Part I. Coordination Chemistry of High-Valent Osmium Polyanionic Chelating Ligand Complexes with Imine, Imido and Nitrido Ligands. Part II. Asymmetric Hydrogenation Using a Resolved Chiral Scandium Hydride Complex Citation Barner, Claudia Jane (1990) Part I. Coordination Chemistry of High-Valent Osmium Polyanionic Chelating Ligand Complexes with Imine, Imido and Nitrido Ligands. Part II. Asymmetric Hydrogenation Using a Resolved Chiral Scandium Hydride Complex. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/tj33-3482. https://resolver.caltech.edu/CaltechETD:etd-06112007-095040 Abstract Part I: The chemistry of high-valent osmium polyanionic chelating ligand complexes with imine, imido and nitrido ligands is explored. The synthesis and x-ray diffraction study of Os( η ⁴-HBA-B)(NC₆H₅NHC₆H₄) is reported. A neutral, octahedral Os(VIII) intermediate complex is postulated. The synthesis and x-ray diffraction study of [OsN( η ⁴-HBA-B)]- is reported. The mechanism of formation is discussed. These complexes represent two of the first examples of non-planar amide ligands in transition metal chemistry. The synthesis of a five-coordinate imido complex, Os(N-C₁₀ H₁₆), is reported. The formation of the first osmium-osmium quadruply-bonded complex, [Os( η ⁴-HBA-B)]₂, is reported. Reactions with the nitrido complex, [OsN( η ⁴-HBA-B)]⁻,and the methylimido complex, OsNMe( η ⁴-HBA-B), are also reported. Part II: The asymmetric hydrogenation of the prochiral substrate, methylene indane, is induced by the resolved chiral scandium complex, Me₂Si( η ⁵-C₅Me₄)( η ⁵-(C₅H₃(1R,2R,5S)-menthyl)ScH, in substoichiometric quantities. Methylene indane is hydrogenated to produce methylindane with an enantiomeric excess of 34% and a maximum turnover of four. 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: Grubbs, Robert H. (chair) Bercaw, John E. Gray, Harry B. McKoy, Basil Vincent Defense Date: 25 September 1989 Funders: Funding Agency Grant Number ARCS Foundation UNSPECIFIED Caltech UNSPECIFIED W. R. Grace Fellowship UNSPECIFIED Record Number: CaltechETD:etd-06112007-095040 Persistent URL: https://resolver.caltech.edu/CaltechETD:etd-06112007-095040 DOI: 10.7907/tj33-3482 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 2546 Collection: CaltechTHESIS Deposited By: Imported from ETD-db Deposited On: 28 Jun 2007 Last Modified: 18 Jan 2022 22:18 Thesis Files PDF - Final Version See Usage Policy. 2MB Repository Staff Only: item control page

PART I: COORDINATION CHEMISTRY OF HIGH-VALENT OSMIUM POLYANIONIC CHELATING LIGAND COMPLEXES WITH IMINE, IMIDO AND NITRIDO LIGANDS PART II: ASYMMETRIC HYDROGENATION USING A RESOLVED CHIRAL SCANDIUM HYDRIDE COMPLEX Thesis by Claudia Jane Barner Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1990 (Submitted September 25, 1989) ii To My Dad, Who never thought he'd live to see this day, and sadly, didn't. And To My Mom, Who never gave up on me. iii Acknowledgements I would like to thank my present advisor, John Bercaw, for all the good advice and patience he has shown me (and for letting me bring Willy to school). He is truly a god. I would also like to thank my former advisor, Terry Collins, for giving me the chance to come to Caltech and for support during the time I was in his group. Thanks also to Harry for being a great and knowledgable friend to me and to Remo. I am very conscious of and grateful for all the help I have received from so many people. The Bercaw group has been quite an experience. Thanks to Van for, well. ... , all of it (the stuff and the thing) and Rocco for lots o' conversation and being a real friend (even though I had the dubious honor of taking over Rocco's line and bench), Leroy (a truly awesome drinking buddy) and Donnie (Dooonnnnieeee) and Marty, for being normal and Pam (good suggestions and good dancing) and Ged (for entertainment) and Al Ray (for raising safety glasses to a fashion statement) and the Trim, for blessed silence, and EOP, for great files, and David (for pep talks and back cracks) and Eric (for weird talks) and Bryan and Roger (for being such good Joes) and and and everyone for everything. And, of course, to the best "suckbuddies" around, namely, Emilio (he made me feel so guilty), Andy (every good line needs a slimy Brit) and Warren (smoke?). Thanks especially to Warren for saving me by doing most of the figures in Part I. Thanks to all who proofed my "vomit- versions." Thanks to everyone who helped with spectroscopy: Scott, Yogi, Pete, Brew, Bob, Rich, Steve, Tracy and Janet. And, of course, I want to thank all the people who really run this institute, like Pat Anderson (an amazingly fashionable and patient lady), Pat Bullard, Virginia, John, Chris, Beth, Sandra, Fran, Emily, Margaret (thanks for Willy!), Bill, Don and Dan, Harvey (you always had a smile, Harvey), Jeannete, Carol and Mary in the Graduate Office. iv I am deeply grateful for the funding I have received at Caltech from ARCS Fellowship Foundation, Dean's Fellowship, W.R. Grace Fellowship, Perkins Loan Program, The Division of Chemistry and Chemical Engineering at Caltech and Mom. I also want to thank all my friends who have made my life rich in groups like World Work, the Environmental Study Group, the Moon Lodge and 0.O.M.E .. Finally, I want to thank my family for all the love I have been given for so many years. Thanks, Mom, for listening to me, supporting me and believing in me. And thanks to Remo, who went to so many Tuesday Nighters and seminars and did so much homework in the Kiwi office and waited so many times for me to "fill my traps" and thought that was the way most kids lived. And thanks to my best friend, Scott, for calming me down, fixing me dinners reading me stories, feeding me calcium, and who has made a big difference. And lastly, thanks to Harry, Sunny, Bump, Red, Boo, Sophie, Micky, Willy and Benny, just for being there. V Abstracts Part I: The chemistry of high-valent osmium polyanionic chelating ligand complexes with imine, imido and nitrido ligands is explored. The synthesis and x-ray diffraction study of Os(1'7 4 -HBAB)(NC5H5NHC5H4) is reported. A neutral, octahedral Os(VIII) intermediate complex is postulated. The synthesis and x-ray diffraction study of [OsN(1') 4-HBA-B)r is reported. The mechanism of formation is discussed. These complexes represent two of the first examples of non-planar amide ligands in transition metal chemistry. The synthesis of a five-coordinate imido complex, Os(N-C10H15), is reported. The formation of the first osmium-osmium quadruplybonded complex, [Os(17 4-HBA-B)]2, is reported. Reactions with the nitrido complex, [OsN(1'7 4HBA-B)r, and the methylimido complex, OsNMe(11 4-HBA-B), are also reported. Part II: The asymmetric hydrogenation of the prochiral substrate, methylene indane, is induced by the resolved chiral scandium complex, Me2Si(17 5-C5Me4)(1') 5-(C5H3(1 R,2R,5S)-menthyl)ScH, in substoichiometric quantities. Methylene indane is hydrogenated to produce methylindane with an enantiomeric excess of 34% and a maximum turnover of four. vi Table of Contents Page Acknowledgements iii Abstracts iv Table of Contents vi Part I: Coordination Chemistry of High-Valent Osmium Polyanionic Chelating Ligand Complexes With lmine, lmido and Nitrido Ligands Chapter One: Introduction to Oxidation Chemistry and Ligand Design References 12 Chapter Two: Section 1-- Reactions of 0S(TJ 4-HBA-B)(PPh3)2 To Form Imine, lmido and Nitrido Complexes 14 Chapter Two: Section 11--Reactions of Osmium Nitrido and Methylimido Complexes 56 Experimental 72 Table 2.7: 1 H NMR Data 81 References 84 Part II: Asymmetric Hydrogenation Using a Resolved Chiral Scandium Complex Chapter One: Introduction 88 References 95 Chapter Two: Asymmetric Hydrogenation With Me2Si (TJ 5-C5Me4) (TJ 5-C5H3(1 R,2A,5S)-menthyl) ScH 96 Experimental 119 Table 2.1: 1 H NMR Data 123 References 125 1 PART I: COORDINATION CHEMISTRY OF HIGH-VALENT OSMIUM POLYANIONIC CHELATING LIGAND COMPLEXES WITH IMINE, IMIDO AND NITRIDO LIGANDS CHAPTER ONE: INTRODUCTION TO OXIDATION CHEMISTRY AND LIGAND DESIGN 2 Oxidation chemistry is currently an important area of research because of its main role in many industrial processes, including such applications as the Wacker process in which ethylene is oxidized to acetaldehyde 1 , the Mid-Century process 2 for the production of terephthalic acid and the Oxirane process3 for the epoxidation of olefins, among others. In addition, oxidation of hydrocarbons from petroleum feedstocks is an important area of research in the production of commodity chemicals and fuels. 4 Oxidative processes also play a role in drug design and in many biological systems. 5 Different approaches have been taken towards developing more efficient oxidizing capabilities such as enzymatic oxidations 6 , reactions with molecular oxygen 7, the use of non-metal oxidants such as peroxides 8 and oxidizing transition metal complexes. 9 Even though metals, and in particular, transition metals, play a crucial role in oxidation chemistry, this area has received relatively little attention from the inorganic research community. The chemistry of low-valent transition metals is much more developed than the chemistry of high-valent complexes. The main goals of oxidation chemistry are to produce better oxidizing agents that perform selective transformations, characterize completely the active oxidizing agent and elucidate the operative mechanisms behind these reactions. Associated with these objectives are several critical problems. Catalytic oxidation processes tend to rely on metal-based heterogeneous systems which are extremely hard to study, and the catalysts tend to be poorly characterized. 10 In addition, most of the oxidation catalysts employed are not selective. It is frequently difficult to control the oxidation resulting in either over-oxidation of the substrate or a mixture of products which then have to be separated. These kinds of problems are perfectly illustrated in the oxidations using Cr(VI) and Mn(VII) as the oxidants that are among the most common reagents employed in organic chemistry. 11 ·12 The major source of the non-selectivity is that the metals perform the first oxidation and are then themselves reduced, but the reduced metal species can also act as an oxidation catalyst. Oxidations with Cr(VI) and Mn(VII) yield an overall reduction of three- or four-electrons at the metal center, whereas an organic substrate is 3 usually oxidized by only two electrons. Ultimately, a chain of oxidations occurs with a number of oxidized products resulting, as illustrated in Equation 1.1. (CH ) CHOH 3 2 + Cr(VI) (CH ) C:O 32 1rr-• ·; CHOH + 2Cr(IV) 2H+ + 2Cr(VI) , 2(CH ) COH + 2Cr(UI) + 32 2(CH ) C::O + 2Cr(V) . + 32 2H+ 2Cr(V) ~ 2\-\ • (1.1) 2(CH ) COH 3 2 -1- . Cr(IV) + Cr(IV) + Cr(VI) Although this particular reaction is selective for the conversion of isopropanol to acetone, any other substrate produces a mixture of products. 11 The Collins group believes that one of the main reasons for the lack of selective, wellcharacterized oxidizing agents is the limited nature of ligands available that will stabilize transition metals in high oxidation states. The scope of ligands capable of stabilizing high oxidation state metal complexes has traditionally been limited to oxo, nitrido and halide ligands. 13 More recently, ligands such as porphyrins 14 and salen-type 15 ligands have been developed. 11 While work in these areas is important, it is necessary to develop new types of ligands that can stabilize high oxidation state metal complexes, designed to undergo only twoelectron reductions and serve as selective oxidizing agents. The Collins group approaches this problem by designing and synthesizing Polyanionic Chelating (PAC) ligands that will allow the development of high oxidation state inorganic complexes for use as oxidizing agents. A major goal has been to establish the design criteria for the kinds of ligands that would be compatible with high oxidation state chemistry. The necessary features include: (1) resistance to oxidation; (2) the formation of five- and sixmembered metallacycles upon coordination; (3) negative charge sufficient to counter the positive charge of the metal center; (4) chemically innocent binding sites that are resistant to hydrolysis and other displacement reactions; (5) ease of derivatization to vary the oxidizing 4 power of the complex and to incorporate steric bulk or chiral centers and; (6) convenient synthesis. In accordance with these criteria, the design and synthesis of polyanionic chelating (PAC) ligands such as H4HBA-Et, 1, H4CHBA-Et, g, H4HBA-B, J, H4CHBA-DCB, 4, H4HMPA-B, §, and H4HMPA-DMP, §, have been accomplished (see Figure 1.1). These ligands have been successfully complexed with a variety of metal centers. Complexes of Ni, Cu, Co, Mn, Cr and Os have been fully characterized• and a number of interesting reactions have been discovered.16a-f The desired mode of bonding to transition metals is shown in Figure 1.2. The ligands coordinate through the fully deprotonated amido and hydroxyl groups as tetraanions and form either five or six-membered chelating rings with the metal center. Figure 1.2. The desired mode of coordination of a PAC ligand One of the first PAC ligands to be synthesized was g. After an extensive investigation, it was discovered that the ethylene bridge is susceptible to oxidation and participates in a cascade of degradation processes. One of the metal species was found to be a catalyst for the electrochemical oxidation of alcohols. 17 A scheme showing the remarkable degradation of this ligand is illustrated in Scheme 1.1. Because the ethylene bridge is susceptible to oxidation, a new set of ligands, ;a and 1, were designed to eliminate this possibility by replacing the ethylene backbone with a phenylene bridge. These new ligands provided a wealth of reactions (see Scheme 1.2) when coordinated 5 ~~ 1 H4 t-eA-Et Cl :p_CI ~no 0 i Cl H4 oeA-Et o c~ o J:~ Cl~:p-CI a Cl Cl 4 H4 HM-B ~~ ~ H4. . .A-8 H.. BBA-Et • 1,2-bia(2-hydroxy H4 oeA-CCB t!:+ § H4tt.N-CIMP "do)ethue B.. CBBA-Et a 1,2-bia(3,l-dichloro-2-h7draxybeuamido)ethane B.. BBA-B • 1,2-bia(2-hydroxybeuamido)bemene H.. CHBA-DCB s 1,2-bia(3,l-dichloro-2-h7droxybemamido)4,5-dichlorobemene B4 BMPA-B a 1,2-bia(2-hyciroxy-2-methylpropanamido)benzene B,.BMPA-DMP a 2,4-bia(2-hydroxy-2-methylpropanamido)- 2,4-dimethylpa.tan-3-one n1■r• 1.1. Commonly used polyanionic chelati111 (PAC) li1a11ds. 6 ·•-- , ... ¥ Scheme 1.1. Oxidative and hydrolytlc decomposition of trans-Os( ,, 4-CHBA-Et) (PY)2, 7 [<;t~>r /0 (N/, ·os·I .-f~ N~ I 'CN+ '--o o, (N~,N o.· 0~ l'P p/ 1~l• ':""3 doc,e (d'I~ 'o /0 i) P~, t-Bln: ii) e,2 N~ t;;~-::~I o' 'N N.J + /'o c:✓--~-") ~ l'N 'o l'oi+ '--o TFA ©::~H IO ~ ~i~ 0\ (N,,,_ !,,,CN+ @ c~"}:) OPPtt3 N-1-::i c;10 ~ ~ ('.,-1-:-.N) ··01· o" I'o PPhs ~ !~ @ i@ tWN) o' l'o ~ Scheme 1.2. Synthesis of Os(IV) complexes. ~ (;f) ~ 8 to osmium and led to the discovery of three isomers for octahedral PAC ligand complexes and the suspicion that the ligands are not as innocent as was first hoped. The PAC ligands J and 1 can exist in three geometries labeled cis-a, cis-{J and trans, as depicted in Figure 1.3. 18 L N:-1-:::N) ''M' d'l'o ' L trans cis-a Figure 1.3. Nomenclature for the three possible isomers of a PAC ligand. Three reasons for these isomerizations have been discovered: The ligand rearranges in response to the electronic requirements of the metal center as a result of structural constraints imposed by the other ligands, and in response to steric repulsion. A detailed study of the thermodynamic relationship among these isomers was made (see Scheme 1.3). 19 These isomerizations led to the discovery that in the cis-a and cis-{J forms the amide groups are nonplanar. PAC ligand complexes are the first examples of non-planar amide ligands in coordination chemistry. 18 During the course of these studies, however, it also became clear that the PAC ligands J and 1 were capable of resonance forms that would reduce the oxidation state at the metal center,as shown in Figure 1.4.18 9 L c+oI L 1~ ~ .o:rov Ef L ,~Q l 1~ Ef • -0.SIV L oi-~ I L tEf • •l.77V L 2· "~ I L Scheme 1.3. Thermodynamic relationships between cis and trans isomers in different oxidation states. 10 0 >°< Q 0 Oo"-w':ab ,, 'o < ~ ~M~ l I >°< . d:M':~ ~t.=ba ,,M,0 0 ~ Q 0 Figure 1.4. Resonance structures of a coordinated PAC ligand. To eliminate the non-innocence of the PAC ligands J and 1, the PAC ligands were redesigned again, and the aromatic ligand "arms" were replaced by an aliphatic bridge between the hydroxyl group and the amide carbonyl. The final PAC ligands, _g_ and§ fulfill all of the criteria outlined above. Cobalt complexes of§. were found to catalyze the epoxidation of olefins such as styrene (see Equation 1.2). 16c Studies with these new ligands are ongoing. (1.2) .... ... The work reported in Part One of this thesis is part of the larger work of the Collins group. The overall major goals are to synthesize and explore the chemistry and oxidizing abilities of new high oxidation state transition metal complexes using rationally designed PAC 11 ligands. The specific goals of the research reported here are to prepare high oxidation state complexes containing unsaturated nitrogen ligands and determine whether or not oxidizing capabilities exist. 12 References 1. Parshall, G.W., Homogeneous Catalysis, John Wiley & Sons, New York, 1980, 101-104. 2. Dumas, T., Bulani, W., Oxidation of Petrochemicals: Chemistry and Technology, John Wiley & Sons, New York, 1974, 39-47. 3. (a) Sharpless, K.B., Verhoeven, T.R., Aldrichimica Acta, 1979, 12, 63. (b) Landau, R., Sullivan, G.A., Brown, D., Chemtech, 1979, 602. 4. Dumas, T., Bulani, W., Oxidation of Petrochemicals: Chemistry and Technology, John Wiley & Sons, New York, 1974. 5. Spiro, T.G.,Ed.,Metal Ion Activation of Dioxygen, John Wiley & Sons, New York, 1980. 6. Johnson, J.M., Halshall, H.B., Heineman, W.R., Biochemistry, 1985, 24, 1579-1585. 7. Parshall, G.W., Homogeneous Catalysis, John Wiley & Sons, New York, 1980, 185-207. 8. Sheldon, R.A., Kochi, J.A., Metal-Catalyzed Oxidation of Organic Compounds, Academic Press, New York, 1981, 17-32. 9. Sheldon, R.A., Kochi, J.A., Metal-Catalyzed Oxidation of Organic Compounds, Academic Press, New York, 1981. 10. Collins, T.J.,Ed., Report of the International Workshop on Activation of Dioxygen Species and Homogeneous Catalytic Oxidations, Galzignano, Italy, June 28-29, 1984. 11. Wiberg, K.B.,Ed., Oxidation in Organic Chemistry, Part A, Academic Press, New York, 1965, 69-184. 12. Ibid., 1-68. 13 13. Cotton, F.A., Wilkinson, G., Advanced Inorganic Chemistry, John Wiley & Sons, New York, 1980. 14. Smith, K. M.,Ed., Porphyrins and Metalloporphyrins, Elsevier, New York, 1975. 15. Siddall, T.L., Miguara, N., Huffman, J.C., Kochi, J.K., J. Chem. Soc. Chem. Commun., 1983, 1185. 16. (a) Spies, G.H., Ph.D. Dissertation, California Institute of Technology, Pasadena, California, 1985. (b) Treco, B.G.R., Ph.D. Dissertation, California Institute of Technology, Pasadena, California, 1987. (c) Collins, T., Ozaki, S., Richmond, T., J. Chem. Soc. Chem Commun., 1987, 803. (d) Collins, T., Gordon-Wylie, S., J. Am. Chem Soc., 1989, 111, 4511. (e) Collins, T.J., Santarsiero, B.D., Spies, G.H., J. Chem Soc. Chem. Commun., 1983, 681. (f) Anson, F.C., Collins, T.J., Gipson, S.L., Keech, J.T., Krafft, T.E., lnorg. Chem., 1987, 26, 1157. 17. Krafft, T.E., Ph.D. Dissertation, California Institute of Technology, Pasadena, California, 1985. 18. Peake, G.T., Ph.D. Dissertation, California Institute of Technology, Pasadena, California, 1987. 19. Keech, J.T., Ph.D. Dissertation, California Institute of Technology, Pasadena, California, 1987. 14 PART ONE: COORDINATION CHEMISTRY OF HIGH-VALENT OSMIUM POLYANIONIC CHELATING LIGAND COMPLEXES WITH IMINE, IMIDO AND NITRIDO LIGANDS CHAPTER TWO: SECTION 1- REACTIONS OF Os(77 4 -HBA-B)(PPh3)2 TO FORM IMINE, IMIDO AND NITRIDO LIGAND COMPLEXES 15 Introduction and Background There has been a great deal of interest recently in multiply-bonded ligand .complexes; for example, carbene (CR2) 1 and carbyne (CR) 1 complexes have been the focus of intensive research efforts. Transition metal oxo ("oxene" (0) complexes are widely used in oxidation chemistry and research continues in this area. 2 Less attention has been directed to imido ("nitrene") (NR) 1 ,3 and nitrido ("nitryne") (N)· 1 .4 complexes. lmido ligands are isoelectronic with the oxo ligand and like the oxo ligand, exhibit strong 71"-bonding to metals, making these ligands particularly well suited for high oxidation state chemistry. An interesting feature of the oxo and imido ligands is the possibility of forming a triple bond to the metal for which the bonding modes are illustrated in Figure 2.1. R I + +N ..o 111 Il -M -M terminal linear imido terminal bent imido ~ •"ciII M oxo N Il M nitrido Figure 2.1. Bonding modes for oxo, imido and nitrido ligands. The terminal linear bonding mode for the imido moiety is the one most commonly observed in crystal structures. The terminal bent mode can be expected only when a linear bonding mode causes the electron count to exceed the eighteen-electron rule. 3 Bisimido Project--lntroduction PAC ligands have the ability to form unusual complexes. The efficient donor abilities of the PAC ligands stabilize metal complexes in unusually high oxidation states, and the reaction chemistry available to these novel compounds has not yet been fully explored. The formation and subsequent reaction chemistry of unsaturated metal-nitrogen bonded PAC complexes 16 should be particularly interesting. The first area of inquiry was centered around the efforts to prepare a neutral Os(VIII) bisimido PAC ligand complex. In 1956, Kobayashi reported the first imido complex (t-Bu-N)Os03. 5 Since then, many imido complexes, most of them with Group IVB and VIIIB metals, have been prepared, which are thermally stable and in many cases air stable. Relatively few bisimido complexes have been reported which is in sharp contrast to the tremendous wealth of their dioxo analogues. Fourcoordinate bisimido complexes are known for osmium, chromium, molybdenum, tungsten and platinum.6 To date, only one octahedral bisimido complex has been characterized.7 The bonding for a bisimido compound is quite interesting because there cannot be two full triple bonds in a cis structure (utilizing only the metal d1r orbitals); the two amphoteric ligands thus lead to several possible bonding modes shown in Figure 2.2. 7 A A I N Il A . Mo=N: ~ ./ R N/ "N IllMo=N ....... / R Mo=N-F 12. II Q Figure 2.2. Bonding modes for bisimido complexes. Crystallographic and variable temperature NMR studies are in better accord with the delocalized structure, §. In addition to the interesting structural questions, there are several other motivations for pursuing the synthesis of bisimido complexes. These considerations, which will be further discussed below, include: (1) an effort to extend the basic reactions available to the PAC ligand complexes and to probe the effects of strong 1r-donors on PAC ligand isomerization in high oxidation state complexes, (2) preparation and isolation of the first neutral, octahedral Os(VIII) complex, and (3) development of bisimido complexes for use as nitrene (:NR) transfer reagents for organic substrates. 17 A noteworthy structural feature of the PAC ligands is the ability to isomerize in response to the electronic requirements of the metal (see Chapter One). Trans, cis-cx and cis-/3 diasteriomers of PAC ligand complexes have all been characterized (see Figure 1.3). It has been established by an examination of the electrochemistry of these compounds that the trans form is the least donating and the cis-cx geometry is the most electron-releasing. 8 It would be very interesting to determine the PAC ligand geometry in an octahedral bisimido complex. Although the metal center would be formally do (OsVIII), the imido ligands are among the strongest 11"-donor ligands known, and the PAC ligand may not be required to isomerize to the most donating (cis-cx) form. The strength of the imido 11"-donation may give rise to unusual reactivities as well as to interesting structural features such as the capability of functioning as nitrene transfer reagents. To date, all of the Os(VIII) compounds known invariably contain 0-2 , F- and/or N-3 ligands.9 Isolation and characterization of a bisimido PAC ligand osmium complex would represent the first example of a well-characterized neutral octahedral Os(VIII) compound. Such a compound is of particular interest considering that the Os(VIII) species, Os04, is a powerful oxidizing agent used extensively in organic chemistry. It was hoped that the electronic properties of the compound could be tuned by varying the nature of the PAC ligand. This assumption was based on electrochemical data such as are shown in Table 2.1. 8 (Volts vs. Fc+/Fc) III/II IV/III "V"/IV trans-Os('r14-HBA-B)(PPh3)z -1.66 -0.44 trans-Os(114-0IBA-DCB)(PPh3)z -l.72a -0.46 Compound trans-OsC4(PPh3)2 -0.12 +0.71 +0.59 +1.34a [0sO612- -1.19 +0.74 a Couple is irreversible. Table 2.1. Formal potentials for osmium complexes 18 The preparation of a neutral Os(VIII) complex with PAC ligands that could be tuned for specific oxidizing abilities was one of the main thrusts in this project. Bisimido complexes can potentially function as a source of nitrene transfer reagents for use in organic synthesis. Metal imido complexes may offer advantages over nitrene fragments generated in organic media which are not always very selective. 10 Although monoimido complexes in general tend to be fairly inert and react only under forcing conditions, 11 Schrock and Nugent have reported metathesis-like reactions for tantalum imido complexes (see Equations 2.1 and 2.2, respectively). 12 ,13 (2.1) (2.2) PhCH=NPh >=NR' + + "TaOCl 3(THFh" "Ta" Bisimido complexes, however, are expected to be more reactive.7 For example, the series Os0n(NR)4-n (n=2, 1) was found by Sharpless and co-workers to be quite reactive towards olefins (see Scheme 2.1). 14 Vicinal oxyaminations and diaminations proceed smoothly in moderate-to-good yields with electron-deficient olefins. Such transformations are potentially very useful to synthetic chemists. Unfortunately, the Sharpless system appears to be limited to electron-poor olefins in the trans configuration. With the synthesis of PAC bisimido complexes, it was hoped that the generality of this reaction could be increased and extended to include other substrates such as ketones, dienes, a,,8-unsaturated ketones and thioketones. Many synthetic routes are available to imido complexes. A recent review outlined many of these and the most common syntheses are shown in Table 2.2. 3 The most promising of these synthetic methods appeared to be the organic azide route; this method was explored first. 19 I 0 ,r " 0 R ~ Os / . Os ~~ O R I 0~ '7N a) 0 Nx, , R' R' 0 0 0 II R= PhCH 2 S-, II 0 II II + + \ O=Os II R"OC- 0 0 R' I 0~ ~N b) 0 Os ,7 ~ N •I R' R R'HNX_,,,,R + "Os" R'HN Ph \ R > \ R > ~ >> A' A A '===I R R' R' Scheme 2.1. Oxyamination and diamination reactions. F R" 20 (a) Deprotonation of primary amines MC12Rte2 --+ M(NR) M=O + RNH2 --+ M(NJ M(NR' 2) 2 + Rtli2 --+ M(l (b) Use of silyl.amines M=O + RN ( Si.Me:3) 2 ~ (Nil Cl-M=O + 2Rm ( SiMeJ) --+ 1 + MeJSiNH2~a (c) Alkylation of nitrido oauplexes MIN + RX --+ M(NR); (d) Oxo/imido exchange reactions M-0 + R)PNJt ---+ M(NR) M-<> + lHD---+ M(NR) M-0 + RNS:> --+ M(NR) (e) Addition of chlormnine base (f) M + lHfCl. ---+ M(NP Oxidation of metal c:arbonyls M(CD) 2 + RR>.l~ MC M(00)2 + RNJ (g) Addition of organic azides M+Rt1)~ Table 2.2. Convenient syntheses for transition metal imido complexes. 21 Results and Discussion As a result of the pioneering work of the Collins group in preparing unusual PAC ligand complexes in high oxidation states, a number of high-valent complexes are available for reactivity studies. The chemistry of the osmium complexes is the most developed, largely because the higher oxidation state osmium complexes still exhibit useful 1 H NMR spectra, despite their paramagnetism, which greatly aids in characterization. For these reasons, the osmium complexes were selected to carry out studies involving unsaturated nitrogen ligands. Most of the coordination chemistry with unsaturated nitrogen ligands was explored using a single PAC ligand, (TJ 4-HBA-B), ,1,. This ligand was chosen because it is easily synthesized and allows easy access to many readily synthesized starting material complexes. The synthesis of .1 is straightforward and is shown in Scheme 2.2 . .1 can be complexed to osmium according to Scheme 2.3. Reduction of the dioxo, Os(VI) complex, [K2] [Os(774-HBAB)(0)2], l, with PPh3 leads to the neutral Os(IV) complex, Os(,, 4-HBA-B)(PPh3)2, §_ (see Scheme 2.3). One of the phosphine ligands is quite labile and can be readily displaced, allowing coordination of a variety of ligands. Both phosphine ligands can sometimes be replaced under more severe conditions or by dichelating ligands. Much of this chemistry has been reported elsewhere.8 Here developments of the chemistry of§_ with unsaturated nitrogen ligands is reported. Two equivalents of phenyl azide were added to a degassed benzene solution of§_ under nitrogen at room temperature according to Equation 2.3. 22 0~ ~C-OH 0 0 6-oAc CI-C-C-CI II II no oi r\ // 0-oAc Aco-(0) C-NH HN-C no oi r\ // C-NH HN-C NaQH. aq 0-ai ~-b .4 B.BBA-B s 1,2-bia(2-hyd.roxybeuamido) benzene Scheme 2.2. Synthesis of H4(HBA-B), 1- 0 K2COs(OH)4(0)2l K2 o>-Qi ~a ( N:;;JiN) cf'H'o 0 l ~ PPh3 0 1-;N> Os, PPh 3 , CF 3 C02H K2 I <N:;;ll~N) Os' 0I cf 11 'o 0 I a PPh3 - ff 4 Scheme 2.3. Synthesis of [K2][0s(11 4-HBA-B)(02)], l, and Os(11 -HBA-B)(PPh3)2, ft. I\) w 24 PPh 3 (2.3) d:9 + 2 PhN3 benzene rt PPh3 8 After one hour the green solution turned brown. Thin layer chromatography (TLC) indicated the formation of a new purple compound, the presence of starting material and a brown intermediate complex. The reaction was allowed to continue and over the course of several days became a brilliant deep purple, concomitant with the disappearance of the brown complex. The concentration of the purple product increased steadily for a period of about three weeks. The reaction rate was increased by adding a ten-fold excess of phenyl azide. Heating the reaction mixture produced decomposition. The product was isolated after successive preparative TLC on silica gel plates and elution with methylene chloride. It was necessary to repeat this step several times to achieve an analytically pure sample. Recrystallization from ETOH/hexane afforded deep purple crystals in a 30% overall yield that were suitable for crystallographic study. Proton NMR spectra showed the expected pattern for the (11 4-HBA-Br 4 ligand (four sets of doublets and two sets of triplets) and a very complicated phenyl region ( see Figure 2.3). Despite the absence of infrared (IR) absorption peaks in the characteristic area for transition metal imido complexes (1100-1300cm- 1) 4 , 1 H NMR and analytical data led initially to the formulation of the compound as the desired bisimido species, Os(ry 4-HBA-B)(NPh)2. A crystallographic study was performed to further elucidate the structure and revealed that in fact the compound was the diimine complex,~- Proton decoupling and NOE 1H NMR experiments allowed total assignment of the NMR spectrum with the x-ray structure in hand (see Figure 2.3). The 1 H NMR spectrum of~ is an excellent example of a paramagnetic complex that still displays useful NMR spectra. The proton resonances are spread apart by the paramagnetism; the spectrum is actually easier to assign than if the compound were not paramagnetic. C D F DO D ~:7f 8A F<Q>-o_;;o/s~N'H<K;o C F X) :©(N~~ 1: D A J C B i iE H D I\) <.Tl J G 11 K PPM ll 10 9 8 Figure 2.3. 1H NMR spectrum of~- 1 6 26 X-ray Diffraction Study Of The Diimine Complex, 9 The x-ray structure of the diimine complex,~ is shown in Figure 2.4 and the bond length data are given in Table 2.3. The ligand is in a cis-/3 arrangement and is twisted out of the planar configuration in response to the structural constraints of the diimine ligand. One of the consequences of the PAC ligand's being in the cis-/3 configuration is that there is considerable rotation around one of the amide C-N bonds (there is a barrier of activation for this rotation of 1035 kcal/mol) 15 , causing a disruption of the amide resonance (pictured in Figure 2.5), and causing the amide to be non-planar. :Q: II / c , .. /R R N ( I H Figure 2.5. Resonance in organic amides The discovery that the amide is non-planar is an important observation. There are very few nonplanar organic amides (examples include formamide, 16 lactams such as penicillin 17 and antiBredt bridgehead nitrogen compounds 18 ), and the PAC ligand complexes represent the first examples of non-planar amide ligands. Non-planarity is assessed by a set of parameters, -y, Xe and Xn, first defined by Dunitz and Winkler. 19 -y approximates the angle between the carbonyl carbon p-1r orbitals and the unperturbed amide nitrogen and is a measure of the twist around the C-N bond. Xe and Xn measure the out-of-plane bending around the carbon and nitrogen atoms, respectively. Figure 2.6 shows how these parameters are related to the amide torsion angles and can be derived from crystallographic data. 27 25A 13s~ ~~.,~-..c.",""H' ) 24A 1 39(6; IIA _ -~ t IIC Figure 2.4. Crystal structure of~- 28 DtiT"-NCI: SlGl'IA S\'l'I -"TC!'I -""r0"' SY"I 1. 973i 1. 9457 o.oi• 0 CSi CSi CSi CSi! CSi C5i Nll Nll 011 Nill 010 Nil N31 ii 91 0 1.988' i!Oi:55 1. 9801 i!. OU7 1. 4441 1. 3'40 1. ilJ7i! 1. 35'1 1.4.04 l. 337i l. 3544 1. 3•s 1 l. ?BH 1 4i!94 1. HSS 1. 474. 1. 38'ti 1. 31!!4 1.4391 1. 4i:i!3 1. Jn7 1. ,1!1 7 4 1. 49"3 1.4204 1. 449i! 1. 3997 1. 4i!5i! 1.437i 1. 4850 1. 3745 1 SOii 1. 3994 1 S141 1.4i!O• 1 43i:9 1. 37•3 1.4435 1. 3'50 1. 3415 1.:1~, 1 40li! 1. 31154 1.4027 1. 3343 1.3'8i! 1. 3777 0.019 0.0i9 0.011 0. Oii 0. Oi!l o.o.o 0. 041 0 OoO 0.0tl 0.041 0.04• 0. 050 0 050 0. 041 0.00 0.037 0.055 0 043 0.0&4 o. 05• 0 OSi! 0. ,41 0. 051 0. I)! 3 0.043 0.040 0.049 0. 050 0. 041 0. 041 0 03i 0.044 0 047 O 037 0 01, O.u4!> 0 C41 0 :)49 o.o4• 0.041 &:&~l 0 040 0. 041 0.041 0.037 0. 03i! 0. 031 & 0 0 & g g & 0 Nil 21 ii JI cu 0 0 0 0 Ji 0 0 0 0 0 0 0 0 JI SI 41 Bl 51 71 g EI! ,a 011 108 0 0 10B 111 NlD :u 0 1111 ao Ci!D 0 0 0 C 0 0 0 0 & 0 0 80 0 80 0 0 0 0 0 0 0 0 0 0 0 0 /I •• ••~8 •& '11D ill 30 JD 40 !D oD ?D ID 9D 101) Ni.E NiF i!ll! illl i!i.P ioP i!H i!4E 311 311 ]ii Bl 341 Ull 3111: ...ea ~D a0 ~D (J 0 '.) 0 0 0 ◊ 70 0 0 0 0 0 0 0 ClD 0 0 JD 40 er, SIi all 80 1011 llt Ui 311 ii!! i!U iJi i!I i4E i!U Ult 3U 331; 341 JU 3&1 NU Table 2.3. Bond length data for~- 0 0 (; ,.0 ,. g g C ae (; 0 i!2 la i!2 29 Figure 2.6. Projection down the amide C-N bond of a coordinated amide showing the derivation of the angular parameters. Collins, Peake, et al. have defined a new parameter, -y, according to Equation 2.4, such that there is no distinction between the cisoid and transoid geometry.20 (2.4) -y= (-y) mod 1r Non-planar amides are defined as those that have a -y value greater than 25 ° . 20 Table 2.4 gives the non-planar amide parameters for the diimine complex,~- As can be seen from these data, complex~ contains both a planar and a non-planar amide. The -y for the amide that is twisted up into a cis configuration is 44, while the amide that remains in the ligand plane has a -y value of only 22 °. 30 compound t (des) XN (des) Xe (des) 6 7 -1 22 ..... 44 Table 2.4 Angular parameters for~ IR spectroscopy also reflects the configuration of the PAC ligand in the amide CO stretching frequencies. For PAC ligand complexes of 1, that are in a trans position, the CO stretching frequencies are ca. 1600cm-1 , while the non-planar amide CO stretching frequencies move to lower energy ca. 1650cm-1 , as is shown in Figure 2.7. Another interesting feature of the structure is the ortho-phenylenediimine ligand. The Os-N2 and Os-N3 bond distances of 1.98(2)A and 2.03(2)A are consistent with the formulation of an osmium-nitrogen single bond. (For example, the Os-N distance in (Os(NH3)(Cl3)(PPh3) is 2.136A.) The ortho-phenylenediimine ligand has several possible resonance structures, la, lb and le, as shown in Figure 2.a. 21 I ex: U I R lb I I N ~ la R R R N (X N N I I R R le Figure 2.8. Resonance structures of the phenylenediimine ligand. The crystal structure reveals that there is a significant contribution from the diimine form, le. The C35-C34 (1.33(3)A) and C32-C33 (1.38(4)A) bonds are somewhat shortened compared to 31 1 P~rt~-,:~~~--<~t-m-----------"~-f-~_,-_----t-________...,i_'..........!-++--'-----------,~·--+----------+-~+---/--A".,_"_-+---------j!'------+ - ---t----7 I : -- --- ----------i----t-------4----+:cc-H: !: ! r:1 I I~ Ii I I !!i I' I I : I ' ! 11 -l r u 1700 1600 1500 1400 WAVENUMBER CM- 1 1300 ~1--t-- _ J_~ -i ' 1 ' 1800 {'...(\ ''I 1 iii,, Jo- 7 --p-------f------+--++II :-t,--,-:itt-t-: i rr-+'I+'-'-,---+'~' !! ; 1200 t- - - ~-- ->-- - - - 1100 1000 900 800 -- 4()-- I L--~J(}-I I ! I ;u-- I I I - - ,-- ·l(l-- I 1800 1700 1600 1500 WAVENUMBER CM-l 1400 1300 1200 1100 1000 900 800 !00 Figure 2.7. A comparison of planar and non-planar amide IR stretching frequencies in~ and~- 32 the other phenylene ring bonds. The N2-C31 (1.42(3)A and N3-C35 (1.37(3)A are in between the average N-C single (1.47 A) and average N-C double bond (1.35A) lengths. Mechanism The reaction to form the diimine complex,~. is not entirely unprecedented. In 1981, Gross, lbers and Trogler reported the preparation of the diimine complex, 11. 21 This compound was prepared photochemically by extrusion of nitrogen from the starting metallacyclotetraazadiene, 10 (see Equation 2.5). H hu (2.5) ~c<g@ pt No reaction occurred in refluxing toluene in the absence of light. The diimine product formation requires ortho C-H activation. This step has a precedent in organic photochemistry (see Equation 2.6). 22 (2.6) Trogler proposed that the bisimido intermediate complex undergoes insertion (by one of the imido groups) into an ortho C-H bond, which leads to product formation. It is likely that the osmium diimine, ~. is formed via a similar pathway. The proposed mechanism for the formation of~ is shown in Scheme 2.4. Parts of this mechanism have been previously reported by Gross and Trogler, 23 Smith, Green and Stone, 24 33 PPh3 PPh 3 e±9 PPh3 e:9 - PPh 3 PPh 3 +N 3 Ph e±9 N3 Ph 8 j·PPh 6R-Ph ;h-lN~N / ~ ::JI - N2 N2 / Ph-N,N,:::::::.N ~ ~ ::::,.... -ph - Ph.,....N, _) I ~ I -Ph N\ I + N3 Ph 3 -Ph N~ / N l + N3 Ph -Ph ~ """N'Ph Ph-N,N,::::::.N N~ N - N2 az II: N I Ph 13 -Ph s;::i • ~ Ph__.H-N' . © 0 Scheme 2.4. Mechanism of formation for~- ~/H ~D p:- " '": :,._ I 34 and Cenini and La Monica. 25 A likely first step is phosphine dissociation from -8_. There is a noticeable increase of reaction rate when the starting material is freed from excess triphenylphosphine by washing with copious volumes of hexanes. The phenyl azide could then attack at the vacant coordination site left by the departing phosphine. The strong trans influence of the coordinated azide could help push out the other more reluctant phosphine ligand. lsomerization of the ligand would allow bidentate coordination of the azide. The azide can then either extrude nitrogen to form an unsaturated monoimido complex, which could then react with another equivalent of phenyl azide, or alternatively, attack by a second equivalent of azide on the azido complex can precede the nitrogen extrusion step. In either case, the metallacyclotetraazadiene, 12., is formed. Subsequent extrusion of nitrogen from 12. would form the Os(VIII) bisimido species, 13. It is proposed that it is this intermediate that undergoes C-H insertion followed by a 1,3 hydrogen shift to form the final product, ~Attempts to isolate and identify the brown intermediate complex which we believe to be the neutral Os(VIII) bisimido species failed because of decomposition and its apparent paramagnetism. During the course of the reaction, which was monitored by TLC by removing aliquots of the reaction mixture at regular intervals, the reaction mixture was cooled to -78 degrees and an attempt was made to isolate the brown intermediate. No meaningful spectra were obtained. Attempts to trap the intermediate or to effect a nitrene transfer in situ with a coreactant also failed. Several olefin substrates were used such as diethyl and dimethyl fumarate and styrene. (Sharpless reported that he was able to isolate the metallacycle bisamido intermediate complex from the reaction of dimethyl and diethyl fumarate with the complexes OsOxNR4-x). 14 There was evidence of reaction indicated by color change, but no meaningful products could be isolated. The most probable cause is that the olefins can react with organic azides to form either a triazoline or an imine compound, as well as with the metal complex (see Equation 2. 7).26 35 (2.7) c6 1so·c cq + N-Ph In another attempt to isolate an Os(VIII) bisimido complex, .e_ was treated with trimethylsilyl azide. Trimethylsilyl azide was selected because it is commercially available. Addition of trimethylsilyl azide to a degass·ed benzene solution of§. produced an overnight color change from green to straw-yellow ( see Equation 2.8). PPh3 (2.8) e±9 + benzene rt PPh3 8 14 A pale precipitate fell out of solution, and filtration followed by multiple washings with hexanes produced a yellow powder. Recrystallization from methylene chloride/THF afforded yellow crystals in an overall yield of 80%, suitable for a crystallographic study. The 1H NMR spectrum showed characteristic ligand peaks, aryl phosphine peaks and a noticeable absence of any resonances that could be attributable to a -Si(CH 3)3 group. 31 P NMR showed a single resonance at 8+42 ppm attributable to the phosphinimine phosphorous. The IR spectrum exhibited an intense stretch at 1120cm- 1, which is in the range of osmium-nitrido stretching frequencies (1123-1023cm- 1 ). 4 The structure was assigned as the Os(VI) nitrido compound, H; this was confirmed by the x-ray crystallographic study. X-ray Diffraction Study Of The Nitrido Complex, 14 Osmium has the most extensive nitrido chemistry of all the transition metals. 4 Most of the osmium nitrido complexes known are Os(VI) species. The most salient feature of the nitrido ligand is its very strong 1r-donor capability. Bond length data, force constants, trans-effect 36 studies and electronic spectra indicate that it is the strongest n-donor known. 4 The crystal structure shown in Figure 2.9 illustrates this point. The bond length data are given in Table 2.5. The Os-N distance is 1.64(1) A, which is the expected distance for an Os-N triple bond as compared to [K2][OsNCl5] (OS-N,1.614 A) and [Ph4As][OsNCLt] (Os-N, 1.60 A). In contrast, the Os-N single bond distance is 2.136 A in Os(NH3)(Cl3)(PPH3)2 and the calculated Os-N double bond distance is 1.77 A. 3 An interesting feature of this compound is that the osmium sits up out of the ligand plane by 0.56A, with an average N-Os-L angle of 106.4 °, which is a greater distortion than is usually seen. For example, OsCl4N- and OsCl5N-2 have N-Os-L angles of 104.5° and 96°, respectively. 27 Belmonte has reported a [NBu 4l[Os(N)R4] compound that also exhibits a large distortion: the average N-Os-C bond angle is 107.7° .28 Dubois and Hoffmann have done orbital calculations on nitrido complexes and suggest that this distortion out of the basal plane is caused by the stabilization of the 2e set of orbitals with increasing N-ML angle. 29 The ligand bending causes increased overlap between the p orbitals on the nitrogen and the metal xz and yz orbitals. A more complete discussion of this follows in a later section. The amide ligand linkages are both non-planar, a highly unusual occurrence (vide supra). The -y, Xn and Xe values for 14 are shown in Table 2.6. cmqioud 30 5 25 7 3 0 Table 2.6. Angular parameters for 14 The IR data again reflect the non-planarity of the amides, and the carbonyl stretches are found at 1637cm-1 (vs. ca. 1600cm-1 for planar amides). It is assumed that the non-planarity of the 37 Figure 2.9. Crystal structure of 14. 38 ATOM ATOM cs cs )( OS OS OS N1A N1A DU Ci!A 1A NU DU NU . Del 1A u 7A ZA i!A i!A 3A 7A 4& SA 3A SA •• •• •• 'IA 10A NU NU DU Di.I 11 il!I as 311 41 SB •• ea 'II p p p p 11 11 ti! 13 14 15 i!1 Z1 ii!! u 24 i!S 31 31 Ji! 33 34 4A ,.•• 7A II 10& 101 11 Bl 11 71 .,, i!I 31 41 u oi 711 01 101 N 11 i!1 31 1i! u 13 14 u u i:.i! 21, u l!4 2S u 3l! :;, 33 34 n DISTANCE SIG"'A SYM 1.UOi! 1. '834 1 1 •"89 1.9734 1. 35'0 1 47&3 1 i!374 1 nu 1. su• 1 4044 1. 3 . . . 1. 3US 1. 4i!B3 1. 3771 1. 4495 1. 4317 1. 3•01 1. 4103 1. 3347 1. 3HS 1. 41,()] 1 as•;,o 1. ?SOe 1. OSI 1 •tt-e 1 4035 1 39!: e 1 1 4114 1 •tii! 1 . .se3 1.•384 1. •uo 1.77•3 1 775• 1 7UI 1. n•J 1.4011, 1 4141 1. 4U7 1. 36'• 1. 40Se 1. 4i!31 1.4471 1.4i!'4 1 3111 1. 401' 1.4031, 1. 4211 1 31 . . 1 4010 1 Jii:.4 1 4i!1'1 1 4355 0.010 o. ooq 0.007 0.00'1 0.001 0.015 o.ou o. 015 o.ou 0.017 0.017 0.017 0.017 0.017 o.ou o.ou 0.017 o.ou 0.011 0.018 0.015 o.ou o.ou o.ou 0 OP 0 01a 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 •au ,a, ~ a, o. ,u 0 OH 0 OU o.ou 0.017 o.ou 0.010 oou o. 011 0.010 0.017 o.ou 0018 0.018 0.011 0.017 o.ou o.ou 0 011 0.01' 0.01' 0.011 o.ou 0.017 0. 011 0 019 O.Oi!O 0 Oi!O 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3S 3' ATOM ATOl'I ATOii ANG:.E SIGl'!A SY1 SYl N1A O! X 101.z, 107 c~ 10, H 1¢~. 3~ H.7£ I! 141, n 147 81 11 0~ n 33 0 43C 0.417 0. 0. 41e 0. 341 0. :uc o. )44 0. Ui! 0. 31' 9. 144 0 0 0 0 cc..:l',H, OH Cit. NU. Di.I: NU Di!i OU 0: Ci 0~ a= D~ OS OS 05 ps X X )f N!A N1• NU DZ• Di.A !'fU ... ·~ft ¢ 0 0 ' 0 0 0 C C 0 () 0 0 Table 2.5. Bond length data for li 0 0 0 39 PAC ligand amide groups (a direct result of the osmium atom sitting up out of the plane) reflects the electronic requirements of the Os(VI) metal center. Presumably, the increased overlap of the metal xz and yz orbitals with the nitrogen p orbitals, combined with the increased donating ability of the non-planar amides, allows this formally Os(VI) compound to "feel" electron-rich at the metal center. The counterion, H2N=PPh3 +, of the nitrido complex is also fairly unusual. The P-N bond distance is 1.64(1) A, which is in good agreement with a P-N double bond (for example, P=N = 1.567 Afor PPh3=NPh-p-Br and P=N = 1.641 Afor Ph2 FP=NCH3). 30 The hydrogens on the nitrogen were not located. The other ligand bond distances do not show any unusual features. Mechanism Thermal decomposition of chloroazide (CIN3) or NaN3 in HX is a well known method for nitrido synthesis. 4 This reaction is the first example of a fully characterized nitrido complex formed from the decomposition of an organic azide, although Chatt and Dilworth have also reported the preparation of an uncharacterized "Mo=N" species from (CH3)3SiN3_ 31 . A suggested mechanism for the formation of 14 is shown in Scheme 2.5. A probable first step is the hydrolysis of (CH3)sSiN3 to form the parent azide, hydrozoic acid (HN3). The -Si(CH3)3 group is known to hydrolyze under mild conditions. As previously mentioned, the complex can readily dissociate one of the triphenylphosphine ligands. At this point the parent azide could undergo a Staudinger reaction with dissociated PPh3 to form the phosphinimine, 1.§, which is known to further react with the hydrazoic acid to produce the ion pair, 16. The azide ion, N3-, is then free to attack the unsaturated osmium complex, extrude nitrogen and form the final product, 14. Attack by free azide ion has been proposed by several groups for the formation of Re(V}, Re(lll), 32 Mo(Vl), 33 , Ru(VI) and Os(Vl) 6 nitrido complexes. The following evidence supports this mechanism: 40 2(CH3)JSiN 3 2 H20 + i- N2 [ HN=PPh 3) i 15 + [ H2N=PPh3] + [ N3] • 16 1- PPh N 3, 2 14 Scheme 2.5. Mechanism of formation for 14. 41 (1) The reaction proceeds faster when the starting material is free from excess PPh3, suggesting that phosphine dissociation from §_ is involved. (2) (CH3)3SiN3 + PPh3 ><, (CH3)3SiN=PPh3 (1§.) This reaction is the well-known Staudinger reaction. 34 Phosphines and azides will react to form phosphinimines quite readily and under a variety of conditions; however, this reaction does not occur at room temperature. (3) (CH3)3SiN=PPh3 (1§.) + §. X 14 The phosphinimine reagent, 1§., was independently prepared and does not react with the starting material, §_, to form the final product, 14. (4a) (CH3)3SiN=PPH3 (1§.) + H20/benzene (4b) (CH3)3SiN=PPh3 (1§.) + CH30H/W NR --;), [HN=PPh3} ~ H2N=PPh3 15 does not decompose to the parent phosphinimine under the reaction conditions, but requires acid and a polar solvent. 34 (5). (CH3)3SiN3 + H20/benzene HN3 + ((CH3)3Si)02 This reaction will occur under the reaction conditions. The products were identified by 1 H NMR. (6a). HN3 + PPh3 (6b). HN=PPh3 + HN3 HN=PPh3 [H2N=PPh3t This reaction is also a Staudinger reaction that will occur under the reaction conditions. 34 Note: a reaction that should be performed is the treatment of§_ with HN3 to determine if 14 is produced. 42 (7). The mother liquor of the reaction shows a new 1 H NMR peak that can be assigned to (Me3Si)2O. (The literature value is +0.1 ppm.) (Me3Si)2O (independently prepared) was added to the mother liquor and the peaks were found to coincide. In addition to the NMR evidence, the IR spectrum of the neat mother liquor shows bands corresponding to (Me3Si) 20. Both the aromatic azide (phenyl azide) and the heteroatomic azide (trimethylsilyl azide) failed to yield the desired bisimido complex. The next logical choice of azide is an aliphatic hydrocarbon. Reactions with several aliphatic azides were explored: 1-adamantyl azide, methyl azide, t-butylazido formate and trityl azide. 1-adamantyl azide is commercially available. _a was stirred overnight at room temperature with an excess of 1-adamantyl azide in benzene according to Equation 2.9. e±9 ~ Ad I PPh 3 (2.9) + PPh 3 8 N benzene (2f9 I s + N2t PPh 3 17 The solution slowly changed from green to a chocolate brown. TLC showed that the reaction was not completely clean. Preparative TLC by elution with 3:1 methylene chloride/acetone allowed the collection of a brown compound, the major product (ca. 60%). However, several successive TLCs were necessary to isolate the compound in a pure state, and the overall yield dropped to 10%. The 1H NMR spectrum showed ligand peaks in the appropriate downfield region, along with triphenylphosphine resonances. Adamantyl resonances not corresponding to the starting azide were present in the aliphatic region. The integrations were consistent with a formulation of 1:1 :1 in the PAC ligand, adamantyl group and triphenylphosphine. The 31 P 43 NMR spectrum showed one new resonance at 5 +34 ppm. The IR spectrum exhibited a strong absorption at 1150 cm- 1, which is in the correct region for metal-imido aborptions (11001300cm-1).3 The analysis was acceptable for the monoimido osmium PAC ligand phosphine complex, 1l- Unfortunately, x-ray quality crystals were never obtained as the compound decomposed. The adamantylimido complex is only the second adamantylimido complex that has been reported. Sharpless prepared the compound Os(NR)2(O)2, where R=1-adamantyl. There is an Os=NR IR peak at 1160cm-1 . This reagent was demonstrated to be active in the ammoxidation of olefins. 35 Because the complex 1l was somewhat intractable and could be obtained only in low yield, no further reactivity studies were undertaken. The other aliphatic azides, methyl azide, t-butylazido formate and trityl azide did not produce clean reactions and were not pursued. In another attempt to produce a bisimido complex, perfluorophenyl azide was selected as a reactant because it was expected that by substituting fluorine for hydrogen and thus eliminating the possibility of C-H insertion, it would be possible to isolate the bisimido complex. However, Gross and Trogler reported a remarkable C-F insertion in a similar system to that discussed above (see Equation 2.10). 36 F (2.10) benzene,hu F He was unable to determine the source of the proton on the diimine ligand but speculated that the source was from a free radical reaction with the protiated solvent. Perfluorophenyl azide was prepared according to literature methods, as shown in Equation 2.11 .37 44 (2.11) NaNO2 / HCI vacuum distillation o·c © F §. was stirred overnight in benzene with an excess of periluorophenyl azide according to Equation 2.12. (2.12) + benzene 18 The solution gradually changed from green to black, and a black precipitate fell out of solution. TLC showed the reaction to be relatively clean with essentially one (baseline; Ri=0) component. The black precipitate did not dissolve readily in solvents with the exception of hot DMSO in which it formed a black-purple solution. However, if the black powder was allowed to remain for a long period of time (e.g., overnight) in a solvent such as diethylether, THF, methylene chloride or acetone, it dissolved slowly to form purple compounds that had an Rt of approximately .8. These compounds were subsequently soluble in benzene. There were no 1H NMR spectra associated with any of these compounds. However, the black-purple solution of DMSO was pumped to dryness in vacuo, after which the compound was also soluble in benzene. It was possible to obtain an 1 H NMR spectrum of this compound in DMSO-d 6 . The spectrum showed only PAC ligand resonances along with DMSO and water resonances in a 1: 1: 1 ratio. The IR of this compound exhibited absorption peaks in the region of O-bound DMSO complexes (925,940 cm- 1) 38 as well as characteristic water peaks. This anomalous compound is tentatively formulated as the Os-Os multiply-bonded PAC ligand dimer, 18 (see Scheme 2.6), on the basis of its behavior towards solvents, ligands and the spectroscopic evidence presented below. It is most reasonable to formulate the multiple " " t-bupy- (-qs;;;;;;;;;; Os-J-t-bupy " 21 or t-butyl pyridine Os s " " s Os " 18 S • THF, CH 2 Cl 2 , Et 2O, acetone ~ co 22 S= DMSO no clean products ~ 01 PPha 89 " " DMSO-t-os-os-J-DMSO I\ I\ •2 H20 19a or PPha DMSO 8 20 Scheme 2.6. Synthesis and reactivity of 18. 89 H20 19b 46 bond as quadruple for reasons that will be explained (vide infra). Microanalysis of the original black powder (1]_) before any solvents have been added is in agreement with this formulation. 1Jl is apparently capable of coordinating other solvent molecules into the two vacant coordination sites (see Scheme 2.6). 1H NMR spectra and microanalytical data after the black compound has been dissolved in hot DMSO lead to the formulation of the compound as [Os( 17 4HBA-B)(DMSO)·H2O]x, 19a or 19b (where x = 1 or 2). It is not possible to distinguish between these two possible structures shown in Scheme 2.6. The reaction of 1Jl with triphenylphosphine supports the formulation of 1Jl as the dimeric species. When 1Jl was stirred in benzene with triphenylphosphine at room temperature, it converted cleanly back to the green monomeric starting material,§_ (see Scheme 2.6). There was no evidence at all by TLC, or from the examination of the entire reaction mixture by 1 H NMR, that any other side product was present. 18 was also treated with t-butylpyridine in benzene at room temperature overnight to form a purple compound. The reaction was not very clean, and a pure sample for microanalysis and 1H NMR spectroscopy was difficult to obtain even after repeated preparative TLC (elution with methylene chloride) steps. The compound appeared to be paramagnetic; the ligand and tbutylpyridine peaks were quite spread out, and the integration of the t-butylpyridine peaks was only half of what was needed for the osmium dimeric complex, £1. It may be that upon workup one of the t-butylpyridine ligands was replaced by another solvent molecule (such as DMSO), or that only one t-butylpyridine ever coordinated and that the other site was occupied by a solvent in the subsequent workup procedure, i.e., 22. The compound Os(71 4-HBA-B)(PPh)3(t-Bupy), which was prepared by stirring§_ in t-butylpyridine, was also found to be paramagnetic with a wide range of 1 H resonances. 8 It was not possible to substitute both phosphines fortbutylpyridine by this method, so there is no direct comparison available. The IR spectra for £1 (or 22) was similar to the spectrum of Os(77 4-HBA-B)(PPh3)(t-Bupy). There was an electronic absorption band at 771 nm (12,970 cm- 1), which could mean that the osmium-osmium bond is intact (vide infra). 47 Walton has reported that when the Os-Os multiply-bonded complexes, [Os 2Xa]" 2 , are treated with phosphorous (PR3) and nitrogen (pyridine) donor ligands the metal-metal bond is cleaved (see Equation 2.13a and 2.13b). 39 (2.13a) (2.13b) x......_ X I X I ,......x i·2 _...........os-os......_ + [X I X I 2 X i· 1 X PR 3 x......_ X I X I ,......x i·2 os-os X I x----- [ X [ x_........... I x.__.,,Os-........ I ;x PR EtOH I ......_x X pyridine 2 PY ] •1 PY-. 1 _......x .,,,,,.,,os-........ [ PY I X X X•CI, Br It may be that in the case of 1.§., the stronger a-donation of the PR 3 ligand is required to cleave the Os-Os quadruple bond and that the t-butylpyridine ligand simply isn't a strong enough donor to cleave the Os-Os bond. The proposal that the black compound is an Os-Os quadruply-bonded dimer presents an interesting possibility. There are many examples of Mo-Mo and Re-Re quadruply-bonded dimeric compounds. 40 Despite the fact that the Os(IV) metal centers (d 4-d 4) are isoelectronic with Cr(II), Mo(II), W(II) and Re(III) all of which form quadruply-bonded species, no examples of Os-Os quadruple bonds have appeared in the literature. 40 In fact, even examples of wellcharacterized Os(III) (d 5-d 5 ) triply-bonded dimers are limited to [Os2X8 ]"2 and Os 2(hp) 4Cl 2 (where hp= 2-hydroxypyridine). 39 .41 In addition to these examples, Collman has proposed a triple bond in the metalloporphyrin complex [Os(OEP)] 2 (where OEP= octaethylporphyrin). 42 Electronic absorption spectra of compounds with multiple M-M bonds have been extensively examined. 43 For the metal dimeric compounds with an electronic configuration of a 21r4 52 (this includes Os(IV)-Os(IV)), the lowest energy electronic transition should be the 5-5* transition. 40 The relative energies of the molecular orbitals for a M2li3 compound with a ground state of d 4-d 4 are shown in Figure 2.10. The 5.5* transition has been observed to occur between the near-IA (ca. 16000nm) and the middle of the visible range (ca. 450nm) in all cases 48 ------- - -- --- . . ... ~ --. -<~-" . ,,.• --- ----,,,,,,,,,,,,," ,.. ... c:,l G> C LiJ ; ; ; ,, ; ,, ,, ,, a • - - ' , ,, " ,, a--" ,,. ____ _--- -- btu (d.z_,z) b, 9 (d 1 z-,z) e, (,r*) - b1uC8•) j!_ b?,(a> __ ....JL-1.L eu(,r) -------- --- ---- H Figure 2.10. Relative energies of the molecular orbitals for a M243 compound with a d 4-d 4 ground state. 49 where this ground state configuration exists. 40 Two general trends for these transitions are reported: the intensities are quite low because of the small overlap of the two d orbitals that make up the o bond, and the transition is always observed at a far higher energy than can be calculated. 40 ,43 Although conclusions drawn from the electronic absorption spectra should be supported by x-ray crystallography, it is interesting to note that 1§_ does show absorptions in the region where M-M multiple bonds are found. The electronic absorption spectrum of 1§_ is shown in Figure 2.11 and has an absorption band at 768 nm (13,200 cm-1) in methylene chloride; the absorption spectrum is identical when it is taken in toluene. (The solutions were prepared and used quickly to prevent reaction with the solvent, but when they were examined several hours later, there was no change.) Figure 2.12 shows the spectrum of the DMSO compound, .1Q, which also displays a band at 775 nm (12,903 cm- 1) in methylene chloride. The spectrum of .1Q that has been stirred in THF also shows an absorption in this region, at 712 nm (14,044 cm·1). An analogous quadruply-bonded rhenium-rhenium compound, Re 2 Cla· 2 exhibits a s-s* absorption band at 14,654 cm· 1 . 43 The lower wave number of the absorption band for 1§_ is consistent with osmium being a heavier element. The electronic spectrum of the starting monomeric complex, fl, shows no absorption band in this region. Metal-metal quadruply-bonded complexes have a characteristic vibronic stucture in the o-5* band, as a consequence of metal-metal stretching, that can be seen at low temperature. Accordingly, the low temperature (77 ° K) electronic absorption spectrum was obtained as a 2methyltetrahydrofuran glass and clearly showed vibronic structure (see Figure 2.13a). The shape of the band is similar to the shape of the low-temperature (5 ° K) spectrum of Re 2 C1 8· 2 (see Figure 2.13b) and is characteristic of s-s* excitations into a a antibonding orbital. 43 The bond order of the metal-metal bond is reduced, causing a "displacement-shift" of the potential energy wells (along some coordinate axis). This displacement makes it likely that the strongest absorption would be at higher energy (i.e., 0-4 versus 0-1; see Figure 2.13a). This pattern is observed for [Re 2C18 ]"2 and for 18. The vibronic structure that is observable at low-temperature for 18 has an average progression of 305 cm·1, which is in the middle of the known range (250 50 8/83189 2 : 23 : S6pt11 l'IENSLOT 2 1888 DATA POINTS FRON 988 TO 288 111111 AVERAGE OF 1 SCANCS) READINGS PEH DATUPI = 1 ,75 ~ 0 z ,6 4: ~ i:x: 0 OJ ~ ~ .45 ,3 ,15 8 158 388 458 688 758 WAVELENGTH Figure 2.11. Electronic absorption spectrum of 1§ in methylene chloride. 988 51 cb8.dat 8/83/89 5:88:81p111 l'IEl1SLOT 8 1888 DATA POINTS FROl'I 988 TO 288 lllll AVERAGE OF 1 SCANCS> READINGS PER DATI.BI = 1 1.88 .96 l'.il .84 u z .72 ~ ~ ~ 0 00 ~ ~ .& .48 .36 .z4 .12 9 158 225 388 375 458 525 688 675 758 825 988 WAVELENGTH Figure 2.12. Electronic absorption spectrum of ll in methylene chloride. 52 ,925 .9 .875 0-3 0-4 w .es u z .825 ~ ~ ~ 0 00 ~ ~ 0-2 0-1 o-o-? Average progression= 305 cm- 1 .o .ns ,75 ,725 .7 .675 ,65 788 658 eee 758 988 858 UAUELEHGTH Figure 2.13a. Electronic absorption spectrum at 77 ° K of 16 showing the vibronic structure of the 6-6* transition band. CM~ .sooo 20000 40000 35000 30000 25000 -N -- .l. 250 300 350 400 450 500 NM Figure 2.13b. Electronic absorption spectrum of Re2c1 8-2 at 5 ° K. 550 700 53 cm· 1 - 350 cm· 1) for metal-metal excited state vibrations (for example, the progression for An experiment that needs to be done is to try to prepare the Os-Os quadruply-bonded dimer independently. A reasonable synthetic method is shown in Equation 2.14a. If the complex, §.,for example, is heated in vacuo, it is possible that a multiply-bonded complex will result. This suggestion is based on Collman et al.'s synthesis of the Ru(II) porphyrin dimer, shown in Equation 2.14b. 44 ......--:::::,. L (2.14a) ( ' Os~o ~~~ (o, Os ~o .,,,, ~ I 'o N A• vac I L (2.14b) Ru(OEP)(CO)L PY N/UII 'o ~ + 4L N .....___....., 0 (L= ¢ 3P, py) t ..,......cs, . . . .w,~o Ru(OEP)(py) 2 A• Vac 18 reacted quickly (within 15 minutes) when CO was bubbled into a methylene chloride suspension to form a bright-purple homogeneous solution. Monitoring the reaction by TLC in methylene chloride showed that one clean compound was present. (A blank reaction without CO did not produce the same results.) An IR spectrum was obtained of the crude reaction mixture that showed what was assigned as a CO stretch at 1980cm·1 . The suggested structure, 20, is shown in Scheme 2.6. Unfortunately, after workup, there was no IR peak at 1980cm·1 , no 1H NMR and no meaningful analysis, so the structure cannot really be assigned, merely suggested. The compound appeared to be stable only in the presence of CO. One experiment that needs to be done is to seal 18 in an NMR tube with excess CO and obtain both 1 H and 13 C 54 NMR spectra. It is possible to prepare the compound Os(77 4-HBA-B)(CO)(PPh3) by bubbling CO through a benzene solution of§.. The IR for this compound has a CO stretch at 1985cm-1 .8 The biscarbonyl complex, Os(77 4-HBA-B)(CO)2, was highly desirable because it would have provided another possible entrance into bisimido complexes. Condensing biscarbonyl metal complexes with either RNO2 or RN3 species produces imido complexes. 3 An effort to produce a neutral Os(VIII) complex from 1§. was made by attempting to oxidize it with elemental sulfur. A number of different reaction conditions were explored and although there was evidence that a reaction of some kind had occurred, no clean products could be isolated. Similarly, unsuccessful attempts were made to produce a neutral dioxo species. Further Attempts To Prepare a Bisimido Complex In further efforts to prepare a bisimido complex, other synthetic strategies were explored. Metal-oxo complexes react with phosphinimines to yield imido complexes. 3 The driving force behind this reaction is apparently the formation of the strong P=O bond (130 kcal/mol). The phosphinimine route is a powerful synthetic method, because trimethylsilyl phosphinimine reagent can be facilely prepared from triphenylphosphine and trimethylsilyl azide, and it is easily derivatized to many other phosphinimine reagents. Hydrocarbons will react with trimethylsilyl phosphinimine and exchange an -R group with the -Si(CH3)3 group. For example, although the synthesis of perfluorophenyl azide is long and arduous, perfluorophenyl phosphinimine can be prepared easily by condensing perfluorobenzene and trimethylsilyl phosphinimine. 34 Several attempts were made to prepare imido complexes using this method. K2[Os(17 4 HBA-B)O2] was stirred with phenyl phosphinimine and trimethylsilyl phosphinimine in two separate attempts. In both cases there was no evident reaction. Another attempt was made using a different ligand system. The monooxo complex, OsO(M2-Et), was stirred with various 55 phosphinimine reagents, but there was no evidence of reaction in this case either. No further attempts were made. Another well-known synthetic method for preparing imido complexes involves the reaction of primary amine compounds with metal oxo complexes. Accordingly, the complex K2[0s(17 4-HBA-B)02] was treated with t-butylamine, but no products were isolated. Several other amines were tried with no positive results. This method was also abandoned. Conclusions While the major goal of synthesizing a neutral Os(Vlll) bisimido compound was not achieved, a number of interesting reactions and compounds were discovered. The first example of a neutral octahedral Os(Vlll) compound, formed as an intermediate in the formation of the diimine complex,~ was postulated. Rather than form a stable Os(Vlll) complex, this intermediate underwent an ortho C-H insertion into an aryl ring to form a diimine ligand. In addition, the first example of the formation of a nitrido complex via the decompositon of an organic azide was discovered. These two compounds represent rare examples of non-planar amides complexed to transition metals. The synthesis of the monimido complex, 1l, is only the second example of an adamantylimido complex. Additionally, there is evidence for the formation of the first osmium-osmium quadruply-bonded complex, jj!. 56 PART ONE: COORDINATION CHEMISTRY OF HIGH-VALENT OSMIUM POLYANIONIC CHELATING LIGAND COMPLEXES WITH IMINE, IMIDO AND NITRIDO LIGANDS CHAPTER TWO: SECTION II-REACTIONS OF OSMIUM NITRIDO AND METHYLIMIDO COMPLEXES 57 Introduction Transition metal nitrido and imido complexes are quite interesting compounds because of their proposed role in catalytic processes related to nitrogen fixation 45 , ammoxidation 46 and amination 14 of olefins and possible use as nitrene transfer reagents. 10 Nitrido species have been postulated as intermediates in the reduction of N2 to NH3 by molybdenum (see Equation 2.15). 47 NMP + "Mo" (2.15) NMP =N-melhylpyrrolidine dppe = 1,2-bis-(diphenylphosphino)ethane A nitrido intermediate has also been proposed in the fixation of azide to ammonia by molybdenum (see Equation 2.16). 33 [c15MoV- N= N= N 1· (2.16) A ruthenium nitrido has also been implicated recently in the reduction of a (µ2-NO) bridge to (µ3-NH) and (µ-NH2) in ruthenium carbonyl clusters (see Equation 2.17). 48 (2.17) Ammoxidation and amination by imido complexes have been explored both as models for industrial heterogeneous processes and as catalysts in their own right. Sharpless has 58 succeeded in the diamination and oxoamination of olefins with bisimidos and oxoimidos (see Scheme 2.1). 14 Chan and Nugent have studied molybdenum imido complexes as models for propylene ammoxidation (see Equation 2.18). 49 (2.18) l-W PhCH= N-t-Bu Despite the similarity of nitrenes to carbene species, there has been only one clear example of a nitrido compound's being activated to serve as a nitrene transfer reagent (see Equation 2.19). Groves and Takahashi have succeeded in activating a Mn(V) complex in situ and in transferring the imidotrifluoroacetic acid imido moiety to cyclooctene. 50 (2.19) lul 59 The lack of reactivity by nitrido complexes can be partially explained by an examination of the orbital picture. In general, it is believed that nitrido complexes are unreactive as a consequence of the strong 1r-donation to the metal. 4 Dubois and Hoffmann have performed orbital calculations on five- and six-coordinate ruthenium nitrido complexes. 29 The frontier orbitals are shown in Figure 2.13. For the d 2 systems illustrated, RuCl5N- 2 and RuCl4N-, the orbitals are filled up to the b2 level. Dubois and Hoffmann propose that the energy of the 2a 1 orbital (LUMO) is lowered substantially for the 5-coordinate nitrido by distortion of the basal ligands out of the metal plane. Distortion out of the basal plane is caused by the stabilization of the 2e set of orbitals with increasing N-M-L angle. The ligand bending causes increased overlap between the pz orbitals on nitrogen and the metal dxz and dyz orbitals to form a hybrid orbital (see Figure 2.14). (The four basal ligands are a-antibonding with respect to Md/. The basal plane distortion moves the ligand orbitals off the d/ torus into a node that drops the 2a 1 orbital down to a lower energy.) Figure 2.14. The hybrid orbital formed by the nitrogen pz orbital and the metal dxz and dyz orbitals. As a consequence of the low-lying 2a1 orbital, Dubois and Hoffmann suggest that the 5coordinate nitridos will be fairly good electrophiles. The 2a1 orbital is approximately 50% nitrogen Pz orbital and 50% metal d/ and may act as an acceptor, either by coordination of a donor ligand to form a 6-coordinate complex (see Equation 2.20) 51 or by reaction at the nitrogen (see Equation 2.21 ). 52 (2.20). NMX.f + x- ____,. MX5N-2 60 3a 11 ti,.._N'°H I + l'j z- c1-.iu".....c1 1 .Cl CJ-Ru:_Cl (:I Cl I , "l-43 Ht',. I c1.,.,.I Figure 2.14. The frontier orbitals for five- and six-coordinate ruthenium nitrido complexes. 61 The observed reactivity does not always bear out the theoretical predictions. Although OsNCl4- fails to react with er, it does react with arsines, stibenes and bipy but does not coordinate the incoming ligand in the axial position (see Equation 2.22). 52 OsNCl 4 - will coordinate PPh3 at the nitride to form a phosphine imidate species (see Equation 2.23). It also loses chloride and coordinates triphenylphosphine. 52 N (2.22) L Cl-..... 111..,.......-L Os ~ L = AsPh 3 , SbPh 3 , AsEt 3 or 112 bipy Cl 1...._c, (2.23) Belmonte has reported the alkylation of a nitride with RI or (CH3)303+ which contrasts with Dubois and Hoffmann's predictions (see Equation 2.24). 28 (2.24) Coordination of donor ligand in the trans axial position changes the electronic environment of the nitride complex. The orbital picture for OsNCls- 2 indicates that the 6coordinate nitride is fairly electron-rich. This is consistent with the observed protonation reaction in an analogous molybdenum compound (see Equation 2.25) 31 and the formation of the thionitrosyl complexes shown in Equations 2.26-2.28· 53 62 HX MeOH (2.25) ~------,N,...,.a__,B=Ph,-f-4~ MeOH MoN(S2CNR2h (2.26) (2.27) [:::s t [MoX(NH)(dppe) 2 [BPh4f orS 8 R = CH3, CH3CH2 ReCl2(N)(PR3)2 PR3 • PMe2Ph, PEt2Ph, PMePh2 (2.28) Dubois and Hoffman n's treatment suggests that the osmium nitrido, 14, may be more reactive than other nitridos, considering that it shows a marked distortion. (0.56A out of the basal plane, vide supra). Accordingly, a number of reactions were explored with both electrophilic and nucleophilic reagents. Results and Discussion As a preliminary step to studying the reaction chemistry of 14, it was decided to exchange the cation, [PPh3=NH2]+, for [NEt4t, to avoid any unwanted reactions with the cation. 14 was stirred in an acetone/methanol mixture with >1 equivalent of NEt4CI for 24 hours (see Scheme 2. 7). After rotary evaporation of the solvent, the solid was dissolved in THF from which the yellow mother liquor could be decanted from a white precipitate. The white precipitate was identified as PPh3NH2CI by 1H and 31 P NMR. Very pure [NEt4] [OsN(17 4 -(HBAB)], 23, was collected after preparative TLC (elution with methylene chloride/THF) in 80% yield. PPh3 I N [PPh3NH21·1osN(r{HBA-B)] 14 NEt4OTf + "Os" 27 e.±9 N I + ~ f (Et4Nt[OsN(r{HBA-B)] -NE40TI CH3 N 1 '2:9 " "Os" 29 JH2 " ~H3 H,; ;H N '2:9 '2±9 PPh 3 ~ . e.±9 H 1· H2 Decomp. I a, PY u) 28 X ~fins I .. I s No clean products I HC>TI NEt4OTf + I p~ ------.:i-on I-½' CH30TI co (/ 2S • PPh 3NH2CI "______-Y 26 ~Js::J ~ NEt4CI CH3I HOTI No Reaction/ Decomp. CH 3 I N H N ~Js::J ~ + NEt.iOTf 24 Scheme 2. 7. Reaction chemistry of 23. ~Js::J '<:..V I PY 30 23 + + NEt4 I + 0 I CH3I 64 There is always a small amount of 14 recovered from this reaction despite the excess of NEt4CI. 1H NMR showed the expected coordinated PAC ligand peaks, and the IR spectrum showed the Os=N peak shifted by 53cm- 1 to 1067cm- 1. In an effort to model the nitrogen fixation work, 45 attempts were made to protonate, hydrogenate and alkylate 23. It was hoped that 23 would be very well behaved and follow an ideal protonation reaction pattern like the one shown in Scheme 2.8. In fact, these reactions proved to be extremely problematic. When 23 was exposed to mineral acids such as HCI, the solution turned from yellow to pink immediately. This pink compound was very difficult to characterize; however, when a mild base, i.e., K2CO3, was added to the pink solution, the solution returned to yellow and the starting compound was recovered quantitatively (see Scheme 2.7). Addition of one equivalent of triflic acid, (HOTf), at -78 ° to a methylene chloride solution of 23 produced a cloudy solution from which the white precipitate, identified by 1 H NMR as NEt4OTf, could be filtered away from the mother liquor (see Scheme 2. 7). After preparative TLC (elution with methylene chloride), an 1 H NMR spectrum of the major component showed it to be a single compound, with PAC ligand resonances and a new singlet in the alkyl region (51.1 ), which integrated as 1 proton vs. 12 for the PAC ligand set of peaks. There were no NEt4 peaks and presumably this is a neutral PAC ligand complex ( compound 24, Scheme 2.7.) IR showed no evidence for N-H activity, but the Os=N band at 1067cm-1 disappeared and a new band at 111 0cm-1 (the Os=N region) appeared. It was not possible to obtain a meaningful microanalysis. In another effort to protonate the nitrido complex, 23 was heated in a sealed tube of d5benzene with one equivalent of HOTf under hydrogen pressure at 110 ° for two days. The 1 H NMR spectrum showed that the PAC ligand was coordinated and intact, and a new resonance appeared at 69.7. However, upon workup the only product that could be identified was NEt4 OTf. Similarly, 23 was sealed in a pressure bottle under hydrogen pressure without HOTf. Again, the 1H NMR spectrum exhibited coordinated PAC ligand peaks, a new resonance at 59.7 65 23 N -2 '2±9 N +L I '2±9 I X H N -1 L -1 (219 H N (219 I I L L -1 L = some coordinating ligand Scheme 2.8. Idealized reaction scheme for modeling nitrogen fixation. 66 and another new peak at 51.2. Upon workup, no meaningful products other than NEt4OTf could be isolated in this case either. On the basis of Dubois and Hoffmann's orbital treatment that 6-coordinate nitrido complexes are fairly electron-rich (vide supra), it was hoped that coordination of an axial ligand would facilitate the protonation of 23. It appears, however, that rather than coordination of PPh 3 at an axial site, the preferred site of reaction is at the nitrido moiety. 23 was stirred with PPh3 in an acetone/methylene chloride mixture with gentle heating. The yellow solution changed to a deep red-brown within .5 hours. A red-brown powder was collected by stirring the reaction mixture in hexanes. An 1H NMR spectrum showed coordinated PAC ligand, PPh3 and NEt4+ resonances in a 1:1:1 ratio. 31 P NMR showed a new peak at 5+37 vs. an internal standard of PPh3 set at 5-4_ The PPh3 that was added as an internal standard was oxidized to O=PPh3 (5+26). The Os=N absorption at 1067cm-1 disappeared and there a new peak appeared at 1142cm- 1, which is in the right range for phosphine imidato osmium complexes (P=N range=1050-1200cm- 1). 52 The suggested structure of this complex, 25, is shown in Scheme 2.7. Upon attempting to obtain a purer sample for analysis by preparative TLC, the sample completely decomposed to numerous uncharacterizable compounds. Similarly, the crystals that were sent for crystallographic study decomposed before the data were collected and the xray structure obtained was of the decomposition product, [Na] [OsN(77 4-HBA-B)J-2 O=PPh3. There has been only one other report of the preparation of a phosphine imidato osmium complex. Pawson and Griffith prepared a series of Os(NPR3)(PR3)2X3 compounds from [OsNX4r and the appropriate phosphine reagents (see Equation 2.23). 52 They reported that these compounds were quite unstable as well and were unable to obtain any NMR spectra or crystallographic data. Another attempt was made to protonate 23 in the presence of a coordinating ligand. Accordingly, 23 was treated with HOTf in the presence of triphenylphosphine. There was only a single product that was identified as!! (see Scheme 2. 7). 67 Similarly, it seemed reasonable to expect that coordination of a CO ligand in the axial site might render the nitrido more susceptible to protonation. A dry THF solution of 23 was heated and CO was slowly bubbled through the solution for several hours (see Scheme 2.7, 26). During the course of this reaction the solution gradually changed from yellow to a rich orange. The THF was removed by rotary evaporation, leaving an orange oil. The IR of this crude material was very similar to 23, with the exception of two additional bands at 1700 and 1775cm1. The Os=N band was shifted to 1090cm-1 . The 1 H NMR spectrum showed the PAC ligand resonances to be shifted slightly upfield from 23. The product decomposed upon being exposed to vacuum, losing the IR bands at >1700cm-1 , and was not further characterized. Similarly, when a hot THF solution of 14 was treated with CO over the course of several hours, the solution changed from yellow to orange. The 31 P NMR spectrum showed a new resonance at 5+37 (vs. 5+42 for 14). The IR showed bands at 1725 and 1780cm-1 , and the nitride band was shifted to 1072cm-1. 1H NMR spectrum showed the PAC ligand peaks to be shifted slightly upfield. Attempts to crystallize and further characterize this product were also unsuccessful. When 23 was heated in methylene chloride in the presence of one equivalent of HOTf and CO, the yellow solution slowly (over several hours) turned orange. An IR of the crude reaction mixture was quite similar to that of 26 (see Scheme 2.7). Subsequent isolation of the several products by preparative TLC (elution with CH2Cl2ffHF) produced a compound that was neutral (i.e., showed no NEt4+ peaks), possessed an intact and coordinated PAC ligand, but showed no IR bands >1700cm- 1 and could not be further characterized (see 27, Scheme 2.7). Attempts to coordinate an axial ligand to 23 failed either (in the case of triphenylphosphine) because it produced disubstituted complexes or (in the case of CO) because the product was extremely unstable (vide supra). It was, however,possible to coordinate a nitrogen donor ligand in the axial site. 23 was heated in a neat solution of pyridine and there was an immediate color change from yellow to orange (see Scheme 2.7). The solution was poured into hexanes and an orange powder was collected. Recrystallization from methylene chloride and hexane yielded yellow crystals. 1 H NMR showed coordinated PAC 68 ligand and pyridine as well as NEt4+ resonances. The IR spectrum showed the nitrido peak shifted from 1067cm-1 to 1080cm-1. Microanalysis was in excellent agreement with the formulation of the compound as [NEt4][OsN(py)(77 4-(HBA-B)], 28, (Scheme 2.7). Contrary to the expectation that a 6-coordinate complex would be more susceptible to protonation, when 28 was treated with HOTf, complete decomposition resulted and no products were isolated. Only one report of a protonation of a nitrido complex has appeared in the literature. 53 Bishop, Chatt and Dilworth reported the protonation of MoN(N3)(dppe)2 with concentrated aqueous mineral acids, HX, to give a compound that they identified as MoX2(NH)(dppe) 2 (see Equation 2.25). They failed to find any IR bands assignable to N-H or 1 H NMR peaks attributable to the N-H proton. They also failed to report any Os=N band in the IR. Somewhat more convincing was the reaction of MoN(N3)(dppe)2 with sodium tetraphenylborate in methanol to give a compound assigned the structure [Mo(NH)(OMe)(dppe)2][BPh 4] (see Equation 2.25). This compound did show an IR band at 3345cm- 1 , but could also be formulated as [MoN(HOMe)(dppe)2[BPh4]. Belmonte reported that in the case of [RuN(CH 2SiMe 3)4r, reaction with HX resulted in the protonation of the M-C bond, while in the case of OsNR 4-, reaction with HX resulted in a new compound OsNR2X2 (where R= neopentyl). 54 This work however, has yet to appear in the literature. Fortunately, 23 proved to be more amenable to alkylation than protonation. 23 was stirred at room temperature for 1.5 hours in a methylene chloride solution with one equivalent of methyl triflate (MeOTf) (see Scheme 2. 7). The solution slowly turned from yellow to darkorange. A dark-yellow product was collected after column chromatography (elution with THF/hexane). Yellow-orange crystals were obtained by recrystallization from acetone. 1 H NMR (DMSO-d 6) showed the presence of a coordinated PAC ligand and a resonance at 84 that was assigned to the N-CH 3 and an absence of NEt4+ resonances. An IR spectrum showed the absence of the nitrido peak at 1067cm-1 and a new peak at 1105cm-1 that was assigned as Os=N. The compound was formulated as the neutral complex, 29, (see Scheme 2.7). 69 Microanalysis was in excellent agreement with this formulation (the presence of 2.5 H2 O molecules was quantified on the analytical sample by 1 H NMR). An attempt was also made to methylate the nitride with CH3I. Although NEt4I was isolated and identified by 1 H NMR, no other product was isolated. At the time 29 was prepared, it was the first example of a clean methylation of a nitrido compound. Subsequently to this, Belmonte et al. reported the successful alkylation of [NBu4][OsN(CH2SiMe3)4] with Me3OBF4 to give Os(N=Me)(CH2SiMe3)4 (see Equation 2.4). She also reported the x-ray structure of this compound. 28 In addition, Belmonte reported that the alkylation proceeded smoothly with alkyl iodides, RI, (where R= methyl, ethyl and propyl). 55 Despite the failure of 28 to protonate, it seemed reasonable to expect that it might be sufficiently activated to allow methylation with a weak methylating agent such as CH 31. To test this, 28 was stirred in methylene chloride at room temperature with CH3I .5 hours. The solvent was removed in vacuo and the compound was redissolved in THF from which a white solid precipitated. This solid was identified by IR and 1 H NMR as NEt4I. A yellow solid was collected from the mother liquor after the solvent was removed in vacuo. Preparative TLC allowed the collection of several different products. One of the products was identified as 23. Another product was identified as N-methylpyridinium iodide by comparison with an independently prepared sample. There was at least one other Os product that could not be further identified, which may be the methylimido pyridine complex 30. In addition to the protonation and alkylation studies, numerous attempts were made to produce neutral oxidized nitrido species with oxidizing agents, including Br2 , MCPBA, pyrBr3, tbutylhydrogen peroxide, hydrogen peroxide and elemental sulfur (see Scheme 2.7). None of these reactions produced clean products, although the nitrido definitely did undergo reactions with a wide variety of reagents. 70 Reactions of 29 One of the major goals of this project was to prepare imido complexes for use as nitrene transfer reagents on organic substrates. With this in mind, the monoimido complex, 29, produced from the methylation of the nitrido was treated with a number of olefinic substrates. 29 was sealed in NMR tubes with olefins such as styrene, cyclohexene, ethylene and dimethyl fumarate and heated in constant temperature baths at a range of temperatures (60 ° 110 °) for varying lengths of time (.5 hour-2.5 days). Acetone-d5, benzene-d5 and chloroform-d 1 were employed as solvents. In all cases there was either no evidence of reaction or the reaction mixture decomposed to a mixture of products from which no meaningful data could be extracted. An attempt was also made to hydrogenate the methylimido complex. 29 was sealed in an NMR tube under hydrogen pressure and heated in a constant temperature bath (80 ° -11 O O ) and monitored by 1 H NMR for variable reaction times (.5 hour-2 days). Three different solvents were tried: benzene-d6, chloroform-d1, and acetone-d6. In acetone-d6, after .5 hour at 100 °, the disappearance of the N-CH3 peak at 54.2 coincided with the appearance of a new singlet resonance at 53.3 and a new set of coordinated PAC ligand resonances. Further heating O caused decomposition. In benzene-d6 after heating for 2 days at 11 O the disappearance of , the N-CH3 resonance at 54.3 coincided with the appearance of a new triplet resonance at 53.0 and a new set of coordinated PAC ligand resonances. It may be that these new resonances were due to a coordinated methylamine ligand and that the triplet pattern was only discernible in d6-benzene. In neither case was it possible to assign any N-H resonances. These NMR tube reactions indicate that this line of research is worth pursuing and that 29 may be useful as a model for nitrogen fixation studies. Numerous attempts were also made to produce either a neutral Os(VIII) compound with oxidizing agents such as MCPBA, TBHP, H202 and benzoic acid or to coordinate a halogen at 71 the open axial site with reagents such as pyridinium tribromide, Br2 and ammonium halides, that might have allowed another path to bisimido complexes. None of these reactions gave clean results. Conclusions It has been shown that the nitrido complex, 23, is somewhat more reactive than other nitrido complexes per Hoffman n's predictions, but is more versatile than expected. The nitrido will coordinate a donor ligand at the nitrido site, (PPh3), and at the axial site (pyridine), and contrary to Hoffmann's expectations, the nitrido complex will also alkyl ate (and perhaps protonate). The 6-coordinate species, 28, does not appear to be more amenable to alkylation (or protonation) than the 5-coordinate species, 23, at least when the axial ligand is pyridine. Although the monoimido complex, 29, does not appear to be useful as a nitrene transfer reagent or as a precursor to neutral Os(VIII) or bisimido complexes, it may be a useful model for nitrogen fixation studies. 72 Experimental Materials Acetone, benzene, di-n-butylether, diethylether, ethanol, hexanes, methanol, and toluene were reagent grade (Aldrich, Baker, Mallinckrodt, M.C.B. or U.S.I.) and were used as received unless otherwise noted. Methylene chloride and tetrahydrofuran (Baker, Mallinckrodt) were distilled from calcium chloride prior to use. 2-Phenyl azide (10% hydrocarbon solution (P.F.& B), trimethylsilyl azide (P.F. & B.), p-toluidine (M.C.B.), dimethyl fumarate (Aldrich), tbutylamine (Aldrich), triethylamine (Aldrich), dimethyl malonate (Aldrich), styene (Aldrich), cyclohexene (Aldrich), ethylene (Matheson), t-butylpyridine, (Aldrich), pyridine (Aldrich), osmium tetroxide (99.8%, Alfa), potassium carbonate (Mallinckrodt), tetra-n-butylammonium chloride (Kodak), methyl iodide (Aldrich), potassium hydroxide, (Baker), hydrogen (Matheson), sodium hydroxide (Baker), triethylamine (M.C.B.), triflic acid (Aldrich), methyl triflate (Aldrich), trifluoroacetic acid (98.5%, M.C.B.), carbon monoxide (Matheson), triphenylphosphine (Aldrich), perfluorophenyl hydrazine (Aldrich), sulfur (Alfa) and sodium nitrite (Aldrich) were all used as received unless otherwise noted. Oxalyl chloride (Aldrich) was freshly distilled before use and o-phenylenediamine (Aldrich) was recrystallized from toluene prior to use. The silica gel used in column chromatography was 60-200 mesh. K2Os(OH) 4 (O2) 16 , trimethylsilyl phosphinimine 32 and perflourophenyl azide 37 were prepared according to literature procedure. 73 Physical Measurements 1H NMR spectra were recorded at 90 MHz on a Varian EM-390 spectrometer or a Jeol FX90-Q spectrometer, at 270 MHz on a Bruker WM 270 spectrometer, at 400 MHz on a Jeol GX400 spectrometer, or at 500 MHz on a Bruker WM-500 spectrometer. 1H NMR chemical shifts are reported in ppm(o) vs. tetramethylsilane with the solvent (acetone-d5) 62.04, CDCl 3 67.24, benzene-d5 67.14, CD2Cl2 65.35, DMSO-d5 62.50) as an internal reference. 31 P NMR spectra were recorded at 36.28 MHz on a Jeol FX90-Q spectrometer or on a Bruker WM 270 spectrometer. 31 P chemical shifts are reported in ppm(o) vs.an 85% phosphoric acid external standard or an internal standard of PPh3 set at 5-4_ Infrared spectra were recorded on a Beckman IR 4240 spectrophotometer. Elemental analyses were obtained at the Caltech Microanalytical Laboratory. Electronic spectra were obtained on a Cary 14 spectrophotometer. Raman spectra were attempted on a Spex Ramalog II, at a power of 50 mW- 1 W, irradiating with an argon laser at 488 nm. X-ray Crystallography Structure Determination of .e_ Data collection: The intensity data were collected on a Enraf-Nonius CAD4 diffractometer with Mo Ka radiation and a graphite monochromator (.XO. 7107 A) and averaged 4809 reflections. Unit cell parameters: a=9.26 A, b=16.03 A, c=17.88 A, a=95.4 °, t,=94.5 °, -y=98. 7 °. The atomic positions of the atoms were derived from the Patterson map. Subsequent Fourier and difference Fourier maps revealed all non-hydrogen atoms. Atomic scattering factors were taken from Stewart, Davidson and Simpson for H, and the International Table for X-ray Crystallograph for all others. 74 Several cycles of full matrix, least-squares refinement minimizing~ w(F 0 2 -(Fc/k) 2) 2 , w= a-2(Fa2), on all non-hydrogen parameters yielded R= ~ IIF 0 1-1Fcfkll/lF0 I = 0.091 and GOF ={~ w(Fc2-(Fc2/k)) 2/n 0 -np)} 112 = 1.18, where n 0 was the number of reflections and np was the number of parameters; all atomic coordinates were in one block and the scale facior and the isotropic Gaussian ellipsoids were in the other. Calculations were carried out on VAX11/780 and VAX 11/750 computers using the CRYRM crystallographic system. X-ray Crystallography Structure Determination of .H. Data collection: Oscillation and Weissenberg photographs indicated monoclinic symmetry. The intensity data were collected on an Enraf-Nonius CAD4 diffractometer with Mo Ka radiation and graphite monochromator (>..-0.7107 A) and averaged 5703 reflections. Unit cell parameters: a=9.56 A, b=23.68 A, c=14.38 A, ,8=93.26 °. The systematic absences led to the assignment of the space group P2 1/n (oko absent fork odd, hol absent for h+1 odd). The atomic positions of the atoms were derived from the Patterson map. Subsequent Fourier and difference Fourier maps revealed all non-hydrogen atoms. Atomic scattering factors were taken from Stewart, Davidson and Simpson for H, and the International Tables for X-ray Crystallograph for all others. Several cycles of full matrix least-squares refinement minimizing~ w(F 0 2 - (Fcfk) 2) 2 , w= a-2(Fa2), on all non-hydrogen parameters yielded R = {~IIF 0 I-IFcfkll/lF 0 I} = 0.090 and GOF = {~ w(F 0 2 - (Fc2/k)) 2/n 0 -np)} 112 = 2.04, where n0 was the number of reflections and np was the number of parameters; all atomic coordinates were in one block and the scale factor and the Gaussian ellipsoids were in the other. Calculations were carried out on VAX11/780 and VAX11/750 computers using the CRYRM crystallographic system. 75 Syntheses All reaction were carried out in air unless otherwise noted. H 4 HBA-B, .1,- 2-acetylsalicyclic acid (50.0 g, 0.27 mol) was heated in an excess of oxalyl chloride under a nitrogen atmosphere for 4 hours. Excess oxalyl chloride was removed in vacuo. The residue was dissolved in methylene chloride (30 ml) and the methylene chloride removed in vacuo. This step was repeated several times. The residue was then dissolved in methylene chloride (100 ml) and chilled in an ice bath. The chilled solution was added dropwise to a stirred, chilled methylene chloride solution of o-phenylenediamine (15.0 g, 0.138 mol). The mixture was stirred 1 hour and an excess of triethylamine was added and stirred another .5 hours. The mixture was treated with 6M NaOH (100 ml) and the organic volatiles were removed on a rotary evaporator. The remaining aqueous solution was decanted from the undissolved organic organic residue. The organic residue was dissolved in a small amount of acetone and treated with a an additional portion of 6M NaOH solution (100 ml). The aqueous solutions were combined and slowly acidified with concentrated HCI. The precipitate was collected, washed with water, and recrystallized from acetone/water. Yield: 78.3 g (81%). Anal. calcd. for C20H15N2O4: C, 68.98; H, 4.63; N, 8.04. Found: C, 69.04; H, 4.75; N, 8.14. 1 H NMR: see Table 2.7. [K2] [Os(r, 4 -HBA-8)(02)] • H20, 1,- A solution of K2[Os(OH)4(O2)] (2.01 g;5.46 mmol) in methanol (500 ml) was added to a solution of .1 (1.9 g; 5.46 mmol) in THF (200 ml) that was continuously stirred. The solution was heated on a hotplate (ca. 40 °; .5h). After cooling, the clear brown solution was evaporated to dryness on a rotary evaporator. The residue was dissolved in ethanol, di-n-butylether was added and the ethanol was removed on a rotary evaporator to give brown crystals. The product was filtered and washed with diethyl ether. Yield: 3.60 g (100%). Anal. calcd. for C20H12N2K2 O4Os• H2O: C, 36.25; H, 2.13; N, 4.23. Found: C, 36.35; H, 2.40; N,4.07. IR (Nujol): 1600 cm- 1 , vamide; 822 cm- 1 , voxo. 1 H NMR: see Table 2.7. 76 Os(11 4 -HBA-B)(PPh 3 ) 2 , !!,. I (1.02 g; 1.58 mmol) and PPh3 (2.0 g; 6.32 mmol) were dissolved in a mixture of ethanol (25 ml) and THF (45 ml). Excess CF3CO2H (.5 ml) was added and the solution was stirred at room temperature (10 min), after which the solution turned from brown to dark-green. Ethanol (20 ml) was added and the solution and the THF was removed on a rotary evaporator to give dark green crystals. The product was filtered and washed with ethanol and hexanes. The solid was recrystallized from dichloromethane/ethanol containing PPh3 (250 mg). The final product was filtered and washed with ethanol and hexanes. Yield: 1.23 g (73%). Anal. calcd. for C55H42N2O4OsP2 •.5(H2O), • .5(CH3CH2OH): C, 62.74; H, 4.26; N, 2.57. Found: C, 62.67; H, 4.44; N, 2.45. IA (Nujol): 1600 cm- 1, vamide. 1H NMA: see Table 2. 7. Os (11 4 -HBA-B)(NHC5H 4 NC5H 5 ), ft,. 600 mg of !i in 100 ml of benzene in a 250 ml flask fitted with a septum cap were subjected to three freeze-pump-thaw cycles. 2.4 ml (4 eq.) phenyl azide were added via syringe. After stirring at room temperature 7-20 days, during which the solution changed from green to brown and finally to purple, the solvent was removed in vacuo. A THF/MeCl2 solution of the resultant purple solid was chromatographed on TLC plates by elution with MeCl2. This step was repeated several times. Slow recrystallization from ETOH/hexanes yielded x-ray quality crystals. Yield: 30%. Anal. calcd. for OsC32H22O4N4 : C, 53.62; H, 3.09; N, 7.82. Found: C, 53.49; N, 3.19; N, 7.65. IA (Nujol): 1650, 1600 cm-1, vamide. 1 H NMA: see Table 2.7. [PPh 3 =NH 2 ][0sN(11 4 -HBA-B)], 14. 100 mg of !i in 50 ml of degassed benzene in a 150 ml flask fitted with a septum cap were subjected to three freeze-pump-thaw cycles. Under a nitrogen blanket, 0.022ml (2 eq.) of Me3SiN3 were added via syringe. The solution was allowed to stir at room temperature overnight, during which it changed from green to pale yellow and a precipitate fell from solution. The yellow precipitate was filtered and washed with hexanes. Recrystallization from MeCl2 yielded large x-ray quality crystals. Yield: 80%. Anal. calcd. for.OsC3aH29O4N4P: C, 55.20; H, 3.53; N, 6.78. Found: C, 55.01; H, 3.67; N, 6.44. IA (Nujol), 1637 cm-1 , vamide; 1120 cm-1 , 11Os=N. 1 H NMA: see Table 2. 7. 77 Os (N-C 10H 15)(17 4 -HBA-B)(PPh3), rr. A benzene solution of 500 mg of§ was subjected to three freeze-pump-thaw cycles and 738 mg of 1-adamantyl azide in benzene were added via syringe. The solution was allowed to stir at room temperature for 1 day during which the solution changed from green to a rich brown. The solvent was removed in vacuo and a brown powder was collected and washed with hexanes. The brown powder was purified by preparative TLC by elution with 3:1 MeCl2/acetone and collected in 60% yield. Successive TLC steps were necessary to obtain a pure product at 10% overall yield. Anal. calcd. for OsC4aH43O4N3P: C, 60.94; H, 4.48; N, 4.44. Found: C, 61.36; H, 4.92, N, 4.90. IR (Nujol): 1602 cm-1, 11amide; 1180 cm-1 , 11Os=N. 1H NMR: see Table 2.7. [Os(17 4 HBA-B)12, 18. To a degassed benzene solution of 2 g of§, 2 ml of perfluorophenyl azide were added via syringe. The solution was stirred at room temperature for 2 hours, during which the green solution became black. The reaction was monitored by TLC and allowed to continue until the starting material had all disappeared. The solvent was removed on a rotary evaporator and dried in vacuo. The black residue was insoluble in all solvents with the exception of hot DMSO and could not be mulled for an IR spectrum. No 1H or 19 F NMR spectra could be recorded of the black residue. Anal. calcd. for Os2C40H24OaN4 : C, 44.94; H, 2.26; N, 5.24. Found: C, 44.94; H, 2.93; N, 4.39. [Os(17 4 -HBA-B)(DMSO)]2•2H2O, 19. ~ was dissolved in hot DMSO and allowed to cool slowly. Black crystals were collected and dried in vacuo. Anal. calcd. for Os2C44H40O12N4S2: C, 41.90; H, 3.20; N, 4.44. Found: C, 41.95; H, 2.91; N, 4.43. IR (Nujol): 1650, 1602 cm-1, 11amide; 925, 940 cm-1, 11O-bound DMSO. 1H NMR: see Table 2. 7. "[Os(17 4 -HBA-B)(CO)12", 20. CO gas was slowly bubbled through a suspension of 50 mg of~- Over the course of three hours, the solution changed from a black suspension to a brilliantpurple homogeneous solution. The reaction was monitored by TLC (elution with MeCl2) and was complete after three hours. TLC showed the presence of a single, purple compound. The solvent was removed on a rotary evaporator. Further workup involved drying the purple 78 compound in vacuo, which apparently pumped away the CO ligands. No analytical or 1 H NMR data were obtained. IR (Nujol) on crude purple material: 1695, 1635 cm-1, vamide; 1980 cm-1, vM-CO. [Os(77 4 -HBA-B)(t-Bupy)x(solvent)2-x12 (x=2 or 1), 21 or 22. 75 mg of 18 were allowed to stir in a benzene suspension with an excess of t-butylpyridine. The solution slowly changed from black to brown over the course of several hours. The solvent and excess t-butylpyridine were removed in vacuo. Attempts were made to recrystalize the residue from DMSO and other solvents. Anal. calcd. for Os2C5aH50OaN5: C, 52.01; H, 3.76; N, 6.27. Found: C, 52.71; H, 4.27; N, 5.57. [NEt4 ][0sN(77 4 -HBA-B)], 23. 700 mg of 14 were stirred in a acetone/MeOH solution with 250 mg (2 eq.) of NEt4CI at room temperature for 24 hours. The solvents were removed on a rotary evaporator and the yellow residue was dried in vacuo. The solid was redisolved in THF from which a white solid (PPh3=NH2Cl) precipitated. The yellow mother liquor was decanted away from the solid and dried in vacuo. Pure 23 was obtained by preparative TLC by elution with MeCl2/THF in 80% yield. IR (Nujol): 1600, 1619 cm- 1, vamide; 1067 cm- 1, vOs=N. 1H NMR: see Table 2.7. "Os(NH)(77 4 -HBA-B)", 24. A degassed MeCl2 solution of 23 was cooled to 78 ° and one equivalent of HOTf was added via syringe. The solution was allowed to warm to room temperature, and a white precipitate fell out of solution immediately. The yellow mother liquor was decanted away from the solid and rinsed with MeCl2 and THF. The solvent was removed in vacuo, leaving a yellow residue. The white precipitate was identified as NEtOTf by 1 H NMR. The yellow solid was purified by preparative TLC by elution with THF/MeCl2. Only the major fraction (Rt=1) was collected. No analytical data were obtained. IR (Nujol): 1620, 1600 cm-1, vamide; 1120 cm-1, vOs=N. 1H NMR: see Table 2.7. 79 NEt4 0s(N=PPh3)(77 4 -HBA-B), 25. An acetone solution of 300 mg of 23 was gently heated on a hot plate (ca. 45°) for .5 hour, during which the solution changed from yellow to red-brown. TLC (elution with THF/MeCl2) showed there to be several products. The solution was poured into hexane and the resultant brown powder was collected by filtration. Preparative TLC (elution with acetone/MeCl2) allowed the collection of the major product (Rt= .7) in 60% yield. This step was repeated several times with an overall yield of about 30%. Anal. calcd. for OsC45H47N4O4P : C, 58.71; H, 5.03; N, 5.95. Found: C, 52.65; H, 3.52; N, 4.58. IR (Nujol): 1601, 1617 cm-1, !/amide; 1142 cm· 1, 1/P=N. 31 P NMR: 5+37 (s), vs. 5-4 (PPh 3). 1H NMR: see Table 2.7. The powder was recrystallized from acetone which afforded x-ray quality crystals for which the structure was analyzed as NaOsN(77 4-HBA-B) • 2 O=PPh3. It was subsequently determined by 31 P NMR that PPh3 was oxidized spontaneously to O=PPh3 by 25. "NEt4 0sN('7 4 HBA-B)(CO)", 26. CO was allowed to bubble slowly through a dry THF solution of 411 mg of 23 for several hours at room temperature as the solution slowly changed from yellow to orange. The solvent was removed on a rotary evaporator, leaving an orange oil. IR (Nujol): 1600, 1615 cm· 1, vamide; 1765, 1700, !/CO(?); 1090 cm· 1, vOs=N. Further workup procedure, including in vacuo drying, resulted in the loss of the IR bands at >1700, and the compound was not further characterized. [NEt 4 ][0sN(77 4 -HBA-B)(py)], 28. 400 mg of 23 were stirred in neat pyridine and gently heated on a hot plate (ca. 40°). There was an immediate color change from yellow to deep-orange. The reaction solution was poured into hexanes and a brown-orange powder was collected. The solid was filtered and washed with hexanes and recrystallized from MeCl2 and hexane. Yellow crystals were obtained in 60% yield. Anal. calcd. for OsC33H37O4N 5 : C, 52.30; H, 4.92; N, 9.24. Found: C, 52.12; H, 5.03; N, 8.96. IR (Nujol): 1602, 1590 cm·1, vamide; 1180 cm-1, l/Os=N. 1 H NMR: see Table 2.7. 0s(NCH 3)(77 4 -HBA-B), 29. To a dry MeCl2 solution of 300 mg of 23, 44 µI (1 eq.) MeOTf were added via syringe and allowed to stir at room temperature for 1.5 hours. The solution turned 80 from yellow to orange and would turn brown if left for longer periods of time. The solvent was removed in vacuo, the residue was redissolved in THF and subsequently columnchromatographed by elution with THF/hexanes. An orange solid was isolated as the major product in 80% yield. Yellow-orange crystals were obtained from acetone. Anal. calcd. for OsC21H15O4N3: C, 44.76; H, 2.68; N, 7.46. Found: C, 44.27; H, 2.73, N, 7.27. IR (Nujol): 1600, 1635 cm·1, vamide; 1105 cm·1, vOs=N. 1H NMR: see Table 2. 7. [NEt4 ][0sN(77 4 -HBA-B)(py)], 30. 100 mg of 28 were stirred in neat Mel at room temperature. After .5 hour, a yellow precipitate was formed. The solvent was removed in vacuo and the residue was analyzed by TLC (elution with THF), which showed the presence of several products. The residue was dissolved in THF from which a white powder was collected and identified by 1H NMR as NEt4I. Preparative TLC under the same conditions allowed the collection of three products. One of the products was identified as 23. One of the other products was identified as N-methylpyridinium iodide by comparison with an independently prepared sample. No other products were completely identified. 81 Table 2.7. 1 H NMR data Compound Solvent PAC Ligand 1 acetone-d5 7.6-8.0 (m, 4H) Ligand 7.2-7.5 (m, 4H) 6.7-7.0 (m, 4H) z acetone-d5 6.4-6.9 (m, 8H) 8.3 (dd, 2H) 9.1 (dd, 2H) ~ CDCl3 ·4,33 (dd, 2H) (PPh3) ·3_95 (dd, 2H) 8.28 (d, 12H) 1.33 (d, 2H) 6.99 (t, 12H) 4.27 (dd, 2H) 7.36 (t, 6H) 6.99 (m, 2H) 8.19(m,2H) ~ CDCl3 8.34 (d, 2H) (NHC5H4NC5Hs) 8.05 (d, 2H) 6.9 (d, 1H) 7.3 (m, 2H) 6.45 (t, 1H) 7.05 (m, 4H) 6.36 (d, 1H) 6.83 (t, 2H) 6.24 (d, 1H) 7.05 (m, 3H) 11.65 (b, 1H) 82 Compound Solvent PAC Ligand Ligand 14 acetone-d5 9.0 (dd, 2H) (PPh3) 8.28 (dd, 2H) 7.26 (m, 15H) 6.87-7.23 (m, SH) 1l acetone-d5 (PAC+ PPh3) (N-C10H15) 9.11 (d, 2H) 1.79(m,16H) 8.80 (d) 8.25-8.18 (m) 7.41 (m) 6.96 (m) } = 25 H 21 CDCl3 8.05 (m) (DMSO) 7.5 (m) 2.5 (b, 12H) 7.02 (m) (H2O) 6.55 (m)} = 12H 2.87 (b, 4H) (PAC +,t-Bupy) (DMSO) 7.3 (d, 2H) 2.5 (b,12H) 6.5 (t) (i-Bupy) 6.2 (m) 2.05 (s, 4.5H) 5.6 (m) 4.9 (dd) 1.8 (d) 1.2 (m) .95 (m)} = 15H 83 Compound Solvent PAC Ligand Ligand 23 acetone-d5 9.15 (dd, 2H) (NEtt) 8.3 (dd, 2H) 3.15 (q, SH) 7.6 (m, BH) 1.2 (td, 12H) 8.94 (dd, 2H) (N-H)? 8.15 (dd, 2H) 4.9 (b, 2H) 24 DMSO-d5 7.0 (m, BH) acetone-d5 9.1 (dd, 2H) 1.2 (s, 1 H) 8.2 (dd, 2H) 7.1 (m, BH) acetone-d5 (PAC+ PPh3) (NEt4+) 9.12 (dd, 2H) 3.4 (q, BH) 8.27 (dd, 2H) 1.2 (td, 12H) 6.86-7.91 (m, 23H) 28 acetone-d5 (PAC+ py) (NEt4+) 9.1 (dd, 2H) 3.15 (q, BH) 8.45 (m) 1.12 (td, 12H) 8.25 (dd) 7.65 (m) 7.0 (m) } = 15 H acetone-d5 9.16 (dd, 2H) (N-CH3) 8.28 (dd, 2H) 4.3 (s, 3H) 7.27 (m, BH) 84 References 1. (a).Schrock, R., Science, 1983, 219, p.13. (b). Nugent, W., Mayer, j_, Metal-Ligand Multiple Bonds, John Wiley & Sons, New York, 1988. 2. Griffith, W., Coord. Chem. Rev., 1982, 5, 459. 3. Nugent, W., Haymore, 8., Coord. Chem. Rev., 1980, 31, 123. 4. Griffith, W., Coard. Chem. Rev., 1972, 8, 369. 5. Kobayashi, C. Clifford, A., C.S. Abstracts, 130th National Meeting of the A.C.S., Atlantic City, New Jersey, Sept. 1956, 50R. 6. Griffith, W., Pawson, D., J.C.S. Dalton, 1973, 1315. 7. Haymore, 8., Matta, E., Wentworth, R., J. Am. Chem. Soc., 1979, 101, 2083. 8. Peake, G.T., Ph.D. Dissertation, California Institute of Technology, Pasadena, California, 1987. 9. Cotton, F.A., Wilkinson, G., Advanced Inorganic Chemistry, John Wiley & Sons, New York, 1980. 10. Cenini, S., La Monica, G., lnorg. Chim. Acta, 1976, 18, 279-293. 11. Goeden,G., Haymore, B., lnorg. Chem., 1983, 22, 157. 12. Rocklage, S.M., Schrock, R., J. Am., Chem., Soc., 1980, 102, 7809. 13. Harlow, R.L., Nugent, W. J. Chem. Soc. Chem. Commun., 1978, 578. 14. Chong, A., Oshima, K., Sharpless, K.B., J.Am. Chem., Soc., 1977, 99, 3420. 85 15. Stewart, W.E., Sidall, T.H., Chem. Rev., 1970, 70, 517-551. 16. Costain, C.C., Dowling, J.M., J. Chem. Phys., 1960, 32, 158-165. 17. Woodward, R.B., Recent Advances in /3-Lactam Antibiotics, Chemical Society, London, 1977, 167-180. 18. Coqueret, X., Bourelle-Wargnier, F., Chuche, J., J. Org., Chem., 1985, 50, 910-12. 19. (a) Dunitz, J.O., Winkler, F.K., J. Mo/. Biol., 1971, 59, 169-82, and references therein. (b) Dunitz, J.O., Winkler, F.K.,Acta Crystallogr., Sect. B., 1975, 831, 251-63. 20. Collins, T.J., Coots, R., Furutani, T., Keech, J., Peake, G., Santarsiero, B., J. Am. Chem. Soc., 1986, 108, 5333. 21. Gross, M., lbers, J., Trogler, W., Organomet., 1982, 1, 530-535. 22. Patai, S., Ed., The Chemistry of the Azido Group, lnterscience, London, 1971, 260-1. 23. Gross, M., Trogler, W., J. Organomet. Chem., 1981, 209, 407. 24. Smith, J., Green, M., Stone, F., J.C.S. Dalton, 1972, 1805. 25. Cenini, S., La Monica, G., lnorg. Chim. Acta, 1976, 18, 279. 26. Cowman, C., Trogler, W., Mann, K., Poon, C., Gray, H.B., lnorg. Chem., 1976, 15, 1715. 27. (a) Phillips, F., Skapsi, A., J. Cryst., Mo/., Struct., 1975, 5, 83. (b) Bright, 0., lbers, J., lnorg. Chem., 1969, 8, 711. 28. Belmonte (Shapley), P., Own, Z.Y., Huffman, J., Organomet., 1986, 5, 1269. 29. Oubois,D., Hoffmann, R., Nouv. J. Chim., 1977, 1, 479. 30. Corbridge, 0., Phosphorous, Elsevier, Amsterdam, 1980, 26. 86 31. Chatt, J., Dilworth, J., J. Indian Chem. Soc., 1977, 54, 13. 32. Chatt, J., Falk, C., Leigh, G., Paske, R., J. Chem., Soc.(A), 1969, 2288. 33. Bereman, R., lnorg Chem., 1972, 11, 1149. 34. Gololobov, Y., Zhmurova, I., Kasukhin, L., Tetrahedron, 1981, 37, 437. 35. Sharpless, K.B., Patick, D., Truesdale, L., Biller, S., J. Am. Chem., Soc., 1975, 97, 2305. 36. Gross, M., Johnson, C., Maroney, M. Trogler, W., lnorg Chem., 1984, 23, 2968. 37. Hazeldine, R., J. Chem. Soc., 1962, 4966. 38. Evans, I., Spencer, A., Wilkinson, G., J. Chem. Soc. Dalton, 1973, 204. 39. Fanwick, P., Fraser, I., Tetrick, S., Walton, R., lnorg. Chem., 1987, 26, 3786-91. 40. Cotton, F., Walton, R., Multiple Bonds Between Metal Atoms, John Wiley & Sons, New York, 1982. 41. Cotton, F., Thompson, J., lnorg. Chim. Acta, 1980, 44, L247. 42. Collman, J., Prodolliet, J., Leidner, C., J. Am. Chem. Soc., 1986, 108, 2916-21. 43. Trogler, W., Gray, H.8.,Accts. Chem. Res., 1978, 11,272; H.B. Gray, personal communication. 44. Collman, J., Brauman, J., Iverson, B., Sessler, J., J. Am. Chem. Soc., 1981, 103, 2450. 45. Chatt, J., Dilworth, J., Richards, R., Chem. Rev., 1978 78, 589-625. 46. Grasselli, R., Burrington, J., Bradzil, J., J. Chem. Soc.,Dalton Trans., 1981, 205-223. 47. Brulet, C., van Tamelen, E., J. Am. Chem. Soc., 1975, 911. 87 48. Johnson, B., Lewis, J., Mace, J., J. Chem. Soc., Chem. Commun., 1984, 186. 49. Chan, D., Nugent, W., lnorg. Chem., 1985, 24, 1424. 50. Groves, J., Takahashi, T., J. Am. Chem. Soc., 1983, 105, 2073. 51. Pawson, D., Griffith, W., lnorg Nucl. Chem. Lett., 1974, 10, 253. 52. Pawson, D., Griffith, W., J. Chem. Soc. Dalton, 1975, 418. 53. Bishop, M., Chatt, J., Dilworth, J., J. Chem. Soc., Dalton Trans., 1979, 1. 54. Belmonte, P., personal communication. 55. Belmonte (Shapley), P., J. Organomet. Chem., 1987, 335,269. 88 PART II: ASYMMETRIC HYDROGENATION USING A RESOLVED CHIRAL SCANDIUM HYDRIDE COMPLEX CHAPTER ONE: INTRODUCTION 89 Introduction Asymmetric synthesis has been an active research area and a field full of challenge for the chemist since the discovery of stereoisomerism in the late 19th century. 1 Amino acids, carbohydrates and nucleotides all exist in nature in predominantly one enantiomeric form. 1 Despite the prevalence of chiral compounds and their recognized importance, it was not until the advent of asymmetric hydrogenation in the mid-1950's that chemists began to be able to produce enantiomerically pure compounds catalytically without enzymes. 2 Transition metal complexes have played a major role in the development of asymmetric hydrogenation catalysts. 2 Early synthetic approaches produced compounds that catalyzed asymmetric reactions that achieved only modest enantiomeric excesses (e.e.) of 10-15%. 2 Since that time a great deal of research effort has been devoted to this area and a degree of sophistication has been achieved; for example, chiral organic transformations mediated by transition metal complexes have been incorporated into the total synthesis of complex natural products. 3 An advantage of using transition metal complexes for asymmetric synthesis is that the reactions can be "tuned" by varying the ligand environment of the metal, making it possible to produce different catalysts for specific substrate reactions. The greater part of the work in this area has been with chiral phosphine ligands, a field that was opened up by two simultaneous events: the discovery of a rhodium hydrogenation catalyst by Wilkinson in 1965,3 and the development of a method to synthesize a wide variety of chiral phosphines by Mislow. 4 These discoveries led to the synthesis of hundreds of chiral phosphine complexes of rhodium or other metals for use as asymmetric hydrogenation catalysts and in other asymmetric reactions .5 Although the chiral phosphine complexes have achieved success as asymmetric catalysts, there are some disadvantages associated with using chiral phosphines as ligands. In order for the efficient transfer of asymmetry to the substrate to occur, the chiral ligands must remain bound to the metal during the stereodifferentiating step; phosphine ligands have traditionally been 90 recognized as being quite labile. 6 In an effort to prepare more robust asymmetric catalysts, new chiral ligand types have been sought.7 One approach that has been taken involves incorporating chirality into common transition metal ligands such as the cyclopentadienyl (Cp) moiety.7 Cp ligands offer at least one advantage over phosphine ligands in that they form strong T/ 5-bonds to the metal. Another advantage is that chirality can be incorporated into Cp ligands, using cheap and readily available natural products from the "chiral pool." Chiral Cp menthyl, neomenthyl (Cpmen,1, Cpnmen,g)• 9 ·10 ,11 ,12 and phenylmenthyl (Cppmen,~_)6, Cp camphor(Cpcam,1) 6 •13 , Cp pinene (Cppin,§) 13 and Cp verbenone (Cpver,_g_) 14 (see Figure 1.1) complexes have been prepared with titanium, iron or ruthenium. ~ 1 ¢io ~- HsC CHs g ~ HsC-tCH1 c,He ~ Figure 1.1. Chiral Cp ligands. Several of the chiral Cp complexes have proved to be active asymmetric hydrogenation catalysts. For example, Cesarotti et. al. reported that (Cpmen or Cpnmen) 2TiCl 2 , l, and §_, (see Figure 1.2) will hydrogenate 2-phenyl-1-butene to 2-phenylbutane in the presence of a 91 cocatalyst, UAIH2(OR}2, with a modest e.e. ranging from 1-28% ( see Table 1.1 ). These catalysts do not appear to be very general, as the other prochiral olefins tested, a-ethoxy and amethoxystyrene, were not hydrogenated. 11 TIC 11 TiC 11 2 2 1 ~ Figure 1.2. Chiral Cpmen or Cpnmen catalysts. Cataly1t Temp. O.Y. and aha. conf. of 2-phenylbutane (OC) 7 8 LiAIH2( OCH2CH2 OCH3)2 LiAlH 2 (0tBut )2 -5 +5 +20 +30 +40 +20 22 28 15 16 10 7 (S)-(+) (S)-(+) (S)-( +) (S)-( +) (S)-{ +) 15 16 11 (S)-{+) (S)-(+) (S)-(+) (B)-(-) Table 1.1. Data for the asymmetric hydrogenation of 2-phenyl-1-butene. Paquette et al. have reported that (CpPi")2Ti Cl 2, ~. or (Cpcam)2 TiCl2, 1..Q, (see Figure 1.3) will hydrogenate a:-ethylstyrene or 2-ethyl-1-hexene to 2-phenylbutane or 3-methylheptane, repectively, in the presence of n-butyllithium or UAIH2(OCH2CH2OCH3}2, with an e.e. ranging from 0.1-34% (see Table 1.2).13 92 1Q Figure 1.3. Chiral (CpPi")2TICl2 and (Cpcam) 2Ti 2 catalysts. I enanti0111eric excess Cabs. config.> of activation Method r-, 9 10 200c 200c -200c 200c oetc -200 Red-Al .11-SuL f .11-SuLt Red-Al Red-Al .11-SuLtb ~ o. 1 ca> ~5H5 1. 7 C.S.) 2 C.S.) 6.9 (.S,) 2.2 (.S,) 14.1 <a> 23. 1 <a> 34 Table 1.2. Data for the asymmetric hydrogenation of a-ethylstyrene and 2-ethyl-1-hexene. Vollhardt et al. have achieved a higher e.e. for the hydrogenation of a-ethylstyrene to 3methylheptane using the catalysts (Cppmen)2TiCl2, 11, or (Cpcam)2 TiCl2, j_g, in the presence of n- 93 ·:~ H~ r H~Clz TICla H3C 11 CHs 12 Figure 1.4. Chiral Cppmen or Cpcam catalysts. The optical yields* reported were between 22 and 34%, the highest yet achieved for the asymmetric hydrogenation of simple unfunctionalized olefins (see Table 1.3). 6 H1 , cotolytt, tolue11e cat. precunor inductn optical time temp, °C time yield," 11 lb 20 20 0 40h 20min 40min Sh 22 25 12 2min 2min 2min -20 33 34 Table 1.3. Data for the asymmetric hydrogenation of a-ethylstyrene. *Note: Optical yield, optical purity and enantiomeic excess (e.e.) are commonly used synonymously; however, the term e.e. is defined as %R-%S = e.e.% and is more correct for determinations made by NMR; optical purity (o.p.) is defined as o.p. = [a]/[amax] · 100% and is more correct for determinations made by polarimetry. 1 ·3 ·5•7 The terms are used here as the original authors reported them. 94 While these reactions mark the beginning of a fruitful area of research in asymmetric hydrogenation catalysts prepared with chiral Cp ligands for simple olefins, there is much room for new developments. The stereoselectivity that has been achieved so far, while improving, is still not in the range that would make these catalysts useful for commercial processes. Reported here are further efforts to prepare a useful catalyst. The main goal of this research project was to prepare a resolved chiral Cp scandium hydride complex and to explore its capability in asymmetric hydrogenation of an unfunctionalized olefin. 95 References 1. Koenig, K., The Applicability of Asymmetric Homogeneous Catalytic Hydrogenation, "Asymetric Synthesis," 1985, 5, pp.71-101. 2. Bosnich, B., Fryzuk, M., Asymmetric Synthesis Mediated by Transition Metal Complexes, "Topics in Stereochemistry," 1981, 12, pp.119-154. 3. Osborn, J., Jardine, F., Young, J., Wilkinson, G., J. Chem. Soc. A, 1966, 1711. 4. Korpi um, 0., Lewis, R., Chickos, J., Mislow, K., J. Am. Chem. Soc., 1968, 90., 4842. 5. Brunner, H., Enantioselective Synthesis of Organic Compounds with Optically Active Transition Metal Catalysts in Substoichiometric Quantities, "Topics in Stereochemistry," 1988, 18, pp.129-24 7. 6. Halterman, R., Vollhardt, P., Organomet., 1988, 7, 883-892. 7. Kagan, H., Chiral ligands for Asymmetric Catalysis, "Asymmetric Synthesis," 1985, 5, pp.1-39. 9. Cesarotii, E., Ciani, G., Sironi, A., J. Organomet. Chem., 1981, 216, 87-95. 10. Cesarotti, E., Kagan, H., Goddard, R., Kruger, C., J. Organomet. Chem., 1978, 162, 297-309. 11. Cesarotti, E., Ugo, R., Vitiello, R., J. Mo/. Cata/., 1981, 12, 63-69. 12. Cessarotti, E., Chiesa, A. Ciani, G., Sironi, A., J. Chem. Soc. Dalt. Trans., 1984, 653. 13. Paquette, L., McKinney, J., McLaughlin, Rheingold, A., Tet. Lett., 1986, 27, 5599-5602. 14. Moriarity, K., Rogers, R., Paquette, L., submitted for publication to J. Organomet. Chem., 1989. 96 CHAPTER TWO: ASYMMETRIC HYDROGENATION WITH 97 Introduction Permethylmetallocene complexes of lanthanides and group 3 metals of the general formula (Cp *)2MH (where M =Sc, La ,Sm, Lu or Nd and Cp * =(TJ 5-C5Me5)) have been shown to function as hydrogenation catalysts for simple olefins (see Equations 2.1 and 2.2) .1 ·2 (2.1) ~ + H2 M • Lu, Sm, Nd, La (2.2) Cp*z$c-CH3 + ===:/ - CH4 Cp*zSc~ -~ 1 + ~ Cp* 2Sc-H + ==::=/ Cp* 2 S c ~ The scope of reactivity of these compounds has been somewhat limited by the steric constraints imposed by the large Cp * ligands. In an effort to reduce steric bulk, a new class of Cp ligands was synthesized with a dimethylsilylene bridge linking the two Cp ligands. 3 This approach was successful in producing a smaller ring centroid-metal-centroid ring angle (the angle is smaller by 14.7° relative to Cp*2ScCH3) and in opening up a "wedge" of reactivity in the coordination sphere of the metal. 2.4 This structure was established by an x-ray diffraction study (see Figure 2.1 ). 4 Several SiMe2-bridged Cp2 derivative ligands have been prepared and complexed to scandium, and the chemistry has been extensively explored. 4 It has been demonstrated that butylC5H3)2(H))2 ((DpScH)2), will promote the catalytic dimerization of a:-olefins and the catalytic cyclization of a,w-diolefins. When the cyclization of a:-w reactions is performed in the presence of hydrogen, the final products are hydrogenated (see Scheme 2.1.). 4 98 • 2.258A Figure 2.1. The molecular structure of (Me2Si(,, 5-C5Me5)2)ScCH(SiMe3)2 showing the open "wedge." 99 [DpScHk [DpScHk For [DpScHJ2, Aa 33% and B,,. 66% For [OpSc(H)PM&:3), A. 45% and 13: 54% 55% 45% Scheme 2.1. (DpScH)2 and OpScH(PMe3) will dimerize, and cyclize a-olefins and a,w-olefins, respectively. It is the purpose of the present work to show that it is possible to extend this reactivity to include the asymmetric hydrogenation of a prochiral olefinic substrate by incorporating chirality into one of the Cp ligands. 100 Results and Discussion An "open-wedged" chiral scandocene complex was prepared by designing a ligand that utilizes a Me2Si-bridge and incorporates an available resolved organic compound (1 R,2R,5S)menthyl, from the chiral "pool." The synthesis of this ligand is straightforward and is shown in Scheme 2.2. Reduction of 2,3,4,5-tetramethyl-2-cyclopentenone with LiAIH 4 yields 2,3,4,5tetramethyl-2-cyclopentanol. Subsequent dehydration of the alcohol produces 1,2,3,4tetramethylcyclopentadiene in moderate yield. The lithium salt, 1, is prepared by deprotonation with n-butyllithium and treated with dichlorodimethylsilane to give Me2Si(C5Me4H)CI, g, which can be isolated as an oil. This starting material allows the synthesis of an asymmetric [Me2SiCp2] ligand. The second Cp ring to be added contains the chiral natural product, (1 R,2R,5S)-menthyl. This ligand can be synthesized readily, as shown in Scheme 2.2. Commercially available (1 S,2R,5S)-(+)-menthol is first converted to the (1 S,2R,5S)-(+)-menthyl tosylate derivative with p-toluenesulfonyl chloride. It is subsequently treated firstly with LiCp to produce the (1 R,2R,5S)-menthylcyclopentadienyl compound (Cpm), J, and secondly, treated with n-butyllithium to give the lithium salt, 1- 1 and g can be condensed to produce the mixed ring ligand, _g_, which is deprotonated to give the dilithium salt, §. §. can be complexed to scandium according to Scheme 2.3 to give the scandocene complex, I, that has a chloride as well as a solvent (THF) of coordination present. A more convenient starting material,§, can be prepared from I by metathesis of the chloride with the bulky lithium reagent, (Me3Si)2CHLi (see Scheme 2.3). The presence of the chiral menthyl group provides the opportunity for two diasteriomers (§ and §') to form, which can be distinguished by the inequivalency in the 1 H NMR spectrum of the two trimethylsilyl (ScCH(SiMe3)(SiMei)) groups. The integration of the four trimethylsilyl methyl resonances indicates that the diasteriomers are formed in roughly equal proportion with only a slight excess of one (8%). 4 101 0 + n-Buli + liCp + p-TsCI OH OTs /---+ n-BuLi Scheme 2.2. Synthesis of ,6.. Me,Sp Iµ'), \-A D .e + Sc:Cla"3Tlf / Cl ~ Me2Si, 'Sc ""THF +~Sitv1e-.ih _ LiC1 A( . Me2S1, Si~ Sc__/· f"siMe3 H l !! (and!!' __. 0 I\) Scheme 2.3. Synthesis of!! and !!'. 103 In related work, Brintzinger et al. have reported the formation of diasteriomeric zirconium complexes with dimethylsilylene-bridged, symmetrically disubstituted Cp rings, as shown in Figure 2.2. 5 They were able to separate the racemic and meso isomers by fractional recrystallization and carried out an investigation of the effect of alkyl Cp-substituents on the isomeric ratios. R R rac RIR2- rac/meso C(CH 3 h H 1/1 mesa Si(CH 3 h H 1/1 C(CH 3 )iC6H5 H 1/1 C(CH 2)sC6Hs H 1/1 C(CH 3 ) CH 3 2/1 Figure 2.2. Racemic and meso isomers of dimethylsilylene-bridged zirconocenes. When§. and§.' are stirred in benzene under 4 atmospheres of hydrogen overnight, a pair of diasteriomeric isomers of the scandium hydride complex are formed,~ and ~• (see Scheme 2.4). These are easily distinguished by the inequivalency of the three downfield protons on the Cpm ring. When allowed to stir for longer periods of time (up to two weeks), one of the hydride diasteriomers converts to the other, which will be (arbitrarily) assigned as~- If rapid stirring does not occur, or if there is insufficient hydrogen present, a small amount of a third complex appears (see Scheme 2.4). The 1 H NMR spectra for the mixture of diasteriomers and the resolved diasteriomer are shown in Figure 2.3a and 2.3b, respectively. ¥ Me,Si, SiMe, /ff + ~ , wemight Sc-H Me2Si-. Me,Si#.H !lc--;?(SiMe, + H fl. I 9 "1Q." + Me,s#.H I+~ __J 9 Scheme 2.4. Synthesis of ~ and ~• and the conversion of ~• to ~- ca 2 weeks _._ 0 .ll,,. 105 a.) 7 .o 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 PPN 1.5 2.0 1.0 0 .0 b.) 8.00 7.00 6.00 s.oo 4.00 PPM 3.00 2.00 1.00 0.0 Figure 2.3a and 2.3b. 1 H NMR spectra of both isomers~ and ~• (a), and only one isomer, ~· (b). 106 It is not known whether~ and iare monomeric or dimeric; however, the related complex [DpScH]2 is formulated as a dimeric species based on a molecular weight determination made by ebulliometry. 4 "[OpScH]x" is not stable in the absence of a trimethylphosphine ligand. 4 Unfortunately,~ is not stable in solution in the absence of hydrogen for the required time period to make a molecular weight determination. It is formulated here as a monomer because of the extremely bulky Cpm ligand, but the possibility exists that it is in fact a dimeric species. The hydrogen-promoted conversion to one scandium hydride diasteriomer is quite fortuitous and yields a resolved chiral complex without the necessity of fractional recrystallization, chromatography or other techniques for diasteriomeric separation. In an effort to understand exactly how this resolution occurs, mechanistic studies were undertaken. It was initially proposed that the interconversion of ~·to ~ involved a simple H Sc-toCp migration. Were this the case, then if~ and~• were stirred under 4 atmospheres of deuterium, deuterium would be washed into the ring protons on the Cpm ligand (see Scheme 2.5). Accordingly,~ and~• were allowed to stir under a deuterium atmosphere for an extended period of time (up to 2.5 months) with periodic recharging of deuterium. The reaction was monitored by 1 H and 2 H NMR. The isomerization was considerably slowed, and the only incorporation of deuterium that was seen was into the isopropyl group of the menthyl moiety. This does not appear to have anything to do with the isomerization process and is probably a secondary reaction. In addition to the deuterium labeling experiment, a small amount of acid was added to the reaction to determine whether acid catalysis played a role. The addition of even trace amounts of HCI produced an entirely different product and this was not pursued. A( Me&¥ M e , S i ~ H(D) 2 • Sc ~ ,H(D) I Me2Si~···~ "S:c ,. 9.' ,Hx:_ H(O) ,,,... ,,.... ..... 0 ...... ll 4(_H Me,sw . .•·~ J !! ~ H j D J Scheme 2.5. The mechanism initially proposed for the isomerization of~· to ~- 108 In an effort to determine whether light might play a role in assisting one of the Cp ligands to undergo homolysis and subsequent rotation and recoordination, three experiments were done under controlled conditions. A mixture of~ and ~• were sealed in NMR tubes in benzene-d5 under 4 atmospheres of hydrogen. One tube was exposed to lab light only. A second tube was kept in the dark, and the third tube was irradiated for 2.25 hours with a 1000 Watt lamp. The irradiated sample showed no isomerization, but did show considerable hydrogenation of the benzene-d5 solvent. The samples that were kept in the dark and exposed to normal lab light began to isomerize in an identical fashion. Thus, it does not appear that a light-induced Sc-Cpm homolytic cleavage mechanism is operative. At this point, there is no positive evidence for an explanation of the mechanism of conversion from ~•to~; what is known is that hydrogen is necessary for the conversion, the mechanism does not involve an H Sc-to-Cp migration, it is not acid-catalyzed and light does not promote the isomerization. A suggested mechanism for the isomerization is shown in Scheme 2.6. For the isomerization of i to~ to occur, the Sc-Cpm bond in ~· must be broken and the ligand allowed to rotate about the dimethylsilylene bridge-Cpm bond. Recoordination of the opposite face of the Cpm ligand affords the other diasteriomer, ~- With the resolved chiral scandium hydride complex, ~. in hand, it was necessary to determine if~ would promote the same kinds of olefin dimerization, cyclization and hydrogenations as (DpScH)2 and OpScH(PMe3). Accordingly, ~ was treated with 20 equivalents of propane in a sealed NMR tube, as shown in Equation 2.3. (2.3) excess ~ [9] £ Me 2Si. Sc-H ¥._H Me2Si----... ~ ~~i/f(H ~ i ..... 0 (0 Scheme 2.6. A mechanism showing the steps that must occur for~• to convert to ~- 110 By comparison with authentic samples, it was established that~ will dimerize and hydrogenate propene to 2-methyl-1-pentene and 2-methylpentane. Apparently 2-methylpentane was formed as a consequence of the decomposition of~- With the reactivity established, it was necessary to select a prochiral substrate to determine if it would be possible to achieve chiral control over these reactions. Initially, the prochiral diolefin 3-methyl-1,5-hexadiene was selected as the substrate of choice; ~ was treated with 20 equivalents of the substrate according to Equation 2.4. (2.4) + Comparison of the 1 Hand 13 C NMR spectra with authentic samples showed the products to be 2-methyl-exo-methylene cyclopentane and 3-methyl-exo-methylene cyclopentane, as expected. Unfortunately, the choice of 3-methyl-1,5-hexadiene as a test substrate presented many problems. The ring-closure step can occur from either face, creating one chiral site; however, the chiral center in the 3-methyl position may be too far removed from the reaction site to exert any great influence, and the subsequent decomplexation of the substrate produces an identical compound (see Scheme 2. 7). On the other hand, hydrogenation after the cyclization step, which is also possible from either face, creates another opportunity for chiral control. Unfortunately, there are numerous possible chiral products (see Scheme 2.7). A further problem with this system is that the only way to establish that any steric control over the products has occurred is by polarimetry. Although this method is commonly used to establish enantiomeric purity, it is not ideal. Enantiomeric purities are established by comparison with samples of known rotation, but small amounts of contamination can substantially change the rotations leading to spurious results. 5 Separating enantiomers by use 9. + ~ # Sc I Sc Sc \ /', Sc/ '•··•... - ScH + + ScH I~ + ScH - ScH i /', Sc.,,..,.- '•·--.. + + ..... ..... ..... H2 - ScH .. ----;:- + + H2 + - ScH ,,,, ~-... + Scheme 2. 7. The numerous possible chiral products from the hydrogenation of 3-methyl-1,5hexadiene by ~- 112 of lanthanide shift reagents and establishing the enantiomeric purity by integration is considered much more reliable. 5 Furthermore, the only rotational standard available for this system is (+)-3-methyl-cyclopentanone. Because of these problems, another substrate was sought. Divinylbenzene was selected as the next substrate to examine. Cyclization of divinylbenzene would create a chiral site with direct participation by the metal center ( see Scheme 2.8). Unfortunately, divinylbenzene is not available commercially in a pure state, and all attempts to further purify it failed. Attempts were also made to synthesize divinylbenzene from the corresponding diketone and the appropriate Wittig reagent. It was not possible to obtain sufficiently pure divinylbenzene by this method either. Fortunately, however, it was not necessary to start with divinylbenzene. As shown in Scheme 2.8, the product of the cyclization of divinylbenzene is methylene indane, and it is the hydrogenation of this that is the stereodifferentiating step of the reaction. The intermediate complex .11 is identical whether or not the starting material used is divinyl benzene or methylene indane, so it is quite possible to start the asymmetric reaction with methylene indane rather than with divinyl benzene. Accordingly, methylene indane was synthesized from the corresponding ketone by treatment with Wittig reagent (see Scheme 2.8). Numerous attempts were made to produce very pure methylene indane. Success was achieved by never allowing the reaction to rise above room temperature. The purity of the methylene indane was assayed by GC and IR. The reaction of~ with methylene indane under 4 atmospheres of hydrogen proceeded smoothly to yield methyl indane. A maximum of four turnovers was achieved. In order to determine that the hydrogenation was accompanied by steric control, it was necessary to employ a silver/ chiral lanthanide combination of NMR shift reagents to distinguish the enantiomers. Methylindane is an ideal candidate for this kind of analysis because it has an aryl group that can bond strongly to a silver shift reagent (which will in turn bond to a chiral lanthanide shift reagent), and the methyl group on the indane ring is unique. 113 0 PPh 3 =CH 2 + ()j co -ScH Sc + ~ +9 # Sc Scheme 2.8. The asymmetric hydrogenation of methylene indane by 9... 114 In 1981 and 1984, Mannschreck and Offermann reported the silver/lanthanide-induced 1 H NMR splittings of enantiomeric signals for various alkene and arene hydrocarbons. 6 •7 (6,6, 7, 7,8,8,8-Heptafluoro-2,2,-dimethyl-3,5-octane-dionato)silver(I) (Ag(fod}), _1_g (see Figure 2.4) was demonstrated to form complexes with arenes and alkenes, which are in an equilibrium in solution that is fast on the NMR time scale. 8 The silver complexation shifts and spreads proton resonances to lower field. Chiral recognition is achieved by the addition of a chiral lanthanide shift reagent such as (+)-tris(3-heptafluorobutyryl-D-camphorato)ytterbium(III) (Yb(hfbc)s), 13 (see Figure 2.4). Yb(hfbc)s bonds to the silver complex of the alkene or arene and imparts chiral recognition despite the "dilution" of the optically active environment around the substrate. 6 n-C 3F7 'OI O)fb/3 Figure 2.4. 1 H NMR shift reagents, _1_g and 13. The initial studies to determine feasibility of distinguishing the enantiomers of methylindane by 1H NMR were done with the commercially available racemate of methylindane. After much experimentation with different lanthanide reagents and combinations of reagents and concentrations, optimal splitting of the enantiomeric methyl groups was achieved by using a CDCl 3 solution that was 0.1 M in methyl indane, 0.05 M in _1_g and 0.2 M in 13. The 1 H 500 MHz spectra of pure methylindane and the mixture are shown in Figure 2.5a and 2.5b, respectively. In order to determine the enantiomeric purity of the methylindane produced from the reaction of~ with methylene indane, it was first necessary to separate the organic product from the organometallic complex. The first attempt was made by distilling the volatiles away from the 115 a.) - I.I II.I le.• I.I 1,1 1,1 I.I 7.1 7.t a. ■ I.I I.I w 4.1 ◄ .I I.I J.t I.I I.I t.l I.I I.I 1,1 1,1 W 4.1 ◄ .I J.I I.I I.I I.I 1.I .. ... -., b.) - 1.7 It.I II.I 1,1 1,1 I, ■ I.I I.I 1.1 7,1 1.1 .I l,I -.I Figure 2.5a and 2.5b. The 1 H NMR spectrum of methylindane (a), and methylindane in solution with _lg and 13 (b). 116 reaction mixture; this failed because of the decomposition of the ligand into volatile organic fragments that were carried over by the distillation. It was necessary to quench the reaction mixture with air, reduce the solvent amount at 0°C, and run the solution through a silica gel plug or on silica gel plates by elution with a 70/20 mixture of hexanes and diethylether. Methylindane has an Rt value of .9 with this solvent system and can be easily recovered from silica gel. The recovered methylindane was dissolved as a 0.1 M solution in CDCl3 with 0.05 M 12_ and 0.1 M 13. (In this case the optimal separation was given by these ratios versus the methylindane standard, which was optimized at 0.2 M 1-,;1, vide supra). The 1 H NMR of the methyl region is shown in Figure 2.6. The best enantiomeric selectivity achieved was 67-33%, which is an e.e. of 34%, which is comparable to the best e.e.'s reported by other systems for asymmetric hydrogenations of simple olefins(see Chapter 1). Commercial methylindane was dissolved as a 0.1 M solution in CDCLJ with 0.05 M Ag(fod) and 0.1 M Yb(hfbc)s, and 1 H NMR spectra were obtained under the above conditions. The enantiomeric methyl groups of the methylindane integrated as 49-51%. In another control experiment, (DpScH)2 was also reacted with methylene indane under hydrogen pressure, subjected to the same work-up procedures and dissolved as a 0.1 M solution in CDCl3 with 0.05 12 and 0.1 M 13. The methyl peaks of the enantiomers integrated as 50.2-49.8%. These experiments clearly indicate that~ does impart steric control to the hydrogenation of a prochiral olefin. 67: 33 % ..... ..... -..J l f s..7 PPM Figure 2.6. An e.e. of 34% for the hydrogenation of methylene indane by ~. as shown by 1 NMR. 1.li 118 Conclusions It has been demonstrated that a resolved chiral scandium hydride complex, ~. can be synthesized from readily available starting materials, utilizing a chiral ligand available from the natural product "pool." It has been shown that this complex will induce the asymmetric hydrogenation of a prochiral olefinic substrate. An enantiomeric excess of 34% was determined by a powerful 1 H NMR technique utilizing a silver/lanthanide co-reagent system to separate enantiomeric signals. Although an enantiomeric excess of 34% is comparable to the best of other values reported for the asymmetric hydrogenation of a simple olefin, it is not high enough to be of use commercially. These kinds of processes become commercially feasible only as the enantiomeric excess approaches 90-100%. It may, however, be possible to increase the selectivity of the catalyst, ~. by optimizing conditions. In some cases, both temperature increases and decreases have been reported to result in greater sterioselectivity. 10 119 Experimental Materials All manipulations were carried out by using either high vacuum or glove-box techniques. 11 Hydrogen, argon and deuterium were purified before use by passing through a MnO over vermiculite and activated 4A sieves column. Propene (Matheson) was degassed by several freeze-pump-thaw cycles prior to use. Benzene, toluene, petroleum ether, 3-methyl-1,5hexadiene (Aldrich), and perdeuterobenzene were vacuum-transferred from LiAIH 4 or 4A sieves prior to use. Diethyl ether, tetrahydrofuran and perdeuterotetrahydrofuran were stored over benzophenone ketyl and vacuum-transferred before use. PPh3CH3Br was recrystallized from isopropanol before use. Methanol, n-butyllithium, scandium oxide, indanone, dimethyldichlorosilane, Yb(hfbc)3, Ag(fod) and (1S,2R,5S)-(+)-menthol were used as received. The methylindane (Dixon) that was used as an NMR standard was used as received. LiCH(SiMe3)2 12, ScCl3 • 3THF 13, [DpScH]2 4 and (1 R,2R,5S)-menthylC5H4 14 were all prepared according to literature procedures. Physical Measurements 1H NMR spectra were recorded on Varian EM 390 (90 MHz) and Brucker WM500 (500.13 MHz) spectrometers. Elemental analyses were performed by the analytical facility at the California Institute of Technology. Gas chromatogragh analyses were carried out using a PerinElmer 8410 instrument and a RSL-150 (Alltech) column. Syntheses Me 2 Si(C 5 Me4 )CI, 2_. C5Me 4HLi (10.0 g, 78.1 mmol) was suspended in tetrahydrofuran (150 ml) and cooled to -78 °C. Me 2SiCl 2 (9.7 ml, 80 mmol), was added via syringe against an argon counterflow. The mixture was warmed to room temperature and stirred for two hours. The 120 solvent was removed in vacuo and replaced with petroleum ether (100 ml). Lithium chloride was removed by filtration, and the petroleum ether was removed in vacuo, leaving the product as a colorless liquid that solidified upon cooling to -40°C. Yield: 16 g. (95%). 1 H data: see Table 2.1. Me2Si(C 5 Me 4 )((1 R,2R,5S)-menthylC 5 H3)Li2, §.. (1 R,2R,5S)-MenthylC5H3Li (2.93 g, 13.9 mmol), dissolved in tetrahydrofuran (40 ml) was added to C5Me4SiMe2CI ((3.0 g, 14 mmol) dissolved in tetrahydrofuran (20 ml). The solution was stirred for twelve hours, and the solvent was removed in vacuo and replaced with petroleum ether (50 ml). The lithium chloride was removed from the solution by filtration and the petroleum ether was removed in vacuo, leaving an oily liquid. Diethylether was added (80 ml), the solution was cooled to -78 ° C, and nbutyllithium (1.6 M solution in hexanes, 30 ml, 48 mmol) was added. The solution was warmed slowly to room temperature and allowed to stir for a further three hours. The solvent was removed in vacuo and replaced with petroleum ether (100 ml). The product was collected as a white solid, washed with fresh petroleum ether and dried in vacuo. Yield 5 g (98%). 1 H NMR data: see Table 2.1 Me2Si(C5 Me 4 )((1R,2R,5S)-menthylC 5 H3)ScCl•THF, I- § (10.24 g, .026 mmol) and ScCl3•3THF (9.57 g, .026 mmol) were dissolved in toluene (150 ml). The mixture was refluxed overnight; the solvent was removed in vacuo and replaced with petroleum ether. The solution was filtered to remove lithium chloride and the solvent was removed, leaving a white powdery residue, which was dried in vacuo. Yield: 13.8 g, (60%). 1H NMR data: see Table 2.1. Me2Si(C5 Me5 )((1 R,2R,5S)-menthylC 5 H3)ScCH(SiMe3)2, !!- (1.0 g, 20 mmol) and (Me 3Si) 2CHLi • 0.5Et2O (0.4 g, 20 mmol) were stirred for four hours in toluene (30 ml). The solvent was removed in vacuo and replaced with petroleum ether (50 ml). Lithium chloride was removed from the solution by filtration, and the solvent was removed in vacuo and replaced with diethylether. Cooling to -78 °c afforded yellow crystals of the product. Yield: .30 g (70%). 121 Anal. calcd. for C33H59Si3Sc: C, 67.75; H, 10.16. Found: C, 67.79; H, 9.66. 1 H NMR data: see Table 2.1. Me2Si(C5Me4)((1R,2R,5S)-menthylC5H3)ScH, ~. ~•- ~ (0.5 g, .85mmol) was dissolved in benzene (25 ml) and the solution was loaded in a thick-walled glass bomb. One atmosphere of hydrogen was admitted to the evacuated vessel at -196°C (ca. 4 atmospheres at room temperature), and the system was allowed to warm slowly to room temperature. The solution was agitated rapidly overnight (for isolation of both diasteriomers) or for as long as two weeks (for isolation of only one diasteriomer). The reaction was monitored by 1 H NMR. The solvent was removed at 0°C (to prevent decomposition), leaving a yellow solid, which was dried in vacuo and quickly transferred to the freezer in the glove-box for storage. Yield: .35 g (87%). Anal. calcd. for C25H42SiSc: C, 76.07; H, 9.90. Found: C, 71.52; H, 9.17. 1 H NMR data: see Table 2.1. Dimerization of Propene . .9, (0.04 g, .076 mmol) was dissolved in .5 ml benzene in a seal able NMR tube. 269.7 Torr (20 eq.) was subjected to three freeze-pump-thaw cycles and then admitted to the evacuated NMR tube. The NMR tube was subsequently flame-sealed. The solution was stirred overnight, cracked open and the volatile products vacuum distilled away from the organometallic product. 1 H NMR data: see Table 2.1. Cyclization and hydrogenation of 3-methyl-1,5-hexadiene. The procedure for the dimerization of propene was followed using .04 g (.076 mmol) .9, and .205 ml (10 eq.) of 3methyl-1,5-hexadiene. The same procedure was followed for the hydrogenation experiment with the exception that one atmosphere of hydrogen was admitted to the NMR tube at -196 ° C. 1 H NMR data: see Table 2.1. PPh 3=CH 2. 34 g (.112 mol) of PPh3CH3Br were loaded in the glove-box with 3.0 g (.125 mol) oF NaH into a flask. 250 ml of THF were cannulated onto the solids at -78 ° C. The resultant slurry 122 was allowed to warm slowly to room temperature a further 3.5 hours. The solution was filtered, the solvent removed and the solid product was dried in vacuo. Yield: 25.8 g (84%). Methylene indane. To 4.79 g (.017 mol) of PPh3=CH2 in a THF solution, a THF solution of 2.29 g (.017 mol) indanone was added at -78°C. The solution was allowed to warm to room temperature and stirred for three days. The THF was removed at O°C in vacuo and the resultant solids were extracted multiple times with petroleum ether. The petroleum ether was removed at 0°C, leaving the product as a viscous oil. 1 H NMR data: see Table 2.1. Hydrogenation of methylene indane by (DpScH)2 .. 05 g (.144 mmol) of (DpScH) 2 was dissolved in 2 ml of benzene in a thick-walled glass vessel. .025 ml (5 eq.) of methylene indane was added. One atmosphere of hydrogen at -196 ° C was admitted to the evacuated vessel. The solution was agitated briskly overnight. Most of the solvent was removed in vacuo, the vessel was opened and the remaining solution was loaded onto silica gel thin layer chromatography (TLC) plates and eluted with a 70:20 solution of hexanes/diethylether. The organic products were collected as a band with an Rtvalue of ca. 0.9. Hydrogenation of methylene indane by~- The procedure described above was followed, using 0.02 g of~ and 0.016 g (4 eq.) of methylene indane. Preparation of NMR samples with shift reagents. A solution of CDCl3 that was 0.1 M in substrate, 0.05 M in 12 and 0.1 M in 13 was centrifuged to settle particulate matter. Clear solution was taken from the top of the centrifuge tube and used for the 1 H NMR studies. 123 Table 2.1. 1 H NMR Data. Compound Assignment o (ppm) Me2Si(C5Me4)CI, g C5(C_tfa)4Si(CH3)2CI 1.?(s) ; 1.9(s) (90 MHz, C5D5) C5(CH3)4Si (C.!::fa)2CI 0.12(s) Me2Si(C5Me4)((1 R,2R,5S)-menthylC5H3)Li2, §. (500 MHz, THF-da) (C!:b)2Si 0.39(s) C5(C.!::fa)4 2.1 (s) ; 1.9(s) Cs.!::fa 5.85(m) ; 5.77(m); 5.73(m) Me2Si(C5Me4)((1 R,2R,5S)-menthylC5H3)ScCI • THF (500 MHz, THF-da (C.!::fa)2Si 0. 71 (s) ; 0.68(s) 0.66(s) ; 0.64 C5(C.!::fa)4 1.90(s) ; 1.88(s) ; 1.85(s) 1.83(s) ; 1.81 (s) ; 1. 78(s) (2 signals under THF-da) Cs.!:::b 6.27(t) ; 6.08(t) ; 5.58(t) 5.51 (m) ; 5.38(t) Cs.!::fa 7.22(m); 7.16(m); 6.13(m) 5.87(m) ; 5.63(m) ; 5.34(m) Cs(C.tfa)4 2.02(s) ; 2.0l ; 1.98(s) 1.87(s) ; 1.81 (s) ; 1. 78(s) 124 1.72; 1.67 SiCH(C.!::b)2 0.32(5) ; 0.27(5) 0.08(5) ; 0.04(5) Si(Ctb)2 0.55(5) ; 0.82(5) 7.08(t) ; 7.01 '(t) ; 6.54(t) 6.37'(t) ; 5.80'(t) ; 5.70(t) 2.25(5) ; 2.24(5) ; 2.23'(5) 2.17'(5); 2.12(5); 2.08'(5) 1.69'(5) ; 1.68(5) Si(Ct!.s)2 0.95(5) ; 0.93'(5) 0.68'(5); 0.60(5) Set!. Methylindane Standard 3.36(b) ; 3.31 '(b) (1.28, 1.29)(d) (500 MHz, CDCl3) 0.1 M methylindane + 0.05 M 1,g + 0.1 M 13 (500 MHz, CDCl3) (1.71, 1.73)(d) (1.67, 1.68)(d) 125 References 1. Jeske, G., Lauke, H., Mauermann, H., Schumann, H., Marks, T., J. Am. Chem. Soc., 1985, 107, 8111. 2. Thompson, M., Baxter, S., Bulls, A.A., Burger, 8., Nolan, M., Santarsiero, 8., Schaefer, W., Bercaw, J., J. Am, Chem. Soc., 1987, 109,203. 3. Frendick,C., Mintz, E., Schertz, L., Marks, T., Organomet., 1984, 3, 819. 4. Bunel. E. Ph.D. Dissertation, California Institute of Technology, Pasadena, California, 1989. 5. Wiesenfeldt, H., Reinmuth, R., Barsties, E., Evertz, K., Brintzinger, H., J. Organomet. Chem., 1989, 369, 359-370. 6. Kagan, H., Fiaud, J., New Approaches in Asymmetric Synthesis, "Topics in Stereochemistry," 1978, 10, pp.201-2. 7. Offermann, Mannschreck, A., Tet. Lett., 1981, 22, 3227. 8. Offermann, W., Mannschreck, A., Organ. Mag. Res., 1984, 22, 355. 9. a.) Wenzel, R., Sievers, R., Anal. Chem, 1981, 53, 353; b.) Solodar, J., Petrovich, J., lnorg Chem., 1971, 10, 395 .. 10. Nogradi, Mihaly, Stereoselective Synthesis, VCH, Weinheim, 1987, p.73, and references therein. 126 11. Burger, B. Bercaw, J., "New Developments in the Synthesis, Manipulation and Characterization of Organometallic Compounds," Wayda, A.L., Darensbourg, M., ACS Symposium Series, 1987, 387. 12. Schrock, R., Fellman, J., J. Am. Chem. Soc., 1978, 100, 3359. 13. Manzer, L. lnorg. Syn., 1982, 21, 135. 14. Cesarotti, E., Kagan, H., Goddard, R., Krueger, C.,J. Organomet. Chem., 1978, 162,297.