b c a b c b a c b a c b a a b b c f <{J> ca> c + d e b a Assignment Cp* Sc(THF)H 2 a f sc-@a Cp* ScH 2 a Cp*2Sc-C 6 H4 CH 3-.Q CQmQQUnd 6.0 5.3 CQugling 0.139 1.88 s 0.91 1.88 s 0.97 q 0.18 t 2.08 s 3.43 ID 1.25 ID 4.80 s = 129 1 J = 120 1J 6.8 6.8 1.90 s 7.3 (calculated) 1.26 a 2.12 s 1.77 s 6.32 a 7.20 ID 6 (J;!Qm) 400 MHz, C7D9 -80 0 C, c n 7 8 500 MHz Comments CJ' N b a b c a a d c b 118.60 11.19 45. 8 Cb road) 22. 76 24.58 a CC ) a ( ceH 3 ) 5 > c Cp* 2 ScCH 2 cH 2 CH 3 b 1.88 s 1.40 t 0.86-1.00 m a b c + a Cp* 2 ScCH 2 CH 2 CH 3 a -0.046 b b a 0.049 1.035 b 0.91 0.077 1.035 o {1212ml a b a b a AfHiignment Cp* 2 Sc 13 cH 213 cH 2 D a Cp* 2 sc 13 cH 2 13 cH 2 D a CQIDJ;2Q!JDC Cp * 2 Sc 13 cH 2 13 cH 3 = 121 6.6 6.6 = 130 1 J = 119 1J = 129 1 J = 120 1J 1J CQUJ;2ling 1 J = 130 -ao 0 c, -56 -ao 0 c, 13c t1Hl, c B 400 MHz, ~ 500 MHz 400 MHz, C7D9 400 MHz, c n 8 7 400 MHz, C7D9 CQmments [\) --..:I b c d b a b b b b 1.77 s 7.93 d t 7.65 d t 1.03 t d 6.58 m 1.84 s 2.06 d a b + d s s d d 1.92 1.80 2.01 3.93 1.90 s 3.o d 7.15 q 6coom> a a' b d a b + D c Assiqrunent 5.1, 1.5 7.2, 1.5 7.2, 1.2 12.7 9.8 15.6 12.7 12.7 Coupling 0 90 C, c n 7 8 -50 0 C, c n 7 8 25 C, c n 6 12 0 Comments 90 Mhz. * Unless otherwise specified, spectra were taken in c 6 n6 at ambient temperature at The chemical shifts are reported in relative to TMS or residual protons of the solvent. Coupling constants are reported in Hz. a Cp* 2 Sc(C,N-~ 2 -C 5 H 4 N) a Cp* 2 ScCH 2 CH=CH 2 Comoound co {\) 29 recorded using Kratos MS25 and Dupont 4592 mass spectrometers. Analyses were performed by Bernhardt Analytical Labo- ratories or the Caltech Analytical Laboratory. Molecular weights were determined using the vapor phase osmometry technique developed by Signer and described by Clark.[27] Structure Determination of Cp2Sc(C,N-~ 2 -CsH 4 N). Single crystals, grown from a toluene solution at -80°C for two days were mounted in glass capillaries under N2. Oscillation and Weissenberg pho- tographs indicated orthorhombic symmetry (a= 16.297(4)A, b = 9.684(3)A, c = 14.486(6)A, V = 2286.3(12)A3 , Z = 4). rapidly with x-ray exposure. The crystals decompose After 8--10 hours the reflection in- tensities begin to decrease, and after 24 hours no diffraction is observed. Intensity data were collected on a locally-modified Syn- tex P2 1 diffractometer with graphite monochromator and MoKa radiation at a fast scan speed (8.37°/min in 20); an entire quadrant (+h,+k,±i) to 20 = 30° was collected in 8.5 hours with 8--20 scans (.D.20 = 2.8° plus dispersion) and indicated no observable deterioration. Systematic absences in the intensity data indicated space group Pna2 1 • The intensities were reduced to F;; the form fac- tors for Sc were corrected for anomalous dispersion. The coordinates of the scandium atom was derived from the Patterson map, and a series of Fourier maps were used to determine the coordinates of the remaining non-hydrogen atoms. Hydrogen atom po- sitions were determined by assuming idealized geometry and by locating peaks in the difference map. The pyridyl ligand is disordered and results in the complete superposition of two pyridyl groups of equal population. To model 30 this, all atoms in the pyridyl group were given full populations, except for the two atoms bound to Sc, which were each modelled as 50% carbon and 50% nitrogen atom. Least-squares refinement of the coordinates of all non-hydrogen atoms with anisotropic Gaussian displacement parameters, Uij 's, an isotropic secondary extinc- tion parameter, and hydrogen atoms with fixed coordinates and isotropic U's (U = 0.09.i1 2 ) resulted in S (goodness--of--fit) {:Ew[F~-(F; /k) 2 ] 2 /(np)}1/:l • 1. 74, R (E II Fa I - I Fe II/EI F0 I for I > 0) • 0.045, and R' (R for I > 3ur) • 0.036; final shift/error < 0.01. Final atom coordinates and Uij'S are given in Table 3, hydrogen atom coordinates in Table 4, bond lengths in Table 5, and angles in Table 6. The final value of the secondary extinction parameter is 1.55(16)x10- 6 . The atoms numbered 1 and 2 are the pyridyl atoms bound to scandium. Atoms C3-C6 are the remaining atoms (carbons) of the pyridyl ring. C11-C15 and C31-C35 are cp• inner ring carbons, and C21-C25 and C41C45 are the methyl groups associated with them, respectively. Structure determination of 2 3 Cp* .5..Q.CH ~ Single crystals, grown from pentane at -ao 0 c for two days, were mounted in glass capillaries under N2 • The crystals were screened by examination of their oscillation photographs and a satisfactory one was carefully centered on a Syntex P2 1 diff ractometer equipped with a graphite-monochromated MoKa radiation. An orthorhombic cell was found and cell dimensions were obtained from a least-squares fit to the setting angles of 15 reflections (various forms of 6 independent reflections) with 20°<26<26°. Systematic absences observed in the data of hOO, h = 2n + 1; OkO; k = 2n + 1; and 001, 1 = 2n + 1 are C(43 C(44) C(45) C(42~ C(34) C(35) C(41) cC~32~ 33 C(23 C(24) C(25) C(31) C(22~ C(l4) C(l5) C(21) cC~l2~ 13 C(5) C(6) C(ll) :m gm Sc 28578(7) 17558(39) 18949(43) 13564(59~ 6534(54 5100(52) 10543(58) 30262(46) 30997(44) 23404(45) 17865(49) 22167(46) 36505(57) 38246(51) 21350(54) 8820(50) 18637(61) 43661(37) 41579(42) 36338(46) 35092(43) 39511(39) 50417(47) 45133(63~ 33535(60 30558(53) 40348(59) % -1816(11) -10404(77) 1576(79) 6826(94~ -317(131 -13007(118) -17920(00) 4534(77) 16736(77) 19636(70) 0056(86) -513(76) -1082(110) 26819(87) 32221(72) 8693(W) -12926(91) -7041(67) -400(78) -9461(86) -21194(73) -19737(63) -2484(102) 13014(89~ -6858(124 -34019(00) -30615(94) J,I 18697 63 14513(81) 8270(62) -16638(51) -11647(54) -7498(55) -10054(55) -15617(53) -23613(63) -11429(67~ -1872(65 -8205(73) -20048(67) 828(76) 0052(60) 14017(50) 8634(60) 432(77) -5893(65) 12332(77~ 23858(70 11543(79) -6777(77) 16781~61~ 0 6026(51) 10236(54) % 401(44 376(46) 429(40) 391(44) 1029(73) 776(65) 711(64) 876(68) 419(45~ Un 415(7) 400(42) 646(63) 953(72~ 1381(94 1254(93) 842(62) 541(54) 476(48) 432( 46) 558(56) 433(47) 1111(86) 722(57~ 553(44 1033(69) 722(63) 552(47) 537(55) 711(57) 462(51) 472(44) 1202(77) 731(63) 1346(100) 769(62) 806(65) U11 ; 317(6) ·'13(44) 376(42) 590(60) 505(61) 397(55) 540(57) 549(56) 452(55~ 525(56 396(51) 572(56) 767(62) 617(56~ . 889(54 451(57) 1000(79) 370(39) -2(7) -56(37) 42(44) 119(58~ 171(75 -228(62) -206(54) 130(43) -8(41~ 66(45 73(52) 16(52) 197(71) -66~50) 234 49) 179(53) -88(58) 42(34) -38~43) 120 45) -30(39) 144(36) 67(54) -17(60) 265(67) -71(53) 314(58) 427 47 433(56) 508(51) 562(59) 939(69~ 789(84 925(73) 695(62) 525(49) 661(61~ 326(47 632(61) 605(48) 934(65) 1014(76) 543(65) 1235(85) 1165(00) 507~51~ ul3 31(11) 156(40) 131(47) 22(54) 36(74) 462(77) 143(54) 96(42) 132( 45) 85(40) 101(46) -9(47) 114(58) 371(57l -18(51 307(63) -76(54) 171(61) 139(56~ 153(48 174(48) 51(64) 451(67) 45(59) 118(63) 412(63) -142(68) U13 -17(10) -48(39) -50(41) -11(53~ 152(51 -113(61) -141(53) 15(42) -36(44~ -50(42 -00(44) -228(49) 195(55) 50(54) -57(59) -136(53) -287(59) -22(57) -225(45~ 18(41 -85(49) -11(65) 103(47) -506(67~ -46(56 -103(61) -150(67) U12 U33 300(7) 509(48) 470(49) 552(62~ 630(61 891(82) 615(62) 415(46) 5 Atomic coordinates (x 10 ) and gaussian 4 amplitudes (A2, x 10 ) for all non-hydrogen 2 cttoms of Cp* Sc(C,N-ry -c H N). 5 4 2 Table 3. U.q 888(34 005(29) 949(30) 925(29~ 374(3) 471(18) 497(21) 698(28~ 839(33 847(35) 666(25) 502(21) 478(21) 461(20) 462(23) 504(21) 813(29) 759(25) 744(25) 803(30) 836(28) 482(19) 539(22) 479(22) 400(22) 502(18) 842(28) \...V I-" 32 Hydrogen atom coodinates (x 10 4 ) for 2 Cp* Sc(C,N-~ -C H N). 2 5 4 Table 4. :t H3 H4 H5 H6 H 211 H 212 H 213 H 221 H 222 H 223 H 231 H 232 H 233 H 241 H 242 H 243 H!251 H 252 H 253 H 411 H 412 H 413 H 421 H 422 H 423 H 431 H 432 H 433 H 441 H 442 H 443 H 451 H 452 H 453 1507 222 0 933 3260 4080 3490 3770 3400 4330 2510 1640 1760 720 650 560 1770 1410 2280 4880 5410 5400 4480 5070 4180 3780 3100 2870 2810 3500 2740 3530 4220 4460 JI % 1571 2015 398 2302 -1856 1644 -2717 483 -300 -2951 380 -2421 -1140 -2171 3310 -1641 3100 -611 2170 -1301 3550 79 2940 289 3820 -581 1310 -391 -110 -901 1310 -1411 -2090 -1551 -1170 -2341 -1770 -2401 -10 -1101 -1130 -741 370 -341 1970 819 1489 1150 1560 1809 -1000 2809 70 2449 -1470 2519 -3310 1669 -4160 1219 -3720 649 -3620 -771 -2710 -1271 -3770 -511 33 () Distance CA) A.t..om At.Qm Distance CA) 1 2.162(7) Cll Cl2 1.390(11) 2 2.183(8) ClS 1.414(11) Cll 2.502(8) C21 1.533(12) Cl2 2.496(8) Cl3 1.403(12) Cl3 2.491 (7) C22 1.533(12) Cl4 2.505(8) Cl4 1.414(11) C31 2.511(8) C23 1.504(11) C32 2.494(8) ClS l.414Cll) C33 2.504(8) C24 1.498(12) C34 2.492(8) Cl5 C25 1.480(12) C35 2.487(8) C31 C32 1.392(11) At.Qm At.mn Sc Cl2 Cl3 Cl4 2 1.330(10) C35 1.404(11) C6 1.393 Cl2) C41 l.534Cl2) 2 C3 1.388(12) C33 1.414(11) C3 C4 1.366(14) C42 1.507(13) C4 cs 1.390(15) C34 1.393(11) cs C6 1.353(14) C43 1.518(13) C35 1.400(11) C44 1.505(12) C45 1.490(13) 1 C32 C33 C34 C35 34 Table 6. Bond angles for Cp* 2 2 Sc(C,N-~ -C 5 H4N>. ,o, At...oJn Angle CENT2 142 Sc 2 35. 7 (2) 2 Sc 71.3(3) cs C6 120.6 (8) 1 Sc 120.6 (8) 1 SC 73.0(5) C3 C4 118.8(8) C3 C4 cs 120.0 (9) C4 cs C6 119.S (9) At...oJn At..!2m CENTl Sc l 2 A.t.mn .MJ2m At..QJl Angle t 0 l .MJ2m Mmn Mmn Angle t 0 l ClS Cll Cl2 108.6 (7) C35 C31 C32 108.5 (7) C21 124.3(7) C41 124. 8 (7) C21 126.0 (7) C32 C41 12S.9(7) Cl3 108. 5 (7) C31 C33 107.6(7) C22 128.2 (7) C42 124.9(7) C22 122.8 (7) C33 C42 127.2(7) Cl4 107.8(7) C32 C34 107.7(7) C23 126 .1 (7) C44 123.9(7) C23 l2S.9(7) C34 C44 128.l (7) ClS 107 .9 (7) C33 C35 108.6 (7) C24 126.7(7) C44 126.0(7) C24 125.0 (7) C35 C44 l2S.l (7) Cll 107 .l (7) C34 C3l 107.6(7) C25 126 .o (7) C45 124. 9 (7) C2S 125.7(7) C45 127 .l (7) Cl2 Cll Cl2 Cl3 Cl2 Cl3 Cl4 Cl3 Cl4 ClS Cl4 Cll ClS C31 C32 C33 C34 C35 35 unique for the space group P2 1 21 21 • Five octants of data Chkl, hkl, hkl, hkl and hkl> were collected with 4°<28<40° and one octant, hkl, with 40°<29<50°. The data were merged to give 2076 independent reflections, of which 1859 had F0 >0 and 1143 F 2 >3U(F 2 >. Three check reflections monitored ev0 0 ery 97 reflections showed a linear decay of 1.5% in F over 230 hours required to collect the data. The data were cor- rected for this decay, Lorentz and polarization factors were applied and a Wilson plot was used to put the data on an approximately absolute scale. Variances, u 2 CI), were assigned intensities on the basis of counting statistics plus an additional term, C0.02 r> 2 , to account for additional errors proportional to the intensity. A Patterson map gave the scandium coordinates and successive structure factor-Fourier calculations located the remaining atoms. Many cycles of least squares refinement, first with isotropic and then anisotropic thermal parameters finally converged with R = 0.15 for all reflections with F0 >0 and 0.112 for reflections with F 2 >3 CF 2 >. Hydrogen atoms were introduced, based on 0 0 difference maps calculated in the planes where they were expected. These maps were unequivocal, as the methyl carbon atoms have very large anisotropic temperature factors, but in most cases 3 reasonable locations for the hydrogen atoms were obtained. These were optimized to tetrahedral geometry 0 with a C-H distance of 0.95 A and the hydrogen atoms were included in subsequent calculations as constant contribu- 36 tions to the structure factors, with isotropic thermal pa•2 • Six more cycles of least squares, rameters, B, of 10.0 A the hydrogen atoms being repositioned two times, concluded the refinement: the final R values are 0.113 for all data and 0.72 for the strong data. 2.98. The goodness of fit (S) is In the final least squares cycle no parameter shifted more than one-half its standard deviation. A final difference map showed no excursion greater than ±0.44 e/A 3 • Crys- tal data are given in Table 7. Hydrogen atom coordinates given in Tables, and final heavy atom coordinates and u .. 's 1] are given in Table 9. Bond lengths and angles are given in Table 10. Atoms numbered Cl-CS and Cll-ClS are Cp* inner ring carbons, and atoms numbered C6-Cl0 and Cl6-C20 are the methyl groups attached to them, respectively. C21 is the methyl ligand bound to scandium. Preparation of Cp* 2 sccl. LiCp* (9.0 g, 63 mmol) scc1 3 CTHF) 3 (10.0 g, 27 mmol), and 250 mL of xylene were placed in a 500 mL round bottom flask equipped with a reflux condenser. The slurry was refluxed for one day and approxi- mately 30 mL of solvent was then distilled from the reaction flask. The slurry was refluxed for a second day and 30 mL of solvent was again removed. The reaction was refluxed for a third day and all of the volatiles removed in vacuo. remaining solid was transferred to a sublimator. The Bright yellow crystalline Cp* 2 ScCl (8.0 g, 80%) was collected by sublimation for four days at 120 °c (10-s torr). The molec- 37 Table 7. Crystal data for Cp* sccH • 2 3 formula: Sc C21 H33 f. wt. 330.45 Space Group P212121(#19) V • 2033(1) A o3 z '"' 4 0 a a:: 8.502(2)A 0 b a:: 11.095(4) A (T • 21°C) Pc• 1.079(1) g/cc 0 c .. 21.554(7) A 0 .>. MoKii = o. 71073 A F(OOO) .. 720.0 e µrmax .. 0.12 Crystal size 0.24 x 0.34 x 0.51 mm 4 Table 8. Hydrogen atom coordinates (x 10 > of Cp* ScCH • 2 3 All Hydrogen atoms were assigned isotropic thermal parameters, B, of 10.0 A2. They are numbered according to the carbon atom to which they are attached. Atom x y z Atom x y z H61 H62 H63 H71 H72 H73 H81 H82 H83 H91 H92 H93 HlOl H102 H103 -1840 -1896 -1820 1446 28 1717 4522 3997 5070 4743 3650 4463 845 -663 786 -551 -1647 -339 905 79 -258 -94 -1245 -1362 -2406 -2906 -1657 -3150 -2354 -2054 964 512 249 -333 -504 -725 -114 -473 110 854 1374 1455 1475 1511 1926 Hl61 H162 H163 H171 H172 Hl73 H181 Hl82 H183 H191 H192 Hl93 H201 H202 H203 -1166 -2248 -1505 2824 1048 1723 4735 3932 4286 3197 1473 2206 -984 -2039 -1706 3067 2262 1738 3881 4229 3430 1756 2639 3054 -119 -510 592 -545 291 622 1054 1465 862 1291 1293 761 2134 2598 1922 2620 2762 3106 2245 1839 2529 H211 H212 H213 5118 4215 4340 967 2141 1091 982 808 330 38 4 Final atomic coordinates (x 10 ) and gaussian amplitudes (x 10 3 ) for Cp* ScCH • 2 3 a Table 9. Atom Sc Cl . C2 C3 C4 cs C6 C7 CB C9 ClO Cll Cl2 Cl3 Cl4 ClS Cl6 Cl7 Cl8 Cl9 C20 C21 x y 2029(2) 249(8) 1362(17) 2718( 14) 261S(l 7) 1060(19) -lSOO(ll) 1113(29) 4244(19) 4009(22) 4Sl( 38) 6S(ll) 1416(13) 24S8(10) 1604(17) 169(1S) -13S9(12) 1792(14) 4001(14) 217S(26) -1291(1S) 42S9(13) 49S(l) -913(8) -602(10) -1018( 11) -1S97(9) -1602(9) -8S6(14) 100(10) -919(13) -2203(13) -2361(14) 18S9(10) 2489(7) 19S9(12) 9Sl(lO) 977(9) 227400) 3624(8) 2394(1S) lSl (12) 263(11) 1293(9) z 1131(1) 579(6) 17S(4) 353(6) 86 7 (7) 1061(5) 576(10) -404(5) -79(8) 1168(11) 1537(6) 1S69(5) 1568(4) 1982(5) 2239(4) 196S(6) 1200(6) 1190(5) 2178(6) 2729(4) 216S(6) 757(5) Ueq b 53(0.3) 72(3) 78(3) 94(4) 100(4) 90(4) 313(11) 334(10) 241(7) 315(10) 367(11) 82(3) 72(3) 84(4) 93(4) 92(4) 15S(4) 151(4) 218(6) 284(9) 183(5) 156(4) Atom U11 U22 U33 U12 U13 U23 Sc 46( l) 30(5) 113(8) 77(9) 128(12) 157(10) 58(7) 7S0(44) 189(13) 354(23) 843(52) 63(6) 104(8) 46(7) 126(11) 121 (10) 111(8) 217(12) 104(10) 610(33) 220(13) 125(9) 57(1) 70( 7) 66(6) 89(9) 50(6) 67( 7) 289(20) 153(12) 200(16) 114(12) 130(11) 75(7) 49(6) 133( 10) 98(9) 65(8) 151 (10) 72(6) 344(22) 18S(l3) 132(10) 162(10) 57(1) 116(8) 56(6) 116(9) 123(10) 47(6) 594(31) 100(8) 334(19) 477(30) 128(10) 109(8) 62(6) 71(6) 54(6) 89(8) 204(11) 163(9) 206(13) 56(6) 198(11) 180(10) -3(1) 1(4) -22(7) -3( 7) 12(7) -62(8) 43(9) -187(20) -73(12) 137(15) -179(22) 0(7) -12(6) 8(6) 51(8) -4(7) 58(7) 9(8) 16(12) 182(20) -70(10) -69(9) 1(1) 3(6) -31( 6) 11 ( 9) -77(10) 29(8) -73(11) -131 (16) 15S(l5) -270(25) 193(21) -1(6) 19(6) -17(5) -14(6) 61(8) -53(9) 86(11) -37(9) -40(13) 136(11) 88(8) -10(1) -50(6) 12(6) - 5S( 7) 4(6) -1(6) -358(23) 4(9) -160(15) -79(17) -52(9) -23(7) -16(5) -54(7) -3(6) -6(6) -83(10) 11(7) -192(15) 0(7) -78(9) -87(9) Cl C2 C3 C4 cs C6 C7 CB C9 ClO Cll Cl2 Cl3 Cl4 Cl5 Cl6 Cl7 Cl8 Cl9 C20 C21 39 Table 10. Bond lengths Cp* scCH 3 • 2 0 0 (A) and angles Atom Atom Dist. (A) Atom Atom Sc Cl C2 C3 C4 2.4BO(lO) 2.460(11) 2 .444(13) 2 .441 (13) 2.473(12) 2.443(10) 2.460(9) 2.476(11) 2.467(11) 2.4S2(12) l.333(1S) 1.463(1S) 1.48B(lB) 1. 300(17) l.4BS(l 9) 1. 2B4(18) 1.601(20) 1. 3B6(18) 1. S09( 23) 1.42S(24) 1. 344(14) 1.302(15) 1. 52005) 1. 388(14) l.S33(14) 1.444(16) 1.461(17) 1. 356(16) 1.462(19) 1. S34(17) 2.243(11) C2 Cl cs Cl C2 C3 C4 cs Cll Cl2 Cl3 Cl4 Cl5 Sc Cll Cl2 Cl3 Cl4 ClS C2 cs C6 C3 C7 C4 CB cs C9 ClO Cl2 Cl5 Cl6 Cl3 Cl7 Cl4 Cl8 ClS Cl9 C20 C21 cs cs C9 C9 Cl ClO ClO ClS Cl6 Cl6 Cl3 Cl7 Cl7 Cl4 Cl8 Cl8 ClS Cl9 Cl9 109. S(lO) 121.S(9) 128.6(10) 108.9(9) 127.2(9) 123.B(9) 104.7(9) 128.3(10) 126. 6(10) lOS. 6(10) 12S.5(11) 128.B(ll) ClS Cll C20 C20 111. 3(11) 126.S(ll) 121.2(10) C3 C2 C3 C6 C6 C3 C7 C7 C4 cs CB Cl Cl2 Cll ClS Cll Cl2 Cl3 Cl2 Cl3 Cl4 Cl3 Cl4 Cl5 Cl4 Angle( 0 ) cs C2 C4 Atom C4 cs cs for 105.4(9) 134 .1(10) 119.B(lO) 110.1(10) 12S.7(11) 124.2(11) 111. 9(12) 121. 5(11) 126. 3(12) 109.1(12) 122.7(13) 12B.2(13) 103.S(lO) 124.S(13) 130.S(12) Cl C4 C3 (0) Cll Sc-Cp*Cl) centroid 2.173(12) Cp*l Sc Cp*2 144. 6(10) Sc-Cp*(2) centroid 2.169(11) Cp*l Sc C21 106.6(8) Cp*2 Sc C21 108.9(B) 40 ular weight in benzene was 365 g/mole (350 calculated for a monomer). 360(s). 10.10. IR data Ccm-l): 2715(w), 1030(m), 800(m), 430(s), Anal. calc. for c 20 H30 c1sc: C, 68.471 H, 8.771 Cl, Found: c, 68.051 H, 8.771 Cl, 10.12. Preparation of Cp* 2 ~ 3 i Cp* 2 ScCl (5.0 g, 14 mmol) was dissolved in 75 mL of toluene. LiCH 3 (8.01 mL, 1.78 M in Et 2o> was slowly added to the solution with stirring. The reaction was stirred for 30 minutes and the solvent removed in vacuo. Approximately 30 mL of petroleum ether was distilled onto the solid and the resulting slurry was filtered. The solid was washed with 10 mL of petroleum ether twice. The filtrate and washings were combined and concen- trated to 10 mL. The solution was cooled to -78 °c, precip- itating the pale yellow product, Cp* 2 ScCH 3 • The product was isolated by cold filtration (2.6 g, 55%). Cp* 2 sccH 3 was stored at -10 OC to prevent decomposition. The molecular weight in cyclohexane was 324 g/mole (331 calculated for a monomer). IR data Ccm- 1 >: 2730 Cw), 1490 Cm), 1114 Cs), 1028 Cs), 800 Cw), 605 Cm), 580 Cm), 420 Cvs, br). for C21H33SC: c, 76.331 H, 10.07. Anal. calc. Found: c, 76.011 H, 9.98. Preparation of Cp* 2 ScCH 2D and Cp* 2 sc 13 cH 3i These com- pounds were prepared in a manner similar to that described for Cp* 2 sccH 3 , except LiCH 2D and Li 13 cH 3 , respectively, were used in place of LiCH 3 • 41 Preparation of Cp* 2 Sc(THF)CH 3 ~ Cp* 2 ScCH 3 (1.0 g, 3.0 mmol) was dissolved in approximately 15 mL of THF and the THF removed in vacuo. Five mL of petroleum ether was added to the resulting solid and the solution cooled to -78 °c, precipitating a solid. This slurry was filtered to give a pale yellow product (0.99 g, 81%) IR data (cm-l): 2710(w), 1490 (m), 1340 (m), 1295 (w), 1242 (w), 1100 (s), 1020 (s), 922 (m), 865 (s), 670 (m), 570 (m), 460 (m), 430 (s), 340 (s). Anal. calc. for c 25 H37 osc: C, 74.59; H, 10.27. 74.53; H, 10.17. Preparation of Cp* 2 ~ 2 ~ 6 n 5 ~ Found: c, Cp* 2 ScCl (2.0 g, 5.7 mmol) was dissolved in 20 mL of toluene. A solution of 0.830 g (6.3 mmol) of KCH 2 c 6H5 dissolved in 20 mL of cold THF was added to the stirred Cp* 2 scCl solution slowly via cannula. After addition of the benzylpotassium solution was complete, the solvent was removed in vacuo, followed by addition of 20 mL of petroleum ether. The resulting suspen- sion was filtered to give a yellow solution and a white solid. The solid was washed with two 10 mL portions of petroleum ether and the combined petroleum ether solutions were concentrated to 5 mL. Cooling the solution to -78 °c caused thebright yellow Cp* 2 sccH 2 c 6 H5 to precipitate (1.34 g, 58%). Cp* 2 sccH 2 c 6 H5 was stored at -10°c to prevent de- composition. The molecular weight in benzene was found to be 460 g/mole (407 calculated for a monomer). IR data (cm- 1 >: 3050 (w), 2710 (w), 1595 (s), 1490 (s), 1200 Cs), 42 1030 (m), 935 (m), 870 Cw), 790 (m), 740 (s), 705 (s), 660 (w). Anal. calc. for c 27 H37 sc: C, 79.76; H, 9.17. 79.67; H, 9.12. Preparation of Cp* 2 £c..=.c 6 n 5 ~ Found: C, Cp* 2 ScCl (2.5 g, 7.6 mmol) was dissolved in 20 mL of petroleum ether. LiC 6 H5 (3.11 mL, 2.4 M in a cyclohexane-Et 2 o solution) was added slowly to the petroleum ether solution, and the reaction mixture stirred for 30 minutes. All of the solvent was removed in vacuo and 20 mL of petroleum ether was added. The solution was filtered and the resulting solid washed twice with 10 mL portions of petroleum ether. The filtrate and washings were combined and concentrated to 5 mL. On cooling to -78 °c the pale yellow product precipitated and was collected by filtration Cl.71 g, 60%). IR data (cm- 1 >: 3430(w), 2730(w), 1590 (w), 1563 Cw), 1485 Cw), 1312 (w), 1233 (w), 1165 (w), 1067 Cm), 1025 (m), 983 (w), 738 Cw), 722 (s), 708 Cs), 462 (s), 418Cs). Anal. Cale. for c 26 H35 sc: C, 79.56; H, 8.99. Found: C, 79.58; H, 8.77. Preparation of Cp* 2 ~ 6 H 4 .c.H. 3 -o, -m, -p. The synthesis of Cp* 2sc-c 6 H4 cH 3 -n is described; the -Q and -m isomers are prepared in the same manner, except the appropriate tolyllithium reagent is substituted for LiC 6 H4cH 3 -n. Cp* 2 ScCl C0.420 g, 1.2 mmol) reacted with LiC 6 H4 cH 3 -n C0.18 g, 1.8 mmol) at room temperature in toluene solution (15 mL). Af- ter 12 hours the toluene was removed and replaced with 10 mL 43 of petroleum ether. The petroleum ether solution was fil- tered and the resulting solid washed twice with petroleum ether. The combined filtrate and washings were concentrated to ca. three mL, and cooled to -78°c. The off-white product was isolated by filtration of the cold petroleum ether slurry (0.285 g, 59%}. IR data (cm- 1 >: 2715(w}, 1578(w}, 1300 Cw}, 1252 (w}, 1222 (m}, 1208 (w}, 1160 (w}, 1140 (w}, 1079(m}, 1053(w}, 1030(s}, 1016(s), 793(w), 774Cs), 720(w), 549(m), 480(s), 417(s). H, 9.17. Anal. calc. for c 27 H37 sc: c, 79.77; Found: C, 79.79; H, 9.08. Preparation of Cp* 2 Sc(THF)H. Cp* 2 ScCH 3 (1.0 g, 3.0 mmol) was dissolved on 20 mL of THF. One atmosphere of hy- drogen was admitted to the reaction vessel and the solution stirred for five hours. The THF was removed in vacuo and replaced with seven mL of petroleum ether. This solution was cooled to -78°c, causing the pale yellow Cp* 2 sc(THF)H to precipitate. g, 70%). The product was isolated by filtration C0.82 The molecular weight in cyclohexane was found to be 408 g/mole (388 calculated for a monomer). IR data (cm- 1 >: 2710 (w), 1390 Cs), 1155 (m), 1140 Cm), 1020 (s), 855 (s), 730 (s), 655 (m), 510 (m), 365 (m). C, 74.19; H, 10.12. Anal. calc. for c 24 H39 osc: Found: c, 74.08; H, 10.14. Preparation of Cp* 2 ~ 2 QI. 3 ~ A 50 mL thick walled glass bomb was charged with Cp* 2 sccH 3 (0.280 g, 0.85 mmol), 10 mL of toluene and four atmospheres of H2 • This toluene 44 solution was stirred for ten minutes at room temperature and then cooled to -78°c. The hydrogen was removed and one equivalent of ethylene (375 torr in 42.2 mL) was condensed into the cold bomb C-78°c>. The bomb was allowed to come to room temperature and the toluene solution transferred to a flask. The toluene was removed in vacuo and replaced with 10 mL of petroleum ether. This solution was filtered and the filtrate concentrated to ca. three mL. Cooling this so- lution to -78°c precipitates a yellow solid, which was isolated by filtration C0.160 g, 55%). IR data Ccm- 1 >: 2714 (m), 2593 (m), 2503 (m), 2440 (m) ,1493 (s), 1172 Cw), 1066 (m), 1025 Cs), 973 Cm), 844 (s), 807 Cm), 722 (w), 660 (m), 618(m), 568(s), 480(s). H, 10.24. Anal. calc. for c 22 H35 sc: c, 76.71; Found: c, 76.66; H, 10.22. Preparation of Cp* 2 sc 13 cH 213 cH 3 and Cp* 2 sc 13 cH 213 cH 2.Ih_ Cp* 2 sc 13 cH 2 13 cH 3 was synthesized by substituting 13 c 2 H4 for nonlabeled ethylene in the procedure described for Cp* 2 sccH 2 CH 3 • Cp* 2 sc 13 cH 213 cH 2D was prepared by substituting labeled ethylene for unlabeled ethylene and using deuterium rather than hydrogen in the described synthesis of Cp* 2 sccH 2 cH 3 , and the reaction time for deuterium with Cp* 2 sccH 3 was five minutes. Preparation of Cp* 2 ~ 2 .CH. 2 .CH. 3 ~ We were unable to pre- pare this compound in a manner analogous to the preparation 45 active solvents to allow its isolation by cooling a concentrated solution. In a typical experiment 500 mg of Cp* 2 sccH 3 Cl.5 mmol) and 10 mL of toluene were loaded into a 50 mL glass bomb. The bomb was charged with 4 atm of H2 and the solution stirred for 10 minutes. The toluene solution was cooled to -78°c and the H2 removed in vacuo. (1. 5 Propene mmol) was added to the cold toluene solution, and the bomb allowed to warm to room temperature slowly. The toluene was removed in vacuQ and replaced with 5 mL of petroleum ether. This solution was cloudy on some occasions (due to a small amount of H2o in the H2 ), when this occurred the solution was filtered. The petroleum ether was removed in vacuo and the resulting solid was the product (90-95% pure, 440 mg, 80% yield). The 500 MHz 1 H NMR spectrum of Cp* 2 ScCH 2cH 2 cH 3 shows a single Cp* resonance Cl.86 ppm, 30H), a triplet for the a-cH 2 group Cl.40 ppm, J = 6.6 Hz, 2H) and two overlapping multiplets for the p-cH 2 and CH 3 groups C0.86-1.00 ppm, SH). The 13 c spectrum of this compound is reported in Table 2. were not obtained. Satisfactory C,H analysis IR data Ccm-l): 2735(w), 1306(w), 1160 Cbr, w), 1060 (sh), 1028 Cs), 862 Cw), 805 Cw), 750 Cm), 723 Cm), 666 Cs), 618 Cm), 580 Cm), 505 Cm), 412 Cvs). Preparation of Cp* 2 sccn 3-cH 2CH=CH 21........ This compound was prepared in a manner analogous to Cp* 2 sccH 2 cH 3 , except allene was used in place of ethylene1 0.220 g of Cp* 2 scCl yielded 0.160 g of Cp* 2 ScCH 2CH=CH 2 (67%). IR data (cm-l): 46 3090 (w), 2725 (w), 1554 (s), 1258 (s), 1155 Cw), 1085 (w), 1028(s), 785(s), 755(rn), 705(s), 635(sh), 618(sh}, 595(rn}. Anal. calc. for c 23 H35 sc: c, 77.49; H, 9.90. 77.25; H, 10.00. Preparation of Cp* 2 Sc(C,N-n 2 -c 5H4.NL&. Found: c, Cl.O g, 3 3.0 rnrnol) was dissolved in 25 rnL of benzene and 0.25 rnL (3.1 rnmol) of pyridine was added. Cp* 2 sccH The solution was heated to re- flux for 30 minutes and the solvent removed in vacuo. The product was dissolved in 5 rnL of Et 2 o and the solution cooled to -78°c giving the crystalline product, which was isolated by filtration and washed once with -78°c Et 2o (0.71 g, 60%}. IR data Ccm- 1 >: 2720(w), 1573(s), 1535Cw}, 1414(m}, 1252(w), 1216 (s), 1150 (w), 1020 Cw), 985 Cw), 757 (s), 720 (s), 405 (m). Anal. calc. for c 25 H34 NSc: C, 76.31; H, 8.71; N, 3.56. Found: c, 76.53; H, 8.69; N, 3.38. R2 Uptake by Cp*2~2~6H5 (eg 7). Cp*2ScCH2C6H5 C0.607 g, 1.50 mmol) was dissolved in ten mL of benzene in a 50 mL thick walled glass bomb. The bomb was then charged with 2.65 rnmol of H2 and the solution stirred for one hour. The benzene solution was cooled to l0°c and the amount of unused H2 measured with a Toepler pump. Initially, 1.26 rnrnol of H2 was collected, indicating that Cp* 2 ScCH 2 c 6 H5 had taken up 0.93 equivalents of H2 • The solution continued to give off hydrogen slowly, well after it should have been degassed. This is most likely due to a secondary reaction of 47 Cp* 2 ScH with c 6 H6 , which yields H2 (vide infra). This sec- ondary reaction may also account for the low (0.93 equivalents rather than 1.0 equivalents) level of H2 uptake observed. Determination of the Half-life of the Reaction of 2 ~ scca 3 with H2 • Cp* 2 ScCH 3 (60 mg, 0.018 mmol) was dis- solved in two mL of toluene in a 15 mL thick walled glass vessel and the solution cooled to -78°c. The solution was stirred rapidly as 2 atm of H2 was put over the solution. The stirring was continued for 7 minutes and all of the gas not condensed at -196 0 C CH 2 and CH 4 > was removed from the reaction vessel by Toepler pump. The H2 was removed from the gas by passage of the gas over a heated CuO catalyst and then through a -196°c trap, leaving only CH 4 in the gas phase. The amount of methane was 0.014 ± 0.0005 mmol. Thus 77% of the original Cp* 2 sccH 3 reacted to give CH 4 , indicating that the half-life of the reaction is 4-5 minutes. The NMR spectrum of the reaction solution after removal of H2 shows that Cp* 2 ScCH 3 accounts for 25% of the scandium species present (Cp* 2 ScH and Cp* 2 sc-c 6a 4 ca 3 account for 75%), consistent with the Toepler experiment. ~ scH 2 + CH 3.Ii. A thick walled glass bomb was charged with Cp* 2 sccH 2c 6 H5 (0.050 g, 0.12 mmol), three mL of toluene and four atmospheres of H2 • The solution was stirred for one hour and the excess H2 removed. CH 3 I (0.36 mmol) was 48 condensed into the bomb at -196°c. As the toluene solution was warmed to room temperature gas was evolved, which was transferred to a gas bulb with the Toepler pump. tral analysis showed the gas was methane. Mass spec- The toluene solu- tion was taken to dryness and the resulting yellow solid identified as Cp* 2 scr by NMR analysis. Cp* 2 Sc(THF)H (0.020 g, 0.0S2 3 mmol) was transferred into an NMR tube mounted on a ground 2 ~ Sc(THF)H + CH .I..... glass joint and fitted with a teflon needle valve adapter. a6-benzene (ca. 0.4 mL) was distilled into the tube, followed by O.OS2 mmol of CH 3 r and the tube was sealed with a torch. Immediately upon warming the tube to room tempera- ture a reaction occurred, yielding only Cp* 2 Sc(THF)I and CH 4 C1 H NMR). ~ 2 scCC.N-ry 2 -CsH 4 N> + n2 • An NMR tube was prepared as described above with 0.025 g Cp* 2 Sc(C,N-ry 2-CsH 4 N> (0.064 mmol) dissolved in 0.3S mL of d12 -cyclohexane. Prior to sealing, the entire tube was cooled to 77 K and 700 torr of n 2 admitted to the tube and it was sealed. After two days the resonance assigned to the ortho proton is no longer observed, while the other protons are observed at the same ppm and relative intensity as in the perprotio analogue. ~ 2 scCC.N-ry 2 -CsH 4 N> + ds-pyridine. A sealed NMR tube was prepared with 0.02S g Cp* 2 Sc(C,N-ry 2 -CsH 4N> C0.064 mmol) and 0.026 mL dS-pyridine (0.32 rnrnol, S equivalents) dis- solved in 0.35 mL c 6n6 • After two days at room temperature the pridyl ligand was replaced by a1 -pyridine in the 1 H NMR spectrum, while the Cp* resonance remained unchanged. 50 REFERENCES 1. L. Pauling; The Nature of the Chemical Bond, third edition, Cornell University Press,, Ithica, New York, (1960) chapter 13. 2. G. Wilkinson, J.M. Birmingham J. Arner. Chern. Soc (1954) ].§, 6210. 3. F.A. Hart, A.G. Masey, M.S. Saran J. Orgrnet. Chern. (1970) .21., 147-154. 4. R.S.P. Coutts and P.C. Wailes; J. Orgrnet. Chern. Cl970) 2..5., 117-122. 5. J. Holton, M.F. Lappert, D.G.H. Ballard, R. Pearce, J.L. Atwood and W.E. Hunter; J. Chern. Soc •• Dalton Trans. (1979) 54-61. 6. ibid., 45-54. 7. L.E. Manzer J. Orgrnet. Chern. (1976) ll..Q., 291. 8. D.M. Roddick, M.D. Fryzuk. G.L. Hillhouse, J.E. Bercaw Organornetallics (1985) ~' 97-104. P.T. Wolczanski, J.E. Bercaw Accts. Chern. Res. (1980) ..2A_, 121. J.E. Bercaw AdY.a.. Chern. Ser. Cl978) No. 167, 36. c. McDade, J.E. Bercaw ;L_ Orgrnet. Chern. Cl985) ~, 281-315. 9. ~, J. w. Lauer and R. Hoffmann; J. Amer. Chern. Soc (1976) 1729-1742. 10. L.E. Manzer; J. Orgrnet. Chern. (1976) .l.1.9., 291-294. 11. G.W. Parshall Accts. Chern. Res. Cl975) _a, 113-117. 12. P.M. Curtis and R.F. Porter; Chern. Phys. Lett. (1976) .3.1,153-155. M.P. Irion and K.L. Kornpa; J. Chern. Phys. (1982) 26.., 2338-2346. 13. R.K. Harris, B.E. Mann "NMR and the Periodic Table"; Academic Press: New York (1978). 14. J. w. Akitt; "NMR and Chemistry, An introduction to the fourier transforrn-rnultinuclear era."; Chapman and Hall: London and New York, (1983) chapter 4. 15. If the averaging is simple, the following equation should hold: 51 + 6observed = where the observed chemical shift in toluene-g 8 is 6.65 ppm, [Cp* 2 scHJ is 0.13(2) 1:1 and CH 2 J is 0.013(3) 11.· 16. J.P. Collman and L.S. Hegedus; Principles and Applications of Organotransition Metal Chemistry, University Science Books, Mill Valley, Calif., (1980) pages 133-139. 17. M. Brookhart and M.L.H. Green; J. Orgmet. Chem. (1983) .2..5..Q. ' 3 9 5 • 18. Ti(Me2PCH,CH2PMe2)Cl3CH3: z. Dawoodi, M.L.H. Grenn, V.S.B. Mtetwa and K. Prout; J. Chem. Soc., Chem. Commun. (1982) 1410-11. Ti(Me2PCH2CH2PMe2)Cl3CH2CH3: z. Dawoodi, M.L.H. Grenn, v.s.B. Mtetwa and K. Prout; J. Chem. Soc., Chem. Commun. Cl 982) 80 2-3. 19. M.L.H. Green, personal communication. 20. P.L. Watson; J. Amer. Chern. Soc (1983) 1.0..5., 6491-6493. 21. o. Eisenstein anad Y. Jean; J. Arner. Chern. Soc (1985) l.Ql., 1177-1186. 22. R.H. Marvich and H.H. Brintzinger; J. Au!er. Chem. Soc (1971) _u, 2046. 23. T.L. Brown, o.w. Dickerhoof, D.A. Bafus, G.L. Morgan; Rey. Sci. Instrum. (1962) .ll, 491. 24. LiCp* was prepared in a manner similar to that employed by previous workers, except low boiling petroleum ether was used in place of 1,2-dimethoxyethane and solid LiCp* was isolated by filtration of the reaction mixture. The solid was washed several times with petroleum ether, and dried .in yacuo. J.M. Manriquez, D.R. McAlister, E. Rosenberg, A.M. Shiller, K.L. Williamson, S.I. Chan, J.E. Bercaw J. Arner. Chem. Soc (1978) l..QQ, 3078-3083. 25. L.E. Manzer, Inorg. Syn. (1982) 21, 135. ScCH 20> 6c1 3 was prepared by dissolving sc 2o1 in 6 M HCl and removing the excess HCl by rotary evaporation. 26. F.w.s. Benfield and M.L.H. Green; J. Chem. Soc., Dalton Trans. (1974) 1324-1331. 27. E.P. Clark; Ind. Eng. Ed., Anal. Ed. (1941) l.,l, 820. 52 C-H Bona Activation and a-Bona Metathesis by Perrnethylscandocene Complexes. 53 Introduction The carbon-hydrogen bond is the most common unit in carbon based molecules, thus it is no surprise that activation of C-H bonds is a very active area of study for organometallic chemists. Several organometallic compounds have been reported that will activate C-H bonds homogeneously, and the mechanisms of these reactions are of great interest. Selective activation of C-H bonds could convert simple hydrocarbons into valuable feedstocks for industrial processes. These reactions are also of interest because the C-H bond is one of the strongest in organic compounds and has historically been the most inert, making selective functionalization of these bonds a great challenge. The mechanisms of C-H activation by group VIII metal complexes are fairly well understood.Cl] The active metal species is generally electron rich and coordinatively unsaturated, and the C-H bond reacts by oxidative addition to the metal center. This reaction is often followed by re- ductive elimination of the newly formed hydride ligand with a leaving group (either intra- or intermolecularly) .[21 This mechanism works well for some low valent transition metal complexes with readily accessible higher oxidation states;C31 however, many do (oxidatively inert) complexes also activate C-H bonds. [41 In every case the do complex 54 possesses a leaving group that is replaced by the carbon of the activated C-H bond (eq 1). M-R + C-H M-C + R-H (1) The mechanisms of these reactions of do complexes have not been investigated as intensively as their group VIII counterparts. The mechanistic studies of C-H bond activation by do metals have centered largely on reactions for which an intramolecular C-H bond is activated (eg. eq 2) .[51 (2) Labeling studies show that the alkyl leaving group reacts directly with the C-H bond being activated,[5, 61 arguing against both hemolytic bond cleavage and activation of a different C-H bond, followed by rearrangement to the observed products (eg. for reaction 2, the mechanism could have involved initial a-abstraction, rather than direct gamma-C-H activation). Activation parameters obtained from kinetic measurements indicate that the transition states for these intramolecular reactions are highly ordered.[4, 51 Based on these results, a four-centered transition state, (4-C), has been proposed. 55 -6 +6 .....• R· ...... M. _..:H 6+ ·c·;,C4-c> Unfortunately, experiments to directly determine the degree of charge localization in the transition state (the magnitude of 6} have not been reported. Moreover, very little is known about the mechanisms of intermolecular C-H bond activation by these complexes. The transition states are thought to be similar to (4-C} based on analogy. Using Cp* 2 ScR CR = hydride, alkyl, aryl} we have studied the inter- and intramolecular activation of C-H bonds by do metal complexes. These complexes react readily with arene and alkane C-H bonds as well as the CH 3 groups of their own Cp* ligands. Examination of the kinetics and product distributions of these reactions allows a fairly detailed picture of the transition state of C-H activation reactions to be drawn. These C-H bond activation reactions are part of a broad class of reactions we term "a-bond metathesis" Ceq 3}. Cp* 2 Sc-A + (3} R-B R = hydride, alkyl, aryl; A-B = H2 , C-H, X = halide c-x, x-c; 56 A general mechanism is proposed for this metathesis reaction. In all cases we have found, the rate of this reaction is strongly dependent on the type of bonding between A and B Cs, sp, sp 2 , or sp 3 > and the orientation of A-Bin the transition state. 57 RESULTS AND DISCUSSION HID Exchange Reactions Catalyzed by Permethylscandocenehydride Complexes. H/D exchange reactions involving C-H bonds are one of the most studied forms of C-H activation. Both hetero- and homogeneous metal complexes catalyze this reaction, with aromatic C-H bonds generally being the most reactive.[l(b), 71 The higher reactivity of aryl C-H bonds is usually attributed to prior coordination of the arene's cloud. [8] Permethylscandocene hydride complexes catalyze H/D exchange for a diverse group of C-H bonds. As with other cat- alyst, the H/D exchange reaction is rapid for arene C-H bonds. H/D exchange between Cp* 2 ScH and c D cD is slow 6 5 3 at low temperatures; no exchange is observed over a period of hours at -6o 0 c. However, when Cp* 2 scD is dissolved in C6H6 at 25°C and one equivalent of CH 3 I (based on scandium) is added to this solution after 5 minutes, the product methane is >95% CH 4 (500 MHz NMR), indicating that the exchange reaction was complete or near complete, (eq 4). 5 min. Repeating this reaction in C6D6 yields only CH 3 D. A upper limit for the half-life of the exchange reaction of one minute is thus established at 25°c. Cp* 2 scCTHF)H also cat- 58 alyzes this exchange, but at a much slower rate. Above 50°c, the THF hydride complex exchanges rapidly with arene solvent; however, for comparison to the base free hydride, the exchange was examined at 25°c. When a c 6n 6 solution of Cp * 2 sc(THF)H is let stand at room temperature for 1.5 hours and then treatea with CH 3 1, only CH 4 is observed by NMR ( <5% CH D), indicating that very little exchange has oc3 curred Ceq 5). 90 min. A Cp * 2 sc(THF)I + CH 4 (5) lower limit for the half-life of this reaction is 1000 minutes. The large difference in rate for Cp* 2 ScH versus Cp* 2 Sc(THF)H suggests that the THF ligana must dissociate prior to H/D exchange. A large difference in rate was also observed in the reactions of Cp* 2 ScH and Cp * 2 scCTHF)H with H2 Cvide supra). When toluene is the organic substrate usea for HID exchange, both positional preference Cortho, meta or para to CH 3 ) and the relative rates of aryl versus alkyl exchange can be measured. The positional preference for exchange was examined by reacting toluene-~ 8 with H2 CCp* 2 ScH catalyst) ( eq 6) • (6) 59 In the initial stages of this exchange reaction {2-3 turnovers), the meta and para positions show equal rates of exchange, while the ortho exchange is very slow. Heating the reaction to so 0 c for 48 hours leads to an unchanging ratio of protons ortho:meta:para, this thermodynamic ratio is 39:41:20, respectively (500 Mhz 1 H NMR). All other H/D exchange catalysts show less than statistical exchange ortho to the methyl group.[7, 9) H/D exchange into the methyl group also occurs but has a rate 60-70 times slower than that observed for aryl C-H exchange. Although continued heating does not change the ratio of protons in the different ring positions, it does increase the degree of H/D exchange into the toluene methyl group. The ratio of aryl versus methyl exchange becomes richer in methyl exchange than statistically predicted (5:3 for c6n5cn 3 >, and eventu- ally reaches thermodynamic equilibrium at 5.0:4.0 Caryl:methyl). Most of the H/D exchange experiments were monitored by 1 H NMR, and the progress of the exchange measured by the decrease in intensity of the resonances due to the exchanging organic substrate relative to a nonexchanging reference cyclohexane or cyclooctane). (~ For this effect to be marked a significant amount of deuterium must be exchanged into the substrate, which is difficult to do if the only deuterium source is o 2 • Only 5-10 equivalents (based on Cp* 2 ScH) can be conveniently put into an NMR tube. For this reason all 60 of the HID exchange reactions were done in c 6 n6 solvent; this gave each experiment nearly 400 equivalents of deuterium, in the solvent alone. The rates of H/D exchange into all of the substrates other than arenes are slow enough that substrate H/D exchange is rate limiting rather than regeneration of the catalyst. To prevent decomposition of the catalysts all of these experiments were done with four atmospheres of n2 over the reaction solutions. Under these con- ditions H/D exchange proceeds as shown in Scheme 1. C-H C-D L "--"' ~er~ X-H fast Cp* 2 ScH X-D x = D, aryl Scheme 1 Permethylscandocenehydride complexes catalyze intrarnolecular H/D exchange for the Cp* ligands and the a-posi tion of the THF ligands at 80°c. The fl-position of THF is not exchanged at this temperature over a period of many days. The methyl groups of Cp* 2 ScH exchange at a rate of roughly 2 turnovers/hour (80°c>, while the Cp* ligands of the THF adduct, Cp*2sc(THF)H, show a slower rate of exchange of 1 turnover/hour. Since free and coordinated THF are in rapid equilibrium, Cp*2 scCTHF)H acts as a catalyst for the intermolecular H/D exchange of the a-positions of THF, even 61 though this exchange is almost certainly intramolecular. In an experiment where t€n equivalents of THF is added to Cp* 2 ScH, the rate of H/D exchange into THF is 2 turnovers/hour, while that of the Cp* methyl groups is onl y 1 turnover/hour. Intermolecular H/D exchange is observed for several 1° CH 4 , Si(CH 3 > 4 , .cH.3CH 2.cH.3 and P(CH 3 ) 3 [101 readily undergo H/D exchange promoted by Cp* 2 scH. Cp* 2 Sc(THF)H C-H bonds. also catalyzes this exchange for these substrates, but in all cases the rate is much lower than that observed for Cp* 2 ScH under comparable conditions. The rates of H/D ex- change for CH 4 and Si(CH 3 > 4 are nearly the same, and the predominant exchange products at low conversion are CH 3o and Si(CH 3 )n(CH 2o> 4 _n' indicating that there is one exchange per interaction with the metal complex. In contrast, Ptc1 4 2[111 and zirconium alkoxides [121 show multiple exchanges per interaction with the metal complex for HID exchange. Intermolecular H/D exchange is also observed for several 2° C-H bonds, but the rate for this exchange is generally very slow compared to inter- or intramolecular exchange of 1° C-H bonds. Examples of slowly exchanging 2° C-H bonds are found in CH 3.cH.2cH 3 , the p-positions of THF, and cyclopentane. The rate of H/D exchange for cyclopentane was measured at 120°c and found to be 0.06 turnovers/hour (1.4 turnovers/day).[131 Two cases of readily exchangable 2° C-H 62 groups are the a-positions of THF and cyclopropane. When cyclopropane is treated with Cp* 2 sco in c 6o 6 , H/D exchange occurs smoothly at a rate comparable to that observed for CH 4 or SiCCH 3 > 4 • Ring opening has not been observed. A scale can be constructed that represents the relative ease with which Cp* 2 scH will catalyze H/D exchange for a particular substrate, and is shown below: [14] H2 >> arenes >> Cp*, a-THF, P{CH3>3 > Cp*(Cp * 2Sc{THF)H) > CH 4 , Si(CH 3 >4 , c 3H6 Ccyclopropane), c 6 H6'CHJ > .CH.3 CH 2.CH. 3 >> /J-THf, C5H10Ccyclopentane), CH3~2CH3. Several trends can be seen in this scale. First, the fastest rates of exchange are observed for those R-H bonds with the most s character. This can be seen in the series H2 Cs) >> aryl-H{sp 2 > >> C-H(sp 3 >. This trend will be discussed in detail in following sections. Second, intramolecular H/D exchange is more facile than intermolecular exchange among C-H bonds with similar s character. This is expected since the intramolecular exchange should have a far lower entropy of activation than intermolecular exchange. Third, the steric bulk of the substrate is very important. The wedge that the hydride sits in is very restrictive, and 2° carbons may not be able to fit into this wedge to react with the active site. Thus intermolecular H/D exchange at 2° carbons would be expected to be slower than at 1° ones. 63 The obvious exception is cyclopropane, which we feel is accelerated by the increased s character in its C-H bonds (151 (vide infra). Thermochemistry of Scandiumhydrjde and -phenyl Com- plexes. In addition to catalyzing H/D exchange between H2 and benzene, Cp* 2 ScH reacts reversibly with benzene to form Cp* 2 scc 6H5 and liberate H2 (eq 7) . (7) This equilibrium is evident if isolation of the hydride complex is attempted. When Cp* 2 scH is generated in c 6 H6 solu- tion and the solvent then removed slowly in vacuo, the product obtained is largely Cp* 2 scc 6 H5 • The equilibrium is driven to the right by removal of H2 with the arene solvent. If H2 is added back to the system, Cp* 2 ScH is reformed. The experiment to measure the equilibrium constant for this reaction was initiated with Cp* 2 ScH in benzene solvent. We were unable to approach this equilibrium from the right side of equation 7 (Cp* 2 scc 6H5 + H2 > under our experimental conditions, because of the difficulty in accurately adding a very small amount of H2 to an NMR sample. We are confident, however, that we are observing a true equilibrium, since a given sample may be cycled between 2s 0 c and ao 0 c and give consistent values for the equilibrium constant at each tern- 64 perature.The equilibrium constants and JG 0 values at several temperatures are given in Table 1. Table 1. Equilibrium constants and JG 0 values for the equilibrium between Cp* 2 scH and Cp* 2 scc 6 H5 in benzene solution Ceq 7) • Temperature (OC) 6 Keq Cx 10 ) 2.8 + 0.4 4,G°Ckcal/mol) 5.6 + 0.7 - 7.10 + 0.8 47 13 + 2 7.19 + 0.10 66.5 23 + 4 - 7.22 + 0.13 80 36 + 5 7.20 + 0.13 6 25 - 7.15 + 0.9 From the temperature dependence of the equilibrium constant, we were able to determine JH 0 and Js 0 for this reaction. Plotting ln(Keq) versus l/T gives a straight line with a slope equal to -JH 0 /R and an intercept of Js 0 /R (Figure 1), from which we calculated a JH 0 of 6.7 ± 0.3 kcal/mol. From JG 0 and JH 0 we calculated a Js 0 of -1.5 + 0.1 e.u •• For similar C-H bond activation reactions it has been postulated that the free energy of equilibrium is equal to the enthalpy of equilibrium and that the entropy of equilibrium is negligible.[161 Our results support this postulate, since JG 0 averages 7.2 kcal/mol for reaction 1, 65 -13-r-~~-,-~~---,~~~.-~~~~~-.-~~--.~~~-..-~~~ 0.0028 0.0030 0.0032 (Temperature) 0.0034 -1 Figure 1. van't Hoff plot for the equilibrium between Cp* 2 ScH and Cp* 2 scc 6H5 in c 6H6 solution. 66 with 6.7 kcal/mol of this being attributed to the enthalpy of the reaction. Using 110.9 kcal/mol for the C-H bond strength in benzene, £171 104.2 kcal/mo! for the H-H bond strength,£181 and 1.5 kcal/mole as the heat of solution of H2 in benzene CJHsoln = Hliq - Hgas> £191 we calculated from the enthalpy of the reaction that the scandium-hydride bond is 1.5 kcal/mole stronger than and the scandium-phenyl bond. Thermochemical measurements made by Marks, . ~ al. ,£201 on a related thorium system CCp* 2 ThR 2 ~ R = hydride, phenyl) showed that a thorium-hydride bond is ca. 7 kcal/mo! stronger than a thorium-phenyl bond. Cp* 2 Sc(THF)H is also in equilibrium with a scandiumphenyl complex Ceq 8) • The equilibrium constants and JG 0 values for this reaction are given in Table 2. Plotting lnCKeq> versus 1/T again yields a straight line (Figure 2) and allows the calculation of 18.9 ± 0.8 kcal/mol for the enthalpy of this reaction. A value of 24.5 ± 2.0 e.u. is determined for Js 0 from the JG 0 and JH 0 values. By comparison of reactions 7 and 8 it is expected that the reaction of Cp* 2 Sc(THF)H should be the more endothermic, since both a scandiumhydride bond and a scandium-oC 4H8 bond must be broken to form products.£211 It is also expected that the entropy for -15 -16 tJ1 Q) -17 ~ ~ rl -18 -19 0 -20-i-~~~.---~~~.-~~--,.~~~-.-~~~-,-~~---j 0.0028 0.0032 0.0030 (Temperature) -1 Figure 2. van't Hoff plot for the equilibrium between Cp* 2 Sc(THF)H and Cp* 2 ScC H in c H solution. 6 5 6 6 68 reaction 8 should be large and positive, since two molecules are reacting to yield three molecules • Table 2. Equilibrium constants and .1G 0 values for the equilibrium between Cp* Sc(THF)H and Cp* ScC H in 2 2 6 5 benzene solution (eq 8) • Temperature (oC) Keq(x J a8> .1G° Ckcal/mol) 25 0.25 + 0.03 11.7 + 0.1 35 0.9 + 0.1 11.3 + 0.1 47 2.9 + 0.4 11.0 + 0.1 56 7.2 + 0.9 10.8 + 0.1 66.5 15 + 2 10.6 + 0.1 80 41 + 6 10.3 + 0.1 - From these measurements, partial potential surf aces for reactions 7 and 8 can be constructed (Figure 3). This fig- ure schematically shows that the predicted energy of the Sc-THF bond is 12 kcal/mol, assuming that the Sc-H bond strength is not changed on forming the THF adduct. This is a gross assumption, and the predicted bond strength of the sc-THF bond should be treated as approximate. It must also be noted that the standard states for the equilibria of • Figure 3 . w Cp2 Sc (THF)H + CsHG 12.1 kcal • Cp 2ScC6 H5 + H2 Partial potential surfaces for equilibria 7 and 8. Cp2ScH + CsH5 s.1 kcal ____ 1l __ --r----- • '°°' Cp 2ScC6 H5 + THF + ~ 70 different. The latter equilibrium produces one mole of THF per mole of Cp* 2 scc 6 n5 , changing the properties of this solution relative to pure benzene. We expect this difference to be small. We are attempting to extend this series to include equilibrations with an sp 3 carbon, but have not yet found a substrate that can compete with intramolecular C-H bond activation. An approximate upper limit can be placed on the strength of a scandium-carbon(sp 3 > bond, however. When Cp* 2 sc-n-butyl is dissolved in a 1:1 mixture of c 6n6 and butane, Cp* 2 scc 6 n5 is formed immediately and is the only scandium product visible in the NMR spectrum after ten minutes. When Cp* 2 scc 6n5 is dissolved in neat butane, the scandiumbutyl complex is not formed, and prolonged heating leads n • As6 6 suming that there be as much as 3% scandium-butyl complex only to intramolecular C-H activation, releasing c and it would not be observed in the NMR, and assuming that ~s 0 for the reaction is small, leads to a scandium-butyl bond strength that is at least 23 kcal/mol weaker than a scandium-phenyl (or scandium-hydride) bond. Metal-carbon and metal-hydride bond strength have been measured for a large number of Co(III) and Mn(I) coordinatively saturated complexes. [22] The differences in energy between the metal- hydride and metal-alkyl bonds are generally 35-40 kcal/mol. The differences that Marks sees in his thorium system are 71 smaller; a thorium-phenyl bond is 18 kcal/mole stronger than a thorium-butyl bond.[201 Jntramolecular C-H Activation by Cp* 2 ~ 3 L Cp* 2 ScCH 3 reacts with benzene to give Cp* 2 scc 6H5 and CH 4 (eq 9). (9) If this reaction is carried out under pseudo first order conditions in scandium CC 6 H6 solvent) and monitored by 1 H NMR, no intermediates or products other than Cp* 2 scc H are 6 5 observed. The reaction is first order in scandium (different initial concentrations of Cp* 2 sccH 3 give identical rates). In order to determine the deuterium isotope ef- fect for this reaction, Cp* 2 sccH 3 was also treated with c 6n6 • Again the reaction appears first order in scandium, but in this case the first order plots (ln(c/c 0 > vs. time) are not linear. Similar results are obtained plotting either disappearance of starting material or appearance of product. Examination of the products of the c 6n 6 reaction (eq 10) shows that more than one mechanism may be involved in this reaction. Cp* 2 scc 6 n5 + Cp*Cp*-~ 1 scc n 6 5 + CH 3 D + CH 4 60% (10) 40% with c n 6 6 3 should produce only CH 3 D,[231 rather than the 60:40 mixture The direct bimolecular reaction of Cp* 2 scce 72 of CH 3D:CH 4 that is observed at ao 0 c. uct of the c 6n6 When the phenyl prodreaction is isolated and its 2 H NMR mea- sured, peaks are observed for the phenyl ligand and a resonance assignable to a Cp* ligand of Cp* scc n • 2 6 5 The proposed mechanisms for the reaction of Cp* sccH 2 3 with c 6n6 are shown in scheme 2. Cp~Sc-C6 D 5 Cp~Sc-CH3 A + CH3D ~ ~"· 46 + CsDs Sc * Cp*Cp*-d 1Sc-CsDs B ! Scheme 2 This scheme consists of two competing pathways. The first pathway, A, has a bimolecular rate determining step, which leads to direct intermolecular C-D activation, forming CH 3 n. The second pathway, ~' has a unimolecular rate determining step, involving intramolecular C-H activation to give a Cp* metallated intermediate, 1, and CH 4 • c 6 n6 to give Cp*Cp*-~ 1 scc 6 o 5 • i then reacts with Ring metallated complexes, similar to i, are common in group IV permethylmetallocene 73 chemistry.[24] As the temperature of the reaction of Cp* 2 sccH 3 with c 6o6 is increased the ratio of CH 3D to CH 4 becomes richer in CH 4 (Table 3), implying that activation of Cp* C-H bonds becomes more prominent at higher temperatures. Ratio of CH 3 o to CH 4 from the reaction of TABLE 3. Cp* 2 ScCH 3 with c 6o 6 (eq 10). TEMPERATURE % CH D % CH 60 76 24 80 60 40 98 53 47 125 50 50 3 4 In order to show that a tuck-in complex was a viable intermediate in this benzene reaction its synthesis and reactivity were investigated. When Cp* 2 sccH 3 is heated in cy- clohexane, CH 4 is released and a crystalline material, X, is produced. 80°c Cp* 2 ScCH 3 X + CH 4 (11) We have been unable to determine the molecular weight of x, but an NMR spectrum can be obtained (Figure 4) that is consistent with one normal Cp* cc 5 ccH 3 > 5 > and one metallated Cp* ligand I 5 (Sc(~ -c I 5 CcH 3 > 4cH 2 >, thus we concluded that X was 74 d b a b Sc I~ Cp d c Figure 4. 90 MHz 1 H NMR spectrum of X. TMS 75 an oligomeric tuck-in complex. Treatment of a benzene slurry of X with H2 causes the solid to go into solution and forms exclusively Cp* 2 scH (eq 12), which is common reactivity for ring metallated complexes.[25] (12) If X is pyrolyzed in c 6o 6 , d 1 -Cp*Cp*Scc 6o 5 is formed (identified by 1 H and 2a NMR, eq 13), which is consistent with pathway ~ of Scheme 2. (13) One piece of evidence argues strongly against X being a simple tuck-in complex. The number of moles of methane gener- ated by reaction 11 (measured by Toepler pump) is only 80% of the number of moles of Cp* 2 sccH 3 present at the start of the reaction. Thus, up to 20% of the scandium in X could still have a methyl ligand. A resonance directly assignable to this methyl ligand has not been observed. Using a higher field instrument (400 MHz) give a spectrum identical to that obtained with a 90 MHz spectrometer (Figure 5). The tuck-in reaction (eq 11) shows the same kinetic behavior that was seen in the reactions of Cp* 2 scca 3 with benzene. The reaction is apparently first order (different initial concentrations of Cp* 2 sccH 3 give the same initial rates) but the first order plots of the data are non-linear, see Figure s. The initial rate of this reaction is (6.1 + Figure 5. ..--I U) -s:: 0 I ::r: u -2.5 -2. 0 -1. 5 0 J I ~ 20 .l 60 80 -3 ) 100 T 120 T p 140 ~~ ~ 160 s 1 ope = 6.1 x 10 -6 sec -1 ~ Time (sec. x 10 40 ~~ I First order plot of the data from the reaction generating the •tuck-in" complex (eq 11) • ("") ......... lo-l 4-l Q) -1.0 -0.5 °' -'l 77 0.8) x 10- 6 sec-l (based on ..Q.a... 75% of one half-life). The reaction mixture turns f rorn yellow to red-brown after only short reaction times and remained this color throughout the course of the reaction. After approximately one half-life solid X begins precipitating and the rate of disappearance of Cp* 2 sccH 3 increases by ..Q.a... a factor of two. This rate continues to increase with longer reaction times. If, in a separate experiment, Cp* 2 sccH 3 is pyrolyzed in c 6n12 with solid X added at the beginning of the reaction, the first order plot is close to linear for over two half-lives, Figure 6, and the initial rate of this reaction is equal to the one observed shortly after X began precipitating in the previous experiment (without added x>. This result strongly suggests that the decomposition of Cp* 2 sccH 3 to give x is autocatalytic. (Cp*-d 15 > 2 sccH 3 was prepared and the kinetics of its thermal decomposition followed. This compound showed the same nonlinear kinetic behavior as the perprotio analogue, 1 6 with an initial rate constant of (3.8 ± 0.4) x 10- sec- • A kH/kD of 1.6 is thus determined for the reaction of Cp* 2ScCH 3 to give X The decomposition of Cp* 2 sccn 3 to give X is a mechanistic nightmare. The products of this reaction apparently catalyze their own formation. And in up to 20% of the reac- tion this catalysis may be incomplete, leaving some of the product with intact scandium-methyl bonds. To further corn- plicate the situation, the rate of the reaction increases Figure 6. 0 ~ 20 60 Time (sec. x 10 40 ,-------,----T--1 ) 80 I 100 120 t . = 15.2 x 10 -6 sec -1 -3 "i"~~ slope I First order plot of the data from the reaction generating the "tuck-in" complex, with added x. -3 j r-1 U) u I -s:: ~ I -2 ::r: u M ......... ~ (!) ~ -1 0 co -'1 79 rapidly after X begins to precipitate, so heterogeneous reactions may also be occurring. The disappearance of Cp* 2 sccH 3 in this reaction is a function of several processes. Thus, the observed deuterium isotope effect is a composite of several deuterium isotope effects, and probably does not reflect directly on an intramolecular C-H bond breaking process. Rather than investigate the mechanism of the tuck-in reaction deeply, we chose to just examine its characteristics, so that Cp* C-H activation can be excluded as a mechanistic pathway in other reactions. The most obvi- ous characteristic of this reaction is the label exchange. When Cp* 2 sccH 3 is reacted with c 6o6 , significant amounts of CH 4 are formed, indicating that activation of Cp* C-H bonds has occurred. The converse labeling experiment can also be carried out. If CCp*-£ 15 > 2 sccH 3 is treated with a substrate bearing only C-H bonds (as opposed to C-D bonds), and only CH 4 is observed, then intramolecular C-D activation is not occurring. is (6.1 The initial rate of the tuck-in reaction Ceq 11) ± 1.0) x 10- 6 sec- 1 , so if a particular reaction has a rate significantly faster than this, then the reaction is most likely not going through activation of a Cp* C-H bond. Lastly, if the activation of Cp* C-H bonds is a significant mechanistic pathway for a reaction involving Cp* 2 sccH 3 , the first order kinetic plots will be decidedly curved, with the rate increasing for longer reaction times. 80 The unimolecular mechanism proposed for the reaction of Cp* 2 sccH 3 with c 6n6 is obviously over-simplified. The pro- posal that only 1 reacts with c 6 n6 to give Cp*Cp*-~ 1 scc 6 o 5 may be wrong, since several similar species may be present. The proposal that CH 4 is eliminated to give a tuck-in complex is still valid, but the structure of this complex is unknown. To determine the isotope effect for an intermolecular C-H activation reaction by Cp* 2 sccH 3 the direct bimolecular reaction of Cp* 2 sccH 3 with benzene CA of Scheme 2) was examined. In order to minimize the amount of the reaction that would go by the tuck-in route, (Cp*-£ 15 > 2 sccH 3 was used. Linear first order kinetics are ob- served for the reactions of d 30 -cp* 2 sccH 3 with both c 6 H6 and c 6n6 solvents, giving initial pseudo first order rates of Cl.4 ± 0.05) x 10- 4 and (5.1 ± 0.1) x 10- 5 sec- 1 , respectively. The rate of the tuck-in reaction was 3% and 8% of the observed rate constants for the C6 H6 and c 6n6 reactions, respectively. The rate of the tuck-in reaction was sub- tracted from each of the two rate constants to give the two corrected rate constants, whose ratio is 2.9 ± 0.4 (k C6H6 /k C6D6 )• c-H Activation by Scandiumalkyl and -aryl Complexes. Methane is the simplest of all hydrocarbons, but has been the most difficult to activate. The first activation of methane by an inorganic complex was demonstrated by Shilov 16 years ago.[26] He found H/D exchange between 81 n 2o/cH 3 cH 2 oon and CH 4 is catalyzed by K2Ptc1 4 • Only re- cently has activation of methane, which led to an isolated metal-methyl complex, been reported. A few group VIII corn- plexes have been found to react with methane to yield rnethylhydride cornplexes,[l{f), 271 while examples of lanthanide and actinide complexes, which give methyl complexes, have been communicated. [281 Cp* 2 sccH 3 activates methane as shown in equation 14. 70°c + + {14) This reaction was done with a large excess of the labeled methane C£ 13 cH 4 1 = 0.5-1.5 M; [Cp* 2 ScCH 3 1 = 0.06 M). A second order rate constant of 1 x 10- 5 M- 1 sec-l was obtained. Watson has measured the rate of this reaction for cornplexes isoelectronic with Cp* 2 sccH 3 , Cp* 2McH 3 {M = Y, Lu).[28{a)1 She found the yttrium complex reacted five times faster than the lutetium one, while the scandium cornpound reacts fifty times slower than the lutetium compound Cky = 2.6 x 10- 3 M- 1 sec- 1 ; kLu = 4.6 x 10- 4 M- 1 sec- 1 >. She proposes a two part mechanism for this reaction, shown in Scheme 3.[291 One pathway of this methane activation reac- tion involves the direct bimolecular reaction of Cp* 2McH 3 with 13 cH 4 , while the second involves a unirnolecular reaction of Cp* 2McH 3 to give a tuck-in complex, followed by its reaction with 13 cH 4 to give Cp* 2M13 cH 3 • 82 Scheme 3 For the lutetium complex the uni- and bimolecular rate constants were reported Ck 1 = 2.3 x 10-s sec- 1 , k = 4.7 x 10- 3 2 1 1 M- sec- >. Yttrium and lutetium have the same ionic radius (Y(III) 0 = Lu(III) = 0.93 A), so steric interactions should be the same for both metals and any differences in their reactivity attributable to electronic differences in the metals. Watson ascribes the greater reactivity of yttrium over lutetium to yttrium's greater electrophilicity. can not be directly compared to Y or Lu. Scandium It is very elec0 trophilic, but its ionic radius (Sc(!!!) = 0.81 A) is much smaller than that of Y or Lu, which acts to amplify any unfavorable steric interactions. It is interesting that both the uni- and bimolecular pathways for the reaction of Cp* 2 LuCH 3 with methane are faster than the total reaction of Cp* 2 ScCH 3 with methane, suggesting that steric interactions BJ decrease the rate of both the uni- and bimolecular reactions. Cp* 2 ScCH 3 reacts with styrene and substituted styrenes by activation of one of the terminal vinylic C-H bonds, forming a scandium styrenyl complex Ceq 15). Cp* 2 ScCH=CHC 6 H 4 X-~ + CH 4 (15) The stereochemistry of the product is assigned trans based on the large coupling constant observed for the two vinylic 3 protons Ci.e. JH-H = 20 Hz for Cp* 2 ScCH=CHC 6 H40cH 3 ).[30l The kinetics of reaction 19 with X = OCH 3 and cF 3 were examined; both reactions are first order in scandium and styrene, and no intermediates are observed during the course of the reaction. The reactions were run with 2-4 times more styrene than scandium, and the data plotted using a standard second order equation [311 Ceg. Figure 7), such that the ob- served rate constants are true second order rate constants. The rates of these reactions were measured at several temperatures, and activation parameters calculated for each reaction (using Arrhenius and activated complex theory). The Arrhenius plots for these reactions are shown in Figure 8, and the rate constants and activation parameters given in Table 4. 84 - 6000 12000 18000 24003 Time (sec.) Figure 7. Representative second order plot for the reaction of Cp* sccH with H C=CHC H cF -E. where 2 2 3 6 4 3 A = [Cp* 2 ScCH 3 ] and B = [H C=CHC tt cF ].[31] 2 6 4 3 85 -7 -8 -9 0.0029 0.0031 0.0033 -6 X = CF 3 -7 -8 -9 Figure 8. 0.0029 0.0031 0.0033 Arrhenius plots (ln(k) vs. 1/T) for the reactions of Cp* ScCH with H C=CHC H X-E (X = OCH 3 , CF 3 ). 2 6 4 3 2 86 Table 4. Rate constants and activation parameters for the reaction of Cp* 2 sccH 3 with substituted styrenes CH 2 C=CHC 6H4X-para: X = OCH 3 , CF ) at 60°C Ceq 19). 3 x OCH 3 CF 3 1 1 4 k2 CM- sec- , x 10 > 12.5 + 0.7 8.5 ±0.5 Ea (kcal/mole) 12.l + 0.3 13.2 ±0.3 log A 4.9 ± 0.2 5.7 ± 0.2 .6Hf Ckcal/mol) 11.5 ±0.3 12.6 ± 0.3 "1 s* Ce.u.) -38 + 3 -34 + 3 These reactions are typified by small enthalpies of activation and large negative entropies of activation, which is very characteristic of associative reactions with highly ordered transition states.[32] Large entropies of activation C-9 to -19 e.u.) are also observed for intramolecular C-H activation by do complexes.[4, 5] Only small substituent effects on the rate of the reaction were observed CCF 3 and OCH 3 substituted styrenes gave very similar activation parameters). Rather than forming styrenyl complexes, Cp* 2 ScH reacts with styrene by insertion of the olefinic group into the Sc-H bond to give a phenethyl complex Ceq 16). Cp* 2 ScH + (16) 87 Cp* 2 sccH 2 CH(CH 3 > 2 reacts with propene to give only Cp* 2 ScCH=CHCH 3 -~ (eq 17). The ry 3-allyl complex, Cp* 2 sccry 3-cH 2 CH=CH >, is stable (vide 2 supra), but is never formed in this reaction. The details of this reaction will be discussed in the following chapter. Cp* 2 sccH 3 reacts very rapidly with the alkynyl C-H bond of propyne to liberate methane and form a scandium propynyl complex (eq 18). (18) Although the kinetics of this reaction were not investigated (due to complicating subsequent reactions,~ inft£), the reaction is complete in minutes at temperatures below o0 c. It is interesting to note that the product of alkyne insertion into the Sc-CH 3 bond (Cp* 2 scCH=C(CH 3 > 2 > is not observed, even though the isobutenyl compound is stable and internal alkynes readily insert into scandium-methyl bonds (vide infra). Cp* 2 ScH also reacts with one equivalent of propyne to give only Cp* 2 scc~CCH 3 (eq 19). (19) Again, the inserted product CCp* 2 ScCH=CHCH 3 > is stable, but never observed. For analogous zirconium and hafnium hydride 88 complexes, rapid insertion of propyne into the M-H bonds is observed, leading to M-CH=CHCH 3 -~ complexes.[331 Cp* 2 ScR CR = methyl, aryl) react cleanly with arenes to give a scandium aryl complex and R-H (eq 20). (20) R = methyl, aryl; X = H, alkyl, amide For reactions of a given arene, Cp* 2 sccH 3 always reacts more rapidly than Cp* 2 scc 6H5 • This is presumably due to increased steric crowding of the reactive site by the phenyl ligand compared to CH • 3 The reactions of Cp* 2 sccH 3 with arenes were studied extensively in order to probe the importance of coordination of the cloud in these C-H bond activating reactions. In the first experiments we examined the rates of activation of C-H bonds in monosubstituted arenes as a function of the substituent (eq 21). (21) Since we were interested in the direct interaction of Sc-CH 3 with the aryl C-H bond, we used (Cp*-£ 15 >2 sccH to minimize 3 the rate of the tuck-in pathway(~ of Scheme 2). The tuckin pathway would not be effected by substituents on the arene, because its rate determining step is activation of a 89 C-H bond of its own Cp* rings. In all cases the first order plots of these reaction are linear <..e.9.L Figure 9) and the amount of CH 3 D (formed by the tuck-in pathway) is 5% or less of the total methanes produced . Under the reaction condi- tions used, substrates with non-aryl protons CC H5 NCCH > 6 3 2 and c 6H5 cn 3 > or fluorides CC 6H5cF 3 > yield only aryl scandium products. Cp* 2 sc-c 6 H4 cF 3 does react with additional c 6H5 cF 3 after prolonged heating, but the products of this reaction are not observed until well after reaction 25 is complete. The rate constants for these reactions at ao 0 c are given in Table 5. Table S. a arene 1 1 5 k2 CM- sec- • x 10 > C6H5 NCCH 3 > 2 C6 H5 CH 3 3.2 + 0.1 C6H6 3.3 + 0.1 c 6H5 CF 3 1.4 + 0.1 3.4 + 0.1 reactions were carried out in c D solutions at [(Cp*-£ 15 > 2 sccH 3 = 0.12 M, [areneJ 6 =1 f.3 M. 0 These 80 C1 Rate constants for the reaction of CCp*-d 15 > 2 sccH 3 with c 6H6 and substituted arenes Ceq 24) at ao 0 c.a 0 5 20 -3 15 Time (sec. x 10 10 ) - 25 - 30 c 6 H5 N(CH 3 ) 2 . Representative first order plot of the data f :om the reaction of 3.0 3.5 Cp* ScCH with 2 3 Figure 9. r-i i:: (/) () I ::r: u M ,........ 4.0 '°0 91 The rate of the reaction is reduced by a factor of two when the substituent on the arene is changed from strongly electron donating (N(CH 3 >2 > to electron withdrawing CCF 3 >. The positional preference for activation of the C-H bonds of toluene by Cp* 2 sccH 3 was examined. The different tolyl isomers Cortho, meta, para) and the benzyl complex are all easily differentiated by 500 MHz NMR. The kinetic and thermodynamic product ratios for the reaction of Cp* 2 scR CR = CH 3 , CH 2 C H and c H cH > with toluene solvent are given 6 5 6 4 3 in Table 6. The first three entries of Table 6 represent the kinetic ratios for the reactions of Cp* 2 sccH 3 and Cp* 2 sccH 2 c 6 H5 with toluene, since authentic samples of Cp* 2 sc-C 6H4 CH 3 -Q and Cp* 2 sc-c 6 H 4 cH 3 -~ do not isomerize on the time scale of these lower temperature experiments cso 0 c>. The kinetic and thermodynamic ratios of tolyl isomers are the same, since heating the ortho tolyl derivative eventually gives an unchanging ratio (entry 7 of Table 6) that is the same as the kinetic ratio. Cp* 2 sccH 2 c 6H5 is one of the products formed in the reaction of Cp* 2 sccH 3 with toluene. The percentages of benzylic product in the reactions of Cp* 2 sccH 3 with toluene are def ini tely low, since Cp* 2 sccH 2 c 6H5 reacts with solvent to give tolyl products at so 0 c (entry 3 of Table 6). Both the kinetic and thermodynamic ratios have a predominance of meta and para tolyl isomers. The amount of meta isomer is always twice that of the para isomer, which coincides with the Cp* 2 ScCH 3 , Cp* 2 ScCH 2 c 6 H5 and 3 15 Cp* 2 ScC H cn -~, 115°c, 36 hrs 60 61 44(72) 52(62) 57(66) m.e.t.a 25 26 >95 14(23) 23(27) 23(27) mu.a 0 0 0 0 39 16 14 benzvl a All of these reactions were carried ogt in sealed tubes in toluene solvent, the given scandium complex was ca. 0.10 M. The relative percentages were determined by cutting and weighing the NMR spectra of the tolyl-methyl and benzylic resonances. The numbers are accurate to ± 2%. Numbers in parentheses represent the selative % of each tolyl isomerdwhen scandium benzyl is formed. Reaction £a. 40% complete. Reaction .Q.a. 75 % complete. 6 4 3 13 so 0 c, 7.5 hrs >95 3 ( 5) Cp* ScCH , 115°C, 12 hrs 2 3 Cp*2ScC6H4CH3-nr 2 Cp* ScC 6 H 4 cH -~, 80°C, 7.5 hrs so 0 c, 7.5 hrs 9 Cll) so 0 c, 7.5 hrsd Cp* sccH , 3 2 Cp* 2 ScCH 2 c 6 H5 , 6(7) so 0 c, 2 hrsc or tho product ratios <relative %)b 2 Sc-C 6 H 4 CH 3 -~, -n with toluene. Cp* sccH , 2 3 Cp* Isomer ratios for the product mixtures from the reactions of Reactant and conditionsa Table 6. l\) '° 93 statistical ratio of these positions, indicating that the rates of reaction at the two positions are equal ((2kpara)/kmeta = 1). The ortho tolyl isomer is presumably less stable for steric reasons.[34) scandium, being very Lewis acidic, is a fairly strong electrophile. For this reason we thought that the reactions of Cp* 2 ScR complexes with arenes may bear a resemblance to electrophilic aromatic substitution reactions, which have been studied extensively by physical organic chemists.[35) These reactions involve attack of an organic or inorganic electrophile on the arene " cloud, followed by a-" rearrangement to give a Whelland type intermediate, which releases H+ to give the substituted arene (Scheme 4).[36) Scheme 4 Electrophilic aromatic substitution has also been studied with organometallic electrophiles. These electrophiles are generally cationic,[37) but anionic ones have also been claimed.[38) The mechanisms of substitution reactions of organometallic electrophiles on arenes are believed to be very similar to those proposed for organic and inorganic ones. The electrophile "-complex (") has actually been ob- served in the reaction of Cp*Re(NO) (CO)+ with arenes.[39) 94 Characteristic reactivity patterns are exhibited by most electrophiles in these substitution reactions, whether the electrophile is organic, inorganic, or organometallic. In nearly every case the reactivity of Cp* 2 ScR with arenes is very different from any of these electrophiles. For example, in electrophilic aromatic substitution reactions, electron donating groups attached to the arene accelerate the rate of substitution, by stabilizing the positive charge on the arene; conversely, electron withdrawing groups act to slow this rate. For organic and inorganic electrophiles the rate difference between c 6H5 NCCH 3 > 2 and c 6H5 CF 3 is generally 10 8-10 10 ,[40] while for organometallic electrophiles this difference is 10 3-10 6 .[38(b), 41] Scandium's rate difference of 2 is very small. Moreover, benzene and toluene react more rapidly than c 6H5 NCCH 3 > 2 , suggesting that electron donation or withdrawal by substituents does not affect the energy of the transition state. Although large substituent effects on the rates of substitution are characteristic of electrophilic attack, they are not diagnostic. Very active electrophiles show small substituent effects, often reacting with toluene and benzene at nearly the same rate.[42] A more diagnostic characteris- tic of electrophilic attack is the positional selectivity for substitution on toluene. All electrophiles (even the 95 very active ones) show a high preference for substitution at positions ortho and para to the methyl group.[37(a)] The meta isomer accounts for <5% of the substitution products. This effect is a result of the methyl group adding stability to the intermediates or transition states with positive charge localized at positions para and ortho to it. For example, Cp*Re(NO) (CO)+ and H2 Ptc1 (Ptc1 is believed to 6 5 be the active electrophile) react with toluene to give tolyl complexes; in both cases the major product results from para substitution ((2kpara)/kmeta = 32 and 18, respectively), with the ortho isomer not being observed. the benzylic carbon is never observed. Substitution at All of the scandium complexes investigated (Cp* 2 ScR; R = methyl, benzyl and tolyl) give tolyl products on treatment with toluene. These tolyl products have statistical distribution between meta and para substitution CC2kpara)/kmeta = 1), and low, but observable, substitution ortho to the methyl group. Moreover, Cp* 2 sccH 3 gives large amounts C>l5%) of the benzylic substitution product (Cp* 2 sccH 2 c 6 H5 > on treatment with toluene. These results, and the observation that the rate of substitution is not strongly effected by electron donating or withdrawing groups, suggest that interaction of sc with the system of the arene is not important in the C-H activation of aryl C-H bonds by Cp* 2 sccH 3 • Deuterium isotope effects are generally not observed in the reactions of organic and inorganic electrophiles with 96 arenes,[43) but an isotope effect of 3.0 was measured for the reaction of H2 Ptc1 6 with benzene. [39) The authors do not address the nature of this effect, but it may be a result of the rate determining step not being formation of the 7r-complex, but rearrangement to the a-complex and its subsequent deprotonation. Cp* 2 sccH 3 has a kH/kD for the reaction with benzene of 2.9 CY.i.d.e supra). From the reactions of Cp* 2 ScR CR = hydride, alkyl, aryl) complexes with hydrocarbons a reasonable picture can be drawn of the transition state for this metathesis reaction. A kH/kD of 2.9 for the reaction of Cp* 2 sccH 3 with benzene indicates that there must be some C-H bond weakening in the transition state. The lack of a significant sub- stituent effect in the reactions of Cp* 2 sccH 3 with either arenes or styrenes suggests that the transition state for this reaction is no more polar than the ground state. Ho- molytic bond cleavage, as in radical chain type processes, is also excluded, since ring opening of cyclopropane is not observed after long reaction times, and because activation of solvent is never observed. When Cp* 2 ScCH 3 is pyrolyzed in c 6n12 , only CH 4 is observed (no CH 3 n>. If Sc-CH 3 bonds were being homolytically cleaved under the reaction conditions, some activation of solvent, giving CH 3 n, would be expected. The large negative entropy of activation for the reactions of Cp* 2 sccH 3 with styrenes suggests that the transition state is highly ordered, and that the rate 97 determining step is an associative one. The transition state that best fits our observations is a nonpolar 4centered one, shown below. R = hydride, alkyl, aryl; A-B = H-H, C-H The formation of significant amounts of Cp* 2 sccH 2 c H5 6 in the reaction of Cp* 2 sccH 3 with toluene suggests that coordination of the rr cloud of the arene is not along the reaction coordinate for these reactions. The lack of substituent effects in the reactions of styrenes and arenes with Cp* 2 sccH 3 are also strong evidence that initial attack of Sc on the rr systems of these substrates is not important in C-H activation. This result is particularly significant, since coordination of the rr system prior to C-H activation is thought to occur in virtually every reaction of organometallic complexes with sp 2 C-H bonds. [441 Reactions of Permethylscandocene Alkyl Complexes with Alkylhalides. In a reaction reminiscent of the reaction of Cp* 2 ScCH 3 with 13 cH 4 Ceq 14), Cp* 2 ScCH 3 reacts with 13 cH 3 I to give a labeled scandium-methyl complex (Scheme 5). 98 Cp* 2 ScCH 3 + ~ Cp* 2 Sc 13 CH 3 Cp* 2 ScI + + CH 3 I / ethane Scheme 5 An additional reaction is observed that leads to coupling of two methyl groups, forming Cp* 2ScI and ethane, which is irreversible. These two reactions proceed at comparable rates, both having half-lives of ~ 30 hours at room ternperature, when 7 equivalents of methyl iodide are used. The reaction leading to Cp* 2 ScI and ethane is definitely the thermodynamically pref erred one. The methyl exchange reac- tion is thermoneutral, since the only differences between the reactants and the products are isotopic changes. The methyl coupling reaction, however, involves cleavage of Cp* 2 sc-CH 3 and r- 13 cH 3 bonds, and formation of Cp* 2 sc-I and CH 3- 13 cH 3 bonds. In general, early transition metal-halide bonds are 20-40 kcal/mole stronger than metal-methyl bonds,[451 and the carbon-carbon bond of ethane is 33 kcal/mole stronger than the carbon-iodine bond of methyl iodide.[161 From these bond strengths an enthalpy of 50-70 kcal/mole is expected for this reaction: with an enthalpy this large it is not surprising that the methyl coupling reaction is irreversible. An alkyl exchange reaction is also observed in the re- 99 (22) and CH r, to 3 3 generate propane, is not observed. Rather, the Cp* sccH 2 3 produced reacts with CH 3 r to give Cp* 2 ScI and ethane. For An alkyl coupling reaction between Cp* 2 sccH 2 cH Cp* 2 sccH 2 cH 3 , alkyl exchange with CH 3 r is more facile than alkyl coupling. Cp* 2 sccH 3 reacts with benzylchloride and benzylbromide to give only the corresponding Cp* 2 scx complex (X = Cl, Br) and ethylbenzene (eq 23). Benzylbromide reacts faster than benzylchloride. + + 25°C X = Cl, Br a-bond Metathesis. In the preceding sections we have shown that a-bond metathesis reactions are common in the chemistry of Cp* 2 ScR complexes CR= hydride, alkyl, aryl). The mechanisms of these reactions are important in understanding the reaction chemistry of scandium complexes. And in a broad sense, these mechanisms will help in understanding the reaction chemistry of all oxidatively inert, highly electrophilic, early transition metal, lanthanide and actinide complexes. From our studies we can draw a fairly detailed picture of the reaction paths for abond metathesis, and discuss factors effecting the rates and product distributions of these reactions. 100 In several cases U-bond metathesis reactions are readily reversible, such as the reaction of Cp* 2 ScH with benzene (eq 7) and H/D exchange catalyzed by Cp* ScH. From the 2 studies of C-H bond activation by Cp* 2 sccH 3 , a great deal has been learned about the mechanism of u-bond metathesis. Kinetic measurements suggest that U-bond metathesis reactions of C-H bonds are associative with highly ordered transition states. Moreover, the observation of a deuterium isotope effect on the rate of the reaction between Cp* 2 ScCH 3 and benzene implies that C-H bond weakening is involved in the rate determining step. Any mechanisms for this reaction that include oxidative addition or reductive elimination steps can be excluded, because scandium has no readily available oxidation states other than III. The pro- posed mechanism, consistent with all the data observations, is shown in Figure 10. The initial interaction is between the bond being activated (A-B) and the LUMO of the electrophilic scandium complex, as in ~. This donation from the A-B bond to the empty metal orbital is similar to an agostic interaction (A-B = C-H). This intermediate does not lie in a well on the potential surface, since it does not build up. We have been unable to observe an agostic ground state for any permethylscandocene-based complexes Cvide supra) , indicating that this interaction should be weak. ~ leads to the 4-centered transition state D, which cleaves to give products (through~·). I I I ,A, \ \ \ t @ I I -c' R-B ® t 1 orbital. [47] = H-H, C-H, C-X, X-C; X = halide. Sc-A+ R-B R = H, alkyl, aryl; ~ The orbital pictured in C and Proposed mechanism for a-bond metathesis. C' is the la A-B Figure 10. -D -c \ 'R1 \ ® Sc-R + A-B--+ cSc3-R ~1sc---B1~ cSc3-A \ I @ A-B ~ I-' 0 I-' 102 The high electrophilicity of the metal center is very important in these a-bond metathesis reactions. The reac- tant A-B bond is activated by interaction with a vacant metal orbital (as in~), and it is the high electrophilicity of the early transition metal that brings about interactions with such weak donor bonds as H-H and C-H. a-bond metathe- sis is formally a [2s + 2sl reaction which is symmetry disallowed. [ 461 By inclusion of the vacant metal orbital, the transition state becomes that of a [2s + 2s + dl reaction, which is allowed. the b 2 orbital.[471 The d orbital may be either the la 1 or The mechanism of Figure 1 is drawn with the la 1 orbital as the interacting vacant metal orbital, because the la 1 is the LUMO of Cp* 2 ScR. It should be stressed that although high electrophilicity of the metal center drives the reaction, it does not greatly polarize or heterolytically cleave the bond being activated. Rather, activation of the A-B bond leads to a 4-centered rather nonpolar transition state. Both thermodynamic and kinetic effects govern the outcome of the a-bond metathesis reaction. The energy of the transition state is strongly dependent on the the ability of the members of this 4-centered interaction to participate in multicenter bridging interactions. Scandium's d orbitals are suitably disposed to allow good overlap with the other groups in B, but the non-metal atoms CR, A and B) must use s and p orbitals for these bridging interactions. The s or- 103 bital, being spherically symmetric, is the best suited for this type of interaction, since its nondirectional nature allows significant overlap with two orbital in very different directions C~.g. the bridging hydrides of B2H6 >. It is, therefore, not surprising that one of the fastest a-bond metathesis reactions is between Cp* 2 ScH and H2 Ckobs > 10 3 M- 1 sec-l at -9s 0 c>. The transition state for this reaction should have the best orbital overlap (hence most bonding and lowest energy) possible for this type of metathesis reaction, since all three of the non-metal atoms utilize spherically symmetric (ls) orbitals in bonding. The reactions of Cp* 2 scR complexes CR = alkyl, aryl) with H2 are also very facile (R = CH 3 , half-life< 5 min. at -7a 0 c>, which may be a result of the high degree of orbital overlap of the two H atom ls orbitals in the transition state. cases where s and p character are mixed degree of s character is important. (~.g. In C-H) , the In the same way that s orbitals bridge better than p orbitals, so a carbon orbital with high s character have better overlap in the bridging position than one with low s character, because the carbon orbital with more s character will be less directional in nature. The result of this is that a-bond metathesis reactions between Cp* 2 sc-R and C-H bonds will be the fastest for C-H with the highest s character Csp > sp 2 > sp 3 >. This trend can be seen in the rates of C-H activation by 104 H-CH 2 c 5 <cH 3 >4 (sp 3 >> and in the rates of H/D exchange catalyzed by Cp* 2 ScH (C-H (sp 2 > > C-H (sp 3 >>. This preference for C-H bonds with higher degrees of s character is also seen in the reaction of Cp* 2 sccH 2 CH(CH 3 > 2 with propene (eq 17). The sp 2 vinylic C-H is chosen exclusively over the sp 3 allylic one, even though the allylic C-H is weaker and activation of it would lead to the stable ~ 3 -allylic complex, Cp* 2 Sc(~ 3 -cH 2 CH=CH 2 >. It is interesting to note that when Cp* 2 ScR (R = hydride, methyl) can react with with orbitals of all p characof HC~CCH 3 > or an sp C-H bond (H-c:ccH >, only 3 the a-bond metathesis reaction is observed, even though ter (~bonds alkyne insertion reactions are facile in this system (~ infra). In the reactions of Cp* 2 ScR complexes (R = alkyl, aryl) with C-H bonds, products resulting from C-C bond coupling are never observed. namic or both. The reason may be kinetic or therrnody- For the reactions of Cp* 2 scR CR = alkyl) with benzene the thermodynamics can be calculated. Since the relative bond strengths of scandium-hydride versus scandium-phenyl bonds are known, a direct comparison of the enthalpies of alkyl-aryl exchange reactions (Scheme 6(a)) and alkyl-aryl coupling reactions (Scheme 6(b)) to be made. 105 + + H-CH 3 Ca) (105) + (b) Scheme 6: alkyl-aryl exchange (a) and alkyl-aryl coupling (b) reactions. The numbers in parentheses represent the bond strength (kcal/mole) of the bond above them. Since both reactions cleave the same bonds in the reactants, the difference in enthalpies of the two reactions is a result of the different energies of the products. From the bond strengths given in Scheme 6, the alkyl-aryl exchange reaction is 2.5 kcal/mole more exothermic than the alkylaryl coupling reaction, ~ the thermodynamically pref erred reaction is alkyl-aryl exchange (assuming that the two reactions have comparable entropies). If the methyl ligand of Cp* 2 sccH 3 is replaced with an n-butyl one, the relative enthalpies of alkyl-aryl exchange (Scheme 7(a)) and alkyl-aryl coupling (Scheme 7(b)) can be calculated for the reaction with benzene. Cp* 2 Sc-C 6 H5 + Cx+l.5) Ca> ( 98) (x) Cp* 2 Sc-H H-C 4H9 + C4H9-C6H5 ( 98) Scheme 7: alkyl-aryl exchange (a) and alkyl-aryl coupling Cb) reactions. The numbers in parenthesis represent the bond strength (kcal/mole) of the bond above them. (b) 106 The bonds formed in the two different organic products (butane and butylbenzene) have equal energies,[171 thus the enthalpies of these two reactions differ by 1.5 kcal/mole, with the alkyl-aryl coupling reaction (Scheme 7(b)) being thermodynamically preferred over alkyl-aryl exchange (again, assuming comparable entropies for the two reactions). In the reaction of Cp* 2 scc 4H9 with benzene, however, the only products formed are Cp* 2 scc 6 H5 and butane; no butylbenzene is observed. The fact that only a single reaction path (Scheme 7Ca)) is operative indicates that there must be a kinetic preference for alkyl-aryl exchange. We do not have reliable bond strength data to calculate the enthalpies of activation of non-arene C-H bonds, so we can not determine what type of reaction (alkyl exchange or alkyl coupling) is thermodynamically pref erred. It is important to note how- ever, that c-c bond formation is never observed in any C-H bond activation reactions by Cp* 2 ScR complexes (R = alkyl, aryl); for example, when Cp* 2 sccH 3 is treated with 13 cH 4 , only methyl exchange is observed, ethane is not formed. The nature of this kinetic preference for exchange over coupling in the reactions of Cp* 2 scR complexes with C-H bonds may be found in the structure of the transition state. The exchange <Dl> and coupling <D..2.> transition states differ only in the orientation of the incoming C-H group. 107 ... R... Sc>·· ····1I "··c·" ..... R... /' sc:·:······-.:·C"·ff '\ /I\ D.l The fact that exchange is the kinetically preferred pathway, indicates that .Ill. must be lower in energy than D.2., which is probably a result of the overlap for the atom in the central bridging position being the most sensitive to s character. In .IU the hydrogen occupies this central position and can bridge the three atoms very aptly with its nondirectional ls orbital. In D.2., however, this central bridging position is occupied by a far more directional carbon orbital Csp, sp 2 or sp 3 ) which can not bridge as well as the hydrogen. A preference for the better bridging ligand in the central position is consistent with calculations done by Steigerwald and Goddard on a related system.[481 They pre- dieted the structure of the transition state (Figure 11) for the H/D exchange reaction between Cl 2 ScH and n 2 • .H. Cl 2 Sc.·) a >'(p .. ··. ,-... fJ : : "n· a= 62° fJ = 75° ,, = 149° Figure 11: the predicted transition state for the H/D exchange reaction between c1 2 scH and o2 • 108 The angle that the central D must bridge (142°) is far greater than that of the lateral D (75°). It is expected for the better bridging ligand to occupy the site with the most obtuse bridging angle. This kinetic preference for .Ill. must be large, because .Dl. has far more severe steric interactions. In .Ill. both of the organic groups are forced into the edges of the wedge of the Cp* ligands, where they are be involved in unfavorable steric interactions. In .IJ.2. one of the organic groups is forced into the wedge, but the other sits is the center, which is the least sterically crowded position in the molecule. These steric effects are most likely responsible for the kinetic selectivity for 1° over 2° C-H bonds. For example, at ao 0 c Cp* 2 sccH 3 reacts with styrenyl C-H bonds roughly two orders of magnitude faster than with arene C-H bonds. In the reaction of Cp* 2 sccH 3 with 13 cH 3 I both alkyl exchange and alkyl coupling reactions were observed (Scheme 8) at comparable rates, with the latter being the thermodynamically favored by roughly 60 kcal/mole. The proposed transition states for these two processes are .ll3. and DA., respectively. 109 .·R ..... ·R.... / sc·;_::.... .':_:r sc .... -...... c- "··c·· ··· ..r ..·' ' /t\ By the previous logic, D.3. should be pref erred over .IM, because the lower energy transition state has the best bridging ligand in the central position. The reason that the D.3. and DJ. are so close in energy may be due to a strong scandium iodine interaction in .J:M. Donation of electron den- sity from one of the iodine lone pairs into the remaining empty orbital on scandium would strengthen the scandium iodine interaction in D4. If the la 1 were the vacant orbital in the transition state,[49] the iodine lone pairs would have poorer overlap with it in D.3. than in .D..4, thus Di would be lowered in energy relative to D.3.. This type of donation would be impossible for hydrogen, so no such stabilization would be expected for C-H bonds. This hypothesis explains the faster rate for benzylbromide over benzylchloride in their reaction with Cp* 2 sccH 3 Ceq 23). The bromine should have more diffuse lone pairs, allowing for greater donation and lowering of the energy of the transition state. The fact that carbon-bromine bonds are weaker than carbonchlorine bonds may also explain the rate difference. 110 EXPERIMENTAL All manipulations were carried out using either high vacuum line or glove box techniques. All solvents were purified by distillation from or CaH 2 titanocene. 13 cH Na followed by vacuum transfer from 12 cH and 4 (Merck or Stohler, ~. 10% 4 90% 13 cH 4 ) was dried over potassium and treated with a heterogeneous nickel phoshine complex to remove H2 and o 2 • All other gaseous reagents were freeze-pump-thawed twice before use. All liquid reagents were dried over 4A molecular sieves, and freeze-pump-thawed twice before use. NMR solvents, d 6 -benzene and d 12 -cyclohexane, were vacuum 0 transferred from activated 4A molecular sieves, then from titanocene. 1 H NMR spectra were run on Varian EM-390 (continuous wave, 90 MHz), JOEL FX-90Q (fourier transform, 90 MHz) and Brucker WM500 (fourier transform, 500 MHz) spectrometers. All spectra of new compounds are listed in Table 7. Infrared spectra were run as nujul mulls on KBr plates and recorded on a Beckman IR-2420 spectrometer. Mass spectra were recorded using a Kratos MS25 spectrometer. Analyses were performed by the Caltech Analytical Laboratory. Many of the reactions in this chapter were carried out in NMR tubes. Any experiment discussed in this chapter and not listed below was carried out in a sealed NMR tube with ~. 20 mg of the scandium complex in .Q.£. I 5 l b c -c 5 CcH > 4 cH 1x 3 2 a b "'-.._.) c d Cp* 2 Sc-CH=CH-C 6 H4 -ocH 3 Cp* Sc-CH=CH-C H -CH 2 \.._,_../ 6 4 3 b c d a a [Cp*Sc(~ c c d b b a c c d b b a c b b a Assignment s s s s 1.88 s 7.90 d 6.35 d 7.51 d 6.92 d 3.45 d 1.88 s 0.01 a 6.43 d 7.58 a 1.20 a 2.21 a 1.97 2.07 1.53 1.50 . 6 (oorn) 1 H and 13 c NMR spectroscopic data.* Cornoound Table 7. 20 20 8 8 20 20 8 8 Co.upling ........ ........ ........ a b c 7.51 m 2.00 s 2.07 s a b 1.88 s 8.19 d 6.37 d 6 (ppm) b c b a Assignment 21 21 Coupling Unless otherwise specified, spectra were taken in c6n6 at ambient temperature at 90 Mhz. The chemical shifts are reported in relative to TMS or residual protons of the solvent. Coupling constants are reported in Hz. * b Cp* 2 scc=:ccH 3 a '-.--I Cp* 2 sc-CH=CH-C 6 H4 -cF 3 Compound l\) p p -11.3 0.35 mL of d 6 -benzene. The spectra were referenced to internal TMS or the protio impurity of the solvent (C 6 D5 H). A measured amount of the appropriate reagent was added to the tube before it was cooled to sealed with a torch. Cyclooctane was 196°c and used as an integration reference (when needed) because of its low volatility and inertness toward reactions with scandium. Data analysis and plotting were carried out with a Compaq Deskpro minicomputer. Weighted fitting techniques were used to calculate slopes and errors [ 501 plots. The error in a given data point was estimated by of the van't Hoff, least squares kinetic and Arrhenius repeated measurement of that data point. Preparation of X.. dissolved Cp* 2 ScCH 3 (0.50 g, 1.5 mmol) was in 5 mL of cyclohexane and the resulting solution heated to a0°c for 4 days. Large yellow crystals of X were isolated from the mother liquor by filtration, and washed several times with petroleum crystals were dried .in vacuo (0.40 g, 85 %) • for C20H29Sc: C, 76.40; H, 9.30. Found: ether. The Anal. calc. C, 75.42; H, 9.04. We were unable to isolate the scandium styrenyl complexes made with para-methoxy- or para-CF 3 -styrene as solids. We were able to isolate Cp* 2 ScCH=CHC 6 H4 cH 3 -..E and its synthesis is described here. Cp* 2 sccH 3 (256 mg, 0.77 mmol) and para- 114 methyl-styrene (91 L, 1.01 equivalents) were dissolved in 4 mL of petroleum ether. for 12 hours and This solution was heated to 63°c concentrated to 2 mL. Cooling the solution to -78°c precipitated 99 mg (30% yield) of the pale yellow product. 1193(w), 835 (m), 1138(w), 7 54 (m), IR data (cm- 1 ): 2710(w), 1503(s), 1102(s), 656 (s), 1055(m), 60 8 (w), Anal. calc. for c 29 H39 sc: 1020(m), 585 (w), 98l(m), 480 (s), 365 (m). C, 80.52; H, 9.09. Found: C, 80.29; H, 9.20. Preparation of Cp* 2 scc:=ccH 3.!.. Cp* 2 sccH 3 (234 mg, 0.71 mmol) was dissolved in 10 mL of petroleum ether and the resulting solution was cooled to -78°c. Propyne (0.78 mmol) was condensed into the solution from a gas bulb, and the reaction vessel was then slowly warmed to room temperature. to 4 mL. The reaction mixture was concentrated Cooling the solution to -78°c precipitated the white solid product (115 mg, 62%), which was isolated by filtration. IR data (cm- 1 ): 2720(w), 2090(s), 1165(w), 1060(w), 1022(m), 965(s), 800(w), 700(w), 655(w), 440(s), 3 7 0 ( m) • Found: An a 1. ca 1 c • f o r C 2 3 H3 3 Sc : C, 7 7 •93 ; H, 9 •3 8 • C, 77.83; H, 9.42. Preparation of d 30 -Cp* 2 ScCH 3 ~ A 50-mL thick walled glasss bomb was charged with 4 g of Cp* 2 sccH 3 (12 mmol), 10 mL of c 6 n 6 (99.5% D) and two atmospheres of n 2 • The bomb was thermolyzed at 80°c for one day and the gas 115 (hydrogen/methane) removed in vacuQ. atmosphere of Approximately one n 2 was then added to 115°c for 24 hours. in y~~UQ and The H2/HD/D 2 and benzene were removed replaced with fresh c 6n 6 (20 mL) and n 2 (1 atm). The bomb was then the rm o 1 y zed at 115 ° C for 2 4 hours • Th is c y c 1 e was repeated with 35 mL of c 6n 6 and n 2 (1 atm). The bomb was cooled to -78°C and 12 mmol of CH 2 c1 2 and one atmosphere of H2 added. After warming to room temperature the benzene solution was stirred for 5 minutes and volatiles removed in vacuo. the resulting Petroleum ether (10 mL) was added to solid and the slurry filtered. The remaining solid was washed twice with petroleum ether, and all of the petroleum ether solutions combined. The petroleum ether was removed .in vacuo and replaced with 5 mL of diethylether. Cooling the diethylether solution to -78°c precipitated 2.73 g (62%) d 30 -Cp* 2 ScCl. The proton impurity of d 30 -Cp* 2 ScCl was measured by 1 H NMR integration relative to ferrocene, Cp*2ScCH3 ws of the yellow product, prepared of its Cp* resonance and found to be <0.2%. from a 30 -Cp* 2 ScCl by a d3 0 method previously discussed (Chapter 1) • Cp* 2 ScD is not stable for long periods as a solid, but can be isolated. Cp* 2 sccH 2 c 6 H5 (33 mg, 0.081 mmol) was dissolved in 0.3 mL of c 6 n 6 and two atmospheres of n 2 put over the solution. After being stirred for ten minutes the benzene solution 116 was cooled to -78°c and the benzene sublimed out of the reaction vessel in vacuo. c 6 a 6 (0.4 mL) was added to the resulting solid and the cloudy solution transferred to an NMR tube. After 5 minutes at room temperature, the tube was cooled to -78°c and 0.098 mmol of CH 3 r was transferred into the tube. The tube was sealed with a torch and warmed to room temperature. out of solution. to Cp* 2 ScI spectrum. As the benzene melted, gas bubbled The 1 H NMR showed resonances assignable and CH 4 ; CH 3 D was not observed in this A small amount of Cp* 2 scc 6 a 5 (5% relative to Cp* 2 ScI) was also observed. We could not find an integration reference that did not undergo H/D exchange under the conditions required to promote H/D exchange into cyclopentane, so the extent of deuterium incorporation was measured by mass spectroscopy. A 2 mL sealable tube was charged with 8 mg of Cp* 2 scca 3 (0.024 mmol), cyclopentane (0.53 mmol), H2 • 50 L of 0.35 mL of c 6o6 and 700 torr of The entire tube was cooled to -196°c and sealed with a torch, giving the sample 3.6 atmospheres of a 2 at 2s 0 c. The entire tube was thermolyzed at 120°c for 26 days, and the reaction solution examined by G. c. mass spectroscopy. By comparison of the mass spectrum of cyclopentane from this reaction to that of normal cyclopentane, it was determined that the cyclopentane from this reaction was 17% enriched in deuterium, giving a turnover rate of 0.06 117 turnovers/day. Measurement of EQuilibrium Constants for EQs. l..fil_. (7) and The sample used to study the equilibrium between Cp* 2 ScH and Cp* 2 scc 6 H5 (Eq. (7)) was prepared in an NMR tube mounted directly on a Kontes teflon stoppered valve. This tube was charged with 18 mg of Cp* 2 sccH 3 (0.054 mmol) and 0.5 mL of c 6 H6 • The entire tube (up to the Kontes valve) was cooled to -196°c, and evacuated. H2 (700 torr) was admitted into the tube and ten minutes allowed for the gas to come to thermal equilibrium. The Kontes valve was closed and the tube warmed to room temperature. The apparatus was removed from the vacuum line and shaken for five minutes to ensure formation of Cp* 2 ScH. The tube was then cooled to -196°c, the hydrogen removed in vacuo, and the tube sealed with a torch. The sample used to study the equilibration between Cp* 2 Sc(THF)H and c 6 H6 (Eq . (8)) was prepared by dissolving an isolated sample of Cp* 2 Sc(THF)H in c 6 H6 solvent in an NMR tube and sealing it with a torch. The 25°C and higher temperature experiments were carried out in a constant temperature oil bath, which was heated with a Precision Scientific Co. thermostated water circulator. In this way the temperature of the oil bath was kept constant to submerged ± 0.1°c. The samples were fully in the bath and agitated to facilitate the 118 equilibration of H2 between the solution and gas phases. The experiments carried out at 6°c were performed in a room cooled to this temperature. At a given temperature, the volume at the solution and the ratio of the two equilibrating scandium species (1 H NMR) were measured. The equation describing the equilibrium constant for Eq. (7) is shown below. = [Cp*2ScC6Hsl [H2l [ Cp* 2ScH] [ C6H6] The relative concentrations of the two scandium complexes were calculated from the ratio of their Cp* peak heights. The concentration of benzene was calculated from its density at the temperature of the experiment. The concentration of hydrogen was calculated using Henry's law and the standard concentration of H2 at the temperature of the experiment.[ To derive the expression for [H 2 lequil' two simultaneous equations were solved (Eqs. (28) and (29)). ( 28) B = ( 29) 119 Where A is the pressure of H2 above the solution, v 1 is the volume of gas space above the solution, R is the gas constant, T is the temperature, B is the concentration of H2 in solution, [H21std is the v 2 is the volume of the solution and concentration of H2 in C6H6 a at temperature T when the pressure of H2 is one atmosphere. [ 51 ] The expression for [H21equil ( B) is given in Eq. ( 3 0) • [ H2Jequ i l = (total mmol H2) v1 [H 2Jstd + RT v2 ( 30) The total mmol of H2 is given by Eq. (31). The (mmol of H2 )initial was calculated using Eq. (30), assuming that only that amount of H2 in the known volume of benzene would remain after evacuation of the sample tube (just prior to sealing). The equilibrium constants for Eq. (8) were determined using analogous calculations. 120 Kinetic Measurements of the Reaction of CP* 2 scCH 3 with C 6 !! 6 and c 6 .Q 6..!.. These reactions were carried out in neat benzene solutions. Sealed NMR tubes were prepared with 25 mg Cp* 2 sccH 3 0.4 mL of benzene and 3 µL of cyclooctane, and thermolyzed at a0°c. The progress of the reaction was followed by 1 H NMR (continuous wave, Varian EM-390), monitoring the integrated intensity of Sc-CH 3 versus internal cyclooctane. The reaction of Cp* 2 ScCH 3 with c 6 H6 gave linear first order kinetic plots while the reaction with c 6o 6 gave nonlinear first order plots. The interpretation of these kinetics is discussed in the text. The ratio of CH 3 D to CH 4 was measured by cutting and weighing the peaks of the 500 MHz NMR spectrum. Reaction of ~ with CD 4 ~ A sealable NMR tube charged with 16 mg of X and 0.40 mL of c 6o12 . tube was cooled to -78°c and evacuated. ~. was The entire CD 4 (400 torr, 1.5 equivalents) was admitted to the tube and the tube sealed. In the initial NMR spectrum, observed between 0 and 0.5 ppm. no signal was After thermolysis at a0°c for eight days, a resonance at ~. 0.15 ppm was observed (CH 4 appears at 0.15 ppm in c 6 o12 ). This resonance grew on further heating, until after 35 days it was >50% CH4• Kinetic Measurements Cp* 2 scca 3 cyclooctane and of Intramolecular C-H (25 mg, and 0.076 mmol), 0.40 mL of c 6 o 12 were loaded C-D 3 µL into a sealable NMR tube, and the tube sealed under 1 atm of N2. 121 The sample was thermolyzed at a0°c. intramolecular C-H activation (Eq. The kinetics of (11)) were followed by 1 H NMR monitoring the decrease in integrated intensity of the Sc-CH 3 resonance relative to internal cyclooctane. The kinetics of intramolecular C-D activation were examined using a directly analogous experiment, except Both of these reactions gave non-first order kinetics. The interpretation of their kinetic behavior is discussed in the text. Kinetic Measurements of the Reaction of d 30 -Cp*2ScCH3 with c 6 ~ 6 and c 6o 6~ These measurements were made in a manner directly analogous to those described for the reactions of Cp* 2 sccH 3 with benzene, was substituted for Cp* 2 sccH 3 • except d 3 0 -cp2ScCH3 The reactions of d3 0 - Cp* 2ScCH3 with c 6 H6 and c 6 o 6 both obey first order kinetics, (1.41 ± giving pseudo first order rate constants of 0.05) x 10- 4 sec- 1 and (5.1 respectively. :t 0.1) x H,- 5 sec- 1 , The true value of kH/k 0 is discussed in the text. Kinetic Measurements of the Reaction of Cp*2ScCH3 with 13 cH 4 ~ Kinetic samples were prepared in 5-mm medium walled NMR tubes (Wilmad Inc., 524-p), which contained 10 mg of Cp* 2 ScCH 3 (0.03 mmol), c 6 o12 (0.5 mL) and 0.65.0 mmol of 13 cH 4 • At 10°c, approximately 20% of the CH 4 122 was in solution; obtained accurate (±5%) with cyclooctane as an concentrations internal were integration Plots of -ln[Sc-CH 3 J versus time should be linear over 2-4 half-lives (depending on the ratio of 12 c standard. to 13 c in the system) in these experiments with excess 13 cH • 4 Typically only data from the first 20-30% of the reaction A value of 1 x 10- 5 M- 1 sec-l was was used. obtained for the second order rate constant. In roughly 20% of these experiments kobs was 3-4 times larger than expected. Repeating the experiment generally gave a value of consistent kobs with those concentrations of methane. observed at other We attribute this rate increase to a small amount of some catalyst promoting the reaction. The identity of this catalyst is unknown, but the reaction chemistry exhibited by the tuck-in complex (X ) ( Y.id.e s..u.n.r..a) candidate. aged s ug g e s t s th a t it ma y be a 1i ke1y This apparent catalysis was most common in sample of Cp* 2 sccH 3 , in which some thermal decomposition could have given rise to a small amount of x. Kinetic Measurements of the Reactions of Cp* 2 scCH 3 with H2C=CHC6~4X-p (X=CF3, OCH3L_ Kinetic samples in septum capped NMR tubes contained 25 mg of Cp* 2 ScCH3 (0 .076 mmol), 0.4 mL of c 6 o 6 and 0.16-0.30 mmol of the appropriate styrene compound. The kinetics of the 123 by 1 H NMR (FT NMR, JOEL FX-90Q), monitoring the disappearance of the Sc-CH 3 signal and the appearance of the methoxy resonance of Cp* 2 ScCH=CHC 6 H4ocH 3 relative to internal FeCp 2 (sublimed twice). The kinetics of the reaction of Cp* 2 sccH 3 with H2 C=CHC 6 H4 cF 3 -p were followed by F 19 NMR (FT NMR, JEOL FX-90Q), monitoring the disappearance of the CF 3 resonance of H2 C=CHC 6 H4cF 3 -p and the appearance of the CF 3 signal of Cp* 2 ScCH=CHC 6 H4cF 3 -p relative to internal c 6 H5 CF 3 • This reaction was followed by 1 H NMR as well, to verify that only Cp*2ScCH=CHC6H4-CF3 was produced. The reactions of Cp* 2 ScCH 3 with both of these styrenes obey second order kinetics, [31) giving linear second order plots (Fig. 7) for over three halflives. The Arrhenius calculated using Table 6. and activation parameters were standard equations and are given in The Arrhenius plots are shown in Fig. 8. Both reactions exhibit small enthalpies of activation and large negative entropies of activation, associative reaction, transition state. indicative of an going through a highly ordered The similarity of the activation parameters for these reactions suggest that substituent effects are unimportant in these reactions. Kinetics of the Reactions with C6g5R (R=N(CH3l2 1___9:!3 1 H, CF3~ of d3 0 -Cp*2ScCH3 Sealed NMR tubes containing 15 mg d 30 -Cp* 2 ScCH 3 (0.041 mmol), 0.35 mL of c 6 o 12 , 0.52 mmol of the appropriate substituted benzene 124 and 2.5 µ L of cyclooctane were thermolyzed at 80 The reaction was followed by 1 H NMR, monitoring the decrease in integrated intensity of the Sc-CH 3 resonance relative to cyclooctane. In all complexes were observed, first order (Fig. 9) cases, only phenyl and the kinetics were pseudo for over three half-lives. accounts for 5% or less of the total methanes produced. The rate constants for these reactions are given in Table 5. with Toluene. A glass teflon was valve reaction vessel with a Kontes charged with 20 mg of Cp* 2 scR (eq. 0.05 mmol) and 5 mL of toluene. The entire vessel was then immersed in an oil bath at the prescribed temperature for the prescribed time period. The toluene solvent was removed in vacuo and the resulting solid dissolved in C6D6. The products of the reaction were identified by 500 MHz 1 H NMR. The sealable The product ratios are given in Table 8. Reaction of NMR was tube Cp* 2 ScCH 3 with charged with Methyliodide. 15 mg A Cp* 2 ScCH 3 (0.045 mmol), 0.4 mL of c 6o6 and 0.27 mmol of 13 cH 3 I. The tube was sealed with a torch and the reaction followed by 1 H NMR. Initially only Cp* 2 sc 12 cH 3 was observed. As the intensity of the Sc- 12 cH 3 signal decreased, resonances for both Sc- 13 cH 3 and 13 c labeled ethane were observed. The 125 approximate half-life for the disappearance of Sc- 12 cH 3 was 5 hours at 2s 0 c. Cp*2ScCH 3 was reacted In a second experiment, 15 mg of with 0.30 mmol of 12 ca 3 r. The resonance of Sc-CH 3 slowly decreased as ethane was being formed. The half-life of this reaction was approximately 30 hours at 2s 0 c, and represents the half-life of only the alkyl coupling reaction, since the alkyl exchange reaction is degenerate in this experiment. Thus, the half-life of alkyl exchange is also roughly 30 hours, because the halflife of the reaction of Cp* 2 ScCH 3 with 13 ca 3 r is the sum of the half-lives of the alkyl coupling and alkyl exchange reactions (Scheme 5). 126 REFERENCES 1. (a) A.E. Shilov, A.A. Shteirnan Coord. Chern. Rev. (1977) (b) D. E. Webster Ady. Orgrnet. Chern. (1977) ,12, 147188. Cc) J.P. Collman, L.S. Hegedus "Principles and Application of Organotransition Metal Chemistry"; University Science Books: Mill Valley, Calif., (1980), chapter 4. (d) G.W. Parshall Cherntech(l984)628-638. (e) A.H. Janowicz, R.G. Bergman J. Arner. Chern. Soc (1983) .l..Q..5., 3929-3939. (f) W.D. Jones, F.J. Feher J. Amer. Chern. Soc (1984) l..QQ, 16501663. 2.A,, 97. 2. G.W. Parshall Accts. Chern. Res. (1970) ..11, 139-145. R. Dicirno, S.S. Moore, A.F. Sowinski, G.M. Whitesides J. Amer. Chern. Soc (1981) .l..Q]., 948-949. S.S. Moore, R. Dicosirno, A.F. Sowinski, G.M. Whitesides J. Arner. Chern. Soc (1980) .l...0..2. 6713-6725. 3. A higher oxidation state does not need to be present in the ground state of the molecule. For example, therrnolysis of cp,TaH 1 (Ta(V)) generates Cp 2TaH Ta(III) by reductive elirninati~n of H • u. Klabunda, G.W. Parshall J. Arner. 2 Chern. Soc (1972) ,ii, 9081-9087. 4. In this class of do complexes we exclude complexes such as Cp 2 TaH 3 , see above reference. For a current review see: I.P. Rothwell Polyhedron (1985) 5. J.W. Bruno, T.J. Marks, (1982) .l.Q.i, 7357-7360. ~' 177-200. v.w. Day J. Amer. Chern. Soc 6. L.R. Chamberlain, A.P. Rothwell, I.P. Rothwell J. Amer. Chem. Soc (1984) .l.Q.6., 1847. 7. J.L. Garnet Catal. Rey. (1971) ,12, 229-268. shall Accts. Chern. Res. (1975) ..8., 113-117. G.W. Par- 8. G.W. Parshall "Homogeneous Catalysis"; John Wiley and Sons: New York, (1980), page 123. J.L. Garnet, M.A. Long, A.B. McLaren, K.B. Petersen J. Chern. Soc., Chern. Commun. (1973) 749-750. 9. K. Klabunde and G.W. Parshall J. Amer. Chern. Soc (1972) .ii, 9081-9087. R.M. Blake, J.L. Garnet, I.K. Gregor, w. Hannan, K. Hou, M.A. Long J. Chern. Soc., Chern. Commun. (1975) 930-932. 127 It is not certain that H/D exchange for PCCH ) is intrarnolecular. An adduct is not formed between cpi 2 ScH and PCCH 3 > 3 but the rate of H/D exchange into the phosphine methyl groups is faster than into to Cp* methyl groups of Cp* 2 ScH. This is similar to the exchange reaction for THb Cwhich does f orrn an adduct) and unlike all of the other 1 C-H bonds tested. 10. 11. A. E. Shilov Pure and Appl. Chern. Cl 97 8) 5.Q., 7 25. 12. W.A. Nuggent, D.W. Ovenall, S.J. Holmes Organornetallics (1982) 2, 161-162. This exchange was not rnonitor5d by NMR, rather the D incorporation after 36 days at 120 C was measured by mass spectral analysis of the cyclopentane, see experimental section. 13. Cp* ScH was the catalyst used for exchange unless H was put on this scale ba~ed on tis facile exchange with 2 the hydride ligand at -95 c Cvide supra) • 14. otherwis~ stated. A. Streitwieeser, C.H. Heathcock "Introduction to Organic Chemistry"; Macmillan: New York, (1976), pages 608- 15. 609. 16. J.E. Bercaw, H.E. Bryndza, L.K. Fong, R.A. Paciello; manuscript in preparation. A.H. Janowicz, R.A. Periana, J.M. Buchanan, C.A. Kovac, J.M. Stryker, M.J. Wax Pure and Appl. Chern. (1984) ..5..2_, 13-23. 17. D.F. McMillen, D.M. Golden Ann. Rev. Phys. Chern. Cl982) 18. K.A. Gingerich J. Chern. Phys. (1971) ..5..i, 3720. n, 493-532. 19. rn.w. Cook, D.N. Hanson, B.J. Adler J. Chern. Phys. (1957) .2.6., 748-751. 20. J.W. Bruno, T.J. Marks, L.R. Morss J. Amer. Chern. Soc (1983) .l.Q2, 6824-6832. 21. Cp* 2 scc 6 H5 does not react with THF. 22. J. Halpern Inorg. Chim. Acta (1985) .J]_, 493-532. The C D used in these experiments was 99.8% enriched in deuteriBrn~ If this reaction had a kinetic isotope effect as high as 10 the expected amount of CH 4 would only be 2%, which would not be observable by NMR. 23. 128 24. A.R. Bulls, J.E. Bercaw; manuscript in preparation. J.E. Bercaw Adv. Chern. Ser. (1978) No. 167, 36. D.K. Erwin Ph.D. Thesis, California Institute of Technology (1979). 25. A.R. Bulls and J.E. Bercaw, manuscript in preparation. 26. N.F. Gol'dehleger, M.B. Tybain, A.E. Shilov, A.A. Shteinrnan Russ. J. Phys. Chern. (1969) .A.J_, 1222-1223. 27. M.J. Wax, J.M. Stryker, J.M. Buchanan, C.A. Kovac, R.G. Bergman J. Arner. Chern. Soc (1984) .l.Q.6., 1121-1122. Activation of methane by transition metals. 28. (a) P.L. Watson J. Amer. Chern. Soc (1983) .l..Q.5., 64916493. (b) C.M. Fendrick, T.J. Marks J. Amer. Chern. Soc (1984) l.Q&., 2214-2216. 29. An analogous mechanism was proposed for the reaction of Cp*2ScCH3 with C6D6. 30. R.M. Silverstein, G.C. Bassler, T.C. Morrill "Spectrometric Identification of Organic Compounds"; Wiley: New York, (1981). 31. J.W. Moore and R.G. Pearson "Kinetics and Mechanism"; Wiley-Interscience: New York, (1981), pages 22-26. 32. C.Y. Chang, C.E. Johnson, T.G. Richmond, Y.T. Chen, w.c. Trogler, F. Basolo Inorg. Chern. (1981), 2.Q, 3167-3172. Q.Z. Shi, T.G. Richmond, w.c. Trogler, F. Basolo J. Arner. Chern. Soc (1984), l..Q.2., 71-76. R. Ugo, A. Pasini, s. Cenini J. Amer. Chern. Soc (1972), _ii, 7364-7370. F. Basolo and R.G. Pearson "Mechanisms of Inorganic Reactions"; second edition, John Wiley and Sons: New York, (1967), pages 403410. 33. c. Mcdade, J.E. Bercaw J. Orgrnet. Chern. (1985) 281-315. ~' 34. The ortho tolyl complex is the least thermally stable of the tolyl isomers. Cp* 2 ScCnH 4 CH 3 -o bsornerized 20% to Cp* 2 scCnH 4cH 3 -rn in the solid state \-20 c, one year), while the meta and para tolyl complexes are stable under these conditions. 35. T.H. Lowry and K.S. Richardson "Mechanism and Theory in Organic Chemistry"; Harper and Row: New York, (1976), chapter 7 and references therein. 36. (a) G.A. Olah, Accts. Chern. Res. (1971) ~' 240-248. (b) J.H. Ridd, Accts. Chern. Res. (1971) ~' 248-253. (c) G.A. Olah, s.c. Narang, J.A. Olah and K. Larnrnertsrna, Proc. 129 Natl. Acad. Sci. Cl982) ll, 4487-4494. 37. (a) J.R. Sweet, W.A.G. Graham Organometallics (1982) 2, 135-141. (b) Y. Aoyama, T. Yoshida, K. Sakurai, H. Hogoshi J. Chem. Soc., Chem. Commun. (1983) 478. 38. G.B. Shul'pin, G.V. Nizova, A.T. Nikitaev J. Orgmet. Chem. (1984) Zli_, 115-153. 39. J.R. Sweet, W.A.G. Graham J. Amer. Chem. Soc (1983) .l..Q2, 3 05-3 06. 40. C.D. Johnson "The Hammett Equation"; Cambridge University Press: Cambridge, England (1973). 41. 25_, G.B. Shulpin, A.N. Kitaigorodskii Zh. Fiz. Khim. (1981) 266-267. 42. G.A. Olah, S.J. Kuhn, s.H. Flood J. Amer. Chem. Soc (1961) .al, 4571-4580. 43. Most electrophiles have formation of the -complex as their rate determining step, and subsequently show no isotope effect, see Lawry and Richardson, ibid •• 44. J.L. Garnet, Catal. Rey. (1971) ..5,, 229-268. G.W. Parshall "Homogeneous Catalysis"; John Wiley and Sons: New York, (1980), chapter 7. W.D. Jones, F.J. Feher J. Amer. Chem. Soc (1984) .l.Q.Q., 1650-1663, and references therein. 45. J.K. Kochi "Organometallic Mechanisms and Catalysis"; Academic Press: New York, (1978), chapter 11 and references therein. 46. R.B. Woodward, R. Hoffmann "The Conservation of Orbital Symmetry"; Academic Press: New York, Cl 97 0) • 47. J.W. Lauber, R. Hoffmann J. Amer. Chem. Soc (1976) .2Ji, 1729-1742. 48. M.L.Steigerwald, W.A. Goddard J. Arner. Chem. Soc (1984) l..Q.2., 3 08-311. 49. For this to be the case the b orbital would have to be the one that initially becomes inv61ved with the substrate. 50. D.P Shoemaker, C.L. Garland, J.I. Steinfeld, J.W. Nibler "Experiments in Physical Chemistry"; McGraw-Hill: New York, Cl 981) , chapter 20. 51. M.W. Cook, D.N. Hanson, B.J. Adler J. Chem. Phys. (1957) .2,6., 748-751. 130 Insertion and Vinylic C-H Bond Activation Reactions of Unsaturated Hydrocarbons with Permethylscandocene Alkyl and Hydride Complexes. 131 INTRODUCTION Early transition metal organometallic complexes are highly reactive toward a wide variety of organic substrates. This high reactivity is thought to be due to the highly electrophilic metal center (usually do) and the coordinative unsaturation of these complexes.£11 Some of the most reac- tive organic substrates are olef ins and alkynes, which, being nucleophilic, often react readily with these organometallic electrophiles. Early transition metal complexes catalyze a diverse series of organic reactions for olef ins, including isomerization, metathesis, hydrogenation, oligomerization and polymerization.£21 In the early 1950's Ziegler and his coworkers discovered that TiC1 4 /Al(CH 2 cH 3 > 3 is a very efficient ethylene polymerization catalyst.£31 His group and others have found similar catalysts that polymerize several other olef ins as well.£41 Understanding the mechanisms of these polymeriza- tion reactions is a great challenge. Industrially useful catalysts are generally heterogeneous and polymerize olef ins too rapidly to make them convenient for mechanistic study. ' Several homogeneous titanium catalysts have been developed that polymerize or oligomerize olef ins at measurable rates.CS] Extensive mechanistic study of these soluble catalysts indicate that the active polymerization catalysts have a Ti(IV) coordinatively unsaturated metal center.Cl Ca), 132 6] Direct observation of individual steps in the polymerization mechanisms of these catalysts is difficult because of competitive reduction of the metal complex and other side reactions leading to deactivation of the catalyst. As a model for this catalyst system, Watson prepared Cp* 2 LuR (R = H, CH 3 ).[7] Like the proposed active site for Ziegler-Natta catalysis, these complexes are highly electrophilic and coordinatively unsaturated. Both ethylene polymerization and propene oligomerization are catalyzed by Cp* 2 LuR. This lutetium system provides evidence for initia- tion, propagation and termination steps proposed for Ziegler-Natta catalysis, and is the first system in which the direct product of olefin insertion into a M-C ~-bond has been observed.[8] Several types of termination step are observed in this lutetium system, including p-hydride and P-methyl elimination as well as intramolecular C-H activation of a Cp* ligand. We have examined the reactions of Cp* 2 ScR CR = H, alkyl) with olefins and alkynes. Previous to our work, the only reported reactions of scandium complexes with olefins were the polymerization of ethylene and oligomerization of propene by Cp2Sc{~ 3- CH2CH=CH2).[9] Although Cp*2ScCH3 is isoelectronic with Cp* 2 LuCH 3 , Sc has a smaller ionic radius than Lu, making steric interactions more severe for Sc. Thus, Cp* 2 ScR complexes polymerize ethylene, but react only stoichiometrically with propene, isobutene and other lJJ olef ins. The products of these reactions are vinylic scan- dium complexes, formed by direct vinylic C-H activation of H2 C=CHR' (R = H, CH 3 >. Allylic scandium complexes are not observed. Our work suggests in addition to P-hydride elimination and chain transfer, direct vinylic C-H activation may be an active termination step for Ziegler-Natta catalysis. 1.34 RESULTS AND DISCUSSION Reactions of alkynes with Cp* 2 sccH 3 and Cp* 2 scH. As previously mentioned, Cp* 2 ScR CR = H, CH 3 ) react with propyne to give RH and Cp* 2 scc:ccH 3 • This scandium propyde complex is a catalyst for the dimerization of propyne (eq 1) • (1) Trimers or oligomers are not observed in the reaction. The likely mechanism for this reaction is shown in Scheme 1. Scheme 1 In the slow step of this mechanism propyne inserts into the scandium-propyde bond, giving a vinyl complex Cy>. In a subsequent fast reaction, Y reacts with the sp C-H bond of propyne to give 2-methyl-l-pentene-3-yne and regenerate 1.35 Cp* 2 scC::CCH 3 • If the activation of an sp C-H bond is sig- nificantly faster than the alkyne insertion step, only dimers would be expected. Similar reactivity and mechanism were reported by Threlkel for several permethyltitanocene complexes. [10] Cp* 2 sccH 3 reacts stoichiometrically with 2-butyne, giving the alkyne inserted product Ceq 2). (2) Prolonged reaction time and heating do not lead to any further insertions. Ethylene polymerization by permethylscandocene alkyl and hydride complexes. At room temperature Cp* 2 ScR com- plexes CR= H, CH 3 , CH 2 c 6 H5 > are active homogeneous catalysts for the polymerization of ethylene. These polyrnerization experiments were done with one atmosphere or less of ethylene over the catalyst solutions. Cp* 2 sccl and Cp* 2 scc 6H5 do not react with ethylene under these conditions. When Cp* 2 sccH 3 and Cp* 2 sccH 2 c 6 H5 were treated with 3 and 10 equivalents of ethylene, respectively, at -78°c and slowly warmed to room temperature, the ethylene was totally consumed by less than half of the scandium complex, suggesting that the chain initiation step (first insertion) is 1J6 slower than the chain propagation step (subsequent insertions). We feel that this is probably due to steric and electronic effects. The scandium-methyl bond is probably stronger than other scandium-alkyl bonds,[11) thus Cp* 2 sc-CH 3 would have a higher barrier to insertion than Cp*2sc-CH2CH2CH3 or Cp*2Sc-CH2(CH2)nCH31 if Sc-C bond breaking is significant in the transition state. Cp* 2 sccH 2 C6H5 should have a more sterically congested coordination sphere than its first insertion product (Cp* 2 ScCH 2cH 2 cH 2 c 6H5 >, making the benzyl complex less open to insertion of ethylene than Cp* 2 ScCH 2 cH 2 cH 2 c 6H5 • Conse- quently, the products of ethylene insertion into Cp* 2 sc-CH 2c 6 H5 would be expected to be better ethylene polymerization catalysts than the parent Cp* 2 ScCH 2 c 6 H5 complex. These polymerization experiments were generally run until the reaction stopped taking up ethylene. At this point the solution was cloudy, due to precipitated polyethylene. The excess ethylene was removed in vacuo, and HCl gas admitted into the reaction vessel to cleave the scandium-polymer bonds. The IR spectrum of this polymer (Figure 1) shows bands typical of polyethylene [12) as well as three bands at 1643, 990 and 910 cm- 1 , which are characteristic of vinyl groups in polyethylene.[13) When the polymer is treated with Br 2 Cgas) the three bands disappear, which is consistent with their assignment as vibrations due to vinyl groups. From the relative intensities of the bands at 910 cm-l (due to RCH=CH 2 > and 1378 cm-l (due to RCH 3 ) and their relative ' I i : 1 I : I I I' ' '' I -1 . --;-'' ' :I - ; __ ' . ~- I .. 1 \· I ' ' - --· .tOOO 1. 1) co :,.. I' :.,' ·1 ± .. i __ : ·1, I · I I' J J600 i I : ' 3MJO Figure. JIOO I '1 , I !100 3000 : J I 1 2ft()I') ' ' I ' 2600 I I : 1.m ' I ~ noo I - ' ! I ~I 1 I 1 .. - I 2000 I ,--- .. ' -----·: '·-·1 :' j I i I •. 1 . __ I I ' I -·1--- - ---·-1- - -r· I ! !- l I ,~ ·J . _, .. __I ~ f J - ;..: I .. I' : ''°° - , I too : I 1600 I I 1500 I 1 W"VfNUMlfl CM' 1 l7'Dn I 11 ,i I ! i I •· - ., · . I ' ' ' I · I 1400 i I_ - / ' I I 1)00 I 1100 I I 1 , .I 1100 I· · ·I I · -··"· ' · t · I I i I . i - :·· - ! !,: ' I ' ,'' - ' -i.. . I ··-·:' --1' . ..I. . I ·--·+-1 · I ' I: I . i ' --'' I . ..I 1· J __ I.. . ..I! . I ·i I t~-' - ~- . ~l--.1 . _;_;,I_ - -.: ·j I ' I' l l 4- · -· ' - : i; . ' ' .. I. I ---L-- T; : !. -.--r-·'' .. . , I . i '' - --!· . l ! I· ' I ' I I I 1 ' ; ' : II I' ·1·· ,- - ·:- ! ; , ' I ' I - f r,.rrT :1-r. . i~ I I 1000 I Ii : I ' I. ' I l ' I ! II j- L I · I I · 1· : . ' 1· I' ;· - I · ( I I . : I I· : ' ; ' I ' ·! --~~ -! I t I~:IL ' ,. ; : : I :· -1- 'i . - . -· . ! I 1' j_ ;__ ! "'° I : , , ... / :. ! ; ""' ;. ... j' f ' ,. ""' ' I '°° ,i .•. '._ I ' ., I ., ' ' '; I I 1'1. i.1- : I. ' '1 , .. I ; I I - 1- . ,I... ; - ....I . j --tt-Ii '· I ·· ! I .ii :I 1:- I ' II,_ :_ i.. l J ' Ij j I, ___ ,-~ ·. 1-·j' i .. ' ;· Ii I '' ' j . ·I I '· i ,·- r-~- -. 1 '."':' ---t- - I ': I, I ' l : ·f' ·: ;I IR spectrum of polyethylene from the reaction of Cp* 2 ScCH 3 with H2 C=CH 2 . J - - - - 1- '' r· I . -, . .. ·- -- - 1--· ! ' I ~"7tI I I ' ' i I .: I · I-~ -- 1 j. J,,_- -, i ·-· · -·i • 001- ·1- t I i -+--~ i · -:--·_ ~- _:-~. ! ,- l- . jJ__ _ , - ~-'i ' l--.J'.----1'" ' --t- ·~-v~ L --·,-. 1 '° , 1 .~ ,--+-+ . -+ - /-~-·· -~ ' I : 1 I; [I:' I I I I :~ rt -:··U\..~X1I·-'l 1 r:~·~ :;1 I i j I-" -.._J 'vJ 138 extinction coefficients [14] the ratio of methyl to vinyl groups was determined to be 1.7:1, respectively. The molec- ular weight of this polymer is 1700 + 300 G/mole, and its polydispersity (Mw/Mn) is 1.7 ± 0.3. This polydispersity is very low,ClSl which is generally found in catalyst systems where chain termination does not occur at an appreciable rate;Cl6l the polymer yields in these systems are one or less polymer chain per catalyst molecule. One mole of Cp* 2 scCH 3 produces 1.7 ± 0.2 moles of polyethylene, indicating that chain termination must be occurring. A low poly- dispersity in a system with significant chain termination is difficult to reconcile. The low molecular weight of the polymer is most likely responsible for this anomaly. The polymer was washed several times with organic solvents during isolation. When these washings were evaporated to dry- ness a sticky polymer residue was left. In washing the polymer, lower molecular weight material was removed that would have made the polydispersity higher. The presence of vinyl groups in the isolated polymer suggests P-hydrogen elimination may be an active chain termination pathway, which would lead to polyethylene with one methyl and one vinyl end group Ceq 3). (3) The polymer prepared from cp 2 sc<~ 3-CH 2CH=CH 2 > (Cp = ry-c 5H5 > has a number average molecular weight of 160,000 to 139 190,000 and a weight average molecular weight of 740,000 to 2,340,000.£81 £a.a. 9. Thus the polydispersity for this polymer is The average molecular weight for this polyethylene is much larger than that found in our system, which may be a result of greater steric interactions for the Cp* system accelerating its rate of chain termination relative to Cp Sc(~ 2 3 -CH 2 CH=CH 2 >. The polydispersity of the two systems can not be directly compared, since the polydispersity of the polyethylene from Cp* 2 sccH 3 is probably anomalously small. If the catalyst/polymer solutions are treated with air, rather than HCl, to cleave the scandium-polymer bonds, a polymer with a carboxylic acid end group is formed <.i......e...... HOOC(CH 2 >ncH 3 >. This polymer gives an IR spectrum with the same bands as polyethylene formed by addition of HCl (Figure 1), as well as several additional large bands at 3400, 1560, 1400, 1090 and 500 cm- 1 , the first four of which are characteristic of a carboxylic acid functionality.£171 The melting points of the polyethylene formed by these scandium complexes range from 118 to 125°c, indicating that the polymer is highly crystalline.[181 The IR spectra show very narrow bands, which are also consistent with high crystallinity and a highly regular polymer chain. Clef in and alkyne hydrogenation catalyzed by Cp* 2 scH. In addition to being a catalyst for ethylene polymerization, 140 Cp* 2 ScH is a catalyst for the hydrogenation of olefins and alkynes at 25°c. Terminal and internal linear olef ins, as well as isobutylene are hydrogenated by H2 in the presence of Cp* 2 ScH. Internal alkynes are hydrogenated to their re- spective alkenes and alkanes. Trisubstituted olefins are not hydrogenated by Cp* 2 ScH. 2-methyl-2-butene does not re- act with Cp* 2 ScH/H 2 after thermolysis for long periods of When 2-methyl-2-butene was treated with Cp* 2 ScD/D 2 , time. however, H/D exchange was observed at the vinylic position ( eq 5) • >=< (5) Stoichiometric reactions of Cp* £c..C.H and Cp* ScH with Cp* 2 ScH + D 2 olef ins. 3 2 Cp* 2 sccH 3 reacts with propene in cyclohexane sol- vent to give Cp* 2 scCH=CHCH 3 -~ and isobutane Ceq 6). (6) The trans configuration of the propenyl ligand was assigned based on the large coupling constant between the two vinylic protons C3JH-H = 19 Hz). Isolation of Cp* ScCH=CHCH -~ has 2 3 proven impossible. Although reaction 7 gives a 85-90% yield by 1 H NMR, removal of the solvent and attempted isolation of the product gives only an impure oil in low yield (10-15%). Treatment of Cp* 2 scCl with LiCH=CH(CH 3 > leads to the same 141 impure oil. We are confident in our assignment of this com- plex as Cp* 2 scCH=CHCH 3 -~, however, because its NMR and IR spectra are very similar to those of analogous zirconium and hafnium complexes.£191 In the presence of excess propene (10 equivalents), reaction 7 appears to go in two stages. Cp* sccH is con2 3 verted to a new compound before an appreciable amount of isobutane is released. This new compound has a single Cp* resonance ca. 0.03 ppm downfield of Cp* 2 sccH 3 , and two doublets at 0.78 and 0.41 ppm with coupling constants of 6 and 8 Hz, respectively. The downfield doublet is roughly three times larger than the upfield one. This spectrum is consis- tent with a scandium isobutyl complex, Cp* 2 sccH 2 CHCCH 3 > 2 , where the isobutyl methyl groups give rise to the downf ield doublet and the methylene group of the isobutyl is responsible for the upfield doublet.£201 Isolation of this interme- diate complex was not possible, since it decomposed on removal of propene and solvent. Attempts to• prepare Cp* 2 sc-i-butyl by alternate routes also met with failure. Treatment of Cp* 2 scCl with LiCH 2 CH(CH 3 > 2 leads only to decomposition of the scandium complex, and treatment of Cp* 2 ScH with isobutylene leads only to a scandium isobutenyl complex Cvide infra). In the second stage of the reaction, the isobutyl complex reacts with propene to give Cp* 2 ScCH=CHCH 3 -~ and isobutane. Propene insertion into the scandium-isobutyl bond does not occur, since the reaction of 142 the inserted product (Cp* 2 sc-2,4-dimethylpentyl) with the C-H bond of propene would give 2,4-dimethylpentane, which is not observed by 1 H NMR or gas chromatography (G.C.). The proposed reaction path is shown in equation 8. 1 [Cp* 2 ScCH 2 CH(CH 3 >2 J (bl Cp* H2 C=CH ce 3 2 ScCH=CHCH 3 -~ + ( 8) CH(CH 3 >3 Under these experimental conditions ([propeneJ/(Cp* sccH J = 2 3 0 10, 2s c> the half-life for the insertion reaction Ceq 8(a)) is~ one hour, while that of the C-H activation reaction Ceq 8(b)) is~ six hours. The scandium isobutyl complex probably reacts directly with a vinylic C-H bond. Reaction with an allylic C-H bond to give Cp* 2 scc~ 3 -cH 2 CH=CH 2 >, followed by isomerization of the allyl ligand to a propenyl ligand does not occur. When Cp* 2 scc~ 3 -cH 2 CH=CH 2 > is py- rolyzed either alone or in the presence of excess propene at ao 0 c, no change is observed in the 1 H NMR spectrum of the allyl complex. When Cp* 2 sccH 3 is treated with only a stoichiometric amount of propene (2 equivalents), the reaction is not as simple as equation 7. Rather, isobutene and Cp* 2 sccH 2 cH 2 cH 3 are observed in addition to the other products Ceq 9). 143 Cp* 2 scCH 3 + 2 H2 C=CHCH 3 ( 9) Cp*2ScCH=CHCH3-~ + Cp*2ScCH2CH2CH3 + H2 C=C(CH 3 > 2 + CH(CH 3 > 3 A similar result is obtained when the reaction of Cp* 2 ScCH 3 with excess propene is carried out at elevated temperatures cso 0 c>. A mechanism that accounts for these observations is shown in Scheme 2. 1_k___ Cp~Sc_)-- ~. Cp 2Sc-H + >= '%~1 d Scheme 2 If k 2 Cpropenel is large relative to k 3 , p-hydride elimination will be a minor pathway. With high concentrations of propene, this may be the case. If the concentration of propene is reduced, however, k 3 may become comparable to k 2 Cpropenel and P-hydride elimination would be observed. .. 144 The effect of temperature on the product distribution is also consistent with the proposed mechanism. A given in- crease in temperature generally increases the rate of a unimolecular reaction more than a bimolecular one, because of a more favorable entropy of activation for the unimolecular reaction. Thus raising the temperature of this reaction would make the P-hydride elimination reaction Cunimolecular} competitive with the reaction to give Cp* 2 ScCH=CHCH 3 -~ (bimolecular} , and isobutylene would observed. Watson has examined the reaction of Cp* 2LuCH 3 with propene.[7] Under reaction conditions similar to those used for reaction 8 ([propenel/[Cp* 2 LuCH 3 l = 10, 22°c>, rapid insertion of propene into the Lu-CH 3 bond is observed, forming Cp* 2 Lu-i-butyl. The rate of this insertion reaction is roughly 200 times faster than the analogous scandium insertion reaction Ceq 8(a}}.[2ll Cp* 2 Lu-i-butyl then inserts a second molecule of propene to give Cp* 2 Lu-2,4-dimethylpentyl, or p-hydrogen eliminates to give isobutene and Cp* 2LuH, which readily reacts with propene to If the decomposition of Cp* 2 Lu-i-butyl is monitored in the absence of propene, several lutetium complexes are observed: consisting of Cp* 2 LuR, where R = CH3, -~ -CH2CH=CH2, -~ -CH2C(CH3)=CH2, CH=C(CH3>2· 3 3 P-methyl elimination is proposed to account for the formation of Cp* 2 LuCH 3 , and labeling studies suggest the intermediacy of Cp*LLc~ 5 -c CcH > tH in the formation of the al- 5 3 4 2 145 lylic and vinylic complexes.[22] In contrast, the reaction of Cp* 2 sc-i-butyl with propene Ceq 8(b)) probably involves direct vinylic C-H activation, since a deuterium isotope effeet (kpropene/kd -propene> of greater than five was ob6 served. [23] If formation of Cp*ScC~ 5 -c 5 CcH 3 > 4 ctt 2 > were rate determining, an isotope effect would not be expected. l'f And, I ) were an intermediate in this reacCp *I Sc ( ~ 5 -c 5 CCH 3 ) 4cH 2 tion and its formation was not the rate determining step, then it should have been observed in the reactions of Cp* 2 sccH 3 with either propene or d 6 -propene. Cp* 2 ScH reacts with propene to give Cp* 2 ScCH=CHCH 3 -~ Ceq 11). (11) Cp* 2 sccH 2 cH 2 cH 3 is a likely intermediate in this reaction, since it is the only scandium species observed in the initial 1 H NMR spectrum. The scandium propyl complex has two modes of reactivity. First it reacts directly with propene to give propane and the scandium propenyl complex. In a second, more rapid reaction, a new compound is formed that has a Cp* resonance shifted slightly downf ield of Cp* 2 ScCH 2 cH 2 CH 3 and a doublet at 0.64 ppm with a coupling constant of 6 Hz. The coupling constant and chemical shift of this doublet are very similar to those observed for the proposed Cp* sc-i-butyl isobutyl methyl groups. 2 The most 146 likely candidate for this new compound is Cp* 2 sc-2-methylpentyl , formed by insertion of propene into the scandium-propyl bond. This new compound slowly reacts with propene to give Cp* 2 scCH=CHCH 3 -~ and 2-methylpentane c1 H NMR). A mechanism consistent with these observations is shown in Scheme 3. Cp~ScH + J Scheme 3 G. c. analysis of the volatiles from this reaction show 2- methylpentane and a small amount <<1%) of higher molecular weight alkanes were formed, indicating that insertion of propene into Cp* 2sc-2-methylpentyl is very slow relative to vinylic C-H activation. 147 Neither Cp* 2 sccH 2 c 6 H5 nor Cp* 2 scc 6 H5 react with propene, except at high temperatures, where the products of their reactions decompose. I 5 I Isolated Cp*Sc(~ -c 5 CcH 3 >4cH 2 > Cdimeric or oligomeric, vide supra) reacts with propene to give a complex mixture of products, but no Cp* 2 ScCH=CHCH 3 -~ is formed. Cp* 2 ScH reacts with trans-butene to give a single scandium species. The expected product was Cp* 2 scCHCCH 3 >cH 2CH 3 , however, treatment of this complex with o2o yields CH 3cH 2 cH 2 cH 2 D, suggesting that the scandium species is actually Cp* 2 sccH 2 cH 2 cH 2 cH 3 • This type of alkyl ligand rear- rangement is common for early transition metal complexes. [241 Cp* 2 scCH 3 reacts cleanly with isobutylene to give methane and a scandium isobutenyl complex (eq 12). Cp * 2 scCH=C(CH 3 >2 CH 4 + (12) Neither a scandium neopentyl complex or neopentane are observed, indicating that isobutylene does not insert into the scandium-methyl bond. pure pale yellow solid. Cp * 2 scCH=CCCH 3 >2 was isolated as a This reaction is analogous to the reaction of Cp* 2 sccH 3 with styrenes Cvide supra), but the rate of this reaction is roughly 50 times slower, possibly due to the greater steric bulk of isobutylene. Cp* 2 ScH re- 148 acts with isobutylene to give the scandium isobutenyl complex and isobutane (eq 12).[25] Cp* 2 ScCH=C(CH 3 > 2 + HC(CH 3 > 2 (12) The reactions of Cp* 2 ScR complexes CR= H, CH 3 , CH 2 c 6H5 > with olefins consist of insertion and C-H activation reactions. In virtually every mechanism for olefin insertion into transition metal-alkyl or -hydride bonds the olefin donates electron density form its rr cloud into a vacant orbital on the metal center prior to insertion. [2, 261 We believe that this is the case for scandium as well, and the ability of the olefin's rr cloud to approach the metal center is strongly dependent on the steric bulk of both the olefin and the ligand on scandium. Similar arguments have been made concerning olefin polymerization by Ziegler-Natta catalysts.[l(a)] Ethylene readily inserts into Cp* 2 sc-H, Cp* 2 sc-CH 3 and Cp* 2 Sc-CH 2 c 6 H5 bonds, but is unreactive toward Cp* 2 sc-c 6H5 bonds. The phenyl ligand presumably blocks the active site from reacting with the ethylene rr cloud. Unfavorable steric effects are probably responsible for propene not inserting into Cp* 2 sc-i-butyl and Cp* 2 sc-2-methylpentyl bonds, while it inserts readily into the less sterically demanding propyl ligand of Cp* 2 sc-cH 2 cH 2 cH 3 • A p-disubstituted alkyl ligand appears to be too bulky to allow interaction with propene's rr cloud 149 (leading to insertion) in this system. In support of this, Cp* 2 ScCH 2 c 6 n5 is unreactive toward propene. The steric bulk of the olefin controlling reactivity is seen in the reactions of Cp* 2 sccH 3 with propene and isobutylene. Propene inserts into the Cp* 2 sc-CH 3 bond, whereas isobutylene does not. The observation of facile vinylic C-H bond activation by scandium alkyl complexes suggests an alternate chain termination pathway in the polymerization of ethylene by these complexes. Because of the significant number of vinyl end groups in the isolated polyethylene, we proposed P-hydride elimination was the primary chain termination pathway. Ac- tivation of a vinylic C-H bond of ethylene by the scandium-polymer complex, however, would also lead to chain termination. The catalyst product of this reaction would be a scandium vinyl complex Ceq 12). Cp*2ScCH2(CH2)nCH3 + H2C=CH2 (12) Cp*2ScCH=CH2 + CH3CCH2)nCH3 The newly formed vinyl complex could then act as a polymerization catalyst, giving a polymer with a vinyl end group. Ziegler-Natta catalysts are very electrophilic and may activate vinylic C-H bonds in a similar manner. Although this termination pathway is applicable to vinyl end groups, it can not form vinylidene end groups, which are often observed in oligomers formed by Ziegler-Natta catalysts. Successive 150 insertion and p-hydride elimination steps are generally used to explain vinylidene end groups.£271 The differences in reactivity of Cp* 2LuR versus Cp* 2 ScR CR= ce 3 , -i-butyl) toward olefins are probably caused by steric effects. Lu has a 0.10 Alarger ionic radius than Sc, making lutetium's active site more open and thereby decreasing steric interactions. Similar steric effects were proposed to explain the rate difference for the reactions of Cp* 2 scce 3 and Cp* 2 Luce 3 with methane (yide supra). Thus propene reacts 200 times faster with Cp* 2 Luce 3 than with Cp* 2 scce 3 , because the rr cloud of the olefin can interact with the Lu atom of Cp* 2 Luce 3 more readily than it can with the Sc atom of Cp* 2 scce 3 • Unfavorable steric interactions probably prevent the insertion of propene into Cp* 2 sc-cH 2 CRCH 3 CR = methyl, propyl) bonds, while propene inserts readily into Cp* 2 Lu-i-butyl bonds. The differing ionic radii of Sc and Lu may also be responsible for the apparent difference in the mechanisms of their C-H activation reactions. Cp* 2 sc-i-butyl appears to react directly with the C-H bond being activated, while Cp*~uc~ 5 -c 5 cce 3 > 4 he 2 > is the postulated intermediate for the reaction of Cp* 2Lu-i-butyl with C-H bonds.£221 The smaller ionic radius of Sc should shorten the M-ce 2 bond CM = Sc, I 5 I Lu) of Cp*Sc(~ -c 5 cce 3 >4ce 2 > relative to I 5 I Cp*Lu(~ -c cce > ce >.£281 Shortening the M-ce 2 bond should 5 3 4 2 151 pull the methylene further below the plane of the Cp* lig5 1 I and, giving Cp*Sc(~ -c 5 CcH 3 > 4cH 2 > greater strain energy than I 5 I Cp*Lu(~ -c CcH > cH >. Thus, while the lowest energy path5 3 4 2 way for C-H activation by Cp* 2 Lu-i-butyl is via I 5 I 5 Cp*Lu(~ Cp*Sc(~ I -c 5 CcH 3 > 4cH 2 >, the greater strain energy of t -c 5 CcH 3 > 4cH 2 > may make this pathway higher in energy than direct C-H activation by Cp* 2 sc-i-butyl. 152 EXPERIMENTAL ~~al Consid~tions. All manipulations were carried out using either high vacuum line or glove box techniques. from All solvents were purified by distillation Na or CaH 2 , titanocene. followed by vacuum transfer from Isobutylene was dried with sodium and stored in a thick walled glass bomb over Na/benzophenone. All other gaseous reagents were freeze-pump-thawed twice and distilled from a -30°c trap just prior to use. All liquid 0 reagents were dried over 4 A molecular sieves, and freezepump-thawed twice before use. and d 12 -cyclohexane, NMR solvents, d 6 -benzene were vacuum transferred from activated 4 A molecular sieves, then from titanocene. 1 H NMR spectra were run on Varian EM-390 (C.W., 90 MHz) and JEOL FX-400Q (F.T., 400 MHz) spectrometers. NMR spectra were run on a JEOL FX-90Q (F.T., spectrometer. Table, 2H 13.8 MHz) All spectra of new compounds are listed in Inf rared (IR) spectra of polyethylene were run as KBr pellets and other samples were run as Nujol mulls on KBr plates. IR spectra were recorded on a Beckman IR- 2420 spectrometer. Analyses were performed by the Caltech Analytical Laboratory. Many of the reactions in this chapter were carried out in NMR tubes. Any experiment discussed in this chapter and not listed below was carried out in a sealed NMR tube with .QJt. 20 mg of the scandium complex in .Q.£. b c + d a Cp* a c b c d 2 ScCH=CHCH3-~ b Cp* 2 ScCH=C(CH 3 >2 a c d a b c c a b b+c+d) a Assignment s s s s 1.90 6.12 2.20 1.57 5.25 d q 2.10 d d 1.90 s 7.0 d q s s s s 1.85 2. 89 1.53 { 1.13 Uw.m1 1 H and 13 c NMR spectroscopic data.* Cp* 2 ScC(CH )=C(CH 3 > 2 3 Compound Table. Jed = 5.4 Jbc = 19.5; Jbd = 1.5 Coupling 500 MHz Comments I-" \JI VJ \ b c d 2 Hf(CH=CHCH 3 -~> 2 b c d e a b c d e c + d + e a b c d a b c d a b Assignment 119.29, 11.42 43.43 34.47 30.70 14.49 7.4 Jed = 5.5 5.24 d q 1.88 d d 1.88 s 0.90 t 0.56-0.61 Jbc = 18; Jbd = 1.5 Jed = 5.5 5.21 d q l.9o d d 1.79 s 5.56 d q Jbc = 18; Jbd = 1.5 Coupling 1.76 s 5.83 d q (ppm) 13C {lH} 400 MHz 400 MHz ref. 19 ref. 19 Comments * Unless otherwise specified, spectra were taken in c 6 n 6 at ambient temperature at 90 Mhz. The chemical shifts are reported in relativ~ to TMS or residual protons of the solvent. Coupling constants are reported in Hz. a Cp* 2 ScCH 2 CH 2 cH 2 CH 3 a Cp* c d a b 2 Zr(CH=CHCH 3 -~> 2 Cp* Comoound +:- t--> \.)\ 155 0.35 mL of d 12 -cyclohexane. The spectra were referenced to internal TMS or the protio impurity of the solvent cc 6 n 11 H). A measured amount of each of the appropriate reagents was added to the tube before it was cooled to +196°c and sealed with a torch. The molecular weight and polydispersities reported in this chapter were measured, using gel permeation chromatography, by Dr. Howard Turner at Exxon Chemical Co., Baytown, Texas. Preparation of Cp* 2 ScC(CH 3 )=C(CH 3 1 2 ~ A solution of 218 mg. of Cp* 2 ScCH 3 3 (0.66 mmol) in 5 mL of petroleum ether was treated with .Qa. 2 mL of 2-butyne. The reaction mixture was stirred for 40 minutes at room temperature and the volatiles removed .i..n. yacuo. (2 mL) Fresh petroleum ether was added to the resulting solid and the clear solution cooled to -78°c. A pale yellow solid product precipitated and was collected by filtration 53%). IR data (cm- 1 ): 1260 (w), 1200 (w), 10 2 0 ( s) , 8 0 0 ( w) , calc. for 2710 (w), 2670 (w), 1610 (m), 1150 (w), 1090 (w), 1060 (w), 1160 (w), 7 8 0 ( w) , 7 4 5 ( w) , 7 2 0 ( m) , c 25 H39 sc: C, (135 mg, 78.09; H, 6 6 0 ( m) • 10.22. Found: An a 1. C, 77.78; H, 10.23. Preparation of Cp* 2 ScCH=C(CHJ1 2 .!.. Cp* 2 ScCH 3 (275 mg., 0.83 mmol) was dissolved in 5 mL of petroleum ether in a viscous walled glass bomb. Approximately one mL of isobutylene was then condensed into the bomb, and it was 156 heated to 80°c for 20 hours. After cooling the bomb to room temperature, the isobutylene was allowed to boil off. After concentrating the solution to 2 mL, it was cooled to -78°c. The white solid product precipitated and isolated by filtration (230 mg, 2715(w), 2695(w), 1584(s), 1080(w), 1060(w), 1022(w), 790(s), IR data (cm- 1 ): 75%). 1498(w), 1260(w), 615(w), 4 4 7 ( m) , 4 2 0 ( s) . An a 1. ca 1 c. for C 2 4 H3 7 Sc: 10.07. C, 77.49; H, 10.01. Found: Dimerization of Propyne. was A sealed 1114(w), 589(w), 500(m), C, 7 7 • 8 0; H, NMR tube was prepared with 26 mg of Cp* 2 ScCH 3 (0.079 mmol), 0.4 mL of c 6o 6 and 0.55 mmol of propyne (7 equivalents). Within 30 minutes at room temperature, all of the propyne had been consumed; methane spectrum. 2-methyl-l-pentene-3-yne, were Cp* 2 ScC CCH 3 the only compounds observed and in the NMR The 2-methyl-l-pentene-3-yne was identified by comparison of its NMR to that of an authentic sample. [ 10] Cp* 2 ScC CCH 3 is most likely the catalyst for this dimerization because the reaction of Cp* 2 ScCH 3 is very fast at room temperature and treatment of Cp*2ScCH3 with one equivalent of HC CCH 3 leads cleanly and quantitatively to Cp* 2 scC CCH 3 • Polymerization of Ethylene. The procedure used for polymerization of ethylene by Cp* 2 ScCH 3 is described here, but Cp* 2 sc(THF)H or Cp* 2 sccH 2 c 6 H5 can be substituted for 157 Cp*2ScCH 3 and the same results are obtained. A solution of 35 mg of Cp* 2 ScCH 3 (0.11 mmol) in 10 mL of low boiling petroleum ether was treated ethylene in ~. 500 mL. with one atmosphere of The pressure was maintained close to one atmosphere ( 0.1 atm) until ethylene uptake ceased (generally 2-3 hours). The excess ethylene was removed .in vacuo and the cloudy to thick suspension was treated with 50-100 equivalents of HCl (gas). After 15 minutes the reaction vessel was opened to the air and the slurry filtered. The resulting polymer was washed several times with petroleum ether, once with acetone, once with water and finally with acetone. The polymer was dried in vacuo. The was yield of polymer 0.30 g. Gel permeation chromatography measurements gave a molecular weight of 1700 ± 300 g/mol and a polydispersity of 1.7 t 0.3. The yield of polymer chains per scandium atom was 1.7 ± 0.2. The melting point of this polymer was 118-120°c. IR data (cm- 1 ): 3080 (w), 2920 (s), 2850 (s), 2660 (w), 2638 (w), 1644(m), 1485(s), 1464(s), 1378(w), 990(m), 910(m), 73l(s), 720(s). The IR spectrum is shown in Fig. 1 and the assignment of its bands is given in the text. The relative intensities of the IR bands of CH 3 groups (1378 cm- 1 ) and vinyl groups (910 cm- 1 ) are 1:4.3, respectively, methyl from which we calculate a groups per vinyl group. [14] ratio of 1.7 We were unable to obtain a 13 c NMR spectrum with an adequate signal to noise level to allow the calculation of the number of methyl or 158 vinyl groups per polymer chain (degree of branching). Both methyl and vinyl groups were observed, however, and their resonances were very small compared to the (CH 2 )n signal, but accurate integration was not possible. The Reaction of Cp* 2 ScCH 3 with Propene (Eg.(8)). A sealed NMR tube containing 20 mg of Cp* 2 ScCH 3 (0.06 mmol), 0.4 mL of c 6 n 12 and 0.64 mmol of propene was prepared. The reaction was monitored by 1 H NMR (90 MHz). The reaction is described in the text and the NMR spectra of Table 1. The attempted synthesis of Cp* 2 ScCH=CHCH 3 -!,_ (similar work up to Cp* 2 scCH=C(CH 3 ) 2 ) at 2s 0 c and 80°c both yield only an impure oil (80-90% product). The IR spectrum of the oil was measured 1565(m), 795 (w), 1530(w), 680 (m), 1250(w), 660 (s), 1150(w), 435 (vs). (cm- 1 ): 2660(w), 1020(s), 985(s), The NMR spectrum of Cp* 2 scCH=CHCH 3 -t is very similar to those reported for the analogous zirconium and hafnium complexes Cp* 2 M(CH=CHCH 3 t)2 (M=Zr, H). The NMR spectra of the zirconium and hafnium complexes are listed along with Cp* 2 ScCH=CHCH3-t in Table 1, for comparison. The Reaction of Cp* 2 ScH with Propene (Eg. (11)). A sealable NMR tube was charged with 20 mg of Cp* 2 ScCH 3 (0.06 mmol) and 0.4 mL of c 6 n 1 2. The 14/20 joint connecting the NMR tube to the stopcock assembly was then 159 secured with several rubber bands. The tube was cooled to -196°c and one atmosphere of H2 admitted to the tube. The stopcock was closed and the tube was warmed to room temperature. The NMR tube assembly was removed from the vacuum line and the solution agitated for five minutes, after which the solution was cooled to -196°c. The H2 was evacuated and replaced with 0.72 mmol of propene. The tube was then sealed with a torch and warmed to room temperature. The initial NMR spectrum showed CP* 2 sccH 2 cH 2 cH 3 as the only scandium species. The intermediates and products of the reaction are described in the text. The volatiles from this reaction were collected by breaking the NMR tube under vacuum and vacuum transferring the contents of the tube to a flask. Heat was applied to the apparatus holding the NMR tube and the vacuum transfer carried out for ~. one hour to ensure all the volatiles would transfer. The React ion of Cp* 2 scH with Trans-Butene. thick walled glass bomb was charged with A 40-mL 330 mg Cp* 2 sccH 3 (1.0 mmol) and 10 mL of methylcyclohexane. of The entire bomb was cooled to -196°c and one atmosphere of H2 admitted. The bomb was warmed to room temperature and the solution stirred for 10 minutes. The bomb was cooled to Trans-Butene (1.08 mmol) was admitted to the bomb at -78°C and the bomb allowed to come to room temperature. The 160 methylcyclohexane was removed and replaced with 2 mL of petroleum ether. Cooling this solution to precipitated a bright yellow solid (115 mg, 30%). -78°c We were unable to get a C,H analysis of this complex, presumably due to its thermal instability. When 14 mg of this complex (0.036 mmol) was treated with 8 "{ L of n 2 o, only CH 3 cH 2 cH 2 cH 2D was observed, suggesting that the complex is The 400 MHz NMR spectrum of CP* 2 ScCH 2 cH 2 cH 2 cH 3 (C 6n 12 ) shows a single Cp* resonance at 1.88 ppm, a triplet at 0.90 ppm with a coupling constant of 7.4 Hz and a broad multiplet 0.56-0.61 ppm. We assign the triplet to the a-CH 2 group and the broad multiplet to the remaining protons of the butyl ligand. spectrum was obtained for Cp* 2 sccH 2 cH 2 cH 3 • A similar A 13 c spectrum is observed that is consistent with Cp* 2 scCH 2 CH 2 CH2CH3. The Reaction of Cp* 2 ScH with Isobutylene (Eg. (12)). A c 6 n 12 solution of Cp* 2 ScH was prepared as previously described and 1.2 equivalents of isobutylene added to it. After 30 minutes at room temperature, Cp* 2 ScH and Cp* 2 ScCH=C(CH 3 ) 2 were in equal concentration ( 1 H NMR). Cp* 2 ScCH 2 CH(CH 3 ) 2 is not observed and the concentration of isobutane parallels the concentration of Cp*2ScCH=C(CH3)2· At the end of the isobutane are reaction only Cp* 2 ScCH=C(CH3)2 and observed. One explanation for these results involves C-H activation by Cp* 2 ScH, rather than by 161 REFERENCES 1. (a) G. Henrici-Olive, s. Olive Chemtech (1981) 746-752. (b) P.T. Walczanski, J.E. Bercaw Accts. Chem. Res. (1980) .2..4_, 121. (c) Chapter two of this thesis. 2. G.w. Parshall "Homogeneous Catalysis"; John Wiley and Sons: New York, (1980). J.P. Collman, L.S. Hegedus "Principles and Applications of Organotransition Metal Chemistry"; University Science Books: Mill Valley, Calif., (1980). R.H. Grubbs "Comprehensive Organometallic Chemistry"; G. Wilkenson, F.G.A. Stone, E.W. Abel, Eds., Pergamon Press: London, (1982), volume 8, chapter 54. 3. K. Ziegler, H.G. Gelert, E. Holzkamp, J. Schneider, M. Sol, W.R. Kroll Angew. Chem. (1960) ~, 121. 4. H. Sinn, W. Kaminsky Adv. Organomet. Chem. (1980) .l..8., 99-149, and references therein. 5. J.c.w. Chien J. Amer. Chem. Soc (1959) ll., 86. D.S. Breslow, N.R. Newburg J. Amer. Chem. Soc (1959) ll., 81. C.E.H. Bawn, R. Symcox J. Polym. Sci. (1959) 14.,139. u. Gianinni, u Zuchini, E. Albizzatti Polm. Lett. (1970) ~, 405. 6. G. Henrici-Olive', s. Olive' Angew. Chem. Int. Ed. Eng. (1971) 2, 105-115. K.H. Reichert, K.R. Meyer Macromol. Chem. <1973) ~' 163. G. Henrici-Olive, s. Olive Angew. Chem. Int. Ed. Eng. (1967) .Q., 790. 7. P.L. Watson J. Arner. Chem. Soc (1982) .l.QA, 337-339, 8. P.L. Watson, G.W. Watson Accts. Chem. Res. (1985) 1.8., 51-56. 9. G.A. Moser "Coordination Polymerization"; c.c. Price, E.J. Vandenberg Eds.,Plenum Press: New york, (1983), pages 193-206. 10. R.S. Threlkel, Ph.D. dissertation; California Institute of Technology, Pasadena, CA (1980). 11. Bulky alkyl groups generally have lower bond strengths than non-sterically demanding ones. J, Halpern Accts. Chem. ~ (1982) ~, 238-244, and references therein. J.W. Bruno, T.J. Marks, L.R. Morss J. Arner. Chem. Soc (1983) .l.Q.S., 6824-6832. 162 12. s. Krirnrn, C.Y. Liang, G.B.B.M. Sutherland J. Chern. Phys. 0956) 2..5., 549-562. 13. ~ L.H. Cross, R.B Richards, H.A. Willis Discuss. Faraday (1950) ~' 235-245. 14. J.C. Woodbrey, P. Ehrlich J. Arner. Chern. Soc (1963) ..6..5., 1580-1584. 15. The polydispersity of a polymer describes how wide its molecular weight distribution is on either side of its average molecular weight. A narrow distribution , in which a high percentage of the polymer chains have nearly the same molecular weight, gives a low polydispersity Crnonodisperse M /M = 1.0). A broad distribution, in which only a few p~rcgnt of the total polymer chains may have the average molecular weight, gives a high polydispersity. 16. L.H. Peebles "Molecular Weight Distributions in Polymers"; Interscience: New York, (1971). M. Szwarc~ Polyrn. Sci. 0 983) .4.9_, 81-86. 17. R.M. Silverstein, G.C. Bassler, T.C. Morill "Spectrometric Identification of Organic Compounds"; John Wiley and Sons: New York, (1981). 18. H. Rudolf, w. Trautvetter, K. Weirauch, "Chernische Technologie" CK. Winnacker and L. Kuchler, eds.), Vol. 5, p. 6 0. Carl Hanser, Munich, 0 97 2) • 19. C. McDade. J.E. Bercaw J. Orgrnet. Chern. 0985) 21..!l, 281-315. 20. The isoelectronic lutetium complex, Cp* 2 LuCH 2CH(CH 1 > , gives coupling constants (J ) of 6.7 and 8.2 Hz ror the 2 methylene and methyl double~~' respectively, of the isobutyl ligand, reference 6. 21. The half-life for formation of Cp* 2 Lu-i-butyl is 0.2 hours for [propenel/[Cp* 2 LuCH 3 J = 1, reference 21. 22. P.L. Watson, D.C. Roe J. Amer. Chern. Soc (1982) .l..QA, 6471-6473. 23. Cp* ScCH reacts readily with d -propene to give Cp* 2 sccH 22CHCC~ 3 > CCD 3 ), at the same rRte as normal propene. The next reaction Cc-D activation as in eq 8(b)) is very slow, and P-deuteride elimination competes, so only an approximate rate of Cp* 2 ScCD=CDCD 3 -~ formation can be determined. P-deuteride elimination appears to be slower than P-hydride elimination from the scandium isobutyl complex, but we have been unable to ·quantify this. 163 24. J. Schwartz, J.A. Labinger Angew, Chem, Int. Ed. {1976) D.B. Carr, M. Yoshifuji, L.I. Shoer, K.I. Gell, J. Schwartz Annals. N. Y. Acad. Sci- 0977) m_, 127. .12, 333. 25. Cp* Sc-i-butyl is not observed during the course of this rea6tion. One explanation for this is that the formation of isobutane is not as simple as insertion followed by C-H activation, analogous to equation 8. Cp* ScH and isobutylene are in equilibrium with Cp* sc-i-~utyl {eq 13), but a second equilibrium is possible {eq2 14), analogous to the equilibrium between Cp* 2 ScH and Cp* 2 ScC 6H5 in c 6 H6 solution Cvide supra). In the present case, th~ H generated by reaction 14 would be expected to react with Cp~ 2 sc-i-butyl in a fast reaction. Thus Cp* 2 sc-i-butyl would not be observed. Cp* 2 ScH + Cp* 2 ScH + - H C=C{CH ) H2 C=C{CH 3 ) 2 ~ ~ 03) Cp* 2 ScCH 2CH{CH 3 ) 2 fast 3 2 2 ~ ~ Cp* 2 ScCH=C{CH 3 ) 2 + H2 04) HC{CH 3 ) 2 05) fast Cp* 2 ScCH 2 CH{CH 3 ) 2 + H2 Cp* 2 ScH + 26. P. Cossee J. Catal. 0964) ,l, 80. E.J. Arlman, P. Cossee ibid. 0964) ,l, 99. K.J. Ivin, J.J. Rooney, C.D. Stewart, M.L.H. Green, R. Mahtab J. Chem. Soc., Chem. ComID..Y.n.a. {1978) 604. R.J. McKinney J. Chern. Soc., Chem. Commun. 0980) 491. 27. G. Henrici-Olive', s. Olive' Angew. Chem. Int. Ed. 0971) l.Q., 105-115. 28. The crystal structure of Cp* LuCC,N-~ 2 -c 5 H N) has been determined. The pyridyl ligand i~ disord2red iA a manner similar to that observed for Cp* 2 Sc{C,N-~ -C H4 N) {~ supra), such that the metal bouna carbon and 5 n!trogen are indistinguishable. The two averaged Sc-C and Sc-N bond lengths are 2.162(7) and 2.183(8) A, while those of the 0 isostructural lutetium complex are 2.274(6) and 2.270(6) A. The difference in Sc versus Lu bond lengths ~lmost exactly match their difference in ionic radii C0.10 A).