Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The reactivity of binuclear rhodium hydrides : fundamental processes involving two metal centres Piers, Warren Edward 1988

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1988_A1 P53.pdf [ 11.75MB ]
Metadata
JSON: 831-1.0060285.json
JSON-LD: 831-1.0060285-ld.json
RDF/XML (Pretty): 831-1.0060285-rdf.xml
RDF/JSON: 831-1.0060285-rdf.json
Turtle: 831-1.0060285-turtle.txt
N-Triples: 831-1.0060285-rdf-ntriples.txt
Original Record: 831-1.0060285-source.json
Full Text
831-1.0060285-fulltext.txt
Citation
831-1.0060285.ris

Full Text

T H E R E A C T I V I T Y O F B I N U C L E A R R H O D I U M H Y D R I D E S : F U N D A M E N T A L P R O C E S S E S I N V O L V I N G T W O M E T A L C E N T R E S By Warren Edward Piers B. Sc., The University of British Columbia, 1984 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF T H E REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in T H E F A C U L T Y OF GRADUATE STUDIES (DEPARTMENT OF CHEMISTRY) We accept this thesis as conforming to the required standard T H E UNIVERSITY OF BRITISH COLUMBIA September 1988 © Warren Edward Piers, 1988 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. 1 further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 DE-6(3/81) ii Abstract Current knowledge of mechanistic organometallic chemistry has resulted largely from the study of mononuclear transition metal complexes. The possibility that different primary organometallic processes involving two or more metal centres may exist has been addressed only recendy. Reactivity studies on a simple, well defined binuclear system ought to provide fundamental insights into the nature of such polynuclear primary processes. The binuclear rhodium hydrides {[R2P(CH2)nPR2]Rh(|i-H)}2 (R = Pr1, n = 2-4, la-lc; R = OPr1, n = 2, Id) were thus reacted with a variety of organic compounds in an attempt to define primary processes involving two metal centres. The reactions of la-c with dihydrogen proceed rapidly to produce fluxional binuclear tetrahydrides whose structure is dependent on the chelate ring size of the diphosphine ligand. The dihydrides also catalyze the hydrogenation of olefins. Two mechanistic pathways for this cycle are proposed to exist as supported by chemical and kinetic evidence. One utilizes binuclear intermediates, the other mononuclear; the latter predominates in the lb catalyzed system (chelate ring size of six) while the former is favoured in the cycle mediated by la (chelate ring size of five). The reactions of la and lb with 1,3-dienes led to the solid (X-ray diffraction) and solution state characterization of binuclear complexes incorporating bridging dienyl ligands in the previously unobserved p>Ti4-o" and p.-T|3-T|3 "partial sandwich" bonding modes. A fluxional process interconverting the two bonding modes was observed spectroscopically in the products of the la/butadiene reaction and a model accounting for this is proposed. Labelling and alternate synthetic studies, as well as the observation of an intermediate at low temperature, support a mechanism for these reactions which involves the dehydrogenation of the dihydrides followed by further reaction of [(P2)Rh]2 with a second equivalent of diene. Bridging amido hydrides of general formula [(P2)Rh]2(|i-NR'CH2R")(|i-H) are produced in the reactions of la and Id with imines (R'N=CHR"). Mechanistic studies iii reveal that initial c-donation of the imine lone pair; of electrons to one Rh followed by %-coordination of the C=N bond to the other precedes formal insertion of the C=N bond into Rh-H. This proposal is consistent with the results of labelling and kinetic studies, but the crux of its support lies in the spectroscopic observation of two intermediates en route to the product amido hydrides. The specific synthesis of cationic |i-r|2-0" imine complexes closely related to one of the proposed intermediates in the reaction was carried out to confirm the plausibility of such an intermediate. Reaction of the amido hydrides with dihydrogen was slow in producing free amine and the hydrogen adducts of la or Id, precluding the use of these dihydrides as catalyst precursors in the homogeneous hydrogenation of imines. Reaction of la and Id with nitriles (R"'C=N) produces ii-alkylideneimido hydride complexes of general formula [(P2)Rh]2(|i-N=CHR'")(^ -H). One derivative (P2, R = Pr1, n = 2; R"' = C H 3 ) has been characterized by X-ray crystallography. Further reaction of these complexes with dihydrogen yield the amido hydrides [(P2)Rh]2(|i-NHCH2R"')(|i-H). No intermediates in these reactions were observed, precluding meaningful mechanistic proposals for this stepwise reduction of nitriles as mediated by two metal centres. iv Table of Contents Page Abstract ii Table of Contents iv List of Tables viii List of Figures ix List of Schemes xiii List of Abbreviations xiv Acknowledgments xvi Chapter 1: Primary Processes in Organometallic Chemistry: Mononuclear vs Polynuclear Systems 1.1 Primary Processes in Mononuclear Organometallic Complexes. 1 1.1.1 Oxidative Addition and Reductive Elimination. 1 1.1.2 Olefin Insertion and B-Elimination. 3 1.1.3 CO Insertion and Deinsertion. 4 1.2 Primary Processes in Binuclear and Polynuclear Organometallic Complexes. 5 1.3 Coordinatively Unsaturated Polynuclear Rhodium Hydrides. 9 1.4 Scope of the Present Work. 16 1.5 References 17 Chapter 2: Synthesis and Fluxional Behaviour of Binuclear Rhodium Polyhydrides and their Use in the Hydrogenation of Olefins 2.1 Introduction. 22 2.2 Synthesis and Properties of Binuclear Rhodium Dihydrides. 22 2.2.1 Synthetic Procedures. 22 2.2.2 Characterization and Properties of the Dihydrides la-d. 26 2.3 The Reactions of la-c with Dihydrogen: The Characterization of Fluxional Binuclear Tetrahydrides. 33 2.3.1 Variable Temperature *H NMR Spectra. 33 2.3.2 Variable Temperature 31P{ lR} NMR Spectra. 36 V Page 2.4 The Hydrogenation of Olefins Using la and lb as Catayst Precursors. 41 2.4.1 The Hydrogenation of 1-Hexene: Results. 41 2.4.2 Mechanistic Studies. 43 2.5 Experimental. 52 2.5.1 General Procedures. 52 2.5.2 Synthesis of Binuclear Di- and Tetrahydrides. 55 2.5.3 Miscellaneous Syntheses. 59 2.5.4 Crossover Experiments. 61 2.5.5 The Hydrogenation of Olefins: Procedures. 63 2.6 References. 64 Chapter 3: Synthesis and Fluxional Behaviour of Binuclear Rhodium Complexes with Bridging Dienyl Ligands: New Bonding Modes for 1,3-Dienes 3.1 Introduction. 67 3.2 The Reactions of [(dippp)Rh]2(H-H>2 and [(dippe)Rh]2(p-H)2 with 1,3-Butadiene. 69 3.2.1 The Reaction of [(dippp)Rh]2(M.-H)2, lb, with 1,3-Butadiene. 69 3.2.2 The Solid State Molecular Structure of [(dippp)Rh]2»C4H6, 6. 72 3.2.3 The Solution State Molecular Structure of [(dippp)Rh]2*C4H6, 6. 75 3.2.4 The Reaction of [(dippe)Rhh(n-H)2, with 1,3-Butadiene. 77 3.2.5 The Solid Smte Molecular Structure of [(dippe)Rh]2»C4H6, 8. 78 3.2.6 The Solution State Molecular Structure of [(dippe)Rh]2»C4H6, 8. 81 3.2.7 The Reactions of [(dippe)Rh]2(|i-H)2 with Isoprene and Piperylene. 88 3.2.8 The Reactions of [(dippp)Rh]2(|i-H)2 with Isoprene and Piperylene. 92 3.2.9 Product Distributions: A Steric Evaluation. 95 3.2.10 The Magnesium Butadiene Reaction. 96 3.3 A Proposed Mechanism of Formation of Binuclear Rhodium Butadiene Complexes. 100 3.3.1 The Proposed Mechanism for the Reaction of Olefins with the Dihydrides [(dippe)Rh]2(|i-H)2 and [(dippp)Rh]2(li-H)2. 100 3.3.2 The Observation of [(dippe)Rh(r|2-C4H6)](^ -H)2[Rh(dippe)]. 103 3.3.3 A Proposed Mechanism. 104 3.4 Theoretical Analysis of the Bonding in the Partial Sandwich Structure. 107 3.5 Experimental. Ill 3.5.1 General Procedures. Ill vi Page 3.5.2 Synthesis of the Chloro-bridged Dimers [(P2)Rh]2(n-Cl)2. 112 3.5.3 Synthesis of Binuclear |i-Dienyl Complexes. 114 3.5.4 Reactions of [(dippe)Rh]2(|i-H)2 and [(dippp)Rh]2(p>H)2 with Isoprene and Piperylene. 120 3.5.5 Mechanistic Studies. 122 3.5.6 Miscellaneous Reactions. 124 3.6 References. 126 Chapter 4: The Reactions of Binuclear Rhodium Hydrides with Imines: Factors Influencing the Insertion of C=N into Rh-H Bonds 4.1 Introduction. 130 4.2 The Synthesis and Properties of Binuclear Rhodium Amido Hydrides. 132 4.2.1 Synthetic Procedures. 132 4.2.2 Spectroscopic Properties. 135 4.2.3 H/D Exchange Processes in the Amido Hydride 18a. 140 4.2.4 The Reactions of the Dihydrides la and Id with Isoquinoline. 144 4.3 Mechanistic Studies. 150 4.3.1 The Reaction of [(dippe)Rh]2(|J.-H)2 with N-benzyhdeneanihne at Low Temperature. 150 4.3.2 The Reaction of [(dippe)Rh]2(|i-H)2 with Benzophenone imine at Low Temperature. 156 4.3.3 Kinetic Studies and a Proposed Mechanism. 160 4.4 Reactions of the Dihydride la with Iminium Salts. 165 4.5 The Relevance of this Work to the Homogeneous Hydrogenation of Imines 173 4.6 Experimental. 176 4.6.1 General Procedures. 176 4.6.2 The Reactions of [(dippe)Rh]2(|i-H)2, la, and [(dipope)Rh]2(|i-H)2, Id, with Imines. 177 4.6.3 Kinetic and Equilibrium Measurements. 186 4.6.4 H/D Exchange Reactions. 189 4.6.5 Reactions of [(dippe)Rh]2(jJ.-H)2, la, with the Iminium Salts 30 and 31. 190 4.7 References. 195 vii Page Chapter 5: The Reactions of Binuclear Rhodium Hydrides with Nitriles: The Stepwise Reduction of C=N at Two Metal Centres. 5.1 Introduction. 198 5.2 The Synthesis and Properties of Binuclear Rhodium Alkyhdeneimido Hydrides. 200 5.2.1 Synthetic Procedures. 200 5.2.2 The Solid State Structure of [(dippe)Rh]2[p:-N=C(H)CH3](n-H), 34a. 204 5.2.3 Fluxional Properties of the p>Alkyhdeneimido Hydride Complexes. 208 5.3 Mechanistic Considerations. 212 5.4 The Reactions of p>Alkyhdeneimido Complexes with Dihydrogen and Related Proces ses. 213 5.5 Experimental. 217 5.5.1 General Procedures. 217 5.5.2 The Syntheses of p>Alkyhdeneimido Hydride Complexes. 217 5.5.3 Reactions of (I-Alkyhdeneimido Hydrides with Dihydrogen. 224 5.5.4 Reactions of the Binuclear Tetrahydrides with Nitriles. 225 5.5.5 Reactions of [(dippe)Rh]2(M.-H)2 with Amines. 227 5.6 References. 227 Chapter 6: Concluding Remarks and Future Prospects 6.1 Conclusions. 230 6.2 Future Prospects. 233 6.3 References. 235 viii List of Tables Page Table 2-1. Selected Bond Parameters for C28H66P4Rh2 (la). 31 Table 2-II. Hydrogenation and Isomerization Rates for the Hydrogenation of 1-hexene by la and lb. 42 Table 2-III. Data plotted in Figures 2-6, 2-7, and 2-8. 64 Table 3-1. Bond Parameters for C34H74P4RI12 (6). 74 Table 3-II. *H and 1 3 C NMR Chemical Shift Comparison Between anti-7, 6, and Typical Transition Metal Diene Complexes. 76 Table 3-IH. Selected Bond Parameters for C32H7oP4Rh2 (8a). 80 Table 3-IV. Values of the Equilibrium Constant, and AG0 for Equation 3-1. 83 Table 3-V. *H and 13C{ lH} NMR Data for the Partial Sandwich Compounds 6, 8b, lib, 13b, and 15-17. 115 Table 3-VI. Coupling Constants in the Partial Sandwich Complexes 6, 8b, lib, 13b, and 15-17. 116 Table 4-1. 31p NMR data for the amido hydrides 18a, 18d, and 22. 138 Table 4-II. Thermodynamic Parameters, Values of AG° and K for the Equilibria Defined by Equations and 4-7 and 4-8. 148 Table 4-in. Rate Constants and Eyring Plot data for la + N-benzylideneaniline ==> 26. 164 Table 4-IV. Rate Constants and Eyring Plot data for 26 ==> 18a. 164 Table 5-1. Selected Bond Parameters for C3oH69P"4Rh2N (34a). 206 Table 5-TI. Carbon-Nitrogen Bond Distances in (i-Alkyhdeneimido Complexes. 207 i x List of Figures Page Figure 1-1. Bonding orbitals of the Rh2(|i-H)2 core in binuclear rhodium hydrides. 15 Figure 2-1. a) 121.4 MHz 3 1 P {itf.} NMR spectrum of [(dippe)Rh]2(H-H)2, la. b) Calculated spectrum based on an AA'A"A"'XX' spin system (see experimental for parameters). 28 Figure 2-2. ORTEP drawing of [(dippe)Rh]2(|i-H)2, la. 30 Figure 2-3. Hydride region of the 400 MHz *H NMR spectrum of the fluxional tetrahydride 5b, recorded at various temperatures. 34 Figure 2-4. 121.4 MHz 31P {1H} NMR spectra of the fluxional tetrahydride 5b, recorded at various temperatures. 37 Figure 2-5. 121.4 MHz 3 1P {lH) NMR spectra of the fluxional tetrahydride 5a, recorded at various temperatures. 39 Figure 2-6. Plot of the rate of hydrogen uptake vs the catalyst concentration in the hydrogenation of styrene using [(dippp)Rh]2(|i-H)2, lb, as catalyst precursor. 48 Figure 2-7. Plot of the rate of hydrogen uptake vs the square root of catalyst concentration in the hydrogenation of styrene using [(dippp)Rh]2((i-H)2, lb, as catalyst precursor. 48 Figure 2-8. Plot of turnover frequency vs catalyst concentration in the hydrogenation of styrene using [(dippp)Rh]2(|i-H)2, lb, as catalyst precursor. 49 Figure 3-1. Previously characterized bonding modes of 1,3-butadiene. 68 Figure 3-2. a) 400 MHz *H NMR spectrum of the partial sandwich complex [(dippp)Rh]2(M.-Ti3-ri3-C4H6), 6. b) 3 1P broadband decoupled, c) 3 1 P broadband decoupled, homodecoupled at Hi. d) 3 1 P broadband decoupled, homodecoupled at H2. 71 Figure 3-3. 162.2 MHz 3 1 P {lH} NMR spectrum of the partial sandwich complex [(dippp)Rh]2(|i-Tl3-Ti3-C4H6), 6. 72 X Page Figure 3-4. a) ORTEP drawing of the partial sandwich complex [(dippp)Rh]2(|i-r|3-r|3-C4H6), 6. b) Side view looking down the C(2)-C(2') axis depicting the 45° twist in the bridging dienyl ligand. 73 Figure 3-5. a) ORTEP drawing of the dienyl hydride complex [(dippe)Rh]2(|i-T|4-a-C4H6), 8a. b) Stereoview of the molecule. 79 Figure 3-6. a) van't Hoff plot for the equilibrium between the dienyl hydride complex 8a and the partial sandwich species 8b in C7D8 solution (linear region), b) The same plot over the full temperature range. 82 Figure 3-7. 121.4 MHz 31P{ 1H} NMR spectra of the equilibrium mixture of 8a and 8b at various temperatures. 84 Figure 3-8. 121.4 MHz 3 1P {^ H} NMR spectrum of the crude product mixture from the reaction of lb with piperylene. 94 Figure 3-9. An orbital interaction diagram for the "full sandwich" species [(PR3)2Rh]2(n-ri4-Ti4-C4H6). 108 Figure 3-10. a) Coordinates for structural distortions from the |i-r|4-ri4 "full sandwich" bonding mode, b) The increased overlap between 713 and the S combination of bi upon rotation of the P2RI1 units about 0. c) The effect of rotation about 0 on the overlap between 1x2 and theAcombinationofbi. 109 Figure 4-1. Possible bonding modes for the amido ligand (I-III) and imines (IV-V). 131 Figure 4-2. 400 MHz *H NMR spectrum of the amido hydride complex 18a. 136 Figure 4-3. a) 121.4 MHz 3 1P {lU) NMR spectrum of the amido hydride 18a. b) Calculated spectrum based on an AA'BB'XX' spin system (see Table 4-1 for parameters). 139 Figure 4-4. 400 MHz *H NMR spectrum of the amido hydride 24a. Peaks downfield of 7 ppm are due to free isoquinoline. 145 Figure 4-5. van't Hoff plot for the equilibrium defined in equation 4-7: trace a, G5D12 : C 7D 8 = 5:1; trace b, G5D12 : C 7D 8 =1:1; trace c, G5D12 : C 7D 8 = 1:5. 147 Figure 4-6. van't Hoff plot for the equilibrium defined in equation 4-8. 149 xi Page Figure 4-7. 162.2 MHz 31P{1H} NMR spectra of the proceeding reaction between la and N-benzylideneaniline as a function of temperature and time, a) Initial spectrum recorded at -70°C and near the beginning of the reaction, b) Spectrum recorded at time = one hour (-70°C). c) Spectrum obtained at the end of the reaction after warming to -20°C. 152 Figure 4-8. 400 MHz !H{31P} NMR spectrum of the intermediate observed in the reaction between la and N-benzylideneaniline (spectrum recorded at -70°C). 153 Figure 4-9. Two possible structures for the intermediate observed at low temperature in the reaction of la with N-benzylideneaniline. 154 Figure 4-10. 162.2 MHz 31P{1H} NMR spectra of the proceeding reaction between la and benzophenone imine as a function of temperature and time, a) Initial spectrum recorded at -70°C and near the beginning of the reaction; formation of the first intermediate of structure 28 or 29. b-e) Spectra recorded as sample was warmed gradually over a period of about 45 minutes; first intermediate slowly disappears, while a second intermediate grows in along with signals for the amido hydride product 21a. f) Spectrum recorded upon completion of the reaction. 157 Figure 4-11. Hydride region of the 400 MHz J H NMR spectrum of the proceeding reaction between la and benzophenone imine. a) Initial spectrum collected at -70°C. b-d) Spectra recorded at =10 minute intervals at -50°C. e) Final spectrum recorded upon completion of the reaction (-20°C). 158 Figure 4-12. Two possible structures for the first intermediate observed at low temperature in the reaction of la with benzophenone imine. 159 Figure 4-13. Second order rate data obtained at various temperatures for the first step of the reaction between la and N-benzylideneaniline (N-BA). 162 Figure 4-14. Eyring plot for the first step in the reaction between la and N-benzyHdeneaniline (N-BA). 162 Figure 4-15. First order rate data obtained at various temperatures for the second step of the reaction between la and N-benzylideneaniline. 163 Figure 4-16. Eyring plot for the second step in the reaction between la and N-benzylideneaniline. 163 Figure 4-17. 121.4 MHz 3 1P {lH} NMR spectra of the cationic \i-r\2-a imine complex 32 recorded at various temperatures. 168 xii Page Figure 4-18. A comparison of the 3 1P NMR spectra of the intermediate observed in the reaction between la and N-benzylideneaniline, 26, and the cationic |a.-rj2-a imine complex 32. a) 121.4 MHz 3 1P {lH} NMR spectrum of 32 (-30°C). b) 162.4 MHz 3 1P {lH} NMR spectrum of 26 (-70°C). 169 Figure 4-19. 162.2 MHz 31p { 1H} NMR spectra of the alkenyl hydride complex [(dippe)Rh]2(H-Tl2-a-CH=CH2)(H-H), isoelectronic with 32. a) 60°C. b) 10°C. c) -70°C. d) Calculated spectrum for spectrum c (see reference 7 for parameters). 171 Figure 5-1. Bonding modes for alkylideneimido ligands to one (I, II), two (HI) and three (TV) metal centres. 198 Figure 5-2. 400 MHz *H NMR spectrum of the ethylideneimido hydride complex 34a. Upon broadband 3 1 P decoupling, the hydride resonance collapses to a broad triplet (upper trace). 202 Figure 5-3. a) 121.4 MHz 3 1 P {lH} NMR spectrum of the ethylideneimido hydride 34d, incorporating the dipope ligand (scale is marked by 100 Hz gradations), b) Calculated spectrum based on an ABCDXY spin system (see experimental for parameters). 203 Figure 5-4. a) ORTEP drawing of the ethylideneimido hydride complex 34a. b) Another view looking dow the Rh(l)-Rh(2) vector, c) Stereoview of the molecule. 205 Figure 5-5. 162.2 MHz 3 1 P pH} NMR spectra of the o-tolylideneimido hydride 36a recorded at various temperatures. 209 xiii List of Schemes Scheme Page Scheme 1-1 2 Scheme 1-2 10 Scheme 1-3 13 Scheme 2-1 23 Scheme 2-2 35 Scheme 2-3 40 Scheme 2-4 44 Scheme 2-5 50 Scheme 3-1 70 Scheme 3-2 77 Scheme 3-3 86 Scheme 3-4 89 Scheme 3-5 90 Scheme 3-6 91 Scheme 3-7 92 Scheme 3-8 93 Scheme 3-9 99 Scheme 3-10 101 Scheme 3-11 105 Scheme 4-1 143 Scheme 4-2 161 Scheme 4-3 170 Scheme 4-4 170 Scheme 5-1 210 Scheme 5-2 211 Scheme 5-3 214 xiv List of Abbreviations A angstrom unit, 10"8 cm atm atmosphere(s) br broad B u n n-butyl group, -(CH2)2CH3 B u l tertiary-butyl group, -C(CH3)2CH3 Bz benzyl group, -CH2C6H5 C Celsius chiraphos S,S-2,3-bis(diphenylphosphino)butane COD cyclooctadiene, C8H12 COE cyclooctene, C8H14 Cp T|5-cyclopentadienyl ligand, C5H5-Cp* t|5-pentamethylcyclopentadienyl ligand, 05(013)5-d doublet dcypp 1,3-bis(dicyclohexylphosphino)propane dd doublet of doublets dipope l,2-bis(diisopropoxyphosphino)ethane dippb 1,4-bis(diisopropylphosphino)butane dippe 1,2-bis(diisopropylphosphino)ethane dippp 1,3-bis(diisopropylphosphino)propane dm doublet of multiplets DME l,2-bis(dimethoxy)ethane dppm bis(diphenylphosphino)methane dppp 1,3-bis(diphenylphosphino)propane dsp doublet of septets dtbpp 1,3-bis(ditertiarybutylphosphino)propane equiv equivalent(s) Et ethyl group, - C H 2 C H 3 g gram GC gas chromatography HOMO highest occupied molecular orbital hr hour Hz Hertz, seconds-1 J coupling constant XV K Kelvin L neutral unidentate donor ligand LUMO lowest unoccupied molecular orbital m multiplet M central metal atom molar Me methyl group, - C H 3 MHz megaHertz min minute mL millilitre mmol millimole mm niillimetre mM millimolar mo molecular orbital mol mole N-BA N-benzylideneaniline, (Ph)N=C(H)Ph NMR nuclear magnetic resonance nOe nuclear Overhauser effect OPri isopropoxy group, -OCH(CH3)2 Ph phenyl group, -G5H5 ppm parts per million PrJ isopropyl group, -CH(CH3)2 q quartet R alkyl, aryl or alkoxy group s singlet sec second t triplet TF turnover frequency, mol product/mol catalyst»unit time THF tetrahydrofuran TMS tetramethylsilane ttt triplet of triplets of triplets w 1/2 width at half height X anionic unidentate donor ligand xvi Acknowledgments My warmest thanks go to Dr. Michael Fryzuk for the enthusiasm and patience he exhibited throughout the course of this work. His insights, teaching ability and hard work have made ours the most satisfying working relationship of my life. I have also benefited enormously from my association (both professional and social) with a fine bunch of colleagues and thus wish to include my thanks to the various forms of wildlife (past and present) indigenous to the Fryzuk lab in room 320 (south). Particularly, I want to thank my friend Dr. Pat MacNeil for all the help she gave me during the first two years of my stint in the Fryzuk group. Also, thanks to Dr. Graham White and Dr. Dave Berg for their time and advice during the writing of this thesis. I also express my gratitude to all members of the fine technical staff here at the Chemistry department of UBC--I couldn't have done it without you! Finally, the people who matter the most. I want to thank my parents (all four of them!) for all their support and encouragement over the years. To my wife Marrian, whose love and patience sustain me, I dedicate this thesis—it is as much hers as it is mine. 1 CHAPTER 1 Primary Processes in Organometallic Chemistry: Mononuclear vs Polynuclear Systems 1.1 Primary Processes in Mononuclear Organometallic Complexes. The observation that organometallic compounds are capable of acting as catalysts for certain chemical transformations has been the main driving force for the study of organometallic materials. The development and study of various homogeneously catalyzed reaction cycles, such as the hydrogenation, isomerization, polymerization, and hydroformylation of olefins, led to the discovery of a few discrete, potentially reversible, primary reaction steps which occur commonly in organometallic chemistry. These reversible primary reaction pairs include reductive elimination and oxidative addition, olefin insertion and B-elimination, and CO insertion and deinsertion, in addition to variations on these three basic types. Mechanistically, each of these reaction pairs have been studied extensively and are quite well understood. 1.1.1 Oxidative Addition and Reductive Elimination. In an oxidative addition reaction,1 a reagent X-Y adds to a metal centre, increasing the metal's coordination number with a concomitant increase in the formal oxidation state of the metal. There are examples of both one electron and two electron additions, the latter being the most common (Scheme 1-1). Mechanisms by which two electron oxidative additions may take place are dependent on the nature of the addendum X-Y. Generally, non polar reagents add via a three centred transition state in which o-donation by the X-Y bond to the 2 Scheme 1-1 Two Electron Oxidative Addition t X Y Y One Electron Oxidative Addition 2LnM(n) + X Y UMf—X + LnM n j J - Y metal is coupled with 7t-donation from the metal into the X-Y antibonding orbitals. Theoretical studies on the cis addition of dihydrogen to Pt(PH3)22 have shown that the transition state is an "early" one in that it resembles the reactants; there is little X-Y bond stretching. This is consistent with the small calculated and observed3 kinetic isotope effect (kn/ko) of 1-2 for this reaction. More polar species such as alkyl halides may add in a similar manner, or via a classic SN2 mechanism in which the metal acts as the nucleophile. The reverse process, reductive elimination 4 involves the loss of two ligands with a corresponding decrease in the metal's formal oxidation state. Usually, the two departing ligands couple, forming carbon-carbon or carbon-heteroatom bonds. Geometrically, the ligands to be eliminated must be cis to each other before reductive eUmination may take place; if they are not, an isomerization must occur prior to loss of X-Y.5 3 1.1.2 Olefin Insertion and B-Elimination. A second primary reaction pair is comprised of olefin insertion and B-elimination reactions (equation l-l).6 The insertion of carbon-carbon double and triple bonds into M-H -L L n M - H r-.c; L n M H olefin insertion {^-elimination N H (1-D L n M or M-C bonds is a necessary step in olefin/alkyne hydrogenation or polymerization sequences, while the microscopic reverse, B-elimination, is the major decomposition pathway for coordinatively unsaturated metal alkyls containing B-hydrogens. Although this reaction pair is commonly invoked in many catalytic cycles,7 olefin insertions and B-eliminations have rarely been observed direcdy, largely due to the facile reversibility of these reactions and their rapid rates when compared to other steps in the catalytic cycle.8 The inability to observe the insertion of an olefin into a M-H bond is also in part due to ground state instability of cis olefin-hydride complexes,9 although observable examples of such complexes have appeared recently.10'11 For example, a thorough study of the reversible insertion of a coordinated olefin into the Nb-H bond of Cp*2Nb(CH2=CHR)H, (R = H, CH3, Ar), showed that insertion and elimination proceed through a planar, relatively nonpolar, cyclic transition state / CHR 6~,' ^ Cp*2Nrx ^ *- Cp* 2Nb \ C H R . . * Cp* 2 NbCH 2 CH 2 R (1-2) H 4 with concerted bond making and bond breaking (equation 1-2).11 In this system, a fractional positive charge develops on the p-carbon, and the hydride migrates more as H" than H +; whether this picture holds true for insertions into the more acidic M-H bonds of late transition metal hydrides is unknown. 1.1.3 CO Insertion and Deinsertion. The migratory insertion of carbon monoxide12 (equation 1-3) into a transition metal alkyl bond is also an important reaction in several catalytic cycles, particularly L n M — C — R (1-3) I 11 hydroformylation, in which olefins are converted to aldehydes. Mechanistic studies have shown that the transformation of an alkyl-carbonyl complex into an acyl complex takes place via a two step sequence.123 The first step involves formation of a coordinatively unsaturated acyl intermediate either through alkyl migration to an adjacent CO ligand (most common pathway13) or by direct CO insertion into the M-R bond. The unsaturated acyl intermediate is rarely observed directly, as it reacts readily with some donor ligand L (usually CO or a phosphine) to yield a coordinatively saturated acyl product. The precise nature of this intermediate is unknown, but its structure is possibly dependent on the type of the solvent. Evidence for a solvated r^ -acyl species exists in some systems,1215'14 while in others, the oxygen of the acyl moiety may occupy the vacant coordination site, forming an T|2-acyl species133-15 (several rj2-acyl complexes have been structurally characterized16). This first step occurs with retention of configuration at carbon17 and is also stereospecific with respect to the stereochemistry at the metal, inversion or retention occurring depending on whether the L nM — R CO K-i LnM — C — R ll O + L 5 first step proceeds via alkyl migration or CO insertion.13 In all cases, the reacting alkyl and CO ligands must be cis to each other and the incoming ligand L always occupies a cis site to the newly formed acyl ligand.12 The reverse process, in which the alkyl moiety migrates from the carbonyl to the metal, is also facile under the right conditions. It is a key reaction in the decarbonylation of aldehydes and acid chlorides as catalyzed by transition metal complexes.18 1.2 Primary Processes in Binuclear and Polynuclear Organometallic Complexes. Although other reaction types, such as ligand substitution reactions or the reactions of coordinated ligands, are known, these three primary reaction pairs serve as the fundamental building blocks in catalytic cycles brought about by organometallic compounds. The mechanisms are quite well understood, but as the above discussion perhaps indicates, the large majority of mechanistic studies have taken place using mononuclear systems. Certainly there are examples of oxidative additions of dihydrogen19 or alkyl halides20 to binuclear complexes, but these reactions proceed via addition to a single metal centre followed by some isomerization process rather than involving both metal centres. The paucity of examples of true "polynuclear reactions" led to a search for transformations mediated by two or more metal centres in the mid 1970's. Interest was focussed on determining whether mechanisms of such processes involved two or more metal centres acting in concert and whether or not a class of primary reaction steps exclusive to binuclear or polynuclear systems could be identified. Strategies utilized in the pursuit of answers to these questions are variations on two basic rationales. The first is based on the fact that some of the primary reactions discussed above were identified through study of decomposition pathways for metal alkyl complexes. Thus, Bergman and co-workers sought a binuclear system containing a metal-metal bond and 6 (1-4) \ Cp one ©"-bound organic ligand attached to each metal centre to examine its thermal decomposition pathways. Such a system was obtained via the alkylation of the radical anions [RCpCo(p>CO)]2eNa+, (R = H, C H 3 ) 2 1 with a variety of alkyl halides and oc,co-alkyldihalides to form the binuclear dialkyls22 and binuclear dimetallacycles23 as shown in equation 1-4. The thermal decomposition of some of these compounds led to the (1-5) Cp dienophile identification of some interesting reactions involving only binuclear intermediates. Particularly interesting was the discovery of a reversible "dimetalla Diels-Alder" reaction in the chemistry of the benzodicobaltacyclohexene complex depicted in equation l-5.23b A second strategy used in the study of binuclear and polynuclear chemistry is to synthesize cluster compounds which are coordinatively unsaturated at one or more of the 7 metal centres in the molecule. As a synthetic target, such electronically and/or coordinatively unsaturated clusters are also desirable because their homogeneous chemistry is thought to be a useful model for heterogeneously catalyzed transformations.24 In addition to this cluster-surface analogy, the few authentic coordinatively unsaturated clusters known have provided fundamental information on primary processes at two or more metal centres. Perhaps the prototype for this class of compounds is the electronically unsaturated trinuclear cluster Os3(CO)io(p>H)2.25 A 46 electron complex, two shy of the requisite 48 electrons needed for electronic saturation in a triangular trinuclear system,26 this complex may be thought of as a (CO)30s=Os(CO)3 moiety bridged by two hydride ligands and an Os(CO)4 fragment. The chemistry associated with this molecule is extensive and largely associated with the hydrogen bridged osmium-osmium double bond.27 The syntheses of the "lightly stablilized" complexes Os3(CO)ioL2 (L = cyclooctene, acetonitrile28; L2 = cyclohexadiene29) expanded the range of substrates for activation by simply making coordinatively unsaturated species accessible through dissociation of the lightly bound stabilizing donors L. Puddephatt and co-workers have exploited a highly coordinatively unsaturated family of trinuclear platinum clusters30 to characterize new reactions and dynamic processes not possible in mononuclear systems;31 indeed these authors have successfully modelled some heterogeneously catalyzed reactions with these systems.32 For example, the heterogeneous (CO)4 Os 8 ( 1 - 6 ) activation of acetylene on platinum (111) has been found to proceed via formation of a M-3_Tl HC=CH complex at low temperatures, which rearranges to a |i3-T|2-vinylidene upon warming. Further reaction with surface bound hydrogen yields a u,3-ethylidyne species.323 Reaction of acetylene with the trinuclear cation [Pt3(ji3-H)(ji-dppm)3]+PF6" models the coordination of acetylene to Pt(l 11) and its transformation to the vinylidene intermediate as shown in equation l-6.32b The (13-T|2-vinylidene intermediate does not, however, rearrange to the 113-ethylidyne species, despite the presence of a hydride ligand. A subsequent study on the reaction of acetylene with the related trinuclear dication [Pt3(N.3-CO)((i-dppm)3]2+[PF6_]2 fully characterized the first M-3-rl2 acetylene complex of a platinum cluster.320 9 1.3 Coordinatively Unsaturated Polynuclear Rhodium Hydrides. First reports of coordinatively unsaturated clusters of the general formula [(L)2Rh((i-H)] n appeared when Green and Bottrill reported the reaction of [P(OMe)3]3Rh(ri3-C3H5) with dihydrogen in 1976 (equation 1-7).33 The reaction produced a red solution which was found to contain a catalyst active towards the hydrogenation of olefins. Although the active ingredient was not isolated, the authors speculated that the initially formed hydrogenation product [P(OMe)3]3RhH disproportionated to [P(OMe)3]4RhH and {[P(OMe)3]2RhH}n, and that the latter complex of unknown nuclearity was the active catalyst. Three years later, Otsuka et al. reported the results of a synthetic study aimed at the preparation of three coordinate, 14 electron complexes of rhodium(I) incorporating bulky phosphine ligands.3 4 While they were successful in this endeavour when employing very bulky phosphine ligands such as P(Bul)3, less bulky ligands also produced binuclear complexes likely through dimerization of initially formed mononuclear rhodium(I) hydrides. For example, reduction of RhCl 3-xH20 with Na/Hg amalgam in T H F in the presence of triisopropylphosphine afforded the complex RhH(PPrJ3)3. In a series of reactions as outlined in Scheme 1-2, the dinitrogen hydrido complex RhH(N2)(PPri3)2 was formed. This complex decomposed in the absence of excess dinitrogen to give a variety of hydride containing products, one of which was formulated as "RhHtPPr^^". Solution molecular (MeO)3P (1-7) "HRh[P(OMe)3]2" + HRh[P(OMe)3]4 10 Scheme 1-2 RhCI3 • 3 H 2 0 RhH(N 2)(PPr' 3) 2 r [RhH(PPr'3)2]2(p:-N2) Na/Hg/THF/N2 PPr1, H 2 N 2 RhH(PPr'3)3 f H 2 ^ PPr'o RhH 3(PPr' 3); •RhH(PPr ) 3) 2 1 ' | + R h H ^ P P r 1 ^ weight measurements showed the complex to be a dimer, but since it was a minor product in the reaction, its chemistry was not explored. The first reported isolable members of this family of coordinatively unsaturated clusters came in 1977 when the syntheses of the binuclear compound {[P(OPri)3]2Rh}2(p.-H)2 and the trinuclear derivative {[P(OMe)3j2Rh}3(|i-H)3, containing the less bulky trimethylphosphite ligand, were detailed by Muetterties and co-workers.35 Considered separately, the rhodium centres in these complexes are formally 16 electron centres, but an electron count of the cluster as a whole reveals that the dimer possesses only 28 valance electrons, while the trimer has only 42, rather than the 34 and 48 electrons, respectively, expected for a fully saturated species.26 A variation on this class of coordinatively unsaturated clusters was implemented with the incorporation of chelating diphosphine ligands. In an attempt to generate 14 electron 11 rhodium(I) complexes analogous to those prepared by Otsuka et al.?* but containing bulky chelating diphosphine ligands, Fryzuk instead found that the mononuclear species presumably generated initially oligomerized to form tetramers, trimers, or dimers, depending on the relative steric bulk of the chelating phosphine employed.36 The binuclear species prepared, l a 3 7 and Id, incorporated the phosphine ligand 1,2-bis(diisopropylphosphino)ethane, dippe, and the phosphinite ligand 1,2-bis(diisopropoxyphosphino)ethane, dipope, respectively. Other binuclear members of this family, lb and lc, were subsequently synthesized as part of this work, incorporating the phosphine ligands l,3-bis(diisopropylphosphino)propane, dippp and 1,4-bis(diisopropylphosphino)butane, dippb, respectively. In general, the complexes incorporating the bidentate phosphine and phosphinite ligands are more resistant to fragmentation than the Muetterties dimer {[P(OPri)3]2Rh}2((J.-H)2, le, (vide infra) and also exhibit a wider range of stoichiometric reactivity towards organic functionalities. While comparisons between the observed chemistry of le and the dimers la-d will be made throughout this thesis, some of the reported chemistry of le will be reviewed here 12 as a reference point. Both the Muetterties dimer and the trinuclear complex {[P(OMe)3]2Rh}3(ji-H)3 were found to be active catalyst precursors for the hydrogenation of olefins and alkynes, but were also prone to undergo fragmentation to catalytically active mononuclear species under certain conditions. Nonetheless, the cluster's integrity was maintained in a number of reactions (vide infra). As an olefin hydrogenation catalyst, the Muetterties dimer le was found to be exceptionally active, with turnover rates exceeding 2 per second at 23°C and a substrate catalyst ratio of=10,000: l . 3 8 This represents an activity 6 times that of Wilkinson's catalyst, [P(C6H5)3]3RhCl, under the same conditions. The observations that le did not fragment when reacted stoichiometrically with dihydrogen and that the resulting tetrahydride reacted with olefin to form alkane and regenerate the dihydride, led the authors to propose a hydrogenation cycle consisting of binuclear intermediates as the major pathway for this reaction (Scheme 1-3) (Scheme 1-3 is discussed in more detail in section 2.4.2). No kinetic evidence was obtained to provide additional support for this proposal. In addition to hydrogenating olefins, the dimer le was active in the hydrogenation of alkynes, initially forming trans olefins stereoselectively.39 This was an important observation in that the usual primary products in the hydrogenation of alkynes are cis olefins; trans olefins are often observed, but are due to secondary isomerization of the initially formed cis products.40 However, the production of trans olefins as the primary products is possible when two or more metal centres are involved in the reaction process.40 The observation that le produced trans olefins as primary products thus supported a hydrogenation pathway mediated by binuclear species. This binuclear pathway was maintained only for about 5 turnovers, however, as the dimer began to fragment under the reaction conditions to a mononuclear catalyst system, and cis olefins began to dominate in the product mixture. The stoichiometric reactivity of le has also been explored to some extent. Reactions with carbon monoxide 4 1 alkynes,40 and isonitriles42 (equations 1-8 to 1-10) all generate binuclear products. In the reactions with alkynes and isonitriles the products reacted further with the organic substrates to yield mononuclear fragmentation products. Reaction of the 13 Scheme 1-3 H — R h ^ - H - ^ R h F j h ^ R h P | v H^ R h P H 2 C = C H R 14 ^Rh Rh + 0 O (Pr'OfcP' H le P(OPr')3 O P \ / k / P ^Rh._ Rh ^ \ p H CO (1-8) O Rh Rti II o (Pr'0)3P. .H .P(OPr')3 P . . . . . H , (Pr'0)3P \ / \ ^P(OPr')3 P . . . . . .P > h RrT R C = C R ^ / . R h . ( 1 . 9 ) / \ / \ R = CH 3 , Ary" P * * ^ ^ " C > p ^ ^ ^ ^ P ( O P r ' ) 3 " T le R N (PHO)3P .H .P(OPr')3 P C. P (Pr'0)3P^ ^P(OPr')3 X P H le N R v / C H 3 ( P H o ) 3 P x ^ c N / P ( O P H , 3 ^ N ^ ^ V H ' ^ V R=Aryl ' R h v ™C ( 1 " U ) (Pr'0)3P^ >(OPr')3 p / H \ isonitrile adducts of l e with dihydrogen stoichiometrically reduces the isonitrile ligand to generate binuclear amido hydride products (equation 1-11).43 In addition, the dihydride 15 reacts with one or two equivalents of phosphite ligand to form mononuclear rhodium(I) phosphite hydrides.42 Finally, le also reacts with three equivalents of 1,3-cyclohexadiene, as shown in equation 1-12, to form one equivalent of benzene and two equivalents of the mononuclear species [PCOPr^ teRhOlS-C^HQ).44 Rh Rh. l e + 3 - H 2 (PHOJaP (PHO)3P (1-12) In addition to chemical reactivity studies, the dimer le was subjected to a rigorous structural analysis in the form of a neutron diffraction study and an Extended Hiickel molecular orbital calculation 4 5 The structure revealed an Rh-Rh distance of 2.65(1)A and, although this is certainly small enough to invoke a metal metal bond, the Extended Hiickel calculation suggested that the bonding framework is best represented as two four-centre two-electron bonds, pictured in Figure 1-1. This notion is supported by Mulliken bond orders '2u Figure 1-1. Bonding orbitals of the Rh2(|i-H)2 core in binuclear rhodium hydrides. 16 found to be only 0.086 for the Rh-Rh interaction but a substantial value of 0.497 (=0.5) for each of the Rh-H bonds. Likely the bonding picture in the binuclear dihydrides discussed in this thesis is similar to that described for le. 1.4 Scope of the Present Work. This thesis describes some of the catalytic and stoichiometric reactivity of the binuclear rhodium(I) hydrides la-d. The dimers la 3 7 and Id 3 6 were first synthesized prior to 1984, when this work began, but the dihydrides lb and lc were prepared during the course of the work described. As a consequence, some of the reactivity of la and Id was also investigated prior to 1984, specifically the catalysis of the hydrogenation of olefins by Id 3 6 and the stoichiometric reactivity of la with olefins37 and 1,3-butadiene.46 However, mechanistic investigation of the latter reactions was carried out as part of the work described in this thesis. Chapter 2 describes a modified synthetic procedure for arriving at the dimers la and Id, developed since the published syntheses of these compounds, and presents a discussion on the characteristics of the four dimers utilized in the chemistry described in the thesis. In addition, the reactivity of complexes la-c towards dihydrogen, as well as a descriptive and mechanistic discussion of their activity towards the hydrogenation of olefins, is presented. Chapter 3 discusses the reactivity of two of these dihydrides, la and lb, with a variety of l,3-dienes.46>47 The full structural characterization of the products of these reactions, as well as a proposed mechanistic pathway for the formation of these products, is presented along with a discussion on the results of theoretical calculations on the bonding in the binuclear product observed in the reaction of lb with 1,3-butadiene. This study was completed in collaboration with Dr. Thomas A. Albright at the University of Houston 4 8 Continuing our examination of the reactivity of these dimers towards small organic molecules, Chapters 4 and 5 deal with the reactions of la and Id with simple imines49 and 17 nitriles, respectively. The details of the mechanism of the reaction with imines were elucidated by following the reactions at low temperature; a kinetic analysis of this reaction is also presented along with a mechanistic proposal. The reactions with nitriles were less amenable to such mechanistic study, but the synthetic work is described in detail as well as the reactivity of the product alkylidene imido complexes with dihydrogen. 1.5 References. 1. a) Halpern, J. Acc. Chem. Res. 1970, 3, 386. b) Vaska, L. Acc. Chem. Res. 1968, 1. 335. c) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles and Applications of Organotransiton Metal Chemistry. University Science Books, Mill Valley, CA. 1987, Chapter 5. 2. a) Low, J. J.; Goddard, W. A. in J. Am. Chem. Soc. 1984,106, 6928. b) Obara, S.; Kitaura, K.; Morokuma, K. /. Am. Chem. Soc. 1984,106, 7482. c) Noell, J. O.; Hay, P. J. / . Am. Chem. Soc. 1982,104, 4578. 3. Zhou, P.; Vitale, A. A.; San Filippo, J., Jr.; Saunders, W. H., Jr. / . Am. Chem. Soc. 1985,107, 8049. 4. Stille, J. K.; Lau, K. S. Y. Acc. Chem. Res. 1978,10, 343. 5. Stille, J. K.; Gillie, A. J. Am. Chem. Soc. 1980,102 , 4933. 6. Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles and Applications of Organotransiton Metal Chemistry. University Science Books, Mill Valley, CA. 1987, p 383-392. 7. Parshall, G. W. Homogeneous Catalysis. Wiley-Interscience: New York; 1980, Chapters 3-5. 8. a) Whitesides, G. M.; Gaasch, J. F.; Stredonsky, E. R. /. Am. Chem. Soc. 1972, 94, 5258. b) Reger, D. L.; Culbertson, E. C. / . Am. Chem. Soc. 1976, 98, 2789. c) Kazlauskas, R. J.; Wrighton, M. S. /. Am. Chem. Soc. 1982,104, 6005. d) Kazlauskas, 18 R. J.; Wrighton, M. S. Organometallics 1982,1, 602. e) Watson, P. L.; Roe D. C. / . Am. Chem. Soc, 1982 104, 6471. f) Komiya, S.; Morimato, Y.; Yamamoto, A.; Yamamoto, T. Organometallics 1982,1, 1528. g) Ozawa, F.; Ito, T.; Yamamoto, T. / . Am. Chem. Soc. 1980,102, 6457. 9. a) Conversely, trans hydride-olefin complexes are quite stable.9b>c. b) Olgemoller, B.; Beck, W. Angew. Chem. Int. Ed. Engl. 1980,19, 834. c) Deeming, A. J.; Johnson, B. F. G.; Lewis, J. / . Chem. Soc. Dalton Trans. 1973, 1848. 10. a) Halpern, J.; Okamoto, T. Inorg. Chim. Acta. 1984, 89, L53. b) Roe, D. C. / . Am. Chem. Soc. 1983,105, 7771. 11. Doherty, N. M.; Bercaw, J. E. /. Am. Chem. Soc. 1985,107, 2670. 12. a) Alexander, J. J. in The Chemistry of the Metal Carbon Bond. Hartley, F. R., Ed.; Wiley: New York, 1985,2, Chapter 5. b) Alexander, J. J.; Kuhlman, E. J. Coord. Chem. Rev. 1980, 33, 195. c) Calderazzo, F. Angew. Chem. Int. Ed. Engl. 1977,16, 299. 13. a) Flood, T. C; Campbell, K. O. /. Am. Chem. Soc. 1984, 106, 2853. b) Flood, T. C; Campbell, K. O.; Downs, H. H; Nakanishi, S. Organometallics 1983,2, 1590. 14. Bergman, R. G.; Wax, M. J. /. Am. Chem. Soc. 1981,103, 7028. 15. Brunner, H.; Vogt, H. Chem. Ber. 1981,114, 2186. 16. a) Roper, W. R.; Taylor, G. E.; Waters, J. M.; Wright, L. / . Organomet. Chem. 1979,182, C46. b) Fachinetti, G.; Fochi, G.; Floriani, C. / . Chem. Soc. Dalton Trans. 1977, 1946. c) Curtis, M. D.; Shiu, K.-B.; Butler, W. M. / . Am. Chem. Soc. 1986, 108, 1550. d) Arnold, J.; Tilley, T. D. /. Am. Chem. Soc. 1986,108, 5355. 17. a) Bock, P. L.; Boschetto, D. J.; Rasmussen, J. R.; Demers, J. P.; Whitesides, G. M. /. Am. Chem. Soc. 1974, 96, 2814. b) Whitesides, G. M.; Boschetto, D. J. / . Am. Chem. Soc. 1971, 93, 1521. c) Whitesides, G. M.; Boschetto, D. J. /. Am. Chem. Soc. 1969, 91, 4313. 19 18. Tsuji, J. "Decarbonylation Reactions Using Transition Metal Compounds", in Organic Synthesis via Metal Carbonyls, Wender, I.; Pino, P. Eds.; Wiley-Interscience: New York, 1977,2, 595. 19. a) Chaudret, B.; Dahan, F.; Sabo, S. Organometallics 1985,4, 1490. b) Bavaro, L. M.; Montangero, P.; Keister, J. B. / . Am. Chem. Soc. 1983,105, 4977. c) Bavaro, L. M.; Keister, J. B. /. Organomet. Chem. 1985,287, 357. 20. a) Puddephatt, R. J.; Payne, N.; Ling, S. S. M. Organometallics 1985,4, 1546. b) Ling, S. S. M.; Jobe, I. R.; Manojilovic-Muir, L.; Muir, K. W.; Puddephatt, R. J. Organometallics 1985,4, 1198. c) Scott, J. D.; Puddephatt, R. J. Organometallics 1986, 5, 2522. 21. Schore, N. E.; Ilenda, C. S.; Bergman, R. G. / . Am. Chem. Soc. 1977, 99, 1781. 22. Schore, N. E.; Ilenda, C. S.; White, M. A.; Bryndza, H. E.; Matturo, M. G.; Bergman, R. G. / . Am. Chem. Soc. 1984,106, 7451. 23. a) Hersh, W. H.; Hollander, F. J.; Bergman, R. G. / . Am. Chem. Soc. 1983,105, 5834. b) Hersh, W. H.; Bergman, R. G. / . Am. Chem. Soc. 1983,105, 5846. c) Theopold, K. H.; Bergman, R. G. Organometallics 1982,1, 1571. d) Theopold, K. H.; Bergman, R. G. /. Am. Chem. Soc. 1983,105, 464. e) Theopold, K. H.; Bergman, R. G. Organometallics 1982,1, 219. f) Theopold, K. H.; Becker, P. N.; Bergman, R. G. / . Am. Chem. Soc. 1982,104, 5250. 24. a) Muetterties, E. L. Bull. Soc. Chim. Belg. 1975, 84, 959. b) Muetterties, E. L. Bull. Soc. Chim. Belg. 1976, 85, 451. c) Muetterties, E. L. Science, 1977,196, 839. 25. Knox, S. A. R.; Koepke, J. W.; Andrews, M. A.; Kaesz, H. D. Am. Chem. Soc. 1975, 97, 3942. 26. Owen, S. M. Polyhedron 1988, 7, 253. 27. a) Adams, R. D.; Selegue, J. P. in Comprehensive Organometallic Chemistry; Wilkinson, G.; Stone, F. G. A.; Abel, E. W. eds. Vol 4, Pergammon: Oxford, 1982, p 1023 ff. b) Kaesz, H. D.; Lavigne, G. Chapter 4 in Metal Clusters in Catalysis, Gates, B. 20 C; Knozinger, H.; Guzci, L. eds. Studies in Surface Science and Catalysis, Vol 29, Elsevier: New York, 1986. 28. Shapley, J. R.; Tachikawa, M /. Organomet. Chem. 1977,124, C19. 29. Bryan, E. G.; Johnson, B. F. G.; Lewis, J. /. Chem. Soc. Dalton Trans. 1977, 1328. 30. a) Manojlovic-Muir, L.; Muir, K. W.; Lloyd, B. R.; Puddephatt, R. J. /. Chem. Soc. Chem. Commun. 1985, 536. b) Ferguson, G.; Lloyd, B. R.; Puddephatt, R. J. Organometallics 1986,5, 344. 31. a) Lloyd, B. R.; Bradford, A. M.; Puddephatt, R. J. Organometallics 1987, 6, 424. b) Bradford, A. M.; Jennings, M. C.; Puddephatt, R. J. Organometallics 1988, 7, 792. 32. a) See references given in 17b and 17c. b) Rashidi, M.; Puddephatt, R. J. /. Am. Chem. Soc. 1986, 108, 7111. c) Douglas, G.; Manojlovic-Muir, L.; Muir, K. W.; Rashidi, M.; Anderson, C. M.; Puddephatt, R. J. J. Am. Chem. Soc. 1987,109, 6527. 33. Green, M.; Bottrill, M. / . Organomet. Chem. 1976, 111, C6. 34. Yoshida, T.; Okano, T.; Thorn, D. L.; Tulip, T. H.; Otsuka, S.; Ibers, J. A. / . Organomet. Chem. 1979,181, 183. 35. Sivak, A. J.; Muetterties, E. L.; Day, V. W.; Fredrich, M. F.; Reddy, G. S.; Pretzer, W. R. / . Am. Chem. Soc. 1977, 99, 8091. 36. Fryzuk, M. D. Can. J. Chem. 1983, 61, 1347. 37. Fryzuk, M. D.; Einstein, F. W. B.; Jones, T. Organometallics 1984, 3, 185. 38. Sivak, A. J.; Muetterties, E. L. / . Am. Chem. Soc. 1979,101, 4878. 39. Burch, R. R.; Shusterman, A. J.; Muetterties, E. L.; Teller, R. G.; Williams, J. M. / . Am. Chem. Soc. 1983,105, 3546. 40. Muetterties, E. L. Inorg. Chim. Acta 1981,50, 1. 41. Burch, R. R.; Muetterties, E. L.; Shultz, A. J.; Gebert, E. G.; Williams, J. M. /. Am. Chem. Soc. 1981,103, 5517. 42. McKenna, S. T.; Muetterties, E. L. Inorg. Chem. 1987,26, 1296. 43. McKenna, S. T.; Muetterties, E. L.; Andersen, R. A. Organometallics 1986,5, 2233. 21 44. Burch, R. R.; Muetterties, E. L.; Day, V. W. Organometallics 1982,1, 188. 45. Teller, R. G.; Williams, J. M.; Koetzle, T. F.; Burch, R. R.; Gavin, R. M.; Muetterties, E. L. Inorg. Chem. 1981,20, 1806. 46. Fryzuk, M. D.; Jones, T.; Einstein, F. W. B. / . Chem. Soc. Chem. Commun. 1984, 1556. 47. a) Fryzuk, M. D.; Piers, W. E.; Rettig, S. J. /. Am. Chem. Soc. 1985, 107, 8057. b) Fryzuk, M. D.; Piers, W. E. Polyhedron 1988, 7, 1001. 48. Fryzuk, M. D.; Piers, W. E.; Albright, T. A.; Rettig, S. J.; Einstein, F. W. B.; Jones, T. /. Am. Chem. Soc, submitted for publication. 49. Fryzuk, M. D.; Piers, W. E. Organometallics 1988, 7, 2062. 22 CHAPTER 2 Synthesis and Fluxional Behaviour of Binuclear Rhodium Polyhydrides and their Use in the Hydrogenation of Olefins 2.1 Introduction. As noted in Chapter 1, the family of polynuclear rhodium hydrides of the formula [(L)2Rh(p>H)]n (n = 2, 3, or 4, depending on the steric bulk of L) has been an important one in the study of coordinatively unsaturated metal clusters. While the first reported binuclear member of this family, [{P(OPri)3)2Rh(u.-H)]2, le,1 contained a monodentate phosphite ligand, our research has centred around analogues of le which incorporate chelating diphosphine ligands of varying electronic properties. This chapter describes a route to complexes la-d, their properties, and reactivity towards dihydrogen. A discussion of the mechanism by which these dimers hydrogenate simple olefins is also presented. 2.2 Synthesis and Properties of Binuclear Rhodium Dihydrides. 2.2.1 Synthetic Procedures. Published syntheses of the dimers la, 2 d,3 and e,1 using the route depicted in Scheme 2-1, were employed, with some variations, for the syntheses of the dimers used throughout this work. An improved synthesis of le, in which the chloro-bridged analogue of le was treated with the mild hydride reagent KBHXOPr^  was reported recently (equation 2-l)-4 23 (Pr'0)3P a PfOPr'ja (OPr^P. H PfOPr'h ^ / N / „ k THF \ / \ / o n Rh Rh + 2KBH(OPri)3 T H F . m Rh W , y \ / \ , / \ H / \ (Prbfep' u P(OPr')3 (OPr')3P^ H P(OPt)a le As shown in Scheme 2-1, the reaction of slightly more than 2 equivalents of freshly prepared allylmagnesium chloride with [(COD)Rh(ji-Cl)]2 proceeds smoothly at 0°C in THF to yield 2 equivalents of the monomelic [(COD)Rh(r)3-C3H5)], 2. The crude product is a yellow-brown oil which may be purified by sublimation under high vacuum (10"4 Torr) at 50-60°C to a probe cooled to -78°C. The bright orange product is thermally unstable at room temperature, and this instability leads to significant decomposition during the sublimation process. In addition, the complex melts at ~30°C.^ Thus, while warming the sublimator la-d 3a-d 24 probe to room temperature prior to isolation of the product, the pure material often melted and dripped off of the end of the probe. This rather maddening situation was exacerbated on particularly warm days. These drawbacks to the purification of 2 led us to investigate the possibility of using the unpurified crude material for the next step in the sequence. The iff NMR spectrum of the crude material showed it to be >95% pure, and subsequent experiments showed that it was sufficiently pure for the next step. Thus, addition of 0.9 equivalents of a diphosphine ligand to a toluene solution of 2 led to solutions of the bis(phosphine)rhodium allyl complexes 3a-d. Although purification of these complexes was also possible, generally the crude materials were carried through to the next step and exposed to 4 atmospheres of pure dihydrogen without purification. After stirring efficiently for 2-3 days, the dihydrides la-b, and d were isolated analytically pure after recrystallization of the reaction residues from toluene/hexanes. The dimer incorporating l,4-bis-(diisopropylphosphino)butane, dippb, was isolated as its tetrahydride hydrogen adduct (vide infra), but could easily be reduced to the dihydride l c via dehydrogenation with an excess of 1-hexene (equation 2-2). Rh 2(dippb) 2(H) 4 ^ ^ ^ ^ ' Rh 2(dippb) 2(H) 2 + (2-2) 5c l c The overall yield (from [(COD)Rh(p>Cl)]2) of this sequence, carried out with no purification of complexes formed en route, was only slightly greater than if 2 was purified before the addition of diphosphine. For example, l a was made in an overall 62% yield when 2 was not purified prior to use; with purification, the corresponding yield ranged from 50-60% depending on the success of the sublimation procedure. Since the yield was certainly no worse than when 2 was purified, proceeding without purification was adopted as the method of choice due to its far greater convenience. 25 A wide variety of bis(phosphine)rhodium allyl complexes may be formed via phosphine displacement of COD in 2 as shown in Scheme 2-1.5 It appears that the hydrogenation of such compounds leading to polynuclear hydrides is a less general reaction. It has been observed that the nuclearity of the complexes produced in this reaction depends largely on the steric bulk associated with the phosphine or diphosphine ligand used.1'3 Thus, when the R substituent in R2P(CH2)2PR2 is methoxy, a tetranuclear cluster is produced, while in the case where R is isopropoxy, a dimeric complex is the major product.3 Ligands with yet greater steric bulk, either through modification of R or additional methylene units in the backbone, preclude formation of polynuclear complexes. For example, when R = Bul and there are three methylenes in the backbone, the hydrogenation of the allyl complex proceeds to give unidentified mononuclear products with much decomposition to rhodium metal observed. This unwillingness to form dimers when bulky R substituents are employed is not surprising in light of the stability of the mononuclear complexes L2RI1H [L = P(But)3 and P(C6Hn)3].6'7 These complexes did not dimerize, whereas when L was P(Pri)3, the 14 electron species was not stable, reacting readily with itself to form a binuclear dihydride or with dinitrogen to form a stable dinitrogen complex (see Scheme 1-2).7-8 The formation of polynuclear products is also precluded when the R substituent is an aryl group. When common diphosphine ligands such as dppp or chiraphos were employed in the sequence shown in Scheme 2-1, the last step failed completely, yielding unidentified products. Likely the dihydride dimers are formed, but decompose via orthometallation reactions. One more pitfall surfaces when proceeding without purification of 2 prior to addition of the bidentate phosphine. Since it is difficult to ascertain exactly how much of 2 is present in the crude product, the exact equivalency of diphosphine cannot be calculated. Unfortunately, completion of the displacement reaction is not generally apparent from the colour changes involved. Addition of 0.9 equivalents of diphosphine based on the weight of crude 2 ensures that no excess phosphine is present, something which must be avoided because the presence of excess diphosphine during the hydrogenation step leads to formation 26 R = P r ' , 3a H2,4 atm. toluene (2-3) of a tris diphosphine dimer upon addition of dihydrogen (equation 2-3). This compound (4) was prepared inadvertantly when, during the preparation of 3a, an erroneously high amount of dippe was added to the solution of crude 2. Exposure to dihydrogen led to a bright orange solution rather than the usual dark brown solution characteristic of the hydrogen adduct of la. !H and 31P{1H} NMR spectra and microanalytical data are consistent with the structure proposed for 4. It thus appears that excess dippe traps an intermediate formed upon hydrogenation of the allyl ligand before two mononuclear fragments can combine to form the desired binuclear products. (Interestingly, 4 is also produced in a slow reaction between la and 0.5 equivalents of dippe). Even when minor amounts of free dippe are present during the hydrogenation reaction, 4 is produced, contaminating the dihydride products due to its similar solubility in hydrocarbon solvents. Care is thus essential to avoid this contaminant. 2.2.2 Characterization and Properties of the Dihydrides la-Id. The dihydrides la-c are all black green solids which form deep green solutions in most common solvents. Dimer Id, incorporating the phosphinite ligand 1,3-bis-(diisopropoxyphosphino)ethane, dipope, is a purple-red solid forming deep red solutions. Each metal centre in these dimers is formally a 16 electron centre; however, the dimer as a 27 whole has only 28 electrons in total and is thus quite electron deficient from an electron counting perspective. All four dihydrides are readily soluble in hydrocarbon and moderately polar solvents such as diethyl ether and THF. They are only sparingly soluble in strongly polar solvents such as acetonitrile, and often react stoichiometrically with such solvents (Chapter 5). Like many transition metal hydride complexes, la-d are unstable in halogenated solvents,9 decomposing mainly to the chloro-bridged analogues of la-d. The NMR spectral data collected for la-d suggest they are isostructural. In the *H NMR spectra, the low field region shows signals for the ligand protons, the patterns manifesting the high symmetry of these complexes. The hydride region reveals complex signals for the bridging hydrides of dimers la-c at -4 to -5 ppm. A less complicated signal for the hydrides of Id comes at -4.3 ppm, and is similar in appearance to the hydride resonance of le. 1 0 The complexity of the hydride signals for la-c when compared to those of ld-e, suggest that, unlike ld-e, the hydrides do not exchange rapidly at room temperature. Indeed, the hydride resonance for la is invariant in appearance from -80 to +60° C, except for variable amounts of line-broadening, and appears as an overlapping triplet of triplet of triplets, which is the pattern expected for a stereochemically rigid structure. The hydrides of ld-e appear as a triplet of quintets, indicating the exchanging hydrides are coupling equally to all four phosphorus nuclei. 3 1P decoupling of the hydride resonances reduces the multiplets in each dimer to a sharp triplet due to coupling to the two equivalent rhodium nuclei, confirming the binuclearity of these complexes. The 31P{1H} NMR spectra of these dimers are comprised of a complex doublet arising from an AA'A"A"'XX' spin system. Figure 2-1 shows the 31P{1H} spectrum of la, along with a computer simulation based on the above mentioned spin system. The complexity arises from long range coupling between phosphorus nuclei on either side of the dimer, four-bond P-P coupling is of approximately the same magnitude as the two-bond P-P coupling. Magnetic communication between these nuclei takes place through the Rh2(|i-H)2, 4-centre, Rh Rh <Y -<>-l a i i i i i — , — 1—— 1 -200 0 200HI ure 2-1. a) 121.4 MHz 3lp {lH} NMR spectrum of [(dippe)Rh]2(M> H)2, l a . b) Calculated spectrum based on an AA'A"A"'XX' spin system (see experimental for parameters). 29 4-electron bond and is a common feature of the 3 1 P spectra observed in many of the complexes discussed in this thesis. In order to firmly establish the molecular geometry of the reactive Rh2(|i-H)2 core of this central compound in our ongoing research on the reactivity of polynuclear unsaturated clusters, a single crystal X-ray analysis of la was performed.11 Figure 2-2 shows the molecular structure of la along with the atom numbering scheme, and Table 2-1 gives some of the more important bond distances and angles. The structure was indeed found to be dimeric, with two crystallographically independent molecules per asymmetric unit, while the unit cell contained 4 molecules of la. Each rhodium centre has distorted square planar geometry, the two planes joined via the edge-shared hydride ligands. Bond distances and angles are all within expected ranges. In particular, the Rh-Rh distances of 2.629(1)A (molecule 1) and 2.6266(8)A (molecule 2) are nearly identical to that found for le, at 2.647A. As expected, the P-Rh-P bond angles of 85.75(5)° are less than those found for le (94.7°). This decrease in the P-Rh-P angle has two effects. First, it increases the H-H separation significandy. This distance is lengthened to about 2.3A in la from 2.0A in le. How much of the difference in reactivity between la and le may be attributed to this tightening of the P-Rh-P angle, with concomitant widening of the H-H separation, cannot be known without more rigorous calculations, but it is not unreasonable to expect that it could affect the relative energy levels of the HOMO or LUMO in these complexes.12 The second effect of a smaller P-Rh-P angle is that of "pulling back" the isopropyl groups on phosphorus thus exposing the Rh2(|i-H)2 core of the molecule to a greater extent than other members of this dihydride series. This is likely the major reason why la and Id react more readily with small organics than lb or le-the reactive site in these molecules is simply more accessible to incoming reactants. Enlargement of this angle due to increased chelate ring size, as is expected in lb, 1 3 pushes the isopropyl groups forward towards the centre of the molecule hampering the approach of these reactants significantly. 30 Figure 2-2. ORTEP drawing of [(dippe)Rh]2(^ -H)2, la. 31 Table 2-1. Selected Bond Parameters for C28H66P4R112 (la). Molecule 1 Distances (A) Rh(l)-P(l) 2.182(1) Rh(l)-P(2) 2.180(1) Rh(l)-H(l) 1.70(5) Rh(l)-H(l') 1.81(5) C(13)-C(23) 1.428(9) Rh(l)-Rh(l") 2.629(1) Angles (deg) P(l)-Rh(l)-P(2) 85.75(5) P(l)-Rh(l)-H(l) 94(2) P(l)-Rh(l)-H(l') 176(2) P(2)-Rh(l)-H(l) 174(2) P(2)-Rh(l)-H(l') 97(2) Rh(l)-H(l)-Rh(l') 97(2) Molecule 2 P(2)-C(21) 1.859(6) P(2)-C(22) 1.849(6) P(2)-C(23) 1.843(5) P(l)-C(ll) 1.831(6) P(l)-C(12) 1.835(6) P(l)-C(13) 1.819(6) Rh(l)-P(l)-C(13) C(ll)-P(l)-C(12) C(ll)-P(l)-C(13) C(12)-P(l)-C(13) H(l)-Rh(l)-H(l') 111.8(2) 103.9(3) 100.8(4) 102.4(3) 83(2) Distances (A) Rh(2)-P(3) Rh(2)-P(4) Rh(2)-H(2) Rh(2)-H(3) Rh(2)-Rh(3) Angles (deg) P(3)-Rh(2)-P(4) P(3)-Rh(2)-H(2) P(3)-Rh(2)-H(3) P(4)-Rh(2)-H(2) P(4)-Rh(2)-H(3) H(2)-Rh(2)-H(3) Rh(2)-H(2)-Rh(3) 2.187(1) Rh(3)-P(5) 2.188(1) 2.180(1) Rh(3)-P(6) 2.183(1) 1.68(4) Rh(3)-H(2) 1.86(4) 1.72(4) Rh(3)-H(3) 1.72(4) 2.6266(8) 85.76(5) P(5)-Rh(3)-P(6) 94(1) P(5)-Rh(3)-H(2) 176(1) P(5)-Rh(3)-H(3) 179(1) P(6)-Rh(3)-H(2) 97(1) P(6)-Rh(3)-H(3) 83(2) H(2)-Rh(3)-H(3) 96(2) Rh(2)-H(3)-Rh(3) 85.49(5) 176(1) 98(1) 99(1) 176(1) 78(2) 99(2) 32 Indeed, the chemistry of lb reveals a reluctance to react with even moderately bulky organics which react readily with la and Id. All four of the dimers discussed above are extremely air-sensitive. Solid samples require only seconds to decolourize when exposed to the atmosphere, decomposing to unknown species. Dihydrides la-b (and presumably lc) are known to be water sensitive, reacting stoichiometrically with one equivalent of H2O as shown in equation 2-4.14b The |i-hydroxo |i-hydrido products are stable to further reaction with water. It has also been observed that la reacts with a variety of other Bronsted acids (equation 2-5) yielding complexes in which the conjugate base of the acid is incorporated into a bridging position. Gas is evolved in this reaction (presumably dihydrogen), suggesting an oxidative addition of the acid followed by loss of dihydrogen through reductive elimination as a plausible pathway for this process. This reactivity towards Bronsted acids has led to some interesting reactivity studies which will not be discussed in further detail here.14 R2 R2 r\ ,\ (CH 2) n Rh .Rh (CH 2) n + HzO THF r\ A (CH 2) n Rhv Rh •p R2 p-R2 v . . . . . (CH 2) n + H 2 (2-4) V \ J R = Pr' n = 2, dippe, la; 3, dippp, lb R2 R2 2^ 2^ R2 R2 R2 Rz X = OPh, OR R = Pr', la SH NH(Ph), NHR, NR2 PH(Ph), P(Ph)2 (2-5) 33 2.3 The Reactions of la-c with Dihydrogen: The Characterization of Fluxional Binuclear Tetrahydrides. 2.3.1 Variable Temperature *H NMR Spectra. The dihydrides la-c react rapidly with dihydrogen to yield the highly fluxional tetrahydride complexes 5a-c, (equation 2-6). Whereas 5c was found to be a stable R 2 R 2 /-'\ / \ /~\ (CH 2) n Rh Rh ( O v _ P / V \ - i R 2 Ra R = Prl , n = 0,1,2 , la-c tetrahydride, the other members of this series, 5a and 5b were stable only in the presence of excess dihydrogen in solution, and could not be isolated in the solid state. This instability is similar to that observed for the dihydrogen adduct of le [{P(OPri)3}2Rh]2(H)4, 5c 1 However, like 5e, both 5a and 5b could be studied and characterized spectroscopically. All tetrahydrides (5a-c) were found to be highly fluxional in solution, but it was possible to fully characterize the processes involved in this fluxionality based on the variable temperature *H, and particularly the 31P{ !H} NMR spectra of these complexes. Qualitatively, the variable temperature *H NMR spectra for 5a-c were very similar to those reported for SeA Figure 2-3 shows the hydride region of the *H NMR spectrum of 5b at various temperatures. At room temperature and above, the hydride region of the spectrum of 5b shows a broad signal very near coalescence, indicating that in this temperature regime, all the hydrides were rapidly exchanging with each other. Exchange with free dihydrogen dissolved in the sample solution was also observed. As the temperature was lowered, this exchange with dissolved dihydrogen stopped, and exchange between the four coordinated Figure 2-3. Hydride region of the 400 MHz *H NMR spectrum of the fluxional tetrahydride 5b, recorded at various temperatures. 35 Scheme 2-2 hydrides was partially frozen out. At about -60°C, a doublet of broad multiplets (-2.35 ppm, TH-P =116 Hz) appeared downfield in the hydride region due to two bridging hydride ligands coupled to a pair of trans phosphorus nuclei. Two other resonances, at coalescence near this temperature, continued to sharpen as the temperature was lowered until finally, at -100°C, they were resolved into peaks due to one other bridging hydride (-5.50 ppm) and a terminal hydride ligand further upfield at -11.15 ppm. These spectra are consistent with a rocking process (Scheme 2-2) which equilibrates the terminal hydride (Ht) and the unique bridging hydride (Hb*)- Thus, the *H NMR spectra of these complexes provides much information regarding the multiple hydride exchanges occurring, but yields no information regarding the stereochemistry about each rhodium centre. A similar process for 5e was reported by Muetterties et al.,1 but the authors were indistinct in their assessment of geometry about the metal centres in the ground state structure containing three bridging hydrides and one terminal hydride. A crystal structure of the moderately stable tetrahydride complex [{P(N(CH3)2)3}2Rh]2(H)4,5f,15 was later obtained 36 (NMe2)3 P H ^ \ ^ ^P(NMe 2 ) 3 'P(NMe2)3 P (NMe2)3 5f by these authors, revealing its structure to be a dimer consisting of one octahedral rhodium (III) centre and one square planar rhodium (I) centre joined by two bridging hydrides. In this tetrahydride, the phosphorus nuclei on the rhodium (III) centre are trans disposed across the RI12H4 plane. This result led the authors to propose a similar geometrical arrangement for the excited state structure of 5e. However, the temperature dependence of both the *H and the 31P{1H} NMR spectra for 5f was minimal, indicating that it was not fluxional over the temperature ranges studied and that perhaps such a structure was not involved in the fluxional process observed for 5e. It is noteworthy that the use of chelating phosphines precludes a structure with trans disposed phosphines. 2.3.2 Variable Temperature 31P{!H} NMR Spectra. The 31P{1H} NMR spectrum of 5b at various temperatures (Figure 2-4) provides clear evidence for the stereochemical arrangement about each metal centre at all temperatures. At room temperature, two doublets were observed at 61.6 (^Rh^-P = 167.3 Hz) and 67.6 (^ RrXirQ-P = 105.8 Hz) ppm (at higher temperatures these signals collapsed into one doublet as all four hydrides exchange with free dihydrogen and each other). The magnitudes of the ^Rh-P coupling constants were consistent with a Rh(I)/Rh(ni) formulation for the dimer.16 As the temperature fell, the doublet due to the phosphorus nuclei coupling to the Rh(I) centre Figure 2 - 4 . 1 2 1 . 4 MHz 31p {1H} NMR spectra of the fluxional tetrahydride 5b, recorded at various temperatures. 38 coalesced and reemerged at -100°C as two doublets of doublets. The only stereochemical arrangement for the ground state structure consistent with this result is an octahedral Rh(ITI) centre, coupled with a trigonal bipyramidal geometry at the Rh(I) centre. As H t and Hb' equilibrate via the rocking motion depicted in Scheme 2-2, the two phosphorus nuclei on the rhodium (I) centre undergo axial/equatorial exchange, while the other two phosphorus atoms remain equivalent at all times. Although the trigonal bipyramidal geometry would undoubtedly be highly distorted from the ideal, the square pyramidal geometry would not give rise to two separate signals in the low temperature limit. The excited state structure is therefore likely tetrahedral at the Rh(I) centre in 5b, although the spectra do not rule out a square planar geometry. The activation energy for this "rocking" process is about 8.6(5) kcal mol"1, which is near the value of 9 kcal mol-1 reported for 5d. 1 The 31P{ !H} NMR spectra of 5c are qualitatively identical to those of 5b as a set, but are broadened and complicated by the fact that the chemical shifts for the two rhodium (III) phosphorus atoms and the downfield resonance for one of the rhodium (I) phosphines are very nearly coincident in the low temperature limiting spectrum. Nonetheless, it is apparent that the two sets of spectra are analogous, and that the same geometries may be invoked for 5c as for 5b. The activation energy for this low temperature ("rocking") process for the larger chelate ring size dimer is about 11.5(5) kcal mol"1. The tetrahydride 5a, however, presents a qualitatively different set of spectra, (Figure 2-5). Since the variable temperature proton NMR spectra for 5a are virtually identical to those of 5b and 5c, the explanation for the simpler patterns observed in both the low and high temperature limiting 31P{1H} NMR spectra probably lies in a geometry change at the rhodium centres in the ground and excited state structures. The spectral behaviour in Figure 2-5 is what might be expected should the rhodium (I) centre in the structures shown in Scheme 2-2 be square pyramidal in the ground state and square planar in the excited state structure. With these geometries the phosphorus nuclei on either rhodium centre are always equivalent during the fluxional movements of the molecule (Scheme 2-3), giving rise to the Figure 2-5. 121.4 MHz 3 1P {lH) NMR spectra of the fluxional tetrahydride 5a, recorded at various temperatures. 40 Scheme 2-3 . • • R h t ^ , H b ' ^ R h . . simpler patterns in the 31P{ ^H) spectra of 5a. Although there are some peculiar aspects to the spectra in Figure 2-5, making direct interpretation difficult, this seems the best explanation for all spectral observations. The fact that there are at least three observable exchange processes occurring in these systems accounts for the spectral complexity. This observed dependence of the geometry about the rhodium (I) centre on the chelate ring size of the diphosphine ligand may be rationalized partially on the basis of steric arguments. On isomerization from the square pyramidal to the trigonal bipyramidal geometry (ground state), steric interactions between the isopropyl groups of the opposing diphosphine ligands are relieved. Similarly, these interactions may be lessened by decreasing the bite angle of the chelating ligand and "pulling back" the isopropyl groups on phosphorus. Evidently the contraction of bite angle by about 10° (vide supra) was enough to render isomerization from square pyramidal to trigonal bipyramidal unnecessary for 5a. In 5b and 5c however, this isomerization is required to stagger the isopropyl groups and relieve the steric tension encountered. The chelate ring size also affects the free energy of activation for 41 the exchange of the terminal hydride and the unique bridging hydride ligands: for the seven-membered dippb system, AG$ = 11.5(5) kcal/mol; for the six-membered dippp sytem, AG* = 8.6(5) kcal/mol; and for the five-membered dippe ligand, AG* = <7.0(5) kcal/mol (estimated). 2.4 The Hydrogenation of Olefins Using la and lb as Catalyst Precursors. 2.4.1 The Hydrogenation of 1-Hexene: Results. The activity of le towards the hydrogenation of olefins1 led us to explore the ability of la and lb to bring about this transformation. Neither proved to be more active than le in this capacity, but we were also interested in probing the reaction mechanism in more detail. la was found to be a reasonably efficient catalyst precursor for the hydrogenation of 1-hexene at ambient temperature and pressure. With a substrate/catalyst ratio arbitrarily set at about 1700, hexane was generated with a turnover rate of about 850-950 hr1 (Table 2-II). Rapid secondary isomerization steps also occurred, with production of cis and trans 2-hexene competing with that of hexane. The catalyst hydrogenated the terminal olefin selectively until it was used up, and then began to hydrogenate the internal olefins, producing hexane at a slightly lower turnover frequency of about 700 hr 1. 1 7 Compound lb was also active as a catalyst precursor for this reaction, but was significantly slower under the same conditions. Turnover rates are shown in Table 2-IL In this reaction, some evidence for the build-up of 3-hexenes was also observed. Notably, while relative rates of isomerization were higher for lb, the overall rate of hexane production was steady, reflecting similar rates in the hydrogenation of terminal and internal olefins or a kinetic requirement for isomerization prior to hydrogenation. Table 2-TL. Hydrogenation and Isomerization Rates3 for the Hydrogenation of 1-hexene by la and lb. b  Catalyst Loss of 1-hexene Products Internal Olefin Hydrogenation hexane r-2-hexene c-2-hexene r-2-hexene c-2-hexene la 2600 850-950 1300-1700 750-1100 520 240 lb 2100 280 1150 700 210 170 aAll rates are given as the turnover frequencies in hr1. bHydrogenation reactions carried out at 1 atm pressure of H2, and 23-25°C. 43 A cursory study of the dippb tetrahydride 5c showed that its activity was not significantly different from that of the dippp dimer lb and it was therefore not studied in any detail. 2.4.2 Mechanistic Studies. Chemical evidence abounds for a mechanistic pathway for hydrogenation which involves solely binuclear intermediates. The tetrahydrides 5a-c react smoothly with one equivalent of olefin to bring about dehydrogenation of the dimers to the dihydrides with concomitant production of alkane (equation 2-7). It was this reaction, observed for 5e, k P.- . H . / Pv H . P V which led Muetterties et al. to propose that the primary pathway for hydrogenation by le involved exclusively binuclear intermediates (Scheme 2-4).1 The reaction of the dihydride le with dihydrogen leads to the tetrahydride 5e which adds one equivalent of alkene to the coordinatively unsaturated Rh(I) centre to produce a binuclear alkene tetrahydride intermediate (in the geometry suggested by the isolation of 5f15); a subsequent migratory insertion and reductive elimination regenerates the dimer le and produces alkane. Application of this proposal to our systems requires a modification of the catalytic cycle shown in Scheme 2-4 to take into account the use of chelating bidentate ligands that cannot span trans coordination (2-7) R = Pr\ n = 0,1,2 l a - c 44 Scheme 2-4 H — R h — ; R h 5e H. H P I Rh I P i ^ P Rh .Rh. .Rh I ^ H ^ H alkene tetrahydr ide H 2 C = C H R alkene tetrahydr ide R h - — H - ^ R h CH 2 CH 2 R C H 3 — CH 2 R Rh ^ R h le 45 sites. An alternate geometry for the binuclear alkene-tetrahydride intermediate fulfills this requirement and is shown below. Further chemical evidence for a binuclear hydrogenation cycle exists in a crossover experiment. When a mixture of equal amounts of la and lb was used to hydrogenate a large excess of 1-hexene, the mixed dimer [(dippe)Rh](|i-H)2[Rh(dippp)], expected should fragmentation/aggregation occur, was not detected after workup. This observation does not totally elirninate the possibility of crossover, but does show that (within the detection limits of 31P{ lH} NMR) crossover is minimal. The possibility of fragmentation induced by the olefin substrate in the absence of dihydrogen may also be eliminated based on chemical evidence. It is well documented that la, 2 and to a limited extent l b , 1 8 react with simple olefins to generate ii-alkenyl hydride complexes of the general formula {[(Pri)2P(CH2)nP(Pri)2]Rh}2[|i-T|2-a-C(R)=CH(R)](M:-H). These complexes react rapidly with dihydrogen to yield alkane and the dihydrides la or lb H H (CH2)n (CHdn + H 2 (CH2)n (CH2)„ (2-8) R2 2^ R = Pr1 , n = 0,1 la-b 46 after removal of excess dihydrogen (equation 2-8). Crossover experiments showed that these two reactions proceed without fragmentation; when a mixture of la and lb were treated with excess ethylene, only the expected vinyl hydride analogs were observed in the product mixture; i.e. no [(dippe)Pvh](p:-r|2-a-CH=CH2)(|i-H)[Rh(dippp)] was observed. When this mixture of vinyl hydrides was subsequently treated with dihydrogen, la and lb were regenerated smoothly upon work-up with no trace of mixed dimer observable by 31P{1H} NMR spectroscopy. Although these vinyl hydride complexes are likely not involved in the proposed mechanistic pathway involving binuclear species (vide infra), it is necessary that the reaction of these complexes with dihydrogen proceeds without fragmentation if the binuclear integrity of the cycle is to be maintained, since la in particular reacts rapidly with olefin substrates at room temperature to form these derivatives. However, since the dihydrides react much faster with dihydrogen than with olefins, under catalytic conditions the vinyl hydride dimers are not likely involved in the cycle subsequent to hydrogenation of the p>alkenyl ligand. This chemical evidence certainly suggests that a pathway which contains only binuclear intermediates exists. It does not, however, totally eliminate the possibility that fragmentation occurs to yield a highly reactive mononuclear species which may (or may not) be responsible for the majority of the catalysis observed. This is a long standing problem encountered in cluster catalysis19' 2 0 and chemical evidence does not provide sufficient grounds to state firmly that catalysis by the cluster is the major mechanistic pathway in the reaction. In order to determine this with confidence it is necessary to carry out kinetic analysis on the reaction and even then it is sometimes difficult to reach meaningful conclusions. One of the kinetic criteria arguing for cluster catalysis is an increase in turnover frequency [TF = mol product/(mol catalyst)(unit time)] upon increasing catalyst concentration.20 If cluster and fragment are in equilibrium, then as the total metal concentration decreases, the concentration of the fragment, relative to that of the cluster, increases. If the cluster were the active catalyst, a decrease in total metal concentration would 47 result in a decreased TF. However, if fragmentation is occurring to yield a catalytically active fragment, then as the total metal concentration decreases, the concentration of the fragment relative to the cluster increases and a plot of TF vs total metal concentration will show an increase in TF with decreasing metal concentration. This test can thus give some indication as to which is the active catalyst in such a system, the cluster or the fragment. For this reason, catalyst concentration studies were carried out using varying amounts of lb as the catalyst precursor. Styrene was used as the substrate to eliminate isomerization side reactions, and the dihydrogen uptake was monitored as a function of time at 30°C. Figure 2-6 shows a plot of the rate of dihydrogen uptake (mols sec-1 L_ 1) vs the concentration of lb. Although it appears to be linear at low [lb], the curvature at higher catalyst concentration suggests fragmentation. The first order behaviour at low [lb] is due to the fact that at such concentrations the dimer is essentially fully fragmented and the reaction is first order in the fragment. A better correlation (R = 0.95) is obtained in when the rate is plotted against [Rh]1/2 (Figure 2-7), providing further evidence for fragmentation. Finally, when the turnover frequency is plotted vs catalyst concentration (Figure 2-8), it is clear that the TF decreases with increasing [lb]. This provides kinetic evidence that, in this system, fragmentation occurs such that the majority of the actual catalysis is done by small amounts of a highly reactive mononuclear fragment. Unfortunately, similar kinetic analysis of the reaction involving la was hampered by this compound's extreme sensitivity to minor amounts of oxygen and water inevitably included under the conditions employed. The scatter apparent in each of Figures 2-6, 2-7, and 2-8 is likely due to this factor as well, particularly at low catalyst concentrations, but lb was generally more well-behaved under the monitoring conditions. Thus, it cannot be said that mononuclear fragments carry out the majority of catalysis in the reaction utilizing la as a catalyst precursor. Indeed, the chemical evidence suggests that the pathway involving binuclear intermediates is the major one in the la catalyzed reaction. The results obtained in 48 5.006-5 0.00e+0 2.00e-3 4.00e-3 6.00e-3 8.00e-3 [Rh] (mols/l) Figure 2-6. Plot of the rate of hydrogen uptake vs the catalyst concentration in the hydrogenation of styrene using [(dippp)Rh]2(|J.-H)2, lb, as catalyst precursor. o.ooe+o H 1 1 1 1 • 1 • 1 • 0.00 0.02 0.04 0.06 0.08 0.10 Square root of [Rh] (mols/l) Figure 2-7. Plot of the rate of hydrogen uptake vs the square root of catalyst concentration in the hydrogenation of styrene using [(dippp)Rh]2(M.-H)2, lb, as catalyst precursor. 49 Figure 2-8. Plot of turnover frequency vs catalyst concentration in the hydrogenation of styrene using [(dippp)Rh]2(p>H)2, lb, as catalyst precursor. the crossover experiments (vide supra) would also be expected if only one of la or lb were fragmenting under the hydrogenation conditions. Based on both sets of evidence (chemical and kinetic) we propose that both pathways exist (Scheme 2-5), but that for lb at least, the one utilizing binuclear intermediates is very slow. Attempts to detect fragmentation of the tetrahydride 5b in the absence of substrate alkenes by deviation from Beer-Lambert behaviour, especially at low concentration, were unsuccessful; plots of absorbance vs [5b] were linear for concentrations spanning a range of 0.14 mM to 0.0012 mM. Thus it can be concluded that the mononuclear pathway for hydrogenation is kinetically much more efficient than the binuclear one since mononuclear fragments have not yet been spectroscopically detected. A number of features of the proposed binuclear cycle (top in Scheme 2-5) are different than that given in Scheme 2-4.1'15 As already mentioned, the dippp olefin tetrahydride 50 Scheme 2-5 Rh ,, .-Rh " H, y \ (i-alkyl intermediate H,C R H H^JLcHR T|:-alkene tetrahydride P.. 1 _.-H ~ Rh-P 1 .--H "Rh-H2C==CHR P 1 ,-H Rh;- ~R CH3 51 complex has trans terminal hydrides on the Rh(HI) centre. The migratory insertion step to form the p>alkyl complex is also in contrast to that in Scheme 2-4; to account for the substantial and rapid olefin isomerization, a bridging, secondary alkyl ligand is invoked. Although speculative, the bridging alkyl is in line with other proposals that involve dehydrogenation of the dihydride dimers la and lb by olefins18 and dienes (see Chapter 3). Literature precedent also exists for stable bridging alkyls in related binuclear rhodium systems.21 In addition, the alkyl must be secondary at some point to explain the olefin isomerization; how this occurs is not explicitiy detailed, but would obviously involve a B-eUmination back to the olefin tetrahydride but with an internal olefin coordinated to the Rh(I) centre. Reductive elimination from the u.-alkyl intermediate generates the product alkane and the dihydride lb. The bifurcation to the mononuclear cycle (bottom Scheme 2-5) is proposed to occur at the tetrahydride species 5b. Rather than bind the olefin, the tetrahydride can fragment under dihydrogen to generate the mononuclear trihydride, [(dippp)RhH3], as shown. This coordinatively unsaturated fragment undergoes the usual olefin binding, migratory insertion and reductive elimination to produce alkane and the 14 electron, monohydride [(dippp)RhH]. The mononuclear cycle can continue by oxidative addition of dihydrogen to this monohydride to regenerate the trihydrido species, or the monohydride monomer can dimerize to lb and thus gain entrance back into the binuclear regime. A possible reason for the fragmentation pathway in the dippp dimer lb system is again the increased steric crowding about the metal centres as compared to the dippe dimer la. The observed differences in the geometries about the rhodium(I) centre in the ground state structures of the tetrahydrides may hamper the alkene binding by 5b, and lead to fragmentation, while with 5a, the binding of the alkene is facilitated. It thus appears that contact between the catalyst and the substrate is significantly hindered in the pathway invoking dimeric intermediates for the larger chelate ring system and that when such contact is made, fragmentation is induced and the mononuclear regime dominates. 52 2.5 Experimental. The beginning of the experimental section for this chapter contains a description of the general techniques used throughout this work. Subsequent chapters will describe pertinent techniques used in the work which that chapter describes, but laboratory procedures and chemicals common to all of the work are described here. 2.5.1 General Procedures. All manipulations were performed under prepurified dinitrogen in a Vacuum Atmospheres HE-553-2 glovebox equipped with a MO-40-2H purifier, or in standard Schlenk-type glassware under argon. The description "reactor bomb" refers to a cylindrical, thickwalled pyrex vessel equipped with a 5 mm Kontes needle valve and a ground glass joint for attachment to a vacuum line. Larger bombs are fitted with 10 mm Kontes valves. "Sealable" NMR tubes are 9" 507 PP NMR tubes with a bl4 socket attached via glass blowing. Hydrated rhodium trichloride was obtained from Johnson-Matthey and used as received to prepare [(COD)RhCl]222a and [(COE)2RhCl]222b by literature methods. The secondary phosphine HPPr^  was prepared by a literature method23 and was used to prepare the bidentate ligand, Pri2PCH2CH2CH2PPri2, dippp.24 The ligands Pr^ PCH^CH^PPr^ , dippe,2 and (PriO)2PCH2CH2P(OPri)2, dipope,3 were also prepared from published procedures. The starting material used in the synthesis of these ligands, 1,2-bis-(dichlorophosphino)ethane, was either synthesized from a published procedure25 or purchased from Strem Chemical Co. The syntheses of the dihydrides la and Id have appeared previously,2-3 but details of a slightly modified procedure are included below. The syntheses of c?2—la, lb, and Id were accomplished simply by exposing toluene solutions of la-b or d to 1 atmosphere of deuterium gas, followed by work-up and recrystallization.2 53 Hydrogen gas used in hydrogenation reactions and in the syntheses of the binuclear hydrides was purified by passing through a glass tower (4" in diameter) packed with MnO (bottom layer) and 4 A molecular sieves separated by a plug of glass wool and a layer of glass beads. The towers were assembled by packing with 200g MnC03 and 1700 mL Vermiculite (mixed) followed by the molecular sieves. Activation was accomplished by passing dihydrogen gas through the tower while heating at about 200°C until the brown colour of the MnCC>3 had turned the characteristic green of MnO. Heating was effected by passing current through a wire wrapped around the tower in spirals separated by about 1". Deuterium gas was obtained from Matheson Gas Products and used without further purification. Molecular sieves were purchased from Fischer Scientific Co. and activated by heating to 200°C under high vacuum (10"2 Torr). Toluene, dimethoxyethane (DME), and diethylether were purified by distillation from sodium-benzophenone ketyl under argon. Acetonitrile and acetone were dried by refluxing over calcium hydride for eight or more hours, followed by distillation under argon. Tetrahydrofuran (THF) and hexanes were pre-dried by refluxing over calcium hydride followed by distillation from sodium-benzophenone ketyl under argon. Melting points were determined on a Mel-Temp apparatus in sealed capilliaries under nitrogen and are uncorrected. Carbon, hydrogen, and halogen analyses were performed by Mr. P. Borda of this department. *H NMR spectral measurements were carried out on one of the following instruments: Varian XL-300, Bruker WP-80, or a Bruker WH-400. All iHf31?} spectra were obtained from the Bruker WH-400. 31P{1H} NMR measurements were carried out at 121.421 MHz on the Varian XL-300 or at 162.21 MHz on the Bruker WH-400 using trimethylphosphite as an external standard (141.0 ppm vs H3PO4). 13C{ *H} NMR measurements were performed at 75.429 MHz on the Varian instrument, using internal solvent peaks as a reference (solvent peaks referenced to TMS at 0 ppm). 2 H NMR spectra were obtained on the Bruker WH-400 at 61.402 MHz. NMR solvents C6D6, C7D8, CD3CN, and (CD3)2CO were purchased from MSD Isotopes, and dried over activated 3A molecular 54 sieves and vacuum transferred. Dg-THF was dried over sodium-benzophenone ketyl and stored under dry nitrogen. Simulated NMR spectra were obtained using the program UBCPANIC.26 AG* values for exchange processes observed via NMR spectroscopy were calculated using the value for the rate constant kc (kc = 7cAvc/V2, where Avc = the chemical shift separation in hertz at coalescence, is estimated roughly from the separation just prior to coalescence temperature, Tc) at the coalescence temperature27 in the Eyring equation: AG* = -RTcln(7tAvch/\2kTc) (R = gas constant; h = Planck's constant; k = Boltzmann's constant; transmission coefficient = 1). Estimates of T c were made visually and therefore have an error of approximately ±3 K. Spin-lattice relaxation times (Ti) were measured using the inversion recovery method;28 spectra obtained in the experiment were analyzed using the Varian XL-300's Ti routine (software version 6.0c). Analyses of hydrocarbon mixtures were done by gas chromatography on a Varian 6000 instrument equipped with FID detectors and a 20% tri-o-cresylphosphate on 60/80 chromosorb-P column (20' x 1/8" SS) operating at 50°C. UV-vis spectral measurements were performed on a Perkin Elmer 552A UV-vis spectrometer. Infrared spectra were recorded on a Nicolet 5DX FTIR instrument. The X-ray crystal structure determination for la was carried out by Dr. F.W.B. Einstein at Simon Fraser University.11 The 1,4-dibromobutane used in the preparation of dippb was obtained from the Aldrich Chemical Co. and distilled under argon prior to use. The substrates 1-hexene and styrene (99%) were also obtained from Aldrich. 1-Hexene was dried for 8 hours over activated 3A molecular sieves, vacuum transferred into a storage vessel, and degassed thoroughly via 3 freeze-pump-thaw cycles. It was then passed through neutral alumina and stored under N 2 in the glove box. Styrene was distilled under argon just prior to use. Ethylene was obtained from Matheson Gas Products and used without purification. Hydrogen gas used for the kinetic measurements on the hydrogenation of styrene was purified by passing through an Englehard Deoxo hydrogen purifier attached to the hydrogen cylinder. 55 2.5.2 Synthesis of Binuclear Di- and Tetrahydrides. Dihydrides: Modified Syntheses of [(dippe)Rh]2(p>H)2, la, and [(dipope)Rh]2(|i-H)2, Id. The complex [(COD)Rh(ri3-C3H5)] was synthesized from 2.5 g (5.1 mmol) of [(COD)Rh(ji-Cl)]2 as previously described.1 After extraction into hexanes and complete removal of same, the crude orange-brown oil was weighed and redissolved in toluene. To 56 this solution was added, with stirring, 0.9 equivalents of dippe or dipope based on the weight of crude 2. This reaction mixture was loaded into a large thick-walled reactor bomb and degassed. After cooling to -196°C, dihydrogen was admitted to 1 atmosphere; the reaction was then stirred for 1 day and repressurized with dihydrogen at -196°C. Stirring was then continued for a further 1-2 days. Removal of excess dihydrogen and toluene, followed by recrystallization from toluene/hexanes (la, 1:1; Id, 0:1) yielded 2.30 g la (3.1 mmol, 62%), A A ' P \ X / H \ X / P Rh Rh A m m M or 2.57 g Id (3.0 mmol, 59%). Parameters used in the simulation of the 31P{1H} NMR spectrum of la: J A A ' = 30.0 Hz; J A A " = 30.0 Hz; JAA'" = 0.0 Hz; J A x = 165.0 Hz; J A X ' = 5.5 Hz. The Synthesis of [(dippp)Rh]2(Lx-H)2, lb. To a solution of [(COD)Rh(r|3-C3H5)] ( 1.00 g, 3.97 mmol) in toluene (25 mL) was added a solution of the ligand dippp (1.09 g, 3.97 mmol) in toluene (10 mL), yielding a golden yellow solution of [(dippp)Rh(rj3-C3H5)]. Toluene and 1,5-cyclooctadiene were removed under reduced pressure, the yellow residues redissolved in toluene, and the solution transferred to a large reactor bomb equipped with a magnetic stir bar. The solution was degassed on a vacuum line, cooled to -196°C, and pressurized with dihydrogen to 1 atmosphere. The reactor bomb was sealed and allowed to warm to room temperature. Efficient stirring was continued for 1 day, after which the reaction vessel was repressurized with dihydrogen at -196°C. After a further day of stirring, the excess dihydrogen and the solvent were removed leaving a black-green solid from which 1.32 g (88.0%) of lb could be 57 obtained after 3 recrystallizations from toluene/hexanes. Mp 225-7°C. *H NMR (CgDg, ppm): CH(CH3)2, 1.83 (dsp, 2 J P = 2.2 Hz, 3 j H = 6.9 Hz); CH(CH3)2, 1.59, 1.22 (dd, 3j P = 12.1 Hz); CH^CH^CH ,^ 0.97 (m); Rh-H-Rh, -6.65 (br m, lJ R h = 35.2 Hz). 31p{lH.} NMR (C6D6, ppm): 50.3 (d, !jRh = 161.0 Hz). Anal. Calcd. for C3oH70P4Rh2: C, 47.37, H, 9.28%. Found: C, 47.49, H, 9.21%. The Synthesis of [(dippb)Rh]2(H)4, 5c. To a toluene solution of [(COD)Rh(r|3-C3H5)] (0.356 g, 1.4 mmol) was added dropwise a solution of dippb (0.409 g, 1.4 mmol) in toluene (5 mL) with stirring. A rapid deepening of the orange colouring of the solution signified the reaction was complete upon addition of the diphosphine. The resulting solution was placed in a 250 mL thick-walled reactor bomb, degassed, cooled to -196°C, and exposed to 1 atmosphere of purified dihydrogen. After 3 days of efficient stirring (the bomb was repressurized with dihydrogen after 1 day) the yellow-orange to yellow-green colour change was complete and the excess dihydrogen and toluene were removed in vacuo. The dark green-brown residue was recrystallized from toluene/hexanes (1:3) at -20°C, yielding 2-3 crops of black-green plates of 5c (0.407 g, 73% from [(COD)Rh(Ti3-C3H5)]). *H NMR (C 7H 8 , ppm, 25°C): PCH2(CH2)2CH2P, 1.75 (m, 8H); CH(CH3)2, 1.68 (dsp, 8H, partially obscured by backbone resonances); PCH2(CH2)CH2P, 1.47 (br m, 8H); CH(CH3)2, 1.35 (dd, 24H, 3 j H = 6.8 Hz, 3 j p = 12.8 Hz); CH(CH3)2, 1.16 (dd, 24H, 3 j H = 7.0 Hz, 3 j p = 12.8 Hz); Rh-H4, -9.93 (br m). *H NMR (C7H8, ppm, -70°C): -7.83, 2Hb (br dm, 2 J P = 119 Hz, U R H » 18 Hz); -12.1, Hb- (br m); -18.74, H t (br m). 31p{lH} NMR (C7D8, ppm, +60°C): 67.5 (d, l J R h = 146 Hz). 31p{lH} NMR (C7D8, ppm, -90°C): 73.9 (br d, lJRh = 111 Hz); 69.3 (br d, URh = 182 Hz). Anal. Calcd. for C 3 2H 7 6P 4Rh 2: C, 48.61; H, 9.69%. Found: C, 48.49; H, 9.80%. 58 The Synthesis of [(dippb)Rh]2(p>H)2, lc. The generation and observation of lc was carried out in an NMR tube. A solution of 5c was dissolved in C 6 D 6 and placed in a sealable 5 mm NMR tube. The tube was fitted to a 180° needle valve adapter and the whole assemblage attached to a vacuum line. After degassing, about 1.1 equivalents of 1-hexene were vacuum transferred into the tube and the tube sealed. *H NMR spectra, recorded immediately after the tube had warmed to room temperature, showed the presence of hexane and lc in nearly quantitative yield. *H NMR (C 6H 6, ppm): CH(CH3)2, PCH^Cr^hCHjjP, 1.80-1.93 (overlapping signals, 16H); CH(CH3)2, 1.57 (dd, 24H, 3 j H = 6.4 Hz, 3 j p = 13.2 Hz); P C H 2 ( C H 2 ) 2 C H 2 P , 1.34 (br m, 8H); CH(CH3)2, 1.19 (dd, 24H, 3 j H = 7.6 Hz, 3 j P = 11.6 Hz); Rh-H-Rh, -8.03 (br m, 2H, l TR h = 34.0 Hz). 3lp{lH) NMR (C 7 D 8 , ppm): 64.7 (d, URh = 166.2 Hz). The Reaction of [(dippe)Rh]2(n-H)2, la, and [(dippp)Rh]2(Li-H)2, lb with Dihydrogen. An appropriate amount of la or lb (25-30 mgs) was dissolved in C7D8 and the solution loaded into a sealable 5 mm NMR tube. The tube was fitted with a 180° needle valve adapter and the assemblage attached to a vacuum line with access to purified dihydrogen. The sample was degassed and dihydrogen was admitted to nearly 1 atmosphere of pressure. Reaction with dihydrogen was immediate as indicated by a rapid deep green to brown colour change. To ensure the presence excess dihydrogen, the sample was cooled to -196°C before sealing. 31p{lH} NMR spectroscopic data obtained are discussed in the text. *H NMR (hydride region of 5a, -80°C, C 7 H 8 , ppm): 2Hb, -7.05 (br dm, 2 J P = 122.1 Hz, ^Rh = 20.6 Hz); Ht/Hb', -13.17 (br m). Further cooling to -95°C results in nearly complete coalescence of this high field signal. *H NMR (hydride region of 5b, -90°C, C 7 H 8 , ppm): 2Hb, -2.35 (br dm, 2 J P = 116 Hz, l T R h = 19.2 Hz); Hb-, -5.50 (br m); H t, -11.15 (br m). 59 2.5.3 Miscellaneous Syntheses. The Synthesis of (Pri)2P(CH2)4P(Pri)2, d«PPb. The procedure used was identical to that used for the ligand (Pri)2P(CH2)3P(Pri)2, dippp.24 Starting with 2.22 g (0.018 mol) of LiP(Pri)2 and 1.05 mL (1.90 g, 0.009 mol) Br(CH2)4Br, 1.98 g (77%) of dippb was obtained. Bp: 100-102°C (0.1 mm Hg). *H NMR (C6H6, ppm): CH2(CH2)2CH2, 1.65 (br m, 4H); CH(CH3)2, 1.59 (dsp, 4H, 3jH = 7.6 Hz, 2 J P = 2.8 Hz); CH2(CH2)2CH2, 1.32 (br m, 4H); CH(CH3)2, 1.02, 1.06 ( overlapping dd, 24H, 3jP = 12.9, 13.3 Hz). 31p{lH} NMR (C6D6): 2.41 (s). Spectral Data for [(dippp)Rh(n3-C3H5)], 3b, and [(dippb)Rh(Ti3-C3H5)], 3c. The compounds [(dippp)Rh(ri3-C3H5)] and [(dippb)Rh(r|3-C3H5)] may be isolated via recrystallization from toluene/hexanes as yellow crystals in 92% and 78.7% yields for 3b and 3c, respectively. Spectroscopic data for [(dippp)Rh(ti3-C3H5)], 3b: *H NMR (C6D6, ppm): Hcent, 4.86 (m, 3j H s y n = 7.0 Hz, 3j H a m i = 12.8 Hz, 2 J R h = 2.0 Hz); H s y n , 3.65 (d); Hanti, 2.19 (dd, J P = 4.8 Hz); CH(CH3)2, 1.92 (br dsp, 3jH = 7.0 Hz), 1.81 (br dsp, 3jH = 7.0 Hz); CH(CH3)2, 1.20 (dd, 3 J P = 14.4 Hz), 1.10 (two overlapping dd, 3JP= 12.9, 13.6 Hz), 1.01 (dd, 3jP= 12.8 Hz); CH 2CH 2CH 2, -1.8, 1.2 (resonances obscured by other peaks). 3ip{iH} NMR (C6D6): 42.9 (d, 1JR h= 188.0 Hz). Anal. Calcd. for Ci8H39P2Rh: C, 51.42, H, 9.35%. Found: C, 51.22, H, 9.33%. *H NMR spectral data for [(dippb)Rh(Ti3-C3H5)], 3c. 1H NMR (C6H6, ppm): 4.80, Hcent (tt, 3 J H a n t i = 11.6 Hz, 3j H s y n = 7.0 Hz); 3.44, H s y n (dd, 2 J H a n t i = 1.1 Hz); 2.10, Hanti (ddd, J P = 5.1 Hz); 1.92, CH(CH 3) 2 (dsp, 3jH =7.0-7.4 Hz; 2 J P » 2.6 Hz); 1.45-1.85, CH(CH3)2, (CH2)4 (overlapping multiplets); 1.18, 1.06, 1.05, 0.98, CH(CH3)2 (overlapping dd, 3 J P = 13.0-13.4 Hz). 60 Spectral data for [(dippe)Rh(H)]2(p.-dippe), 4. R 2 4 When >1.5 equivalents of dippe are added to a solution of 2, exposure of the reaction mixture to dihydrogen leads to formation of 4. This dimer may be isolated analytically pure (in =75% yield) as a bright orange powder via recrystallization from toluene/hexanes. *H NMR (C6D6, ppm): CH(CH3)2, 2.28 (dsp, 4H, 3 j H = 6.8-7.0 Hz, 2 J P * 3 Hz); 2.03 (overlapping dsp, 8H, 3 j H = 6.9 Hz); CH(CH3)2, 1.47, 1.37, 1.33, 1.26, 1.06, 1.03 (overlapping dd, 3 J P = 13.0-16.0 Hz); PCH2CH2P, 1.15 (m, 4H, other backbone proton signals buried under methyl resonances); Rh-H, -5.59 (dddd, 1H, 2Jptrans = 109.3 Hz, ijRh = 26.7 Hz, 2 J p d , ( c h e l a t i n g ) = 17.7 Hz, 2JPc£l(teidging) = 17.7 Hz). 31 P { 1 H } NMR (C6D6): Prra»,(chelating), 102.7 ( l j R h = 149.0 Hz, 2 J P , r f l l w = 302.6 Hz, 2 J P d , = 21.8 Hz); Pm(chelating), 84.1 (m, ^Rh = 144.0 Hz); P;ra«j(bridging), 56.7 (2Jpd5 = 16.1 Hz). Anal. Calcd. for C4 2 H 9 8 P6Rh 2 : C, 50.70; H, 9.93%. Found: C, 50.95; H, 10.10%. 61 2.5.4 Crossover Experiments. The Hydrogenation of 1-hexene with [(dippe)Rh]2(|i-H)2, la, and [(dippp)Rh]2(n-H)2, lb. A toluene solution (10 mL) of la (0.080 g) and lb (0.075 g) was loaded into a small thick-walled reactor bomb. After degassing, 5.0 mL of 1-hexene was syringed into the bomb under a strong flow of dihydrogen. The hydrogenation was allowed to proceed to completion by stirring overnight under 4 atmospheres of dihydrogen. The volatiles were then removed and the residues examined by 31P{ *H} NMR spectroscopy. The only peaks observed were the doublets characteristic of la and lb. Stoichiometric Reactions. A toluene solution (10 mL) of la (0.052 g) and lb (0.046 g) was loaded into a small thick-walled reactor bomb, degassed and exposed to an excess of ethylene. (Ethylene was used because it is the only simple olefin with which lb will react stoichiometrically in convenient reaction times).18 The reaction mixture was stirred until the deep green to red-orange colour change was complete. The solvent was removed and the residues were analyzed by 31P{1H} NMR spectroscopy. Only signals attributable to the unmixed vinyl hydride dimers, i.e., {[(Pri)2P(CH2)nP(Pri)2]Rh}2(M--H)(n-ri2-a-C2H4), n = 2, 3, were observed. This mixture of vinyl hydrides was redissolved in toluene and treated with excess dihydrogen. Analysis of the products of this reaction, after removal of excess dihydrogen and toluene, showed regeneration of la and lb in quantitative yield with no trace of the mixed dihydride dimer observed. 62 Synthesis of the Mixed Dimer [(dippe)Rh(|i-H)2Rh(dippp)]. P \ / \ / Rh Rh P / H \ A mixture of [(dippe)Rh(ri3-C3H5)] (generated from 0.150 g [(COD)Rh(Ti3-C3H5)] and 0.156 g dippe) and [(dippp)Rh(r)3-C3H5)] (generated from 0.150 g [(COD)Rh(ri3-C3H5)] and 0.164 g dippp) was dissolved in 25 mL of toluene and loaded into a 250 mL thick-walled reactor bomb. The solution was degassed and cooled to -196°C; dihydrogen was admitted to 1 atmosphere of pressure. After a 2-3 day period of efficient stirring, excess dihydrogen and toluene were removed under reduced pressure. The residues were redissolved in hexanes (30 mL) and filtered through a pad of Celite. Concentration of the filtrate to 10-15 mL, followed by cooling to -20°C resulted in isolation of a crop of red-black crystals (0.217 g, 49% on mass balance). Analysis by 31P{ lH} NMR spectroscopy revealed the presence of l a and l b (=10% in total) as well as signals expected for the AA'MM'XY spin system in the mixed dimer (=80%, the remaining peaks due to side products). 31P{ *H} NMR (C6D6): 103.7, P(CH2)2P (dm, l J R h = 157.3); 52.7, P(CH2)3P (dm, l J R h = 150.9 Hz). 63 2.5.5 The Hydrogenation of Olefins: Procedures. The Hydrogenation of 1-Hexene with [(dippe)Rh]2(Li-H)2, la> a n d [(dippp)Rh]2(p>H)2, lb. A 250 mL thick-walled glass bomb equipped with a magnetic stir bar was charged with catalyst precursor (la or lb, 0.050 g) and purified 1-hexene (10.00 g, substratexatalyst ratio =1700-1800:1). la reacted immediately with the substrate, producing an orange solution, whereas lb was unreactive and the intense green colour characteristic of the dihydride was maintained. The bomb was then attached to a vacuum line with access to purified dihydrogen and the solution degassed thoroughly via 3 freeze-pump-thaw cycles. When the solution had warmed to room temperature (22-23°C), dihydrogen was admitted to one atmosphere. Samples were periodically removed through the bomb's teflon needle valve port using an argon and hydrogen flushed syringe. The volatiles were vacuum transferred away from the catalyst, and analyzed by gas chromatography. Identification of the peaks in the GC trace was carried out by "spiking" with the pure materials (Aldrich). Kinetic Measurements. Kinetic measurements on the hydrogenation of styrene by lb were made by following the uptake of dihydrogen at constant pressure. The apparatus was identical to that described by James and co-workers.29 Samples for each run were prepared from varying amounts (0.5-4.0 mL) of a stock solution of lb (0.048 g in 25 mL toluene, 1.58 x 10"2 M). The total volume of each sample was made up to 4.0 mL with toluene. Higher concentrations of catalyst were attained by dissolving accurately weighed amounts of lb (0.016 and 0.024 g) in 4.0 mL of toluene. Styrene was then added (0.55 mL; substratexatalyst ratio ranged from =150-3800:1), and the reaction vessel assemblage attached to the constant pressure uptake apparatus. The sample solution was degassed with two freeze-pump-thaw cycles and then 64 Table 2-III. Data plotted in Figures 2-6, 2-7, and 2-8. Rate(x 105)a [lb](x lO3) [lb]0.5(x 102) Turnover Frequency(x 102) mol sec-1 H moll"1 (mol H)0-5 (mol product) (mol cat)-1 sec-1 0.66 0.28 1.67 2.11 1.42 0.56 2.36 2.24 1.87 1.11 3.33 1.48 2.21 0.83 2.88 2.34 2.35 2.22 4.71 0.93 2.68 1.67 4.09 1.41 3.05 4.62 6.80 5.42 4.48 7.22 8.50 5.66 aEstimated error: ±10% allowed to thermally equilibrate in an oil bath thermostatted to 30°C. Once equilibrated, shaking was begun and dihydrogen admitted to 1 atmosphere of pressure. After a brief (2 minute) induction period, gas uptake was monitored by measuring the difference in levels of a liquid of negligible vapour pressure in a capilliary manometer thermostatted together with a gas burette at 25°C. Admission of dihydrogen to the system through a needle valve maintained a constant pressure by raising the mercury level in a calibrated tube of the burette. A travelling microscope was used to follow the change in height of this mercury column. Runs were typically allowed to go for 4-5 turnovers. The reproducibility of each rate value was estimated at ±10% when two runs were repeated. The sole product in the reaction was ethylbenzene. Data obtained and plotted in Figures 2-6,2-7, and 2-8 are given in Table 2-in. 2.6 References. 1. Sivak, A. J.; Muetterties, E. L. / . Am. Chem. Soc. 1979,101, 4878. 2. Fryzuk, M.D.; Jones, T.; Einstein, F.W.B. Organometallics 1984,5, 184. 3. Fryzuk, M. D. Can. J. Chem. 1983, 61, 1347. 65 4. McKenna, S. T.; Muetterties, E. L. Inorg. Chem. 1987, 26, 1296. 5. Fryzuk, M. D. Inorg. Chem. 1982, 21, 2134. 6. Yoshida, T.; Okano, T.; Otsuka, S. /. Chem. Soc. Chem. Commun. 1978, 855. 7. Yoshida, T.; Okano, T.; Thorn, D. L.; Tulip, T. H.; Otsuka, S.; Ibers, J. A. / . Organomet. Chem. 1979,181, 183. 8. Yoshida, T.; Okano, T.; Thorn, D. L.; Otsuka, S.; Ibers, J. A. /. Am. Chem. Soc. 1980,102, 6451. 9. Collman, J. P.; Hegedus, L. S.; Norton, J. R., Finke, R. G. Principles and Applications of Organotransition Metal Chemisty. University Science Books, Mill Valley CA (1987); p 93. 10. Sivak, A. J. PhD thesis, University of California at Berkeley, 1979. 11. Fryzuk, M. D.; Piers, W. E.; Einstein, F. W. B.; Jones, T. Can. J. Chem. submitted for publication. 12. a) For example, Otsuka et al.l2b have shown that the relative energy of frontier orbitals in ML 2 systems changes dramatically with changing L-M-L angle, b) Otsuka, S. / . Organomet. Chem. 1980,200, 191. 13. a) See for example: James, B. R.; Mahajan, D.; Williams, G. M.; Rettig, S. J. Organometallics 1983, 2, 1452. Interestingly, this angle is also in the range of P-Rh-P angles in complexes containing non chelating monodentate ligands.13b (b) Teller, R. G.; Williams, J. M.; Koetzle, T. F.; Burch, R. R.; Gavin, R. M.; Muetterties, E. L. Inorg Chem. 1981,20, 1806. 14. a) Fryzuk, M. D; Jang, M.-L.; Jones, T.; Einstein, F. W. B. Can. J. Chem. 1986, 64, 174. b) Fryzuk, M. D.; Piers, W. E. unpublished results. 15. Meier, E. B.; Burch, R. R.; Muetterties, E. L.; Day, V. W.. / . Am. Chem. Soc. 1982, 104, 2663. 16. It has been found empirically16b and calculated theoretically160 that the magnitude of 1 jRh-P is about 50% larger for complexes with a formal oxidation state of I than for those 66 with an oxidation state of III. b) Meek, D. W.; Mazanec, T. J. Acc. Chem. Res. 1981,14, 266. c) Nixon, J. F.; Pidcock, A. Ann. Rev. NMR Spectrosc. 1969,2, 345. 17. Piers, W. E. B.Sc. Thesis, University of British Columbia, 1984. 18. Fryzuk, M. D.; Piers, W. E. Polyhedron 1988, 7, 1001. 19. Muetterties, E. L.; Krause, M. J. Angew. Chem. Int. Eng. Ed. 1983, 22, 135. 20. Laine, R. M. /. Mol. Catal. 1982,14, 137. 21. a) Schmidt, G. F.; Muetterties, E. L.; Beno, M. A.; Williams, J. M. Proc. Natl. Acad. Sci. USA 1981, 78, 1318. b) Schmidt, G. F. PhD thesis, University of California at Berkeley, 1981. c) Kulzick, M. A.; Andersen, R. A.; Muetterties, E. L.; Day, V. W. / . Organomet. Chem. 1987,336, 221. 22. a) Chatt, J. ; Venanzi, L.M. / . Chem. Soc. 1957, 4753. b) van der Ent, A.; Onderdelinden, A. L. In Inorganic Synthesis ; Wold, A.; Ruff, J. K. Eds.; McGraw-Hill: New York, 1973, Vol 14, p 93. 23. Schellenbeck, P. ; Hoffmann, H. Chem. Ber. 1966, 99, 1134. 24. Tani, K.; Tanigawa, E.; Tatsubo, Y.; Otsuka, S. /. Organomet. Chem. 1985,279, 87. 25. a) Burt, R. J.; Chatt, J.; Hussain, W.; Leigh, G. J. / . Organomet. Chem. 1979,182, 203. 26. PANIC: Parameter Adjustment in NMR by Iteration Calculation. PANIC is a minicomputer version of larger LAOCOON type programs used with the Bruker Aspect 2000 software package. 27. Thomas, W. A. Annu. Rev. NMR Spectrosc. 1968,1, 43. 28. Farrar, T.; Becker, E. D. Pulse and Fourier Transform NMR 1971, Academic Press: New York, p 20-22. 29. a) James, B. R.; Mahajan, D. /. Isr. Chem. 1977,15, 214. b) James, B. R.; Rempel, G. L. Discuss. Faraday Soc. 1968,46, 48. c) James, B. R.; Rempel, G. L. Can. J. Chem. 1966, 44, 233. 67 CHAPTER 3 Synthesis and Fluxional Behaviour of Binuclear Rhodium Complexes with Bridging Dienyl Ligands: New Bonding Modes for 1,3-Dienes 3.1 Introduction. That ccordination of an unsaturated hydrocarbon to a metal complex leads to activation of that hydrocarbon fragment is one of the tenets of organometallic chemistry.1 Simple hydrocarbon units such as olefins, dienes and arene rings are generally unreactive in their native state but upon interaction with a metal become susceptible to attack by rather mild nucleophiles and/or electrophiles.2 The growing prominence of organometallic reagents in organic synthesis attests to this observation.3 Much of our understanding of the activation of unsaturated hydrocarbon ligands by transition metals arises from structural and theoretical studies on mononuclear complexes.4 From these studies, a protocol for predicting5 reactive sites on a coordinated n-system is available for most unsaturated hydrocarbon moieties. With polynuclear complexes on the other hand, much less is known about the activation of unsaturated hydrocarbons. In fact, we are rarely able to predict a priori the binding modes of 7t-ligands to metal clusters let alone sites of reactivity. However, a particularly attractive premise that has grown out of polynuclear cluster chemistry6 is that new ways to activate unsaturated hydrocarbon fragments may be possible if binding to two or more metal centres is achieved. 1,3-Butadiene is a good case in point. This simple acyclic conjugated diene normally binds to mononuclear transition metal centres in the T]4-cw mode7 as in I (Figure 3-1). More recently, the r\*-trans form II has been observed as the kinetic product8 in group 4 68 M M M II III \ M M M IV V Figure 3-1. Previously characterized bonding modes of 1,3-butadiene. metallocenes and as the most stable isomer9 in (Ti5-C5H5)Mo(NO)(n4-C4H6). Certainly in complexes of the type I, reactivity patterns are predictable. With polynuclear metal complexes, the s-trans bridging mode III has been characterized10 as one of the possible binding modes for 1,3-butadiene; for example,103 type in was found in Os3(CO)io(|i-,n2-r|2-C4H6) along with Os3(CO)10(Tl4-C4H6) which exhibits type I binding. The s-cis p:-Ti2-r|2 bridging mode IV is known11 for Cp2Co2(M.-CO)(|i-Ti2-T|2-C4H6); in addition, |i-ri2-r|2 bonding of butadiene in the absence of a metal-metal interaction (type V) has been observed in Cp2Mn2(CO)4(H-Tl2-Tl2-C4H6)12 and [NMe3Et]2[Pt2(Cl)6(li-Tl2-Ti2-C4H6)].13 It is quite clear that all five modes of binding activate 1,3-butadiene in different ways, and types III-V are, of course, only possible with two or more metal centres proximate. This chapter describes the results of a study of the reactions of 1,3-butadiene,14 1-methyl-1,3-butadiene (piperylene), and 2-methyl-1,3-butadiene (isoprene)15 with the two binuclear rhodium hydrides la and lb, incorporating the electron-rich dippe and dippp 69 diphosphine ligands, respectively. Although the only difference between these two starting dihydrides is the size of the chelate ring in the ancillary bidentate phosphine ligand, the product distribution upon reaction with 1,3-dienes varies remarkably. In most cases binuclear compounds are the major products and two new bonding modes for 1,3-dienes to two metal centres are described. A mechanism for these reactions is presented along with a summary of theoretical studies on the nature of the binding of butadiene to two metal centres. 3.2 The Reactions of [(dippp)Rh]2(|i-H)2 and [(dippe)Rh]2(p:-H)2 with 1,3-Butadiene. 3.2.1 The Reaction of [(dippp)Rh]2(p.-H)2, lb, with 1,3-Butadiene. The addition of 5-10 equivalents of 1,3-butadiene to the dippp dimer lb in toluene at room temperature results in a gradual colour change from deep green to orange over a period of 1-2 hours. The major product (50-60%) has the formula [(dippp^h^'C^g, 6, and is obtained as orange crystals from toluene/hexanes at -20°C. Also produced in this reaction is a syn/anti mixture of the mononuclear 1-methylallyl derivatives (dippp)Rh(rj3-l-MeC3H4), 7, in 20-25% isolated yield after sublimation. By 31P{1H} NMR spectroscopy, these are the only two organometallic products generated in the reaction. The ratio of the binuclear complex 6 to the mononuclear derivatives 7 varies from about 2:1 to 1:1 depending on the amount of 1,3-butadiene used; low ratios of 1,3-butadiene to the dippp dimer lb favour the binuclear complex 6. The organic by-products are a mixture of 1-butene, cis-butene, and trans-butene (4:3:1) corresponding to one equivalent based on 6. The reaction of the dideuteride dimer d2-lb with 1,3-butadiene was also examined; no incorporation of deuterium into the binuclear complex 6 was detected by 2 H NMR spectroscopy. Deuterium label was detected in the butene by-products and in the allyl complexes as shown in Scheme 3-1. 70 Scheme 3-1 D The 1H NMR spectrum of 6 (Figure 3-2) consists of three multiplets for the butadiene ligand and a set of resonances for the dippp ligand which is complex but typical of some symmetry in the compound. The 31P{1H} NMR spectrum (Figure 3-3) is of an AA'BB'XX' spin system which indicates that the phosphorus donors of each dippp ligand are inequivalent but both dippp ligands are symmetry related; this pattern is quite characteristic and serves as a useful diagnostic for this particular type of binuclear complex incorporating 1,3-butadiene. From the spectroscopic data alone it is not possible to determine the exact structure and binding mode of the butadiene moiety; however, it is clear that the C4H5 fragment is intact (i. e., none of the C-H bonds have been cleaved) and that the diene itself is cis bound14b (JCtC< = 4.5 Hz).16 Figure 3-2. a) 400 MHz lH NMR spectrum of the partial sandwich complex [(dippp)Rh]2(|i-r|3-r|3-C4H6), 6. b) 3 1P broadband decoupled, c) 3 1P broadband decoupled, homodecoupled at Hi. d) 3 1P broadband decoupled, homodecoupled at H 2. 72 3.2.2 The Solid State Molecular Structure of [(dippp)Rh]2*C4H6, 6. The single crystal X-ray analysis of 6 showed that the 1,3-butadiene was indeed intact and partially sandwiched between the two rhodium centres (Figure 3-4). Each rhodium interacts in an rj3 fashion with the opposite faces of the twisted cis- 1,3-butadiene fragment [torsion angle of 45.0(8)°]. Table 3-1 lists pertinent bond lengths and bond angles. The rhodium dippp moieties are not unusual; the P(l)-Rh-P(2) angle of 96.96(4)° is within the 73 A B P(2) Figure 3-4. a) ORTEP drawing of the partial sandwich complex [(dippp)Rh]2(|J.-Ti3-r|3-C4H6), 6. b) Side view looking down the C(2)-C(2') axis depicting the 45° twist in the bridging dienyl ligand. 74 Table 3-1. Bond Parameters3 for C34H74P4RI12 (6). Distances (A) Rh-P(l) 2.2492(11) C(3)-C(4) 1.517(7) Rh-P(2) 2.1996(11) C(3)-C(4b) 1.43(3) Rh-C(l) 2.139(5) C(4)-C(5) 1.492(7) Rh-C(2) 2.208(4) C(5)-C(4b) 1.38(3) Rh-C(2') 2.297(4) C(6)-C(10) 1.519(7) P(l)-C(3) 1.850(4) C(6)-C(ll) 1.512(7) P(l)-C(6) 1.871(5) C(7)-C(12) 1.515(8) P(l)-C(7) 1.854(5) C(7)-C(13) 1.545(7) P(2)-C(5) 1.850(4) C(8)-C(14) 1.525(7) P(2)-C(8) 1.855(5) C(8)-C(15) 1.508(8) P(2)-C(9) 1.870(5) C(9)-C(16) 1.510(8) C(9)-C(17) 1.534(7) C(l)-C(2) 1.438(7) C(2)-C(2') 1.441(9) Angles (deg) P(l)-Rh-P(2) P(l)-Rh-C(l) P(l)-Rh-C(2) P(l)-Rh-C(2*) P(2)-Rh-C(l) P(2)-Rh-C(2) P(2)-Rh-C(2') C(l)-Rh-C(2) C(l)-Rh-C(2*) C(2)-Rh-C(2') Rh-P(l)-C(3) Rh-P(l)-C(6) Rh-P(l)-C(7) C(3)-P(l)-C(6) C(3)-P(l)-C(7) C(6)-P(l)-C(7) Rh-P(2)-C(5) Rh-P(2)-C(8) Rh-P(2)-C(9) C(5)-P(2)-C(8) C(5)-P(2)-C(9) C(8)-P(2)-C(9) Rh-C(l)-C(2) Rh-C(2)-C(l) Rh-C(2)-Rh' 96.96(4) 166.00(14) 128.15(13) 97.66(11) 96.98(14) 133.68(12) 160.86(12) 38.6(2) 68.5(2) 37.2(2) 118.97(15) 116.9(2) 113.4(2) 100.2(2) 101.9(2) 103.2(2) 119.15(15) 111.6(2) 119.33(15) 102.8(2) 98.8(2) 102.6(2) 73.3(2) 68.1(2) 132.5(2) Rh-C(2)-C(2') C(l)-C(2)-Rh C(l)-C(2)-C(2') Rh'-C(2)-C(2') P(l)-C(3)-C(4) P(l)-C(3)-C(4b) C(4)-C(3)-C(4) C(3)-C(4)-C(5) P(2)-C(5)-C(4) P(2)-C(5)-C(4b) C(4)-C(5)-C(4b) P(l)-C(6)-C(10) P(l)-C(6)-C(ll) C(10)-C(6)-C(ll) P(l)-C(7)-C(12) P(l)-C(7)-C(13) C(12)-C(7)-C(13) P(2)-C(8)-C(14) P(2)-C(8)-C(15) C(14)-C(8)-C(15) P(2)-C(9)-C(16) P(2)-C(9)-C(17) C(16)-C(9)-C(17) C(3)-C(4b)-C(5) 74.7(3) 107.0(3) 120.4(3) 68.0(3) 115.7(3) 115.7(12) 45.3(13) 115.7(5) 116.7(3) 117.6(12) 46.4(13) 110.6(3) 111.1(4) 110.4(5) 111.6(3) 116.4(4) 110.2(4) 116.3(4) 110.8(4) 111.2(5) 110.4(3) 112.4(4) 110.2(5) 131(2) a The symbol"'" denotes atoms symmetry related by an inversion centre. 75 normal range expected17 for six-membered chelate rings. The two [Rh(dippp)] units are skewed with respect to each other as shown in Figure 3-4; the dihedral angle formed by the two planes defined by the Rh, P(l) and P(2) atoms is 75.6(1)°. The Rh-Rh separation of 4.1238(6) A is consistent with the presence of the bridging, partially sandwiched C4H.6 unit inserted between the two metals.18 The dimensions associated with the inner core of this binuclear complex provide some insight into the bonding of this molecule. The carbon-carbon bond lengths of the C4H5 fragment are virtually identical at 1.44 A (C(l)-C(2), 1.438(7); C(2)- C(2)', 1.441(9) A) and are slightly longer than 1.40 A as found in an T|3-cyclooctenyl rhodium derivative.19 The C-C bond lengths found in RhCl(rj4-C4H6)2 are 1.45A for C-C(internal) and 1.38A for C-C(external).20 The diene unit is also twisted about the C(2)-C(2)' bond with a torsion angle of 45.0(8)°, best seen in the view in Figure 3-4b. The rhodium-carbon distances reveal that the rj3 binding mode is asymmetric on each face; Rh-C(l) is 2.139(5) A, Rh-C(2) is 2.208(4) A and Rh-C(2)' is 2.297(4) A. 3.2.3 The Solution State Molecular Structure of [(dippp)Rh]2«C4H6, 6. The solution spectroscopic data for 6 are consistent with the solid-state, partially sandwiched structure; however, the lH and ^CpH} NMR chemical shifts of the diene protons in 6 are more typical of a cis n4- 1,3-butadiene structure than the observed n 3 type structure observed. Table 3-II illustrates this by comparison of the spectroscopic data of 6 to the anti isomer of 7 as well as (t|4-C4H6)Fe(CO)321 and the rhodium (I) butadiene complex [(T|4-QH6)Rh(Cp)].22 In particular, the chemical shift for the protons H a trans to the vicinal protons He is indicative of an s-cis butadiene ligand. These protons characteristically resonate upfield in the 0.5-1.0 ppm region, whereas in Rh(I) allyl complexes such as anti-7, the anti proton H a resonates above 2.0 ppm 2 3 This high field shift for H a is also characteristic of an s-cis bound diene ligand; in an s-trans configuration resonances for protons H a are shifted to 76 Table 3-II. *H and 1 3 C NMR Chemical Shift Comparison Between anti-7,6, and Typical Transition Metal Diene Complexes.  Rh(Cp) *H NMR H a 2.12 ppm H a 0.87 ppm -0.03 ppm 0.34 ppm H s 3.37 ppm H b 2.65 ppm 1.46 ppm 2.45 ppm H s . 4.45 ppm H c 4.73 ppm H c 4.78 ppm 4.89 ppm 4.92 ppm 13C NMR C i 43.1 ppm C i 37.9 ppm 40.53 ppm — C 2 101.6 ppm c 2 60.2 ppm 85.49 ppm — C 3 60.4 ppm lower field [for example, 8Ha= 1.43 ppm in Os3(CO)10(p-'n2-ri2-C4H6)].10a The 1 3 C {lH} NMR spectral data for 6 also suggests solution T | 4 bonding, in that the resonance for C 2 is much closer in chemical shift to a diene C 2 carbon than C 2 in anti-7. However, that the r| 3-T) 3 bonding is maintained in solution is evidenced by the 3 1 P { lH) NMR spectrum in which two signals are observed; in a n.-rj4-r|4 structure, four equivalent phosphines would be expected. In addition, this binding must be quite rigid because neither the *H nor the 3 1 P { *H} NMR spectra of 6 are temperature dependent in the range -80 to +160°C. 77 3.2.4 The Reaction of [(dippe)Rh]2(p>H)2, la, with 1,3-Butadiene. The reaction of 1,3-butadiene (> 2 equivalents) with the dippe hydride dimer la proceeds very rapidly even at -10° C as evidenced by the colour change from dark green to bright orange; yellow crystals of the formula [(dippeJRhk^Hg (8) can be isolated in >90% yield from toluene/hexanes at -30°C. In contrast to the previous reaction of the dippp hydride dimer lb with 1,3-butadiene, binuclear products predominate with la and 1,3-butadiene; only traces of the mononuclear 1-methylallyl complexes (dippe)Rh(r| -l-MeC3H4) 9 are detectable (<1% by 31P{ *£!}). Also produced in this reaction is one equivalent of 1-butene based on the starting dimer la. In analogy to the cfc-lb reaction, the dippe dideuteride ^ 2-la generates 8 with no evidence of deuterium labelling in the dirhodium product; all of the deuterium is incorporated in the 1-butene in the 3,4 positions (Scheme 3-2). The and 31P{ lYL} NMR spectra of 8 are complex and temperature dependent to the extent that unambiguous assignment of its structure by these sporting methods was not Scheme 3-2 P \ Rh 8 excess 2 + d2-la D 78 possible. However, single crystals were obtained from the reaction mixture and subjected to X-ray analysis (vide infra); only by combining the results of this X-ray analysis with the solution spectroscopic data from the dippp reaction, were the structure and dynamic behaviour of 8 unravelled. 3.2.5 The Solid State Molecular Structure of [(dippe)Rh]2«C4H6, 8. The molecular structure obtained through X-ray analysis of a single crystal of 8 is as shown in Figure 3-5. Figure 3-5a shows the structure and atom numbering scheme, while Figure 3-5b is a stereoview of the molecule. There are two dimeric molecules per asymmetric unit, each having similar bond distance and bond angle parameters (Table 3-HI). The butadiene ligand is attached to Rh(l) via a cw-rj4 donation and through a a bond between C(91) and Rh(2), thus bridging the two metal centres in a previously unreported mode of bonding. The Rh(l) centre is of distorted square pyramidal geometry with the bridging hydride occupying the apical position, while Rh(2) is near to being square planar in its geometry.143 The interatomic rhodium-carbon distances in this structure compare favourably with similar parameters found for the u.-rj2-o" vinyl hydride complex [(dippe)Rh]2(M--rj2-o-CH=CH2)(p--H)24 in which the vinyl ligand bridges the two rhodium centres in a mode analogous to C(91) and C(92) in 8. In addition, the Rh-Rh separation of 2.8105(8)A in 8 is just slighdy shorter than that of the 2.8655(5)A distance found in the vinyl hydride complex. Rhodium-phosphorus distances and angles in 8 are also comparable. The bridging hydride ligand was located and refined isotropically and was found to bridge the two metal centres symmetrically, unlike the hydride ligand in the vinyl hydride complex. The C-C distances in the dienyl ligand of 8 are all virtually identical at 1.40A, indicating complete % delocalization over the diene ligand. The presence of a relatively large rj4-a bridging diene ligand and bulky diphosphine Figure 3-5. a) ORTEP drawing of the dienyl hydride complex [(dippe)Rh]2(|i-r|4-a-C4H6), 8a. b) Stereoview of the molecule. 80 Table 3 - i i i . Selected Bond Parameters for C32H7pP4Rh2 (8a). Molecule 1 Molecule 2 Distances (A) Rh(l)-P(l) Rh(l)-P(2) Rh(l)-H(l) Rh(l)-C(91) Rh(l)-C(92) Rh(l)-C(93) Rh(l)-C(94) Rh(2)-P(3) Rh(2)-P(4) Rh(2)-H(l) Rh(2)-C(91) 2.250(2) 2.259(2) 1.82(8) 2.217(7) 2.263(7) 2.179(7) 2.218(8) 2.186(2) 2.241(2) 1.84(8) 2.048(7) Rh(3)-P(5) Rh(3)-P(6) Rh(3)-H(2) Rh(3)-C(101) Rh(3)-C(102) Rh(3)-C(103) Rh(3)-C(104) Rh(4)-P(7) Rh(4)-P(8) Rh(4)-H(2) Rh(4)-C(101) 2.244(2) 2.259(2) 1.73(6) 2.222(7) 2.262(7) 2.171(7) 2.198(8) 2.241(2) 2.184(2) 1.71(6) 2.047(7) C(91)-C(92) 1.42(1) C(101)-C(102) 1.40(1) C(92)-C(93) 1.39(1) C(102)-C(103) 1.41(1) C(93)-C(94) 1.40(1) C(103)-C(104) 1.40(1) Angles (deg) P(l)-Rh(l)-P(2) 85.72(7) P(5)-Rh(3)-P(6) 85.56(7) P(l)-Rh(l)-H(l) 99(3) P(5)-Rh(3)-H(2) 101(2) P(l)-Rh(l)-C(91) 175.6(2) P(5)-Rh(3)-C(101) 175.0(2) P(l)-Rh(l)-C(92) 139.4(2) P(5)-Rh(3)-C(102) 141.2(2) P(l)-Rh(l)-C(93) 107.9(2) P(5)-Rh(3)-C(103) 108.7(2) P(l)-Rh(l)-C(94) 96.8(2) P(5)-Rh(3)-C(104) 95.9(2) P(2)-Rh(l)-H(l) 102(3) P(6)-Rh(3)-H(2) 105(2) P(2)-Rh(l)-C(91) 97.7(2) P(6)-Rh(3)-C(101) 99.2(2) P(2)-Rh(l)-C(92) 104.1(2) P(6)-Rh(3)-C(102) 103.7(2) P(2)-Rh(l)-C(93) 128.9(3) P(6)-Rh(3)-C(103) 128.0(2) P(2)-Rh(l)-C(94) 165.8(2) P(6)-Rh(3)-C(104) 164.6(3) H(l)-Rh(l)-C(91) 83(3) H(2)-Rh(3)-C(101) 79(2) H(l)-Rh(l)-C(92) 117(3) H(2)-Rh(3)-C(102) 112(2) H(l)-Rh(l)-C(93) 123(2) H(2)-Rh(3)-C(103) 119(2) H(l)-Rh(l)-C(94) 92(2) H(2)-Rh(3)-C(104) 90(2) P(3)-Rh(2)-P(4) 87.34(7) P(7)-Rh(4)-P(8) 87.15(7) P(3)-Rh(2)-H(l) 172(2) P(7)-Rh(4)-H(2) 94(2) P(3)-Rh(2)-C(91) 93.8(2) P(7)-Rh(4)-C(101) 175.6(2) P(4)-Rh(2)-H(l) 91(2) P(8)-Rh(4)-H(2) 174(2) P(4)-Rh(2)-C(91) 176.7(2) P(8)-Rh(4)-C(101) 93.6(2) H(l)-Rh(2)-C(91) 88(2) H(2)-Rh(4)-C(101) 85(2) Rh(l)-C(91)-Rh(2) 82.3(2) Rh(3)-C(101)-Rh(4) 82.0(3) Rh(l)-C(91)-C(92) 73.2(4) Rh(3)-C(101)-C(102) 73.4(4) Rh(2)-C(91)-C(92) 132.5(6) Rh(4)-C(101)-C(102) 133.7(6) C(91)-C(92)-C(93) 121.0(7) C(101)-C(102)-C(103) 120.3(7) C(92)-C(93)-C(94) 119.6(7) C(102)-C(103)-C(104) 120.0(7) 81 ancillary ligands leads to a quite sterically congested core in this molecule; this is particularly evident looking at the stereoview of the molecule. In addition to the close proximity to the |i-dienyl ligand of various isopropyl groups, the anti diene proton on C(94) is pointed directly at Rh(2). While the ORTEP diagram implies no direct interaction, a larger substituent in this position may be expected to have more of an effect. These potentially destabilizing steric interactions are discussed in further detail below. 3.2.6 The Solution State Molecular Structure of [(dippe)Rh]2«C4H6, 8. In solution, the complex of the formula [(dippe)Rh]2#C4rl6 is in fact an equilibrium mixture of the dienyl hydride dimer 8a and the partial sandwich complex 8b (equation 3-1). The van't Hoff plot of the temperature dependence of the equilibrium constant is shown in (3-1) (3-2) 82 2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5 1/T(1/Kx1000) Figure 3-6. a) van't Hoff plot for the equilibrium between the dienyl hydride complex 8a and the partial sandwich species 8b in C7Dg solution (linear region), b) The same plot over the full temperature range. 83 Table 3-FV. Values of the Equilibrium Constant, and AG° for Equation 3-1.  Temp, °C Keq- [8b]/[8a] AG0, cal/mol 18.9 0.27 771.4 30.8 0.33 676.1 40.7 0.39 583.4 45.8 0.42 546.0 50.7 0.46 498.5 60.9 0.54 412.4 70.9 0.62 330.7 Figure 3-6; Table 3-IV lists a number of temperatures and the value of the equilibrium constant (defined by equation 3-2) along with the values of AG0. Graphically determined thermodynamic parameters for equation 3-1 are AH0 = 3.25±0.1 kcal/mol, and AS° = 8.50+.0.2 cal/mol. The origin of the complexity of the dippe reaction with 1,3-butadiene was only apparent once the 31P{ lH} NMR spectrum of 8 was measured at 121 MHz and compared to the spectrum of the dippp sandwich 6; initial 31P{ *H} NMR spectra of 8 at lower field (32.4 MHz) were uninterpretable. The diagnostic, temperature independent AA'BB'XX' pattern of 6 mentioned previously was evident in the 31P{ lH) NMR spectrum of 8 along with a more complex temperature dependent pattern; a series of these spectra are shown in Figure 3-7. The !H NMR spectrum of the mixture of 8a and 8b consists of the characteristic resonances for the diene protons of 8b and other broad, featureless peaks associated with 8a; no fine structure or coupling patterns are discernable for 8a even with 3 1P decoupling. A hydride multiplet is evident at -7.6 ppm and is assigned to 8a. All of the *H NMR resonances associated with 8a are temperature dependent. The ^C^H} NMR spectrum of 8 can be assigned. At low temperature, four resonances due to the dienyl carbons of 8a are observed at 156.9 (multiplet, Ci, a-bound to Rh), 123.5 (s, C2), 88.8 (s, C3) and 33.1 ppm (s, C4); in Figure 3-7. 121.4 MHz 31P{ !H} NMR spectra of the equilibrium mixture of 8a and 8b at various temperatures. 85 addition, the two temperature independent multiplets for the diene carbons of 8b are also apparent. As the temperature is raised, the two inner carbons at 123.5 and 88.8 ppm broaden and coalesce to a broad peak at approximately 106 ppm; the remaining two resonances for the terminal carbons are barely observable due to broadening. Because of the temperature dependence and the complexity of the *H, 13C{ ^ H}, and 31P{ *H} NMR spectra for 8a, the solution structure of this dienyl hydride may not have the C4H5 fragment bound in the ii-r]4-a form as found in the solid state. In fact, the only convincing evidence for this formulation is the presence of the hydride multiplet in the *H NMR spectrum which distinguishes this binding mode from the u-r|3-r|3 type for 8b. At low temperatures, this u.-r|4-o" mode is certainly a possibility as it is consistent with all of the spectroscopic data; however, there may be other formulations for 8a. For example, dissociation of one end of the 7t-bound diene to generate a ii-rj2-o" dienyl hydride is possible, and it is known that analogous |i-r|2-a alkenyl-hydride complexes can undergo a "windshield wiper" fluxional process.15- 2 4 However, this type of fluxional behaviour alone is not sufficient to fully account for the temperature dependence of the solution spectra. The variable temperature 31P{ !H) NMR spectra shown in Figure 3-7 require a process or series of processes that make all four phosphorus and both rhodium nuclei equivalent in the fast exchange limit. In the low temperature limit, the complexity of the 31P{ ^ H) NMR spectra is consistent with a structure like the ji-T)4-a mode or perhaps the \i-r\2-o type. Whatever process is invoked must also satisfy the 13C{ lH} NMR results wherein the inner carbons of the dienyl ligand exchange in the fast exchange limit; presumably the terminal carbons are also exchanging in this process. In addition, the observation that the dienyl hydride complex 8a and the partial sandwich 8b are in equilibrium must also be broached. Indeed, the temperature dependence of the solution spectra suggests that this complex 8 is undergoing extensive structural rearrangements. Scheme 3-3 presents two possible scenarios to the explain the exchange processes of the complexes 8a and 8b. The interconversion of the ii-r|4-0" form (A) with the p>T|2-0" mode Scheme 3-3 87 (B) does not exchange any of the carbons on the dienyl unit nor does it make the phosphorus environments identical; however, the two rhodium centres are exchanged. To accomplish the interchange of carbons, two possible transformations are: reductive elimination from A to the cis diene C or hydride migratory insertion to the double bond of B to give F. In both the diene manifold (C, D & E) and the p>butenediyl manifold (F & G), the interchange of rhodium centres, phosphorus donors and carbon nuclei is achieved. There is ample precedent (see introduction) for the proposed structures having a cis bound diene to one metal of a cluster (i.e.; C, C , D & D') and the (x-T|2-r|2-cw form in E. Analogy for the p>ri3-a structure (i. e., F and F') is also available with (CO)5Mn(p:-r|3-a-C4H6)Mn(CO)425 while G has some precedent in [(CO)5Re]2(|i-0"-0"-C4H6).26 The rearrangement of the p>l,2-butenediyl form F (and F') to the u.-l,4-butenediyl structure G allows easy conversion to the partial sandwich product 8b by coordination of the double bond to the two unsaturated rhodium centres. Although the accessibilty of the partial sandwich complex via G might tend to disfavour the diene manifold, it should be noted that E can be converted to G by interaction of the electron density of the Rh-Rh bond with the 1,3-butadiene 7t-system as shown in equation 3-3. This corresponds to a 4 + 2 cycloaddition and has precedent27'28 in related 2 + 2 additions of olefins to photochemically generated Os2(CO)s, a molecule with a formal Os=Os bond. H Rh •X-H Rh (3-3) E G 88 The equilibrium between the dienyl hydride 8a and the partial sandwich 8b favours the dienyl hydride (equation 3-1). The rate of this transformation is extremely slow since sharp signals for 8b are observed at all temperatures, irrespective of the fluxional behaviour of 8a. Indeed, at -20°C, this equilibrium is completely halted (Figure 3-6b). Integration of the respective resonances (^PpH) NMR spectroscopy) of the two species below -20°C shows no shift in equilibrium even after several hours. This could reflect a freezing out of exchange between A and F, thus blocking the exchange pathway between 8a and 8b. 3.2.7 The Reactions of [(dippe)Rh]2([i-H)2 with Isoprene and Piperylene. The dihydride l a reacts readily with isoprene and piperylene, but the reaction times are significantly longer as compared to the analogous reaction with 1,3-butadiene which is rapid even at -20°C. Although shorter reaction times can be achieved by the use of excess diene, this also decreases the proportion of binuclear products. Optimal production of binuclear products with reasonable reaction times are possible with 5-10 equivalents of the diene. The reaction with isoprene is summarized in Scheme 3-4. A syn/anti mixture of the mononuclear 2,3-dimethylallyImodium complexes is obtained as the major product, along with the binuclear isoprenyl-hydride derivatives lOa/lOa' as the minor component; the best ratio of mononuclear to binuclear products is approximately 4:1. Unfortunately, the binuclear complexes could not be separated from the mononuclear dimethylallyl products by fractional crystallization or chromatography and this complicated NMR analysis. Although we could not detect any of the binuclear isomeric isoprenyl sandwich derivative [(chppe)Rh]2(p-T]3-r|3-C5H8), at best we can say that there is less than 10% since this is probably the limit of detectability given the complex pattern of overlapping signals. The complexity of the spectra also precluded meaningful data from 1 3 C and 3 1 P spectra; however, from the variable temperature 1H{31P} NMR spectra the isoprenyl-hydride species, 10a/10a', were observed 89 Scheme 3-4 to be fluxional and apparentiy inter/convert between the two possible diastereomeric forms as shown in Scheme 3-4. Support for this comes from the hydride resonance at -8.34 ppm that appears as a broad triplet (with broad-band 3 1P decoupling) at ambient temperatures; as the temperature is lowered the pattern becomes more complex consistent with two species both having inequivalent 1 0 3Rh nuclei coupled to the bridging hydride ligand. Other resonances downfield were observed, in particular, a doublet at 1.5 ppm is attributed to the methyl group on the isoprenyl ligand; as the temperature is lowered, this doublet broadens and sharpens into two doublets of equal intensity again consistent with the slow exchange limit for the diastereomeric isoprenyl-hydride complexes 10a/10a\ Other resonances were observed to broaden and coalesce, but we have been unable to assign these peaks with certainty. Nonetheless, to explain the rearrangements in the high temperature limit, one must again invoke a process to exchange both rhodium centres and both ends of the isoprenyl ligand, all of which requires substantial bond-breaking and bond-making. One possible process is Scheme 3-5 Rh* = Rh(dippe) 91 shown in Scheme 3-5 and is a modification of the process proposed in Scheme 3-3 to explain the fluxionality of the butadienyl hydride 8a and its conversion to the butadiene sandwich complex 8b (vide supra). Dissociation of one end of the isoprenyl ligand followed by a "windshield wiper" process interconverts the two rhodium centres on any one particular diastereomeric isoprenyl-hydride complex 10a. To interconvert the diastereomers, a binuclear reductive elimination of the hydride and the isoprenyl ligands generates a | i.-Tl 4-isoprene complex similar to C, D, and E in Scheme 3-3. The diene C-H bond can then be cleaved either by oxidative addition or by insertion of one of the double bonds of the isoprene unit into the Rh-Rh single bond followed by a binuclear |3-elimination to regenerate lOa/lOa'. The reaction of [(dippe)Rh]2(|i-H)2, la, with piperylene was more straightforward and is summarized in Scheme 3-6. The binuclear products are now the major components with the mononuclear substituted-allyl complexes spectroscopically observed in only 10-15% 92 yield. The binuclear products are a 4:1 mixture of the syn-piperylene sandwich lib and piperylenyl-hydride diastereomers lla/lla' and can be isolated, as a mixture, in approximately 50% yield with only a slight contamination of the tenacious allyl side products. The behaviour of this mixture in solution is similar to that already discussed for the prototype butadiene reaction; the piperylene sandwich lib is in equilibrium with the piperylenyl-hydride diastereomers lla/lla' but in this case the equilibrium favours the sandwich derivative. This observation is consistent with our analysis of the steric requirements of this reaction (vide infra). Unfortunately, we were unable to obtain an accurate ratio of the diastereomers 11a and 11a' due to the complexity of the spectra. 3.2.8 The Reactions of [(dippp)Rh]2(Li-H)2 with Isoprene and Piperylene. The corresponding reactions of lb with isoprene and piperylene are significantly slower than those described above for la. The reaction with isoprene is particularly sluggish, requiring reaction time of 6-7 days in the presence of a large excess of isoprene. Heating the reaction mixture to 60°C decreases the reaction time to about 16 hours. In both cases, the reaction produces a mixture of syn and anti-l,2-dimethylallyhhodium(I) dippp diastereomers 12 (symanti, 4:96); no binuclear products are observed (Scheme 3-7). Scheme 3-7 lb 12 93 Scheme 3-8 The reaction of the dippp dimer lb with cis/trans-piperylene also generates mononuclear substituted allyl derivatives, but not exclusively; also formed is the binuclear piperylene sandwich complex 13b in approximately 20% spectroscopic yield as shown in Scheme 3-8. The complexity of the product mixture is demonstrated by Figure 3-8 which reproduces the 31P{1H) NMR spectrum of the residues of the completed reaction. In addition to signals due to the dippp piperylene partial sandwich, a plethora of signals due to the various allylrhodium(I) dippp isomers are present. The lH NMR spectrum of this mixture was also very complex, although it was possible to assign unambiguously the resonances for the binuclear product when approximately 90% pure samples were obtained via fractional recrystallization; analytically pure samples could not be obtained. The piperylenyl hydride isomers were not detected. The three 1,3-dimethylallylrhodium(I) dippp diastereomers (i.e., syn.syn; anti,anti; and syn,anti) were detected through comparison of the crude reaction mixture's lH NMR spectrum and a similar spectrum of an independendy synthesized mixture of the three isomers in question. Undoubtedly the remaining signals in the NMR spectrum are due to the syn and anti isomers of l-ethylallylrhodium(I) dippp, although unambiguous assignment of these resonance was precluded due to overlapping signals. Figure 3-8. 121.4 MHz 3 1P {]H} NMR spectrum of the crude product mixture from the reaction of l b with piperylene. 95 The other possible isomer of the piperylene partial sandwich product, namely the anti isomer 13b', was not detected. In light of the twisting about the internal C-C bond in 6 to avoid steric interaction between anti protons, it is not surprising that a methyl group in the anti position would be destabilizing. 3.2.9 Product Distributions: A Steric Evaluation. It is evident from the product distributions in the reactions discussed above that the outcome of these reactions is affected by the steric properties of both the ancillary diphosphine ligand, and the 1,3-diene. The longer reaction times required for the substituted dienes probably reflect a sterically more difficult trajectory experienced by the diene on its way to the reactive Rh2(|i-H)2 core of the dihydride dimers. This was particularly true for isoprene, which required a period on the order of days to react with la and lb. These longer reaction times in the presence of excess diene led to a greater proportion of mononuclear allyl products. A plausible explanation for the nature and distribution of the binuclear products in these reactions is also possible based on steric interactions found in these products. In the reaction of la with butadiene the dienyl hydride 8a is the major product. Inspection of the ORTEP diagram for 8a (Figure 3-5) suggests that there is potential for significant steric Rh* = Rh(dippp) 13b' 96 interactions between the bridging diene ligand and the isopropyl groups on one or both of the dippe ligands should any structural perturbations, such as substitution on the diene or changes in the diphosphine ligand, be implemented. Specifically, potential steric interference may exist between the isopropyl groups on P(3) and the a-bound diene carbon's substituent, Rh(2) and the protons on C(94), or the isopropyl groups on P(l) with carbons (93) and (94) of the diene ligand. Indeed, its instability in solution suggests that these interactions are already significant. Note that in the |i-,n3-r|3 partial sandwich mode of binding, such interactions are relieved. The severity of these interactions would presumably increase as the isopropyl groups are forced nearer to the core of the molecule as a result of a larger P-Rh-P angle, perhaps explaining the absence of a binuclear product incorporating the u.-r)4-0" bonding mode in the reaction of lb with 1,3-butadiene. Similarly, the presence of a methyl group on the terminal carbon of the diene would exacerbate the steric repulsions mentioned above. Thus, in the equilibrium mixture of dienyl hydride and partial diene sandwich, the thermodynamic preference for dienyl hydride in the equilibrium between 8a and 8b is reversed to a preference for sandwich in the equilibrium between lla/lla' and lib. In the isoprene reaction the preference is again for dienyl hydride (in fact, no u,-r*3-T|3 compound is observed), not surprising since a methyl group on the internal carbons of the diene would be expected to have a rninimal effect on the steric interactions in question. 3.2.10 The Magnesium Butadiene Reaction. The unique nature of the binuclear products obtained from the reaction of 1,3-butadiene with the hydride dimers la and lb provided an incentive to examine other possible synthetic routes to these materials. One rational sequence which proved successful was the reaction of "magnesium butadiene" ([Mg*C4H6»2THF]n) with the chloro-bridged dimers 14 as shown in equations 3-4 and 3-5. Typically, the [P2Rh](|i-Cl)2 derivatives were dissolved or suspended in THF, cooled to -10°C, and "magnesium butadiene" added all in one portion. 97 (3-4) (3-5) The products were obtained in excellent yield after workup. This procedure is the method of choice for the preparation of the dippp partial sandwich 6 since the yield is virtually quantitative and no mononuclear allyl side products are obtained. Similarly, the reaction of the dippe rhodium chloro-bridged dimer leads to the equilibrium mixture of 8a and 8b in reasonable yield. The reaction of magnesium butadiene with a number of other rhodium chloro-bridged dimers incorporating chelating diphosphine ligands has generated the binuclear butadiene partial sandwich structures, including those derivatives for which hydride dimers are not easily accessible. For example, the dihydride dimers [(dtbpp)Rh]2(p>H)2 and 98 P v CI P t 1 (-MgCy THF <0 (3-6) Rh R R R R R = Bu', 14C R = C 6 H 1 1 , 14d R Bu' 15 16 [(dcypp)Rh]2(M--H)2 (dcypp = l,3-bis(dicyclohexylphosphino)propane) were not available via normal preparative routes (Chapter 2). However, the chloro-bridged analogues were readily available, and reacted with magnesium butadiene to give the partial sandwich derivatives containing dtbpp, 15, and dcypp, 16, in excellent yield (equation 3-6). The use of other magnesium diene reagents has also allowed the synthesis of binuclear complexes not available via the hydride plus diene route. When magnesium isoprene was reacted with [(dippp)Rh]2(p>Cl)2, (equation 3-7), the isoprene partial sandwich 17 could be obtained analytically pure in 58% yield. This complex, while stable as a solid, thermally decomposed in solution over a period of days to the substituted allyls observed in the reaction THF,-10'C (-MgCy (I/ Rh (3-7) 17 99 Scheme 3-9 Rh* Rh* 14 .ri . Rh / / Rh ..AJ Rh (-MgCI2) MgCI Rh* = Rh(dippe), Rh(dippp) Rh(dtbpp), Rh(dcypp) products products between l b and isoprene, as well as other unidentified products. Perhaps this thermal instability partially explains the poor yield of binuclear products in the lb/isoprene reaction in that decomposition of the binuclear product probably occurs during the long reaction time. Based on literature precedent,29 the mechanism of this reaction probably involves stepwise metathesis of the bridging chlorides by the bifunctional "butadiene" Grignard as shown in Scheme 3-9. This generates the already proposed (Scheme 3-3) intermediate F which can rearrange to G; both F and G are pivotal to the proposed mechanism of formation of binuclear products as discussed in the following section. The generality of this magnesium diene reaction doesn't appear to extend beyond this family of chloro-bridged dimers. Reaction of magnesium butadiene with [(COD)Rh]2(p> Cl)2, [(Tl3-C3H5)Pd]2(|i-Cl)2,30 and [(CO)4Re]2(|i-Cl)231 led to intractable black tar in all 100 three cases. Nonetheless it remains an effective route to the rhodium diphosphine p>dienyl complexes. 3.3 A Proposed Mechanism of Formation of Binuclear Rhodium Butadiene Complexes. 3.3.1 The Proposed Mechanism for the Reaction of Olefins with Dihydrides [(dippe)Rh]2(p-H)2 and [(dippp)Rh]2(p:-H)2. The reactions of simple olefins with la and lb generate p>r|2-c alkenyl hydride complexes of the general formula [(P2)Rh]2[p--r|2-c-C(R)=CH(R,)](p-H).24 Although the syntheses and properties of these complexes are not part of this thesis, common mechanistic features between this reaction and those of la and lb with dienes warrant a brief review of the proposed mechanism by which the dihydrides react with olefins.15 The formation of the observed alkenyl-hydride products from the starting dihydride precursors 1 requires that two equivalents of olefin be consumed: one equivalent to dehydrogenate the dihydride dimers and the remaining equivalent to become the alkenyl-hydride unit. The use of the deuterium labelled complex, [(dippe)Rh]2(|i-D)2, ^ 2-la, and ethylene shows no incorporation of deuterium into the vinyl-hydride product [(dippe)Rh]2()J.-Ti2-a-CH=CH2)(p>H) which is clearly consistent with this dehydrogenation step. In Scheme 3-10 are represented possible mechanisms for the formation of the observed products. The coordinatively unsaturated dihydride dimers must first react with the olefin to generate an olefin complex H which undergoes insertion, producing a p>alkyl JI-hydrido intermediate which undergoes a binuclear reductive elimination to generate the binuclear rhodium(O) dimer I and one equivalent of alkane. The subsequent reaction with the second equivalent of alkene can follow path a or path b. In path a, the olefin coordinates to one metal centre of the dimer I, and the second metal centre activates the vinylic C-H bond by 1 0 1 . - H 1 : Rh(dippe) Rh(dippp) Rh* Rh" Scheme 3-10 l-feC =CHR ^ C = f C ^ , Rh*; R C H 3 C H 2 R •Rh* HaC =CHR Rh* Rh" H CH 2 R •CH 2 „ ' H Rh" H . ,H W I > R Rh Rh* path b H H H < b f c > H Rh* Rh* J path a • H Rh*"" N R V ^ R h * V T - c \ H P-elimination Rh* / Rh* y H ' ' \ J 102 an oxidative addition. This is the most commonly assumed pathway32 for the formation of alkenyl-hydride complexes. An intriguing alternative pathway is presented in b wherein the olefin inserts into the metal-metal bond of the binuclear rhodium(O) dimer I to produce a bridging alkanediyl linkage in J which subsequently undergoes a binuclear B-elirnination to generate the observed products. Both pathways a and b are consistent with all of the labelling studies. The recent stereochemical studies28 on the reaction of, formally, H2C=CH-2 and "(CO)40s=Os(CO)4" show that this process corresponds to a 2 + 2 concerted cycloaddition to generate a diosmacyclobutane derivative. In the mechanism shown in Scheme 3-10, path b also corresponds to a 2 + 2 cycloaddition between the double bond of the olefin and the single bond of I, the binuclear rhodium(O) complex. Our attempts to determine the stereochemical course of the reaction of olefins with purported I using cis-l,2-<i2-ethylene and cis-\,2-d2-propylene were thwarted when isomerization side-reactions scrambled the label. The use of tetrafluoroethylene as a possible trap for a species like J with a B-elimination stabilized 1,2-alkanediyl bridge was unsuccessful since the only products isolated from the reaction with [(dippe)Rh]2(p>H)2, la, or [(dippp)Rhh(H-H)2, lb, were the fluoro-bridged dimers [(P2)Rh]2(p.-F)2. While we have no evidence to distinguish between path a and path b, we have been able to follow the progress of the reaction in the early stages by running the reaction at low temperature. Thus the interaction of 1-butene with the dippe dimer la at -78°C generates an intensely magenta coloured solution which, by 31P{1H} NMR spectroscopy, shows the presence of another species besides signals for la and the diastereomeric butenyl hydride products.24 The four extra peaks are attributable to the intermediate H (Scheme 3-10) in which the olefin is ri2-bound to one of the rhodium centres. Raising the temperature in the presence of excess olefin resulted in the disapperance of the signals due to this intermediate and the completion of the reaction. A similar intermediate has been detected in the reaction of la with 1,3-butadiene (vide infra). The reaction of terminal olefins and ethylene with the dippp dimer lb is so slow that this type of intermediate has never been observed. It is 103 conceivable that the increase of the steric crowding about the inner Rh2(|i-H)2 core in lb hinders olefin coordination even at low temperatures. 3.3.2 The Observation of [(dippe)Rh(n2.c4H6)](p>H)2[Rh(dippe)]. When the reaction of [(dippe)Rh]2(H-H)2 (la) with excess (2-5 equiv) 1,3-butadiene is conducted at temperatures lower than -40°C a transient purple colour is observed. By maintaining the temperature at -80°C a deep purple solution is obtained that persists to about -50°C whereupon it fades to the orange colour of the products. If the excess 1,3-butadiene is pumped off under vacuum at < -60° C the purple solution reverts back to the deep green colour of the starting dihydride, la. At these low temperatures this purple intermediate can be isolated as a dark purple crystalline solid that decomposes to a green-orange material above -20°C, perhaps indicating a disproportionation to la and products. In our hands this thermal instability precluded any solid-state structural analysis but did allow solution spectroscopic data to be collected. The 1H, 31P{ lH} and 13C{ 1H} NMR spectra are indicative of a highly asymmetric structure and consistent with one end of the 1,3-butadiene 7C-bound to one rhodium centre of the hydride dimer as shown in equation 3-8. All four phosphorus nuclei are inequivalent as la K 104 judged by the complexity of the 31P{!H} NMR spectrum. The *H NMR spectra at low temperatures (<-60°C) show six separate proton resonances; although the resonances are broad, a combination of homonuclear and heteronuclear (31P) decoupling experiments established the connectivity and the fact that the diene was cis bound ( 3 JH3 ,H4 = 5 Hz).16 The 13C{!H} NMR spectra also confirm this structure as four carbon resonances are observed. This r|2-bound cis butadiene is rather labile since even at -60°C magnetization transfer between the diene protons on the T|2-bound ligand and free butadiene is observed upon homonuclear decoupling of the dienyl ligand resonances. In the analogous reaction of the dippp dimer, lb, with 1,3-butadiene, the equilibrium lies much futher to the left of the equation and no intermediate corresponding to the purple species described above was detected. 3.3.3 A Proposed Mechanism. In this section, a discussion of the details of the reaction of 1,3-butadiene with the binuclear hydrides la and lb is presented. The observation of binuclear products in both reactions suggests that binuclear intermediates are potentially involved. Although the rates of the reaction of 1,3-butadiene with each of the hydride dimers are quite different, as are the product distributions, it is reasonable to assume a common mechanism. The results of the reaction of the dideuterides ^2-la and d.2-Vo with 1,3-butadiene (Schemes 3-2 and 3-1, respectively) clearly establish that the first equivalent of butadiene merely serves to dehydrogenate the hydride dimers; this is summarized in the top portion of Scheme 3-11. Once the rj2-dienyl hydride intermediate K (observed as the purple intermediate in the dippe system) is formed, insertion of the diene into the bridging hydride to generate the binuclear 1-methylallyl hydride L, followed by a binuclear reductive eh'mination of butene to generate the rhodium(O) dimer I, is plausible. The |i-allyl intermediate has precedent in binuclear palladium complexes of the type [(R3P)Pd]2(|X-C3H5)(ix-Cp).33 In the 105 Scheme 3-11 Rh Rh* Rh^--H " R h * Rh* = Rh(dippe) Rh(dippp) Rh Rh (-MgClj) Rh Rh* 14 Rh Rh M Rh Rh R h * — — R h * Rh 7 \ 7 | Rh Rh || A / Rh Rh E H h / / Rh . A / Rh- Rh R h ' \ Rh Rh Rh Rh Rh* X H Rh* partial sandwich dienyl-hydride 106 dippe sequence, the insertion and reductive ehmination processes must be quite fast since only 1-butene is detected as a side product. By contrast, the analogous dippp reaction must be sufficiently slow to allow isomerization of L to M which upon reductive elimination generates isomerized butenes. In addition, if the reductive elimination from L or M is indeed slow in the dippp reaction, a further insertion of butadiene can be invoked which will lead to the mononuclear allyl products. Excess 1,3-butadiene would be expected to accelerate this fragmentation and indeed, when a large excess was employed, the proportion of mononuclear products increased. The rest of the reaction mechanism hinges on the rhodium(O) dimer I. Reaction of I with butadiene allows access into the diene manifold already discussed in the fluxional process of 8a. Once E is formed, the subsequent transformations to G and on to the partial sandwich derivatives 6 and 8b are analogous to that described in Scheme 3-3. In addition, the p.-l,4-butenediyl dimer G can be rearranged to the |i-l,2-butenediyl dimer F which after B-elimination generates the dienyl hydride species A. Again the importance of the purported intermediate G is apparent since both of the observed types of binuclear products, the dienyl hydride and the partial sandwich, emanate from it. That G can be accessed from the "magnesium butadiene" reaction with the chloro-bridged dimers (Scheme 3-9) is further support for this part of the mechanism. The rhodium(O) dimer I is an interesting molecule. While attempts to synthesize this type of coordinatively unsaturated d9 - d9 dimer have thus far met with failure (vide infra), there is some precedent for its existence in the isolation and X-ray crystal structure of the d 1 0 -d 1 0 platinum dimer [(dtbpp)Pt]2.34 Extended Hiickel calculations on this platinum dimer suggest that, in the planar geometry, the HOMO is antibonding (7cd*).35 Given that I has two fewer electrons (d9 - d9 vs d 1 0 - d10), the rhodium(O) dimer should be stabilized with respect to the platinum(O) dimer electronically. Another theoretical analysis for the hydride dimer [{(PriO)3P}Rh]2(|J.-H)236 gives an analysis for the RI12P4 core which is similar to the Pt2P4 core discussed above. 107 The only attempted route to I that met with any success involved the reduction of the chloro-bridged dimer [(dippp)Rh]2(|i-Cl)2,14b. A cyclic voltammagram of 14b revealed a reduction occurring near the solvent limit. A variety of reducing agents were employed in an attempt to reduce it, but only when Na/K was used did any reaction occur. Workup provided a red crystalline product which decomposed rapidly at room temperature. The equilibrium between the dienyl hydride 8a and the partial sandwich 8b obtained in the dippe reaction is more clearly delineated in this Scheme. Dissociation of the 7t-bond of the sandwich to generate the ii-l,4-butenediyl dimer G, followed by rearrangement to F and subsequent P-elimination produces the dienyl hydride. Using similar arguments to those presented above (section 3.2.9) the effect of the chelate ring size may be steric in origin since the larger six-membered chelate ring of the dippp ligand forces the isopropyl methyls more towards the Rh-Rh core of the binuclear unit than does the dippe ligand. In the dienyl hydride structure the central core is already crowded thus destabilizing this isomer for the dippp system. In the dippe system, the slight thermodynamic preference for the dienyl hydride 8a over the partial sandwich 8b is not easily justified, although the achievement of electronic saturation at one of the rhodium centres in this structure may be a factor. 3.4 Theoretical Analysis of the Bonding in the Partial Sandwich Structure. The unique nature of the partial butadiene sandwich structure found in 6, 8b, and others, prompted a collaboration with Dr. Thomas A. Albright to examine the sandwich's bonding from a theoretical point of view. The following discussion is a summary of this study.14c Extended Hiickel molecular orbital calculations were used to construct an orbital interaction diagram for the "full sandwich" |i-T|4-T)4 butadienyl dimer shown in the middle of Figure 3-9. On the right side of the Figure are symmetric (S) and antisymmetric (A) combinations of the pertinent orbitals of a C2 V Rh(PR3)2 fragment4 while on the left are the ure 3-9. An orbital interaction diagram for the "full sandwich" species [(PR3)2Rh]2(lt-Tl4-Tl4-C4H6). Figure 3-10. a) Coordinates for structural distortions from the (i-T|4-Ti4 "full sandwich" bonding mode, b) The increased overlap between 713 and the S combination of bi upon rotation of the P2Rh units about 9. c) The effect of rotation about 0 on the overlap between 712 and the A combination of bi. 110 four 7t orbitals of butadiene. In |i-T|4-r)4 bonding, 7Ci is stabilized by the A combination of the Rh(PR.3)2 molecular orbital 3ai and 7T-2 is stabilized by the A combination of bi. Finally, 713 and the A combination of b 2 interact slightly for overlap reasons. All the other molecular orbitals are non-bonding and are not shown in Figure 3-9. A striking feature of this MO diagram is the very small energy difference between the HOMO and the LUMO. This near degeneracy signals a possible second-order Jahn-Teller distortion. This is a structural distortion resulting in a second order energy change in the HOMO which widens the gap between the HOMO and LUMO giving an overall stabilization of the complex.4 In the ix-rj4-ri4 "full sandwich" structure, this distortion takes the form of a rotation of each Rh(PR3)2 fragment in opposite directions out of the xz plane by an angle 0, as shown in Figure 3-10a. This rotation "turns on" overlap between the S combination of bi and 713 fr°m n e a r 0, at 0 = 0°, as 6 is increased (see Figure 3-10b); note that overlap between the A combination of bi and 7t2 decreases only slightly (Figure 3-10c). The effect is to widen the energy gap between the HOMO and LUMO significantly, until an optimium geometry of 6 = 48° is reached in which the stabilization achieved is 35.2 kcal/mol over the p>T]4-ri4 "full sandwich" structure of Figure 3-9. Rotation about the C(2)-C(2') axis (<) in Figure 3-10a) further enhances overlap of the orbitals in Figure 3-10b and 3-10c to an optimized geometry of 8 = 48° and 0 = 27°. This second distortion provides a further 13.5 kcal/mol in stabilization over the 8 = 0° and <|> = 0° Li-rj4-T|4 structure. The experimental values of 6 = 37.8° and <|> = 45.0° found for 6 are in reasonable agreement with the optimized parameters discussed above; indeed, the potential energy surface in this region was found to be quite shallow and the experimental structure was only 4.3 kcal/mol less stable than the optimized structure. The calculations showed that, in a formal sense, 7t2 and 7C3 are occupied which should result in equalization of the C-C distances within the |i-butadienyl ligand; this was observed experimentally. In addition, calculated Rh-C overlap populations indicated that each Rh atom is coordinated in an T J 3 mode and that the end carbon (i. e., C(l)) was more strongly bonded than the inner carbons C(2) and C(2'). Ill This reproduces the observed asymmetry of the Rh-C bond lengths in the partial sandwich 6. 3.5 Experimental. 3.5.1 General Procedures. Butadiene was obtained from Matheson Gas Products, condensed into a small reactor bomb and vacuum transferred into reaction vessels from -10°C. Isoprene and piperylene were purchased from Aldrich Chemical Co. (the piperylene came as a 90% mixture of cisltrans isomers, with 10% cyclopentene) and dried over 3A molecular sieves. They were then vacuum transferred into storage vessels, thoroughly degassed via the freeze-pump-thaw routine, and stored in the dark. The deuterium labelled gases cw-l,2-d2-ethylene and as-1,2-^2-propylene were purchased from M S D Isotopes and used without purification. Tetrafluoroethylene and 1-butene were obtained from Matheson Gas Products and also used without purification. [Mg(C4H6>2THF]n37 and [Mg(C5H8)-2THF]n38 were prepared by a literature methods. The ligands R2P(CH2)3PR2, (R = B u t 3 9 and cyclohexyl40) were prepared using literature procedures. Cramer's complex, [(C2H4)2Rh]2(M--Cl)2, was prepared from a literature synthesis.41 The complex [(COD)Rh(T|3-syn,syn-l,3-Me2C3H3)] was prepared stereoselectively using a literature procedure,4 2 and used to synthesize a mixture of diastereomers of [(dippp)Rh(r|3-l,3-Me2C3H3)].23 The X-ray crystal structure of 6 was performed by Dr. Steven J. Rettig of this department; the structure of 8a was determined by Dr. F. W. B. Einstein of Simon Fraser University. Extended Hiickel calculations on the partial sandwich bonding mode were carried out by Dr. Thomas A. Albright of the University of Houston. 1 4 0 Dr. George Richter-Addo is acknowledged for cyclic voltammetric measurements on the chloro-bridged dimer [(dippp)Rh]2(li-Cl)2,14b. 112 3.5.2 Synthesis of the Chloro-bridged Dimers [(P2)Rh]2(n-Cl)2. R h R h > ' C I N -14a R = Prl, 14b R = Bu', 14c R = C 6 H 1 1 l 14d Synthesis of [(dippe)Rh]2(ii-Cl)2, 14a. To a stirred suspension of [(COE)2RhCl]2 (0.25 g, 0.35 mmol) in toluene (40 mL) was added dropwise a solution of l,2-bis(diisopropylphosphino)ethane, dippe, (0.18 g, 0.70 mmol) in 5 mL of toluene. The reaction was stirred for 15 minutes and the toluene removed under reduced pressure. The orange residues were suspended in cold hexanes and the solid collected on a fine porosity frit, yielding 0.23 g (86%) of an orange powder. *H NMR (C6D6, ppm): CH(CH3)2, 2.03 (dsp), 3JH= 7.6 Hz, 2JP= 1.8 Hz); CH(CH3)2, 1.50, 1.05 (dd, 3JP= 14.8 Hz); CH 2CH 2, 0.95 (m). 3lp{lH} NMR (C6D6, ppm): 102.2 (d, 1 J R h = 206.4 Hz). Synthesis of [(dippp)Rh]2(p>CI)2, 14b. To a vigorously stirred suspension of [(COE)2RhCl]2 (0.50 g, 0.70 mmol) in hexanes (50 mL) was added in one portion, a solution of l,3-bis(diisopropylphosphino)propane, dippp, (0.39 g, 1.39 mmol) in hexanes (10 mL). The disappearance of all solid [(COE)2RhCl]2 along with a deepening of the orange colouring of the solution was followed 113 by precipitation of the bright orange product. The product suspension was reduced in volume to 40 mL and cooled to -30°C to complete product precipitation. The product was collected on a fine porosity frit and washed with three 15 mL portions of cold hexanes. (0.55 g, 0.66 mmol, 95.0%). !H NMR (C6D6, ppm): CH(CH3)2, 2.12 (dsp, 3 j C H 3 = 7.6 Hz, 2 J P = 1.8 Hz); CH(CH3)2, 1.61, 1.10 (dd, 3 j p = 14.8 Hz); CH 2CH 2CH 2, 1.51 (dp, 2 J C H 2 = 4.8 Hz, 3j p = 18.0 Hz); CH2CH2CH2, 0.83 (br m). 31P{iH} NMR (C6D6, ppm): 47.68 (d, 1 J R h = 189.2 Hz). Anal. Calcd. for C 3 0H 6 8P 4Cl 2Rh 2: C, 43.43; H,8.28; CI, 8.55%. Found: C, 43.65; H, 8.36; CI, 8.75%. Synthesis of [(dtbpp)Rh]2(|i-CI)2, 14c. To a stirred suspension of [(C2H4)2Rhj2(H-Cl)2 (0.250 g, 0.64 mmol) in THF (50 mL) was added dropwise a solution of l,3-bis(ditertiarybutylphosphino)propane, dtbpp, (0.429 g, 1.28 mmol) in THF (5 mL) over a period of 5 minutes. A deep red solution resulted, which was allowed to stir for 15 minutes before removal of the THF in vacuo . The remaining red powder was washed with two 10 mL portions of cold hexanes and isolated by filtration (0.526 g, 87%). !H NMR (C6D6, ppm): C(CH3)3, 1.53 (d, 3 j p = 12.0 Hz); CH 2 CH 2 CH 2 , 1.31 (br m, 3 j H = 7.6 Hz); CH 2CH 2CH 2, 1.09 (br m). 31p{lH} NMR (C6D6, ppm): 52.99 (d, lJ R h = 197.7 Hz). Synthesis of [(dcypp)Rh]2(p>Cl)2, 14d. To a vigorously stirred suspension of [(COD)Rh]2(|i-Cl)2 (0.250 g, 0.51 mmol) in THF (30 mL) was added in one portion, a solution of 1,3-bis(dicyclohexylphosphino)propane, dcypp, (0.442 g, 1.02 mmol) in THF (10 mL). The disappearance of all solid [(COD)Rh]2((i-Cl)2 along with a deepening of the orange colouring of the solution was followed by precipitation of the bright orange product. The product suspension was reduced in volume to 40 mL and cooled to -30°C to complete product 114 precipitation. The product was collected on a fine porosity frit and washed with three 1 5 mL portions of cold hexanes. ( 0 . 5 2 2 g, 0 . 4 5 mmol, 8 9 . 7 % ) . *H N M R ( C 6 D 6 , ppm): ligand resonances, 0 . 9 - 2 . 1 (br m). 31P{!H} (C 6D 6 , ppm): 3 5 . 3 7 (d, 1 J R H = 1 8 7 . 6 Hz). Anal. Calcd. for C S ^ Q O P ^ R I ^ : C, 5 6 . 4 0 ; H, 8 . 7 6 % . Found: C, 5 6 . 0 9 ; H, 8 . 7 0 % . 3.5.3 Synthesis of Binuclear pv-Dienyl Complexes. 8a 8b Rh* = Rh(dippe) Synthesis of [(dippp)Rh]2(p>r|3-r|3-C4H6), 6: Method A. [(dippp)Rh]2((i-H)2 (0.50 g, 0.66 mmol) was dissolved in toluene (40 mL), placed in a thick-walled reactor bomb, and attached to a vacuum line. The solution was degassed and approximately 5 equivalents (3.29 mmol, 509 mm Hg in 110 mL) of 1,3-butadiene were transferred under vacuum into the reaction vessel. The reaction was stirred at room temperature until the black-green to yellow-orange colour change was complete (4 hours). Table 3-V. lH and "CQH} NMR Dataa for the Partial Sandwich Compounds 6,8b, lib, 13b, and 15-17. H 3 ^ b H 3 -< H 3 V b H 4 H 3 \ h Hi 6, 8b, > 15, 16 H 2 Hi lib, V - R H 5 13b Hi > H 4 17 Compound R Hi H 2 H 3 H4 H 5 CH3 Ca c b Q Cd 6 H 0.87 2.65 4.78 37.9 60.2 8b H 0.75 2.66 4.85 40.0 63.2 l i b CH3 0.95 2.59 4.71 4.93 1.61 2.00 37.6 56.5 67.4 50.5 13b CH3 0.65 2.32 4.57 4.71 b 2.03 36.8 57.2 66.5 48.6 15 H 0.86 2.98 4.90 39.5 60.1 16 H 0.73 2.56 4.79 —_b 59.8 17 CH3 0.80 2.66 4.73 b 0.87 b 42.5 60.8 80.7 34.0 aPPM, Q;D6. S^ignal obscured by ligand resonances and/or impurities. Table 3-VI. Coupling Constants3 in the Partial Sandwich Complexes 6,8b, lib, 13b, and 15-17. 7^ Hi Hj 6, 8b, 15, 16 17 Compound R Jl,2 Jl.3 J2.3 J3.3' J3.4 J4.5 J5.Me Ja.P Ja.Rh Jb.P Jc.P Jd.P Jd.Rh 6 H 3.9 10.8 6.6 4.5 35.0 10.0 18.0 8b H 4.0 10.3 b b 36.5 10.5 19.8 lib CH3 4.5 12.4 6.4 5.2 11.2 5.6 37.2 9.1 21.6 19.3 44.1 10.6 13b CH3 3.2 12.0 6.2 4.4 11.1 5.6 35.2 10.1 c 19.4 36.5 10.4 15 H 3.8 10.4 6.3 4.0 35.9 10.0 16.0 16 H 3.9 10.0 6.3 4.2 c c 18.2 17 CH3 3.9 11.5 ' 6.9 5.8 33.5 10.2 19.2 21.8 35.6 10.6 aHertz. bUnobtainable due to poor resolution. cSignal obscured by ligand resonances and/or impurities. 117 The solvent and excess butadiene were pumped away to produce an oily, yellow-orange crystalline mass, from which 0.33 g (0.42 mmol, 63%) of 6 was obtained upon recrystallization from toluene/hexanes, (1:2). mp 178-179° (dec). *H NMR (CgD6, ppm): diene protons, see Tables 3-V and 3-VI; CH(CH3)2, 2.24 (dsp, 3 J H = 7.2 Hz, 2 J P = 2.4 Hz), 2.08 (dsp,3JH = 7.8 Hz), 1.78 (2 overlapping dsp,3JH = 6.4-7.0 Hz); CH 2CH 2CH 2, 1.90 (m, 3 J H = 4.8 Hz); CH 2CH 2CH 2, -1.26 (m); CH(CH3)2, 1.52, 1.35, 1.12 (overlapping dd, 3 J H = 6.4-7.8 Hz, 3 J P = 12.0-14.4 Hz). 13C{!H} NMR (C6D6, ppm): see Tables 3-V and 3-VI; ligand resonances, 17.5-29.0. 31P{!H} NMR (C6D6, ppm): AA'BB'XX' pattern, PA, 54.4 ( 2 J A B = 40.5 Hz, l J R h = 228.0 Hz); PB, 36.0 ( l]R h = 180.4 Hz). Anal. Calcd. for C34H74P4Rh2: C, 50.24; H, 9.19%. Found: C, 49.99; H, 9.32%. The nature of the 1-methylallyl byproducts was verified by spectroscopic comparison with an authentic sample.23 An analytically pure sample could be obtained by sublimation of the residues obtained from the mother liquor of the above recrystallization at 10"4 mm Hg and 70-80°C to a probe cooled to -78°C. Due to decomposition, the yield was 0.123 g (0.284 mmol, 21.5% based on starting material, symanti = 1:4). lH NMR (C6D6, ppm), : Hcent-4.73 (m); CHsyn(CH3), 4.45 (m, 3JHcent = 7.4 Hz, 3 J C H 3 = 3.9 Hz) CH^nO!^), 3.37 (br d, 3 J H c e n t = 7.2 Hz); CHanti(HSyn), 2.12 (dd, 3JH c ent= 12.2 Hz, 2 J H s y n = 6.4 Hz); CCH3(H s y n), 1.22 (m); CH(CH3)2 and PCH2CH2CH2P, 1.60-2.10; CH(CH3)2, 0.9-2.3 (overlapping dd). 13C{!H} NMR (C6D6, ppm): C 2, 101.6 (s); C 3 , 60.4 (dd); C l f 43.1 (ddd); CH 3, 24.3 (d). 3lP{lH) NMR (C6D6, ppm): 45.5 (dd, 1 J R n = 183.8 Hz, 2 j p = 40.5 Hz); 40.0 (dd, lJRh= 190.7 Hz). Method B. To a THF (10 mL) suspension of [(dippp)Rh]2(-i-Cl)2 (0.053 g, 0.064 mmol) was added 1.1 equivalents of [Mg(C4H6>2THF]n (0.016 g, 0.072 mmol) suspended in 5.0 mL of THF. The reaction mixture was heated at 50° C until clear (30 minutes) at which time the THF was removed under reduced pressure. The orange residues were extracted with hexanes, and the extracts filtered through a Celite pad, resulting in a clear yellow solution. Upon slow evaporation of the hexanes 0.044 g (0.054 mmol, 84.7%) of spectroscopically 118 pure yellow-orange crystals were obtained. Synthesis of [(dippe)Rh]2(|±-H)(|i-rj4-o--C4H5), 8 a : Method A. [(dippe)Rh]2(p>H)2 (0.100 g, 0.136 mmol) was dissolved in toluene (10 mL), and loaded into a small reactor bomb. An excess of 1,3-butadiene was condensed into the reaction vessel and the reaction warmed gradually to room temperature. Immediately at -78°C, the characteristic deep green of the dihydride changed to an intense purple which persisted until approximately -50°C, at which time the solution began to turn orange. Before room temperature was reached the colour change was complete and the toluene and excess 1,3-butadiene were removed in vacuo . The yellow-orange residues were recrystallized from toluene/hexanes, yielding 0.097 g of 8 a / 8 b (91%). It appears that 8 a crystallizes preferentially but, in solution, a mixture of the two compounds is observed. Spectroscopic data for 8 a : *H NMR (C7D8, ppm, room temperature): diene protons, 6.48 (br m, 2H); 2.65 (br m, 2H); CH(CH3)2, 2.23 (br m, 2H); 2.02 (dsp, 3 J H = 7.0 Hz, 2H); 1.83 (dsp, 3 j H = 7.5 Hz, 2H); 1.67 (br m, 2H); CH(CH3)2, backbone, 0.69-1.50; Rh-H-Rh, -8.26 (br m, ^Rh = 26.7 Hz). *H NMR (C7D8, ppm, -70°C): diene protons, 7.49 (br d, J = 16.6 Hz, 1H); 5.56 (br s, 1H); 2.8 (br s, 2H); ligand resonances are all broad, 0.60-2.60; Rh-H-Rh, -8.15 (br m). 3lp{lH) NMR (C7D8, ppm, 40°C): 101.88 (d, 1 J R h = 163.7 Hz); 101.04 (d, !JRh= 171.4 Hz). 31p{lH} NMR (C7D8, ppm, -80°C): three overlapping signals at 101-105 (dd); 99.4 (dd, 2 J P = 25.5 Hz, 1 J R h = 139.5 Hz (see text)). ^COH} NMR (C7D8, ppm, -70°C): CCT(Ci), 156.9 (br m); C 2 , 123.5 (br s); C 3 , 88.8 (s); C 4 , 33.1 (s); ligand resonances, 10.8-28.7. Spectroscopic data for 8 b : *H and ^COH} NMR (ppm, C6D6), see Tables 3-V and 3-VI. Anal. Calcd. for C 3 2H 7 0P4Rh 2,8a: C, 48.99, H, 8.99%. Found: C, 49.03, H, 9.02%. Method B. To a THF (10 mL) suspension of [(dippe)Rh]2(p>Cl)2 (0.070 g, 0.087 mmol) was added 1.1 equivalents of [Mg(C4H6)-2THF]n (0.021 g, 0.096 mmol) suspended in 5.0 119 mL of THF. The reaction mixture was stirred at room temperature for 30 minutes, at which time the THF was removed under reduced pressure. The orange residues were extracted with hexanes, and the extracts filtered through a Celite pad, resulting in a clear yellow-orange solution. Recrystallization from hexanes at -20°C yielded 0.055 g (0.070 mmol, 81.0%) of spectroscopically pure yellow-orange crystals. Synthesis of [(dtbpp)Rh]2(|i-ri3-T|3-C4H6), 15. A solution of [(dtbpp)Rh]2(|i-Cl)2 (0.040 g, 0.043 mmol) in THF (15 mL) was cooled to -30°C. A similarly cooled suspension of magnesium butadiene (0.011 g, 0.049 mmol) in THF (2 mL) was added dropwise to the deep red solution. Upon warming, this red colouring gave way to an orange yellow solution, which was pumped down, extracted with three 15 mL portions of hexanes, and filtered through a Celite pad. The hexane solution was concentrated to about 5 mL and cooled to -20° C. The product was collected as bright yellow crystals, (0.034 g, 87.2%). *H and ^ C p H } NMR (ppm, C6D6): see Tables 3-V and 3-VI. 31P{1H} NMR (C6D6, ppm): P A , 67.43 (dm, lJ R h = 236.2 Hz); P B , 47.88 (dd, 2JP= 24.0 Hz, 1J R h= 183.1 Hz). Synthesis of [(dcypp)Rh]2(|i-T|3-n3.C4H6), 16. A THF suspension of [(dcypp)Rh]2(p>Cl)2 (0.185 g, 0.16 mmol) and magnesium butadiene (0.045 g, 0.20 mmol) were loaded into a reactor bomb and stirred at 70°C until the suspension went clear (0.5-1.0 hour). The resulting orange solution was evaporated to dryness and extracted with three 50 mL portions of hexanes. Concentration of this solution resulted in the precipitation of a yellow powder which was collected on a fine porosity frit yielding 0.162 g (90%). Approximately 5% of this powder was comprised of 1-methlyallyl derivatives which were inseparable from the major product. *H and 13C{ *H} NMR (ppm, C6D6): see Tables 3-V and 3-VI. 31p{lH} NMR (C6D6, ppm): PA, 47.6 (dm, l J R h = 120 228.7 Hz); PB, 47.88 (dd, 2JP= 41.2 Hz, lJR h= 180.3 Hz). Synthesis of [(dippp)Rh]2(p>r>3-r|3-C5H8), 17. A solution of [(dippp)Rh]2(p>Cl)2 (0.110 g, 0.13 mmol) in THF (10 mL) was loaded into a small reactor bomb. The bomb was attached to a vacuum line and the solution cooled to -78°C. A solution of 1.2 equivalents of [Mg(C5H8)-2THF]n (0.038 g, 0.16 mmol) in THF (1 mL) was added via syringe under a strong flow of argon. The reaction was stirred for 15 minutes at -78°C and allowed to warm to room temperature. The THF was removed in vacuo, and the residues extracted with hexanes. After filtration through a Celite pad, the hexanes were pumped off, and the bright orange solid recrystallized from hexanes at -20°C to yield 0.064 g (58%). lH and l^C{lH} NMR (ppm, C6D6): see Tables 3-V and 3-VI. 3lp{lH} NMR (ppm, C6D6): 57.4 (dddd, ! j R h = 230.5 Hz, 2 J P = 37.7 Hz, J R h - = 14.5 Hz); 56.2 (dddd, ! j R h = 247.8 Hz, 2 J P = 34.5 Hz, J R h - = 13.8 Hz, J P = 4.6 Hz); 36.3 (ddd, l j R h = 184.4 Hz, 2 J P = 37.8 Hz, J R h - = 9.7 Hz); 28.5 (ddd, ! j R h = 181.9 Hz, 2 J P = 34.4 Hz, J R h - = 9.8 Hz). Anal. Calcd. for C35H76P4RI12: C, 50.85; H, 9.27%. Found: C, 50.63; H, 9.30%. 3.5.4 Reactions of [(dippe)Rh]2(|j.-H)2 and [(dippp)Rh]2(|i-H)2 with Isoprene and Piperylene. Reaction of [(dippe)Rh]2(p.-H)2 with Isoprene. A solution of l a (0.10 g, 0.14 mmol) in toluene (10 mL) was loaded into a reactor bomb and degassed on a vacuum line. Approximately 10 equivalents of isoprene were then introduced into the system, and the reaction was stirred at room temperature until the green to orange colour change was complete (4-6 hours). The solvent was removed in vacuo to leave 0.110 g (=95% by mass balance) of an orange oil consisting of a 3:2 mixture of syn: anti 1,2-121 dimethylallylrhodium(I) dippe diastereomers (80% by NMR) and about 20% of the equilibrating binuclear isomers 10a and 10a'. *H NMR (C7D8, ppm), (dippe)Rh(rt3-syn-l,2-Me2C3H3): CH s y n(H a n ti), 3.40 (s); CHant i(CH3), 3.07 (m, J C H 3 = 6.4 Hz, J P = 4.0 Hz); Qlanti(Hsyn), 2.37 (s); CCH 3(H a n t i), 2.09 (dt, J P = 4.8 Hz); C-CH3, 1.65 (d, J R h = 1.5 Hz), m NMR (C7D8, ppm), (ctippe)Rh(ri3-anti-l,2-Me2C3H3): CHsyn(CH3), 4.38 (m, J C H 3 = 6.8 Hz); CHsynCHanti), 3.41 (s); CHant i(Hsyn), 2.32 (d, J P = 7.4 Hz); C-CH 3, 1.63 (d, J R h = 1.6 Hz); CCH 3(H s y n), 1.32 (m). *H NMR (C7D8, ppm), lOa/a': broad signals for dienyl protons appear at 5.90 and 2.63 ppm at ambient temperature; C-CH3,1.5 (d, room temp.); 1.6, 1.4 (d, J = 6-8 Hz, -70°C); Rh-H-Rh, -8.36 (br m). Reaction of [(dippe)Rh]2(|i-H)2 with Piperylene. A procedure analogous to that described above for the reaction of la and isoprene was employed. This reaction took 4-5 hours for completion and produced 0.045 g (42%) of a solid consisting of a mixture of the binuclear products, lla/lla' and lib, as well as traces of contaminating mononuclear allyl products. *H and 13C{ *H} NMR, lib see Tables 3-V and 3-VI. 3ip{iH} NMR (C 6D 6, ppm), lib: 102.9 (dm, lJ R h = 239.3 Hz); 96.2 (dm, J j R h = 240.1 Hz); 88.5 (2 overlapping dm, lJRh » 187-193 Hz, 2P). *H NMR (C 6D 6, ppm), lla/lla': dienyl proton resonances at 3.98 and 5.77 (br m, ambient temp.); Rh-H-Rh, -7.83 (m, ! j R h = 25.1 Hz). 31p{lH) NMR (C 6D 6, ppm): 94.8 (d, ^Rhfove) = 167.8 Hz). Reaction of [(dippp)Rh]2(p>H)2 with Isoprene. An analogous procedure to that employed for the reaction of la with isoprene was carried out in this reaction. In this case, the reaction was allowed to stir for 6-7 days at room temperature or 16 hours at 60-70°C prior to workup. The oily yellow-brown product consisted of a 96:4 mixture of antksyn l,2-dimethylallylrhodium(I) dippp. *H NMR (C6D6, 122 ppm), (dippp)Rh(Ti3-anti-l,2-Me2C3H3): CHsyn(CH3), 4.13 (m, J C H 3 = 6.0 Hz); CHsynCHanti), 3.18 (m, 3j p = 1.5-2.5 Hz); CHWH s y n ), 2 - 2 * (d, 3Jp = 7.2 Hz); CcentCH3, 1.74 (d, J R h = 1.9 Hz); CCH3(Hsyn), 1.30 (dd, J P = 2.6 Hz). 3ip{iH} NMR (C6D6, ppm): two overlapping dd centred around =44 (2Jp = 38.2 Hz, lJRh = 168.6 and 160.8 Hz). Reaction of [(dippp)Rh]2(p>H)2 with Piperylene. An analogous procedure to that described above for the reactions of la with isoprene and piperylene was employed. An oily, extremely hexane soluble orange solid was obtained (91% by mass balance) consisting of a complex mixture of l,3-dimethylallylrhodium(I)dippp, and l-ethylallylrhodium(I)dippp isomers, as well as the binuclear product 13b (20% by NMR). 1H and l3C{lH) NMR, 13b see Tables 3-V and 3-VI. ^Pf^ H} NMR (C6D6, ppm): 54.7 (dm, lJ R h = 235.0 Hz); 48.6 (dm, l J R h = 234.8 Hz); 35.3 (2 overlapping dm, i j R h - 152-158 Hz). 3.5.5 Mechanistic Studies. Observation of {[(dippe)Rh(n2.c4H6)](|i-H)2[Rh(dippe)]}, K. bound R = PrJ K unbound A solution of pure la in C7D8 was placed in a sealable NMR tube and degassed on a vacuum line. The sample was then cooled to -78°C in a dry ice/acetone bath and 3 equivalents of 1,3-butadiene were transferred under vacuum into the sample tube. An immediate green to 123 purple colour change was observed; this purple species was stable for at least 8 hours at -78°C and, when the tube was transferred to an NMR probe precooled to -80°C, *H, iSCpH}, and 31p{lH} measurements could be obtained. *H NMR (C 7D 8 , ppm, -80°C): Hcent(unbound> 6.12 (m, partially obscured by free butadiene resonances); Htrans(unb0und)> 5.32 (br d, J = 16.4 Hz); Htran^ bound). 4.45 (br m); Hcent(bound)> 4.08 (br m); Hc;5(b0Und)> 2.85 (br m); Hc/5(free), 2.45 (br m); ligand resonances, 2.1-0.8 (broad); Rh-H-Rh, -5.04 (m), -12.25 (m). ^ C p H } NMR (C 7 D 8 , ppm, -80°C): diene resonances, 140.4 (s); 119.0 (s); 116.8 (s); 101.9 (s). 31p{lH.} NMR (C 7 D 8 , ppm, -80°C): 111.4 (ddd, ^Rh = 174.3 Hz); 106.3 (dd, ^Rh = 131.3 Hz, 2 J P = 21.3 Hz); 105.1 (dd, I J R I , = 156.3, 2 J P = 30.4 Hz); 95.8 (br dd, l T R h = 122.0, 2 J P = 30.0 Hz). Reaction of d2-la and </2-lb with 1,3-Butadiene. A sample of either d2-la or d2-Va (0.10 g) was dissolved in enough C6D6 for two NMR samples (5 mm tube), and placed in a small reactor bomb. 1,3-Butadiene (4 equivalents) was condensed into the reaction vessel, and the reaction allowed to occur. Upon completion, about one half of the solution was transferred under vacuum to a sealable 5 mm NMR tube, and the volatiles then analyzed by *H and 2 H NMR. The remainder of the reaction solution, containing the organometallic products, was analyzed separately using the same techniques. No deuterium was detected in the binuclear organometallic products 6, 8a, or 8b. The deuterium was found in the butenes produced in the reactions, as well as in the mononuclear 1-methylallyl complexes 7 produced in the reaction of lb and 1,3-butadiene. Equilibrium Measurement of the Equilibrium 8a<=>8b. A 0.163 M solution of pure [(<ttppe)Rh]2«(C4H6) (0.064 g; 8.16 x 10"5 mol; 0.50 mL C 7 D 8 ) was sealed in a 5 mm NMR tube under about 0.85 atm of N 2 . The sample was placed in a thermostatted probe for 15 min before pulsing began and the spectra were collected at 124 constant temperature by using a 73° pulse, a 5 second relaxation delay, and broad-band proton decoupling. The isomer ratio was determined from each spectrum by intergration of the appropriate peaks; no correction for nOe effects were made. The possibility of intensity anomalies due to relaxation time differences was ruled out when a separate experiment yielded T i values of between 2-3 seconds for each phosphorus nucleus in the spectrum. 3.5.6 Miscellaneous Reactions. Synthesis of the (dippp)Rh(r)3-l,3-Me2C3H3) Diastereomers. To a stirred solution of (COD)Rh(r| 3-syn,syn-l,3-Me2C3H3) 4 2 in toluene was added dropwise a toluene solution of exacdy one equivalent of dippp. Stirring was continued for 15 minutes at which time the toluene was removed in vacuo . The residues were analyzed via * H N M R spectroscopy and were found to contain the syn,syn and syn,anti diastereomers of (d ippp)Rh( r i 3 - l , 3 -Me2C3H3) in a 1:4 r a t i o ; 2 3 b traces of the anti,anti isomer were also detected. * H N M R (C6D6, ppm), (dippp)Rh(n 3 -syn,syn-l ,3-Me2C3H3): CHcent, 4.58 (t, 3JH anti = H-2 Hz ) ; CHanti(CH3), 2.86 (m, 3 J C H 3 = 5.2 H z , J P = 6.2 Hz) ; CCH3(Hanti), 1.95 (dt, J P = 4.7 Hz) ; ligand resonances, 0.8-1.9. lH N M R (C6D6, ppm), (dippp)Rh(r- 3-syn,anti-l ,3-Me 2C3H3): C H c e m , 4.63 (dd, 3 J H a n t i = H - 4 H z , 3 J H s y n = 7.4 Hz) ; C H s y n ( C H 3 ) , 4.21 (m, 3 J C H 3 = 4.4 Hz) ; CHan ti(CH 3), 2.90 (m, 3 J C H 3 = 2.4 Hz) ; CCH3(syn), 2.01 (dt, Jp = 4.4 Hz) ; CCH3(anti), obscured by ligand resonances; ligand resonances, 0.8-1.9. 125 Reactions of [(dippe)Rh]2(|i-H)2, la, and [(dippp)Rh]2(p>H)2, lb with Tetrafluoroethylene. An identical procedure was employed for both of these reactions. The hydrides (la or lb) were dissolved in toluene and loaded into a small, thick-walled reactor bomb equipped with a 5 mm Kontes needle valve. The vessel was attached to a vacuum line and about 5 equivalents of tetrafluoroethylene were vacuum transferred into the bomb. Reaction was immediate for la, while lb took about 0.5 hours to react completely. The solids isolated had very similar *H NMR spectra to those of the chloro-bridged dimers [(dippe)Rh]2(H--Cl)2, 14a, and [(dippp)Rh]2(|X-Cl)2,14b, and the 31P{1H} spectra consisted of a complex multiplet indicative of the AA'A"A'"MM'XX' spin system expected for the fluorobridged dimers [(dippe)Rh]20/L-F)2 and [(dippp)Rh]2(ji-F)2. Data for [(dippe)Rh]2('i-F)2: *H NMR (C6D6, ppm): CH(CH3)2, 1.90 (br m); CH(CH3)2, 1.49, 1.18 (dd, 3 J H = 7.6, 6.4 Hz, 3 J P = 14.4, 13.6 Hz); P C H 2 C H 2 P , 0.92 (m, 2 J P = 9.2 Hz). 3ip{iH} NMR (C6D6, ppm): 109.1 (m). Data for [(dippp)Rh]2(|a-F)2: -H NMR (C6D6, ppm): CH(CH3)2, 1.90 (br m); CH(CH3)2, 1.52, 1.25 (dd, 3 J H = 7.2, 7.2 Hz, 3 J P = 13.2, 12.0 Hz); P C H 2 C H 2 C H 2 P , -1.4 (partially obscured); P C H 2 C H 2 C H 2 P , 0.75 (br m). ^PpH} NMR (C6D6, ppm): 59.3 (m). Anal. Calcd. for C30H68P4F2Rh2: C, 45.24; H, 8.60%. Found: C, 45.60; H, 9.00%. Reaction of [(dippp)Rh]2(p:-H)2 with Ethylene. Hi A solution of lb (0.075 g, 0.10 mmol) in toluene (5 mL) was loaded into a thick-126 walled reactor bomb equipped with a 5 mm Kontes needle valve. The vessel was attached to a vacuum line and degassed; ethylene was then admitted to a pressure of 1 atmosphere. A slow deep green to red colour change was observed over a period of 2-3 hours. The toluene and excess ethylene were removed under reduced pressure, and the residue recrystallized from minimum hexanes, from which 0.067 g (86%) of the deep red crystals of [(dippp)Rh]2(|i-ri2-a-CH=CH2)(u-H) were isolated. XH NMR (C6D6, ppm): RhCH=CHtHc, 9.79 (m, JH t= 17.5 Hz, J H c = 9.5 Hz, JH-I3c = 132.4 Hz, additional couplings to three phosphorus nuclei, Jp= 3.3, 3.3, 3.8 Hz); RhCH=CHtHc, 5.40 (m, JHt = 3.6 Hz, JH-13C= 148.3 Hz, J P « 4 Hz); RhCH=CHtHc, 4.53 (dd, JH-13C = 151.9); CH(CH3)2, 1.8-2.2 (3 br overlapping dsp); CH(CH3)2 and PCH2CH2CH2P, 1.0-1.6; Rh-H-Rh, -9.76 (m). ^CpH) NMR (C 6D 6, ppm): Rh13CH=13cH2, 201.7 (m); Rhl3CH=13CH2, 69.6 (d, JCc = 35.4 Hz); ligand resonances, 18-32. Anal. Calcd. for C32H72P4RI12: C, 48.85; H, 9.24%. Found: C, 48.89; H, 9.15%. 3.6 References. 1. Collman, J.P.; Hegedus, L.S. Princiciples and Applications of Organotransition Metal Chemistry; University Science Books: Mill Valley, 1980, p 13ff. 2. Davies, S.G. Organotransition Metal Chemistry: Applications to Organic Synthesis; Pergammon Press: New York, 1982, p 116-217. 3. a) Pearson, A. J. Metallo-Organic Chemistry; Wiley: New York, 1985. b) Yamamoto, A. Organotransition Metal Chemistry: Fundamental Concepts and Applications; Wiley: New York, 1986. 4. Albright, T. A.; Burdett, J. K.; Whangbo, M. H. Orbital Interactions in Chemistry; Wiley: New York, 1985. 5. Hegedus, L. S. In The Chemistry of the Metal-Carbon Bond ; Hartley, F. R.; Patai, S. Eds.; Wiley: New York, 1985; Vol 2, p 401. 127 6. a) See for example: Muetterties, E. L.; Burch, R. R.; Shustermann, A. J.; Teller, R. G.; Williams, J. M. /. Am. Chem. Soc. 1983,105, 3546. b) Krause, M. J.; Muetterties, E. L. Angew. Chem. Int. Eng. Ed. 1983,22, 135-48. 7. Marr, G.; Rockett, B. W. In The Chemistry of the Metal-Carbon Bond ; Hartley, F. R.; Patai, S. Eds.; Wiley: New York, 1982; Vol 1, Chapter 9, p 237 and references therein. 8. a) Erker, G.; Wicher, J.; Engl, K.; Rosenfeldt, F.; Dietrich, W.; Kruger, C J. Am. Chem. Soc. 1980,102, 6346. b) Dorf, U.; Erker, G. Organometallics 1983, 2, 462. c) Czisch, P.; Erker, G.; Korth, H.; Sustmann, R. Organometallics 1984, 3, 945. d) Yasuda, H.; Kajihara, Y.; Mashima, K.; Nagasuna, K.; Lee, K.; Nakamura, A. Organometallics, 1982,1, 338. 9. a) Hunter, A. D.; Legzdins, P.; Nurse, C. R.; Einstein, F. W. B.; Willis, A. C. J. Am. Chem. Soc. 1985,107, 1791. b) Hunter, A. D. PhD Thesis, Univ. of British Columbia, 1986. 10. a) Tachikawa, M.; Shapley, J. R.; Haltiwanger, R. C; Pierpoint, C. G. /. Am. Chem. Soc. 1976, 98, 4651. b) Franzreb, K-H.; Kreiter, C. G. Z. Naturforsch., B 1982, 37B, 1058. 11. Vollhardt, K. P. C; King, J. A. jr. Organometallics 1983, 2, 694. 12. Ziegler, M. Z. Anorg. Allg. Chem. 1967,355, 12. 13. Adams, V. C; Jarvis, J. A. J.; Kilbourne, B. T.; Owston, P. G. / . Chem Soc. Comm. D, 1971, 467. 14. a) Fryzuk, M. D.; Jones, T.; Einstein, F. W. B. /. Chem. Soc. Chem. Commun. 1984, 1556. b) Fryzuk, M. D.; Piers, W. E.; Rettig, S. J. /. Am. Chem. Soc. 1985, 107, 8057. c) Fryzuk, M. D.; Piers, W. E.; Albright, T. A.; Rettig, S. J.; Jones, T.; Einstein, F. W. B. /. Am. Chem. Soc, manuscript submitted. 15. Fryzuk, M. D.; Piers, W. E. Polyhedron 1988, 7, 1001. 16. a) A coupling of 3-5 Hz between these magnetically inequivalent nuclei is typical for complexes in which the diene ligand is c/^-bound.16b b) Crews, P. /. Am. Chem. Soc. 128 1971, 95, 636. 17. See for example: James, B. R.; Mahajan, D.; Williams, G. M.; Rettig, S. J. Organometallics 1983, 2,1452. 18. The upper limit on a Rh-Rh bond is about 3.2A. See Cowie, M.; Dwight, S. K.; Inorg. Chem. 1980, 19, 209, and references therein. 19. Vitulli, G.; Raffaelli, A.; Costantino, P. A.; Barberini, C; Marchetti, F.; Merlino, S.; Skell, P. S. /. C. S. Chem. Comm. 1983, 232. 20. Allegra, G.; Immirzi, I. Acta Cryst. 1969, B25, 120. 21. Data for [(C4Fi6)Fe(CO)3]: Bachmann, K.; von Philipsborn, W. Org. Magnet. Res. 1976, 8, 648. 22. Data for [(C4H6)Rh(Cp)]: Nelson, S. M.; Sloan, M; Drew, M. G. B. /. C. S. Dalton 1973, 2195. 23. a) Fryzuk, M. D. Inorg. Chem. 1982,21, 2134. b) The displacement of COD with a diphosphine is known to occur with isomerization of the T|3 allyl ligand.23a 24. Fryzuk, M.D.; Jones, T.; Einstein, F.W.B. Organometallics 1984, 3, 184. 25. Krieter, C. G.; Lipps, W. Angew. Chem. Int. Eng. Ed. 1981, 20, 201. 26. Beck, W.; Raab, K.; Nage, U.; Sacher, W. Angew. Chem. Int. Eng. Ed. 1985, 24, 505. 27. Takats, J.; Burke, M. R.; Grevels, R.-W.; Reuvers, J. G. A. /. Am. Chem. Soc. 1983, 105, 4092. 28. Norton, J. R.; Scott, C. P.; Hembre, R. T. /. Am. Chem. Soc. 1987,109, 3468. 29. Erker, G.; Kriiger, C; Miiller, G. Adv. Organomet. Chem. 1985,24, 1. 30. Tatsuno, Y.; Yoshida, T.; Otsuka, S. Inorg. Synth. 1979,19, 220. 31. Norton, J. R.; Dolcetti, G. Inorg. Synth. 1976,16, 35. 32. Nubel, P. O.; Brown, T. L. / . Am. Chem. Soc. 1984,106, 347A. 33. Werner, A.; Kuhn, A. Angew. Chem. Int. Ed. Engl. 1977,16, 412. 34. Yoshida, T.; Yamagata, T.; Tulip, T. H.; Ibers, J. A.; Otsuka, S. /. Am. Chem. Soc. 129 1978, 100, 2063. 35. Dedieu, A.; Hoffmann, R. / . Am. Chem. Soc. 1978, 100, 2074. 36. Teller, R. G.; Williams, J. M.; Koetzle, T. F.; Burch, R. R.; Gavin, R. M.; Muetterties, E. L. Inorg Chem. 1981, 20, 1806. 37. Wreford, S.S.; Whitney, J.F. Inorg. Chem. 1981, 20, 3918. 38. Yasuda, H.; Kajihara, Y.; Mashima, K.; Nagasuna, K.; Lee, K.; Nakamura, A. Organometallics 1982,1, 388. 39. Tani, K.; Tanigawa, E.; Tatsubo, Y.; Otsuka, S. /. Organomet. Chem. 1985,279, 87. 40. Issleib, K; Muller,D.-W. Chem. Ber. 1959, 92, 3175. 41. Cramer, R. Inorg. Synth. 1974,15, 14. 42. Muller, J.; Stuhler, H.-O. Chem. Ber. 1979,112, 1359. 130 CHAPTER 4 The Reactions of Binuclear Rhodium Hydrides with Imines: Factors Influencing the Insertion of C=N into Rh-H Bonds 4.1 Introduction. The amide ligand, "NRR', is extremely common in the chemical compounds of metal and metalloid elements,1 with a wide variety of homo and heteroleptic compounds known for most main group and early transition metals. Amides of the later transition metals, /. e., the iron, cobalt and nickel triads, were relatively rare prior to the late 1970's at which time the apparent lack of such compounds sparked interest in exploring the coordination chemistry of simple amide ligands with these metals. At present, many late transition metal amides are known2 and the chemistry associated with the amide ligand is becoming clearer. Three bonding modes for the amido ligand are possible (I, II, and III, Figure 4-1). In bonding to one metal centre, the lone pair on the nitrogen atom may or may not be involved in the metal-nitrogen bond. When no 7i-donation of this lone pair occurs as in type I, the nitrogen is pyramidal in geometry with sp3 hybridization; 7C-donation to the metal as in type II bonding renders the nitrogen planar and sp2 hybridized. Donation occurs from the orthogonal p orbital on nitrogen to vacant metal d orbitals; the inability of late transition metals to accept such electron density was thought to be responsible for the lack of stable late transition metal amides.11* While this may account partially for the paucity of homoleptic late transition metal amide complexes, it is clear that the amide linkage to such metals may be stabilized by the presence of softer donors to the metal centre 2 131 FT FT H IV V Figure 4-1. Possible bonding modes for the amido ligand (I-III) and imines (IV-V). A third type of bonding is the bridging mode, type HI. The availability of the lone pair in type I allows bonding to another metal centre; often, this can be suppressed through incorporation of bulky alkyl groups on the amide nitrogen.1 Common methods of synthesizing transition metal amides1'2 do not include the insertion of imine C=N double bonds into transition metal hydrides, although a metal amide would undoubtedly be the outcome of such a process. Such an insertion is likely discouraged by the imine ligand's preference towards bonding to metals through o-donation of the nitrogen lone pair of electrons3 (IV, Figure 4-1) rather than 7t-coordination of the C=N bond as in V. Although T J 2 bonding of imines has not been observed previously in mononuclear systems, additional metal bonding sites make possible new bonding modes for simple organic ligands, as was demonstrated by the bridging dienyl complexes discussed in Chapter 3. 132 Indeed, Tt-donation to adjacent metal centres by the C=N double bonds of a-coordinated diimines (1,4-azadienes) has been observed in many polynuclear complexes;4 however, similar bonding modes for simple imines are unprecedented. In this chapter, the reactions of the dihydrides la and Id with simple aldimines, the ketimine benzophenone imine, and the cyclic imine isoquinoline are described. The products are binuclear amido hydrides in which the (i-amido ligand has arisen from an apparent insertion of the imine substrate into a Rh-H bond of the dihydride. Low temperature studies have show that a more complex process involving both metal centres is involved. Support for this proposal is based on spectroscopic evidence, kinetic studies, and synthetic work aimed at the characterization of derivatives of proposed intermediates in the reaction. Finally, a brief discussion on the relevance of these studies to the hydrogenation of imines is presented. 4.2 The Synthesis and Properties of Binuclear Rhodium Amido Hydrides. 4.2.1 Synthetic Procedures. The dihydrides [(dippe)Rh]2(p>H)2, la, and [(dipope)Rh]2(|J.-H)2, Id, containing chelating diphosphine ligands with a five membered chelate ring, react rapidly with the readily Ph FUPr 1 , l a R = OPr', Id R = Pr', 18a R = OPr i, 18d 133 available aldimine N-benzylideneaniline to afford the binuclear products [(R2PCH2CH2PR2)Rh]2[Mi-(C6H5)NCH2(C6H5)](ix-H) (R = Pr*, 18a; R = OPr*, 18d) in good yield (equation 4-1).5 The products incorporate a bridging amido ligand apparently derived from the imine via a direct insertion of the C=N double bond into a Rh-H bond in the dihydrides. This reaction may be extended to include other simple commercially available imines, with varying steric bulk associated with the substituent on nitrogen (equations 4-2 and 4-3). In each case, the products (19-21a, 19-21d) are analogous to complexes 18a and 18d, and are produced in quantitative yield as determined by 31P{ ^ H) NMR spectroscopy. Isolated yields, though not maximized in some cases, vary from about 50-90% depending on the Rh Rh ^ - « . p ' H N P - — / \ / \ R = Pr!, la FUOPr', Id + N N = C t o l u e n e R'= CH 3 R'= CH2Ph Ph R R CH2 R D \ Rh. ^Rh / Ks ^  \ J •p p-R'= CH3 CH2Ph R = Pr' 19a 20a R = OPr' 19d 20d (4-2) ^ \ s\ / ^ Rh Rh ^ p ' H N p - ^ H, / + N = C . Ph toluene Ph \ Rh Rh / KS N«' \ J (4-3) •P P-R = Pr\ la R = OPr', Id R = Pr', 21a R = OPr', 21d 134 solubilities of the products. The amido hydrides incorporating the dipope ligand were found to be highly soluble even in non-polar hydrocarbon solvents such as hexanes, and thus isolated yields were generally lower for these complexes. Qualitatively, reactions involving Id and imines were slower than those between la and imines; in addition, reaction rates appeared to be influenced by the steric bulk of the substituent on nitrogen in the imine, with bulkier groups impeding the reactions significantly. For example, at roughly the same concentrations, the reaction of la with N-benzylidenemethylamine was more rapid than that with N-benzylidenebenzylamine. Steric bulk on the imine carbon atom had less of an effect; the ketimine benzophenone imine reacted rapidly with both la and Id to form the amido hydride products 21a and 21d (equation 4-3). In all reactions only one equivalent of imine adds to the dihydrides; thus, in most cases, a 2-4 fold excess of imine was employed to facilitate the reactions. The dihydride dimer [(dippp)Rh]2(p>H)2, lb, which incorporates the more sterically demanding dippp ligand, is virtually inert towards reactivity with these imines, with the exception of benzophenone imine. Even at elevated temperatures (=90°C) and use of large excesses of imine, lb failed to react with N-benzylideneaniline or N-benzylidenemethylamine. When such conditions were employed with benzophenone imine, a slow reaction ensued and the dippp amido hydride product 21b was isolated in 77% yield after 16 hours of reaction time (equation 4-4). H (4-4) R = Pr\ lb R = PrJ, 21b 135 The amido hydride products are bright orange (dippe series) or yellow (dipope series) crystalline solids. The complexes are all air-sensitive both in solution and in the solid state. Complex 18a is stable towards water and methanol, however. In addition to moisture stability, 18a exhibits high thermal stability in solution. This will be discussed more fully below in a section dealing with H/D exchange experiments. 4.2.2 Spectroscopic Properties. The amido hydrides were fully characterized via multinuclear NMR spectroscopy and microanalysis. The spectroscopic data for these complexes is comparable to that observed for the related, structurally characterized amido hydride [{(PriO)3P}2Rh]2[p>NMe(p-Cl-C6H4)](|i-H), 22, formed via hydrogenation of a bridging isocyanide ligand (see equation 1-12).6 The similarities imply that all of these amido hydride complexes possess the same square planar geometry at each rhodium centre that compound 22 has, as determined by X-ray diffraction. Although the NMR spectra do not distinguish the proposed square planar geometry from a structure with tetrahedral rhodium centres, certain features of the 3 1P pH} spectra support the square planar structure (vide infra). The NMR spectra of amido hydrides 18-21 contain several diagnostic features; the spectrum of 18a is shown in Figure 4-2 as a representative example. The ligand region of the spectrum tells of the degree of symmetry to be found in the molecule. When the substituents on nitrogen are different, as in 18a, four methine signals and eight four line patterns for the methyl groups are observed for the isopropyl moieties of the diphosphine ligand. In complexes 20a and 20d, in which both substituents are benzyl groups, the complexity of the ligand region is reduced two fold, exhibiting only two signals for the methines and four "doublet of doublets" patterns for the methyls. The 13C{ lH} NMR spectra provide concurring symmetry information in the ligand region of the spectrum. Figure 4-2. 400 MHz *H NMR spectrum of the amido hydride complex 18a. 137 The signal at 5.32 ppm in the *H spectrum of 18a is due to the benzylic protons of the amido ligand. In all of the amido hydride complexes, this signal is a broad triplet which collapses to a singlet upon broadband decoupling of phosphorus. The magnitude of this four-bond coupling is 2-4 Hz, except in compounds 21a and 21d, where more substantial couplings of 5.2 and 8.8 Hz, respectively, are observed. A similar coupling was observed in the signals due to the N-methyl groups in complexes 19a and 19d. In the 13C{!H} NMR spectra, the benzylic carbons (and the N-alkyl carbons of 19a and 19d) resonate in the region 65-75 ppm, appearing as slightly broad singlets (wi/2 =8 Hz) with no apparent fine structure originating from coupling to either phosphorus or rhodium. In all the amido hydride complexes but 18a and 18d the aromatic region of the spectrum is uncomplicated. In 18a, however, the N-phenyl ring experiences restricted rotation about the QpS0-N bond and thus there are five separate environments for the protons on this ring. This phenomenon was also observed in the complex [{(Pr-O^P^Rlitetli-NMe(/?-Cl-C6H4)](|i-H), 22.6 In this complex free rotation was observed at room temperature, but upon cooling of the sample the rotation was frozen out and separate signals for each of the five proton environments were observed. In the spectrum of 18a, the five separate resonances are observed at room temperature, indicating that the barrier to rotation is higher than that calculated for complex 22. Although heating a sample of 18a did result in the coalescence of the signals due to the ortho (8.58 and =7.2 ppm) and the meta (=7.2 and 7.00 ppm) protons of the ring, the high temperature limit was not reached. This corresponds to a AG* of approximately 16.7(5) kcal/mol (100°C) compared to that of 12.0 kcal/mol (-15°C) reported for 22.6 The N-phenyl ring in 18d also experienced hindered rotation but with a lower barrier of AG* = 9.8(4) kcal/mol (-60°C). The differences in the barriers to rotation of the ring reflect differences in the steric bulk of the ancillary ligands in each complex. Since the N-phenyl ring lies in the idealized mirror plane of these molecules, more steric crowding of the molecular core results in a higher barrier to free rotation about the Qpso-N bond. 138 Table 4-1. 3 1 P NMR data for the amido hydrides 18a, 18d, and 22. r B ^ B 1 Parameter3 18a 18d 22b 5A 73.9 175.8 132.7 5B 105.5 206.1 149.7 XJAX 191.3 265.1 297.8 ^ B X 166.9 223.3 247.8 2 JAB 30.3 39.9 82.5 2Jxxr 5.6 3.2 3.0 3JAX- 5.8 6.7 7.8 3JBX' -1.2 -2.4 -3.6 4JAA- 33.5 44.8 85.6 4JAB- 2.9 4.4 6.2 4JBB- -0.9 -3.8 -6.0 a Units: 8 = ppm; J = Hertz. b Data taken from reference 6. Consistent with the above discussion is the observation of ten singlets in the region of the 13C{1H} NMR spectrum of 18a for aromatic carbon resonances, rather that the eight expected should free rotation about the CipSo-N bond be facile at room temperature. The 31P{ !H} NMR spectra of these complexes are diagnostic. The spectrum of 18a, along with a calculated reproduction, is shown in Figure 4-3. Since the two groups of phosphorus nuclei, as well as the rhodium atoms, are magnetically inequivalent, the spin system is AA'BB'XX'. The resulting spectra are complex; nonetheless, simulations enabled the determination of all coupling information for 18a and 18d (Table 4-1) and since the spectral patterns varied insignificantly in each of the a and d series, simulations of the spectra 139 Figure 4-3. a) 121.4 MHz 3lp {lH} NMR spectrum of the amido hydride 18a. b) Calculated spectrum based on an AA'BB'XX' spin system (see Table 4-1 for parameters). 140 of 19-21 were not performed. Some interesting observations can be made upon examination of the data in Table 4-1. As has been observed in other systems,6-7 magnetic communication between opposite sides of the dimer is facile. In particular, the large four-bond J A A ' coupling is noteworthy. Larger even than the two-bond J A B coupling, it is likely due to the fact that the large bridging amido ligand forces P A and P A ' into closer conjunction with the Rh-Rh vector such that they are nearly opposed across this vector. For example, in [{(PriO)3P}2Rh]2[p.-NMe(p-Cl-C6H4)](|i-H), the PA-Rh-Rh' angle was 145.1°.6 This structural feature greatly enhances coupling between these two nuclei. The same rationale explains why the three-bond J A X ' of P A to the distal rhodium is larger than the analogous coupling of P B to Rhx1. Other four-bond couplings (i.e., J A B ' and J B B O are small. 4.2.3 H/D Exchange Processes in the Amido Hydride 18a. The complex 18a exhibits quite remarkable thermal stability in solution. Toluene solutions of 18a may be heated at 120°C for at least four days without spectroscopically detectable decomposition, although the solution darkens gradually throughout heating. Heating at 150°C for 10 hours results in less than 10% decomposition. One might expect that the bulkiness of the p:-amido ligand, which normally destabilizes the bridging mode of ligation for such ligands,lb and the presence of B-hydrogens in the amido ligand would facilitate decomposition pathways for this molecule. Its observed stability is more evidence against original assumptions that late transition metal amide complexes are inherendy unstable due to the mismatching of a hard ligand base and a soft metal acid. Though quite stable towards decomposition under thermal duress, hydrogen-deuterium exchange with the solvent was observed. Exchange was first noted when a sample of 18a in c^ -toluene was heated at 150°C for 10 hours. When the darkened solution's *H NMR spectrum was collected, all signals due to the |i-amido ligand were absent. Complete decomposition was ruled out since the signals due to the diphosphine ligand were identical to 141 those of perprotio 18a. The 2 H NMR spectrum of this sample showed indeed that deuterium had been incorporated into all positions of the |±-amido ligand, while no deuterium was incorporated into the ancillary diphosphine ligand. This last observation is in contrast to exchange processes documented for the dihydride le and the amido hydride [{(PriO)3P}2Rh]2[M--NMe(CH2C6H5)](p:-H)6 in which incorporation of deuterium into the methyl positions of the monodentate phosphite ligand was observed upon H/D exchange with D2 gas. Evidently the constraints imposed by the chelating nature of dippe hinders the methyl groups from interacting with the metal centre in a manner necessary for deuterium incorporation. The H/D interchange process between 18a and ^ -toluene was followed by both *H and 2 H NMR spectroscopy. The exchange reaction was allowed to proceed at about 120°C, a temperature at which decomposition was insignificant. £>g-toluene was used as a deuterium pool because when D2 was used, hydrogenolysis of the ii-amide ligand predominated over exchange and only signals due to (C6H5)ND(CH2C6H5) and the tetradeuteride d4-5a were observed in the 2 H NMR spectrum (this reaction and its implications for hydrogenation of imines are the subject of another section). Under the conditions described above, deuterium was incorporated into the p-hydride and the benzylic position of the p>amide ligand first. Signals at -9.4 and 5.3 ppm in the 2 H NMR spectrum appeared less than one hour after heating began. Eventually, after about 2 days, these positions were completely deuterated. In addition, the ortho positions of the benzyl group's phenyl ring were >75% substituted with deuterium. This was evidenced by the gradual appearance in the *H NMR spectrum of a second doublet at 8.79 ppm attributable to an isotopomer of 18a deuterated in one of the two ortho positions; a corresponding signal in the 2 H NMR spectrum was also observed. The reported H/D exchange between [{(PriO)3P}2Rh]2[|i-NMe(CH2C6H5)](|i-H) and D 2 gas also included selective incorporation of deuterium into the benzylic and ortho positions of the |i-amido moiety.6 142 The H/D exchanges at temperatures around 120"C are more selective than the multiple exchanges observed at 150°C (vide supra). This is suggestive of an intramolecular H/D exchange mechanism similar to that proposed by Parshall8 and a possible mechanism is given in Scheme 4-1. Exchange into the bridging hydride position occurs when the solvent is activated by one rhodium centre to give a transient aryl deuteride species. Deuteride/hydride exchange followed by reductive elimination of ^ -toluene completes the exchange. For clarity, Scheme 4-1 shows only aryl C-D bonds being activated in this step, but the appearance of toluene signals in the proton NMR spectra suggest that the C-D bonds of the dj-methyl group are equally prone to activation. Incorporation of D into the amido ligand is brought about when the benzylic or ortho C-H bonds oxidatively add to one of the rhodium centres; subsequent H/D scrambling followed by reductive elimination completes the exchange. Repetition of the solvent exchange step continues the cycle. Facile substitution into positions four bonds away from the metal centre is a common phenomenon in H/D exchange processes8-9 and is attributed to the ease with which relatively unstrained five membered rings form compared to other metallacycles in the oxidative addition step. The facile incorporation of D into the benzylic position despite formation of a strained three membered metallacyclic intermediate is a reflection of the relative weakness of the benzylic C-H bonds.10 Interchange between the hydride and benzylic or ortho positions appears to be facile only with a large deuterium pool available. For example, when d2-lSa was heated in toluene, no scrambling of the deuterium from the hydride position into the ortho site was observed. Although superficially inconsistent with the proposed mechanism in Scheme 4-1, this observation is a reflection of the fact that the rate of exchange out of the bridging deuteride position into the solvent is much faster than exchange into the ortho positions under these conditions. At higher temperatures the exchange is less discriminate; this is usually indicative of heterogeneously catalyzed exchanges. Likely, elevated temperatures result in production of 143 Scheme 4-1 Ph 144 rhodium metal upon partial decomposition of 18a, which catalyzes intermolecular exchanges between the dg-toluene and the remaining aromatic sites on the u.-amido ligand. It should also be noted that H/D exchange between complex and solvent is not limited to this amido hydride complex. The parent dihydride la (and also lb) will exchange hydrogen for deuterium with C6D6 slowly at room temperature and more rapidly at higher temperatures. Again, this exchange is limited to the hydride position exclusively; no deuteration of the diphosphine ligands is observed. 4.2.4 The Reactions of the Dihydrides la and Id with Isoquinoline. In addition to the reactions described above, the dihydrides la and Id also react with the cyclic imine isoquinoline 23 (equation 4-5). In both cases, the reaction is >95% 24' R = Pr1, a R = OPr\ d Figure 4-4. 400 MHz lH NMR spectrum of the amido hydride 2 4 a . Peaks downfield of 7 ppm are due to free isoquinoline. 146 regioselective, producing both 24a and 24a'. Assignment of the structure of the predominant regioisomer was made based on *H NMR evidence; the *H NMR spectrum of 24a is shown in Figure 4-4. In both of the regioisomeric structures there is one olefinic type proton, expected to resonate upfield of the rest of the |X-amido ligand protons (H shown in the structures in equation 4-5). The doublet at 4.98 is assigned to this proton. The signal's doublet multiplicity is only consistent with 24a, since this proton would be split into a triplet (or a doublet of doublets) if it was adjacent to the methylene carbon, as in 24a'. The ^ CpH) and 31P{1H} NMR spectra were consistent with both structures and had similar features to the spectra of the amido hydrides discussed above. Like complexes 18-21a, d amido hydrides 24a and 24d are quite air-sensitive, but in contrast exhibit thermal instability in solution. In fact, as shown in equation 4-6, the system is a measurable equilibrium at room temperature, the position of which is dependent on the nature of the solvent at constant temperature. Figure 4-5 shows a van't Hoff plot for the equilibrium defined by equation 4-7, at (4-6) R = Pri, l a R = OPr i, Id 23 R = Pri, 24a R = OPr\ 24d K [24a] eq -[1a] [23] (4-7) 147 2.9 3.0 3.1 3.2 3.3 3.4 3.5 1/T(1/Kx1000) Figure 4-5. van't Hoff plot for the equilibrium defined in equation 4-7: trace a, C6D12 : C7D8 = 5:1; trace b, C6D12 : C7D8 = 1:1; trace c, QD12: C7D8 = 1:5. varying solvent compositions. The top trace corresponds to a solvent mixture of d.12-cyclohexane and d^ -toluene in a 5:1 ratio by volume. The next line down is for values of K in a mixture of those solvents in a 1:1 ratio and the last trace for a 1:5 ratio of solvents. Graphically determined thermodynamic parameters, as well as values for K and AG° at various temperatures, are given in Table 4-11. The apparent lack of correlation between the solvent ratio and the values obtained for AH° and AS" indicates that the changes in these parameters upon changes in solvent composition are too small to be accurately determined given the experimental uncertainty associated with the method of measurement used. Nonetheless, the experiment does clearly show the affect on the equilibrium constant with Table 4-IL Thermodynamic Parameters, Values of AG" and K for the Equilibria Defined by Equations and 4-7 and 4-8. Equilibrium la + 23 <==> 24a Id + 23 <==> 24d Solvent Ratio 5:la 1:1* 1:5a 1:1b AH°c -11.9 -10.3 -10.7 -9.2 AS°d -30.4 -26.0 -28.2 -22.9 T«, K, AG o f T K AG" T K AG" T K AG° T K AG° 292 179.5 -3.0 294 83.1 -2.6 294 64.1 -2.4 283 108.8 -2.6 303 83.9 -2.7 303 52.5 -2.4 303 35.5 -2.2 293 64.9 -2.4 308 64.1 -2.6 313 30.3 -2.1 313 18.9 -1.8 303 49.4 -2.3 323 25.5 -2.1 323 18.5 -1.9 323 11.8 -1.6 310 34.1 -2.1 343 8.3 -1.4 333 10.5 -1.6 333 7.5 -1.6 320 19.9 -1.9 330 11.7 -1.6 340 8.0 -1.4 350 4.9 -1.1 a Ratio given is for d/2-cyclohexane:c?s-toluene. b Ratio given is for n-hexane:d#-toluene. c kcal/mol, uncertainty of ± 0.5. d eu, uncertainty of ± 2.0. e°K. f kcal/mol, uncertainty of ± 0.2. 149 changing solvent composition: as the proportion of aromatic solvent (dg-toluene) in the mixture increases, the equilibrium constant as defined above decreases; i. e., aromatic solvents favour the left side of the equilibrium. Intuitively one might expect that a shift to the left upon addition of toluene would be disfavoured entropically if more effective solvation of the free isoquinoline by the toluene is assumed. However, the aromatization of isoquinoline and bond energy changes in the back reaction apparently provide enough of a driving force to account for the observed shift to the left in aromatic solvents (the same effect is observed in benzene). In the analogous reaction with the dipope dimer Id and isoquinoline similar behaviour 1 T « 1 • 1 • 1 • 2.8 3.0 3.2 3 . 4 3.6 1/T(1/Kx1000) Figure 4-6. van't Hoff plot for the equilibrium defined in equation 4-8. 150 K [24d] eq - [1d] [23] (4-8) is observed; Figure 4-6 shows the van't Hoff plot for the equilibrium defined by equation 4-8 in a 1:1 ds-toluene/hexane solvent system. The thermodynamic parameters and values of K and AG° for this equilibrium are also given in Table 4-II. Like in the dippe system, the left side of the equilibrium is favoured by higher proportions of toluene in the solvent mixture, as evidenced by a nearly complete shift to the left of the equilibrium upon dissolution of pure 24 d in benzene. 4.3 Mechanistic Studies. 4.3.1 The Reaction of [(dippe)Rh]2(|i-H)2 with N-benzylideneaniline at Low Temperature. Mechanistically, the reaction of la with N-benzylideneaniline seems straightforward. By simple product analysis, the amido hydrides appear to have resulted from a direct insertion of an imine carbon-nitrogen double bond into a bridging hydride ligand. Indeed, when the dideuterides d2-la or d2-ld were allowed to react with any of the imines, the label appeared Ph Rh Rh + toluene (4-9) Ph d2-la d2-18a 151 as one deuterium in the benzylic position of the pi-amido ligand, while one deuterium was retained in the bridging position (equation 4-9). However, when the reactions were carried out at low temperature and monitored via NMR spectroscopy, an intermediate was observed, the nature of which provides evidence for a more complex process that involves both metal centres. Figure 4-7 shows 31P{1H} NMR (162.21 MHz) spectra of the proceeding reaction between la and N-benzylideneaniline as a function of temperature and time. At -70°C, near the beginning of the reaction (Figure 4-7a), the spectrum consists of the doublet characteristic of the dihydride la and four faint resonances not attributable to the product 18a. After approximately one hour, these signals have grown in almost completely, still with no trace of the peaks due to 18a. (Figure 4-7b). At temperatures at or below -50°C, this intermediate is stable in solution, but is unstable at higher temperatures even in the solid state. Thus, when the sample is warmed to -20°C, the intermediate begins to decompose to 18a, and after about 45 minutes, the amido hydride has formed almost completely (Figure 4-7c). A spectroscopically similar reaction course was observed in the analogous reaction involving Id and N-benzylideneanihne. The !H{31P} NMR spectrum of this intermediate (Figure 4-8) also indicate a highly unsymmetrical structure in that the complex has 16 separate environments for the methyl groups of the dippe ligands as evidenced by the 16 doublets in the upfield region of the spectrum. The aromatic region is also complex, but the broad singlet at 4.90 ppm, integrating to one proton, can be assigned to the imine proton. Notably, it is shifted significandy upfield from the resonance of the free imine at 8.16 ppm. The most important structural information the spectrum has to offer is found in the hydride region in which signals for one terminal hydride ligand (-11.45 ppm, d, ijRh = 18.4 Hz) and one bridging hydride (-13.79, br t, ijRh = 19.6-20.4 Hz) are observed. Although normally a bridging hydride appears downfield of terminal hydride resonances, the multiplicity of each signal upon phosphorus decoupling is _rv_/\_ J li "\ y\ /" Rh Rh la Ph CH?Ph 1 Rh Rti / 18a - 7 0 °C »-70°C - - 2 0 " , C -> i—r—»—r—r—r—r—r—r—r—r—i—r—r—r—i—i—r—<—i r i — i — i — i — i — i — r — i — i — r — » — r — i — r — r — i — i — i — i — r — r — 1 110 100 9 0 8 0 ppm Figure 4-7. 162.2 MHz 31P{1H} NMR spectra of the proceeding reaction between la and N-benzylideneaniline as a function of temperature and time, a) Initial spectrum recorded at -10° C and near the beginning of the reaction, b) Spectrum recorded at time ~ one hour (-70°C). c) Spectrum obtained at the end of the reaction after warming to -20°C. 153 Figure 4-8. 400 MHz ^f 3 1?} NMR spectrum of the intermediate observed in the reaction between l a and N-benzylideneaniline (spectrum recorded at -70°C). 1 5 4 H Figure 4 - 9 . Two possible structures for the intermediate observed at low temperature in the reaction of la with N-tenzyMeneaniline. strong evidence for the assignments made. Again, similar features are observed in the lH NMR spectrum of the intermediate formed from Id and N-benzylideneaniline. The NMR evidence clearly indicates that the first step in this reaction involves the cleavage of the Rh2(p>H)2 core of the dihydride by the imine. Accounting for the required structural features discussed above, two plausible structures for the intermediate are given in Figure 4 - 9 . Initially, structure 25 was favoured due to its similarity to a structural proposal for an intermediate observed in the hydrogenation of [{P(OPri)3)2Rh]2[p>CN-(p-Cl-C6Ff4)](M.-H)2 (equation 4 - 1 0 ) . 6 Structure 26, however, has some analogy to the p>Ti2-o" alkenyl hydride complexes produced in the reactions of la with simple olefins.7-11 In 26, the imine fragment bridges the two metal centres in the same manner as the (i-alkenyl ligand does in the alkenyl hydride complexes. In order to distinguish between these two possible structures, the reaction of la with (C6Hs)N=13CH(C6H5) was carried out. The 31P{ lH} NMR spectrum of the 13C-labelled intermediate was not visibly different from the spectrum of the unlabelled species; /. e., no large P-C couplings were observed. In the ^Cf-H} NMR spectrum, a broad signal (wi/2 - 2 0 Hz) at 6 2 . 4 ppm was observed again with no large P-C or Rh-C couplings apparent. In addition, a large ^ O C - H coupling constant of 151 .3 Hz 155 R N (PHO) 3 P^ / C \ ^ p ( O P r ' ) 3 ,Rh . .Rh (Pr'0)3P^ \ / ^P(OPr') 3 H R = p-CI-C6H5 H / p , x c = N , .P H ^ R h w R h -p 7 H (4-10) (doublet) suggested that the imine carbon remains sp2 hybridized in the intermediate. These results eliminate structure 25 from consideration, since one would expect a large one-bond J R I I - C coupling as well as observable P-C couplings in the signal for the labelled carbon. Also, the sp3 hybridization of that carbon in 25 should have resulted in a 1 Ji3c-H coupling constant similar to the 129.6 Hz observed for the benzylic C-H bond in the final product 18a rather than the observed coupling constant, which was in the same range as the corresponding coupling constant in the free imine (158.3 Hz). Thus, structure 26 appears to be the formulation which best fits all of the data. The large upfield shifts for the resonances due to both the imine proton (vide supra) and the imine carbon (62.4 ppm vs 160.0 ppm for the free imine) can be accounted for by 7c-coordination of the imine C=N bond to rhodium. Upfield shifts of similar magnitude are observed in the resonances for the p" protons and the (3 carbon nuclei in the vinyl hydride complex [(dippe)Rh]2(H-T|2-a-CH=CH2)(H-H).11 156 4.3.2 Reaction of [(dippe)Rh]2([i-H)2 with Benzophenone imine at Low Temperature. Although not apparent from the data described above, the formation of 26 probably takes place subsequent to coordination of the imine via the nitrogen lone pair to one coordinatively unsaturated rhodium centre. The reaction between la and benzophenone imine, however, provides spectroscopic evidence for such a pathway. In a more complex reaction course, a second observable species is formed along with an intermediate analogous to that formed in the reaction of la and N-benzylideneaniline en route to the amido hydride product. A series of 31P{1H} NMR spectra recorded at various times and temperatures during the reaction is shown in Figure 4-10. In contrast with the la/N-benzylideneaniline reaction, it was not possible to isolate each step in this reaction and thus peaks for all components are present during the course of the reaction. Nonetheless, the sequence in which each intermediate appears is clear. Almost immediately after adding a 2-3 fold excess of benzophenone imine at -78°C, the dihydride disappears and four new signals arise (Figure 4-10a). As the reaction proceeds, this species decomposes to a second unsymmetrical (four phosphorus environments) intermediate; the appearance of signals due to 21a occurs virtually simultaneously (Figure 4-10b-e). Upon completion of the reaction this complex mixture of components has all gone to product; as seen in Figure 4-1 Of, the reaction is quantitative. Parenthetically, it should be noted that at low temperatures the signals due to the amido hydride are broad and coalescing presumably due to the freezing out of rotation about the N-C bond in the (i-amido ligand; this spectral behaviour was confirmed by a separate low temperature experiment (AG* = 9.1(3) kcal/mol). Monitoring this reaction in a similar fashion via *H NMR spectroscopy clarifies the process somewhat; in particular, the hydride region of the spectrum is informative (Figure 4-11). Early in the reaction, signals for two bridging hydrides are observed at -3.75 and -5.04 ppm, with large trans H-P couplings apparent (Figure 4-1 la). As the reaction proceeds these 1 1 0 1 6 6 9 0 80pprn 1 2 0 1 1 0 ' 100 9 0 ' 8 0 p p m Figure 4-10. 162.2 MHz 31P{1H} NMR spectra of the proceeding reaction between la and benzophenone imine as a function of temperature and time, a) Initial spectrum recorded at -70°C and near the beginning of the reaction; formation of the first intermediate of structure 28 or 29. b-e) Spectra recorded as sample was warmed gradually over a period of about 45 minutes; first intermediate slowly disappears, while a second intermediate grows in along with signals for the amido hydride product 21a. f) Spectrum recorded upon completion of the reaction. 158 Figure 4-11. Hydride region of the 400 MHz *H NMR spectrum of the proceeding reaction between l a and benzophenone imine. a) Initial spectrum collected at -70°C. b-d) Spectra recorded at =10 minute intervals at -50°C. e) Final spectrum recorded upon completion of the reaction (-20°C). 159 signals disappear, while at -11.51 and -15.69 ppm new resonances grow in (Figure 4-llb-d). These resonances are due to the bridging and terminal hydride ligands of the intermediate analogous in structure to 26. Other signals in the spectra in Figure 4-11 are due to la (-4.63 ppm) and 21a (-7.88 ppm). The second, unknown intermediate, then, is formed prior to the |i-ri2-o" imine species analogous to 26 (i. e., 27) and contains two bridging hydrides and four inequivalent phosphorus nuclei. In addition, a signal at 11.0 ppm is assigned to the nitrogen bound imine proton. While this data doesn't allow for unambiguous assignment of structure, two reasonable proposals are given in Figure 4-12. Structure 28 results upon coordination of the imine to one rhodium centre via the nitrogen lone pair; the observed asymmetry in the complex occurs if there is restricted rotation about the Rh-N bond. Structure 29 is similar in structure to many observed and proposed [(P2)Rh]2(|J.-X)(|±-H)26,12 complexes found in the chemistry of the dihydrides la-e. It represents the species present just prior to Rh-H bond cleavage to form the second observable intermediate containing the terrninal hydride ligand. Ph 28 R = Pr i 29 Figure 4-12. Two possible structures for the first intermediate observed at low temperature in the reaction of la with benzophenone imine. 160 4.3.3 Kinetic Studies and a Proposed Mechanism. Scheme 4-2 shows a proposed mechanistic sequence for this reaction, incorporating all of the results discussed above. In the reaction between la and N-benzylideneaniline the first intermediate was not detected, reducing the process to two kinetically isolable steps. Both steps in this reaction may be conveniently followed kinetically by 31P{1H} NMR spectroscopy and the results of these studies are congruent with the mechanistic pathway proposed in Scheme 4-2. The first step was found to be cleanly second order, first order in both la and N-benzylideneaniline. A graph of second order data at various temperatures is shown in Figure 4-13, while an Eyring plot is given in Figure 4-14. Activation parameters for this step are AH* = 10.1(5) kcal/mol and AS* = -19.8(5) eu. Note that these results do not preclude a rapid pre-equilibrium step forming the first intermediate in unobservable quantities. The second step, found to be first order in the rate of disappearance of the intermediate, was followed after warming the sample to temperatures around -20° C. Figure 4-15 gives a plot of first order data at these temperatures and Figure 4-16 shows the corresponding Eyring plot. Activation parameters for the second step are AH* = 17.3(3) kcal/mol and AS* = -3.6(3) eu. Rate constants at various temperatures for steps one (ki) and two (k2) are given in Tables 4-III and 4-IV, respectively. Kinetic analysis of the reaction involving the dideuteride d2-la with N-benzyhdeneaniUne revealed a small kinetic isotope effect (kn/ko) of 1.41(5) for step one and a more substantial value of 2.24(3) for step two of the reaction. The latter kjj/kD is within the range of 1.4-2.7 predicted for the primary isotope effect on the insertion of ethylene into the Nb-H bond of Cp*2Nb(H)(rj2-C2H4).13 While the two reactions are quite different, if delivery of the hydride to the imine carbon in 26 takes place via the 4-centre transition state commonly invoked for olefin insertion reactions, the kinetic isotope effects should be comparable. Accounting for the kn/kD of 1.41 observed for the first step is more difficult. In step one, the proposed sequence involves a bridging hydride going to a terminal hydride. 161 Scheme 4-2 Rh Rh / V \ Ph N = C Ph Rapid Pre-Equilibrium H R 2 I R2 P>w ,«*H ' '. I ^ P 2 :Rh ' _ j R h - ^ p***^ ^ H * * ^ P R 2 R2 H I 0 R NJL:-/X -i-pr R2 J - * ^ H ^ F h -H k, -70*C Stepl P ^ / \ .N -"Rh. H 'PR. H - * * C - \ Ph k2 -20°C Step 2 Ph CH2Ph <yy<> \ Rh Rh L,/ v \ - -<> 162 < ffl • Z -1.5 -2.0 -2.5 -3.0 -3.5 -4.0 204 K 218 K \ \ 214 K \ s 2 0 9 K n r ^ — 1000 2000 t ime (sec) 3000 Figure 4-13. Second order rate data obtained at various temperatures for the first step of the reaction between l a and N-benzyhdeneaniline (N-BA). 4.5 4.6 4.7 4.8 4.9 5.0 1/T(1/Kx1000) Figure 4-14. Eyring plot for the first step in the reaction between l a and N-benzyUdeneanihne (N-BA). 163 Figure 4-15. First order rate data obtained at various temperatures for the second step of the reaction between la and N-benzylideneaniline. 3 . 8 3 . 9 4 . 0 4 . 1 4 . 2 1/T(1/Kx1000) Figure 4-16. Eyring plot for the second step in the reaction between la and N-benzylideneaniline. 164 Table 4-in. Rate Constants and Eyring Plot data for la + N-benzylideneaniline ==> 26. Temperature3 rate constant,1* lq lnfki/T) in 204 3.13e-3 -11.08 4.90e"3 209 4.39e-3 -10.77 4.78e-3 214 1.14e-2 -9.84 4.67e"3 218 1.26e"2 -9.76 4.59e-3 a Degrees K. b 1-mol"1-sec-1. Table 4-IV. Rate Constants and Eyring Plot data for 26 ==> 18a. Temperature3 rate constant,1* k2 In(k2/T) 1/T 244 1.65e-* -14.21 4.10e-3 248.5 3.24e-4 -13.55 4.02e-3 254 6.92c-4 -12.81 3.94e-3 258 1.20e-3 -12.28 3.88e-3 a Degrees K. b sec"1. Based on zero-point energy arguments, this should result in a kfj/kD of <1, since putting deuterium in a terminal position should accelerate the rate if it is rate determining. However, given the delocalized 4-centre, 4-electron bonding in the Rh2(|i-H)2 core of la, cleavage of the Rh2(|i-H)2 unit upon interaction with the imine likely involves more than simply breaking one Rh-H bond. The observed kn/kD is thus a combination of primary and secondary isotope effects resulting from rupture of the bonding in the dihydride core. In summary, a likely pathway for the production of the intermediate 26 involves coordination by the lone pair on the imine nitrogen to one coordinatively unsaturated rhodium centre in la, followed by it-donation of the C=N double bond to the adjacent rhodium atom 165 with concomitant rupture of the Rh2(p>H)2 4-centre, 4-electron core of the dihydride molecule to form 26. This pathway to 26 is suggested by the observance of the first intermediate in the la/benzophenone imine reaction, with structure 28 or 29. The second step of the reaction, taking place subsequent to generation of 26, may be formally analyzed as an insertion of the C=N double bond into the newly formed terminal Rh-H bond, producing 18a. Although we cannot distinguish unambiguously between migratory insertion of the terminal hydride versus the bridging hydride, the isotope effect is more easily explained by the latter. A migratory insertion of an imine C=N bond into a metal hydride has not previously been observed directly, although such a process has been inferred as a step in the catalytic hydrogenation of Schiff bases by rhodium/phosphorus catalyst systems.14 Mechanistic work on these systems suggest, however, that a hydride transfer mechanism is operative rather than the classical 4-centre insertion mechanism invoked for C=C insertions into metal hydride linkages. A classical insertion mechanism in C=N insertions is rare, probably because of the tendency of imines to bond to metals via the lone pair on nitrogen rather than through any bonding through the 7t system.3 In this binuclear system, however, 7t-donation to a proximal rhodium centre is induced because the lone pair is occupied by the other rhodium atom in the dimer. 4.4 Reactions of the Dihydride la with Iminium Salts. The |i-r|2-o" imine dihydride intermediates 26 and 27 are a potentially interesting group of compounds themselves, because of the unprecedented bonding mode of the |i-imine ligand. Unfortunately, attempts at isolating 26 at low temperature were only partially successful. While it could be precipitated from a concentrated solution in hexanes at -78°C, even as a solid it decomposed to the amido hydride product upon warming. This extreme sensitivity precluded further characterization in our hands. 166 Attempts were also made to trap the intermediate by reacting it with halogenated solvents or ethylene after it had formed at -78°C, the purpose being to remove or modify the terminal hydride ligand. These reactions, however, resulted only in decomposition and no reaction for CClnH4.n and ethylene, respectively. Since these attempts to further characterize 26 met with failure, an alternative strategy was employed in seeking supporting evidence for the structure of the intermediate. The instability of the intermediate is seemingly due to the fact that it is geometrically "set up" to undergo insertion to form the amido hydride. It was therefore reasoned that a cationic compound without the terminal hydride would perhaps be stable and isolable (equation 4-11). (4-11) The observation that the dihydride la reacts with a variety of Bronsted acids to evolve dihydrogen and yield complexes with the conjugate base of the acid in a bridging position (Chapter 2) suggested that la might react with the iminium salt of N-benzylideneaniline to give the desired complex. Protonation of this aldimine with one equivalent of tetrafluoroboric acid (HBF4) led to isolation of the iminium salt 30, which upon reaction with the dippe dimer la, gave a dark purple solution from which purple, analytically pure crystals of the cationic p>imine hydride complex 32 were isolated (equation 4-12). An analogous reaction obtains when the tetrafluoroborate iminium salt of N-benzylidenemethylamine (31) is reacted with la, producing 33. 167 (4-12) R' = Ph, 32 R' = CH 3 ) 33 The NMR spectral data obtained for these complexes is consistent with the structural formulation in equation 4-12. Particularly diagnostic are the 31P{1H) spectra collected at various temperatures; Figure 4-17 shows a series of such spectra for R' = C6H5, 32. At +30°C, two broad doublets of multiplets are observed, similar in shape to the patterns observed for the amido hydride complexes. As the sample is cooled, each doublet coalesces (15°C, AG* = 12.8(5) kcal/mol) and reemerges as two other signals, until at -30°C the low temperature limit spectrum is reached. Notably, this low temperature spectrum is remarkably similar to the spectrum of the intermediate 26, as seen in Figure 4-18. This is good evidence that the two complexes are related structurally and further supports the structural assignment for the |I-T12-G imine dihydride intermediates in the dihydride/imine reactions. That the cationic |i-imine hydride complexes exhibit such fluxional behaviour is not surprising given 168 30°C 20 -15 *-v*wUrW -30 60 •WUt^ Vr/ »**'*M '#->»'i ^ i * ^ * * * ! ******** uv 1 i—I—I—I—I—I—I—1—i—i—r do " 1 — i — i — i — I — i — i — i — r ~ i — i — r 105 100 95 90 PPM BIT Figure 4-17. 121.4 MHz 3 1P {lH} NMR spectra of the cationic \i-T[2-c imine complex 32 recorded at various temperatures. 169 Figure 4-18. A comparison of the 3 1 P NMR spectra of the intermediate observed in the reaction between la and N-benzylideneaniline, 26, and the cationic |i-r|2-o imine complex 32. a) 121.4 MHz 3 1P pH} NMR spectrum of 32 (-30°C). b) 162.4 MHz 3 1 P {lH} NMR spectrum of 26 (-70°C). 170 Scheme 4-3 that the isoelectronic alkenyl hydride complexes alluded to previously undergo a well documented15 "windshield wiper" fluxionality as depicted in Scheme 4-3.7 Indeed, the ^P^H} spectral behaviour of [(mppe)Rh]2(M>T12-a-CH=CH2)(p>H) (Figure 4-19) is similar to that of the -^imine hydrides 32 and 33. By analogy then, the fluxional process proposed for 32 (and 33, R' = CH3) is a similar "windshield wiper" process which equilibrates the four phosphorus environments observed at low temperatures into the two seen at ambient temperature (Scheme 4-4). Of interest in the 1 3 C and lH NMR spectral data are the resonances due to the imine Scheme 4-4 Rh* = Rh(dippe) Figure 4-19. 162.2 MHz 3 1 P pH} NMR spectra of the alkenyl hydride complex [(dippe)Rh]2(|i-Ti2-a-CH=CH2)(p>H), isoelectronic with 32. a) 60°C. b) 10°C. c) -70°C. d) Calculated spectrum for spectrum c (see reference 7 for parameters). 172 carbon and proton atoms in the bridging imine ligand. The carbon signals occur at 93.3 and 102.8 ppm for 32 and 33, respectively. These chemical shifts are somewhat higher than the range of 68-80 ppm observed for the chemical shifts of the f3-carbons in the u.-alkenyl hydride complexes; however, the magnitude of the upfield shifts of these nuclei when compared to the free ligands in both types of compounds is comparable. The resonances in the *H NMR spectrum for the imine proton (7.78 and 7.70 ppm for 32 and 33 respectively) are close to the range of 6-7 ppm observed for the syn-$ proton in the alkenyl hydrides. This confirms the stereochemistry about the C=N bond as being trans in these complexes and thus the reaction of la with the iminium salt is stereospecific with respect to stereochemistry about the C=N bond. The chemistry of these cationic |i-imine hydride complexes is potentially interesting, but largely unexplored to date. General observation reveals their thermal instability in solution, particularly in strong donor solvents. Though moderately stable in THF or acetone for short periods of time (<24 hours), the complex 32 decomposes rapidly in acetonitrile. The fate of the organic ligand in this decomposition is unclear and the organometallic species produced is a highly symmetrical compound which exhibits a sharp doublet in the 3 *P NMR spectrum and a *H NMR spectrum very similar to that of the free dippe ligand. Interestingly, this compound is also generated in the decomposition of 33. Clearly, more study is warranted. The reactivity of these cationic complexes towards nucleophiles is also an area of potential exploration. Of particular relevance to this work is the reaction of 32 with mild hydride reagents. When treated with a slight excess of LiAlH(OBut)3 in THF at room temperature, the amido hydride 18a was isolated in 78% yield (equation 4-13). When this reaction was monitored spectroscopically (31P NMR spectroscopy) at low temperature, however, a variety of species were observed. The major product was still the amido hydride 18a, but other unidentified products were also formed. Intermediates were also detected, including significant amounts of the intermediate 26. The presence of other species is not 173 Ph + [ B F „ r UAIH(OBu')3 THF (4-13) surprising in light of the number of potential sites of reactivity in the molecule. In addition to the two metal centres, direct attack on the |i-imine ligand, at either carbon or nitrogen, is also feasible. In addition the bulk of the hydride reagent employed probably hampered attack at the metal centres. Nonetheless, detection of 26 provides further circumstantial evidence as to the validity of the mechanistic proposal for the reaction of la and Id with imines. The results just described are incomplete and provide the basis for new and potentially important studies on these unique cations. 4.5 The Relevance of this Work to the Homogeneous Hydrogenation of The initial impetus for these studies was to investigate the activity of the dihydrides la-d towards the hydrogenation of imines. The desirability of this transformation is rooted in the fact that homogeneous catalyst systems which bring about the hydrogenation of imines or nitriles under ambient conditions are rare.14-16"19 Most systems reported to date utilize rhodium/phosphine based catalysts, although metal carbonyls (M = Fe,17 Co,18 and Cr, Mo, or W19) in strongly polar solvents have been employed successfully at high temperatures and pressures. The paucity of catalyst systems which will hydrogenate imines without resorting to forcing conditions is likely a reflection of the imine's tendency to bond to metals via the Imines. 174 lone pair on nitrogen rather than through 7C-bonding of the C=N double bond.3 If one draws parallels between imine hydrogenation and the catalytic hydrogenation of olefins then %-coordination of the C=N bond to the metal is a requirement before insertion into a metal hydride bond is possible; o-bonding does not lead to reduction of the C=N double bond. The requirement of 7t-coordination for smooth insertion into an M-H bond necessitates that the lone pair on nitrogen be unavailable for donation to the metal centre. This feature is included in the mechanistic proposal for the hydrogenation of aldimines by a cationic rhodium triphenylphosphine complex. In this proposal, an intermediate (shown below) in which the I N Rh C - H P > H V H imine C=N bond is 7C-coordinated, forms when the lone pair hydrogen bonds with a coordinated alcohol solvent molecule. Hydrogenation using this catalyst system was effected under very mild conditions (25°C, 1 atmosphere dihydrogen).143 In the reactions of la and Id with imines (vide supra), insertion of the C=N bond into the Rh-H bond was very facile in the |i-r|2-o" imine dihydride complex 26 even at low temperatures. This observation provides more evidence that re-coordination is necessary to effect reduction of the C=N double bond in imines. Recall that in the formation of 26 (Scheme 4-2), the lone pair is first donated to one rhodium centre; subsequently %-coordination to the other rhodium centre takes place followed by rapid insertion into the Rh-H bond. The presence of multiple metal bonding sites thus facilitates the hydrogenation of the C=N bond; this factor is also likely a contributor to the effectiveness of heterogeneous 175 catalysts in bringing about this transformation.20 The use of coordinatively unsaturated cluster catalysts allows for the possibility of performing this reaction homogeneously by providing more metal bonding sites. The reduction of the C=N bond via the pathway observed in these binuclear rhodium hydride systems necessarily produces a bridging amide complex. Up to this point, the C=N bond is only partially reduced and to complete the hydrogenation, the amide ligand must be hydrogenolyzed to produce amine and regenerate 5a. While the reaction of 18a with dihydrogen does proceed to produce free amine and the tetrahydride 5a (equation 4-14), unfortunately, the rate of the reaction is prohibitively slow, requiring about 3 days to complete in the NMR tube (with efficient stirring the reaction is faster, but remains too slow for practical purposes). In the presence of excess imine substrate and 4 atmospheres of dihydrogen none of the dihydrides la-b, or Id exhibited significant activity in the production of N-phenylbenzylamine. Raising the dihydrogen pressure to 100 atmospheres had little (4-14) R = Pr1, 5a 176 effect. It thus appears that a pathway involving binuclear intermediates is rendered kinetically inert upon formation of the u,-amido hydride "intermediate", at least in these systems. When the tetrahydride 5a was formed prior to addition of the imine substrate and used as the catalyst precursor, some hydrogenation was observed. Approximately 7 turnovers occurred over a 16 hour period. Eventually, however, the amine production halted due to a build up of 18a. In this reaction, the possibility of catalysis by small amounts of a mononuclear species must be considered; when the mononuclear complex [(dippe)Rh(rj3-C3H5)], the precursor to la, was used as catalyst precursor, hydrogenation was more efficient. The results, however, were not impressive enough to pursue. 4.6 Experimental. 4.6.1 General Procedures. N-benzylideneaniline was prepared via a literature procedure;21 its 1 3 C labelled isotopomer, (C6H5)N=13CH(C6H5), was prepared in the same manner using 1 3 C labelled benzaldehyde obtained from MSD Isotopes. N-benzylidenemethylamine, N-benzylidenebenzylamine, benzophenone imine, and the cyclic imine isoquinoline were all purchased from the Aldrich Chemical Company, distilled under argon prior to use and stored in the glovebox. N-phenylbenzylamine was also purchased from Aldrich and distilled prior to use. Tri-ferf-butoxylithiumaluminum hydride was also purchased from Aldrich. The tetrafluoroborate iminium salts [ ( C 6 H 5 ) HN = C H ( C 6 H 5)] B F 4 and [(CH3)HN=CH(C6H5)]BF4 were prepared via protonation of the corresponding imine with H B F 4 (Aldrich, 85% in Et20) in diethylether. The iminium salts precipitated upon addition of the H B F 4 and after washing twice with 50 mL portions of Et20 were dried under vacuum and stored in the glovebox. Purity was found to be >98% via *H NMR. 177 4.6.2 The Reactions of [(dippe)Rh]2(p>H)2, la, and [(dipope)Rh]2(p>H)2, Id, with Imines. R = Prj, a series R = OPr1, d series R' R" Compound CeH 5 CHgCgHs 18 C H 3 C H 2 C 6 H 5 19 CHgCeHs C H 2 C 6 H 5 20 H CH(CeH5)2 21 R = Pri, 24a R = OPr', 24d \ / r V(Phl2 L« p \ y \ / p Rh Rh / V \ P PR = Pr\ 21b 178 Synthesis of [(dippe)Rh]2[>(C6H5)N(CH2C6H5)](|i-H), 18a. To a stirred solution of [(dippe)Rh]2(|i-H)2, la, (0.164 g, 0.22 mmol) in toluene (15 mL) was added solid N-benzylideneaniline (0.081 g, 0.44 mmol). Within 5 minutes, the deep green to orange color change was complete. The toluene was removed in vacuo and the orange solid remaining was recrystallized from minimum toluene/hexanes (1:2), yielding 0.167 g (82%) of orange crystalline [(dippe)Rh]2[MC6H5)N(CH2C6H5)](-i-H), 18a. lK 3 NMR (C6D6, ppm): C-Hi, 8.74 (d, 2H, 3 J H 2 = 8.0 Hz); C-H4 or C-H5, 8.58 (d, 1H, 3 J H = 8.4 Hz); C-H2, C-H3, C-H6 or C-H2, and C-H4 or C-H5/7.14-7.30 (overlapping multiplets, 5H); C-H6 or C-H 2 7.00 (m, 1H, 3 J H » 7.2 Hz); C-Hg, 6.75 (m, 1H, 3 J H « 7.6 Hz); NCH2C6H5, 5.32 (br t, 2H, 4 J P » 2.0-3.0 Hz); CH(CH3)2, 2.32, 2.26, 1.78, 1.68 (dsp, 2J P= 2.0-4.0 Hz, 3JH= 7.2-8.0 Hz); CH(CH3)2, 1.52, 1.50, 1.13, 1.06, 0.88, 0.85, 0.64, 0.57 (overlapping dd, 3Jp = 12.8-14.4 Hz); CH9CH7. (resonances buried underneath signals for the methyl groups); Rh-H-Rh, -9.41 (ttt, 2Jptrans = 56.4 Hz, 2Jpcis = 12.8 Hz, !jRh = 23.1 Hz). 13c{lH} NMR (C6D6, ppm): Cipso(N), 164.6; Cipso(C), 143.7; other aromatic carbons, 130.3, 128.8, 127.4, 126.3, 126.1, 124.7, 117.8, 111.3; NCH 2 C 6 H 5 , 66.8; CH(CH3)2, 26.8-28.7 (4 m); CH2CH2, 21.8-22.8 (m); CH(C.H3)2, 18.3-22.7 (8 s). 31P{ 1H} NMR (C6D 6, ppm): PcisN = A; PtransN = B; Rh = X: 5A = 73.9; 6B = 105.5. J A A ' = 33.5 Hz; L A B = 30.3 Hz; J A B - = 2.9 Hz; J A X = 191.3; J A X ' = 5.8 Hz; J B B ' = -0.9 Hz; J B X = 166.9 Hz; J B X ' = -1.2 Hz; JXx' = 5.6 Hz. Anal. Calcd. for C4iH75P4Rh2N: C, 54.01; H, 8.29; N, 1.54%. Found: C, 54.30; H, 8.52; N, 1.62%. 179 Synthesis of [(dippe)Rh]2[n-(CH3)N(CH2C6H5)](^-H), 19a. To a stirred solution of [(dippe)Rh]2(p>H)2, la, (0.274 g, 0.37 mmol) in toluene (20 mL) was added neat N-benzylidenemethylamine (0.088g, 0.74 mmol), in one portion. A rapid green to orange color change ensued. The toluene was then removed under reduced pressure, and the resulting orange residue was recrystallized from toluene/hexanes (1:5). The orange crystals were washed twice with cold hexanes to remove excess imine. Yield: 0.262 g (82%). *H NMR (C6D6, ppm): Uortho, 8.35 (d, 2H, ^ Umeta = 7 6 Hz); #meta and Hpara, 7.05-7.18 (m, 3H); NCH2C6H5, 5.10 (br t, 4 J P = 4.8 Hz); NCH3, 3.80 (br t, 3H, 4 J P « 1.5-2.0 Hz); CH(CH3)2, 1.8-2.2 (4 overlapping dsp, 8H, 3jH = 6.9-7.7 Hz, 2 J P = 3.7-4.0 Hz); CH(CH3)2, 1.37, 1.35, 1.33, 1.08, 1.06, 0.88, 0.83 (overlapping dd, 48H, 3jp = 12.8-14.4 Hz); PCH2CH2P, 1.15-1.25 (m, 8H); Rh-H-Rh, -8.92 (ttt, 2Jptrans = 57.2 Hz, 2Jpcis = 12.8 Hz, URh = 22.7 Hz). 13c{lH} NMR (C6D6, ppm): C i p s o , 143.8; other aromatic carbons, 129.3, 126.2, 125.8; NCH3, 75.6; NCH2(C6H5), 56.2; phosphine ligand resonances, 18-28. 31p{lH} NMR (C6D6, ppm): Ptrans(Nh 1 0 L 1 (dm. ^Rh « 152 Hz); ?cis(N), 75.1 (dm, ijRh « 195 Hz). Anal. Calcd. for C36H75P4RI12N: C, 50.77; H, 8.88; N, 1.64%. Found: C, 50.67; H, 9.10; N, 1.59%. Synthesis of [(dippe)Rh]2[M.-N(CH2C6H5)2](p-H), 20a. To a stirred solution of [(dippe)Rh]2(p>H)2, la, (0.045 g, 0.06 mmol) in toluene (10 mL) was added neat N-benzylidenebenzylamine (0.015g, 0.08 mmol), in one portion. The reaction mixture was allowed to stir until the green to orange color change was complete (10 minutes), at which time the toluene was removed in vacuo. The remaining solid was recrystallized from toluene/hexanes (1:2), yielding a crop of bright orange crystals (0.039 g, 69%) after washing with cold hexanes. ! H NMR (C6D6, ppm): H^nho, 8.54 (d, 4H, 3 3 n m e t a = 7-2 Hz); H ^ , 7.21 (m, 4H, ^ npara - 7.5 Hz); ILpflrfl, 7.13 (m, 2H); NCH2(C6H5), 5.44 (br t, 4H, 4 J P = 3.5-4.0 Hz); CH(CH3)2, 2.35, 1.85 (dsp, 8H, 3jH = 180 7.6-8.0 Hz, 2jp = 3.0-4.0 Hz); CH(CH3)2, 1.48, 1.06, 0.73, 0.72 (dd, 48H, 2jp = 12.0-14.0 Hz); PCH2CH2P, 1.1 (br m, obscured by methyl groups); Rh-H-Rh, -9.40 (ttt, 2Jptrans = 57.2 Hz, 2j P d j = 12.8 Hz, l J R h = 22.0 Hz). ^C{^U] NMR (C6D6, ppm): Ci p s o , 144.3; other aromatic carbons, 129.2, 125.9, 125.5; NCH2, 76.1; CH(CH3)2, 26.4-28.7 (m); PCH2CH2, 19.7-21.1 (m); CH(CH3)2, 18.4, 19.0,19.0, 19.6 (s). 31p{lH.} NMR (C6D6, ppm): Ptrans(N), 98.8 (dm, lJ R h » 154 Hz); Pcis(N), 74.7 (dm, lJ R h - 199 Hz). Anal. Calcd. for C42H79P4Rh2N: C, 54.37; H, 8.58; N, 1.51%. Found: C, 54.26; H, 8.80; N, 1.50%. Synthesis of [(dippe)Rh]2[|i-HNCH(C6H5)2](|i-H), 21a. To a stirred solution of [(dippe)Rh]2(p>H)2, la, (0.060 g, 0.082 mmol) in toluene (10 mL) was added neat benzophenone imine (0.022 g, 0.12 mmol), in one portion. The reaction mixture was stirred for an additional 10 minutes after which the green to orange color change was complete. The solvent was removed in vacuo, and the remaining orange solid was recrystallized from toluene/hexanes (1:3), yielding a crop of orange crystals of [(dippe)Rh]2[u.-HNCH(C6H5)2](|i-H) (0.052 g, 70%) after washing with cold hexanes. lH NMR (C6D6, ppm): Uortho, 7.78 (d, 4H, 33nmeta = 7.6 Hz); Umeta, 7.20 (m, 4H); Hparf l, 7.13 (m, 2H); NCH(C 6H 5) 2, 6.28 (br t, 1H, 4 J P = 5.2 Hz); H-N, 4.57 (br s, 1H); CH(CH 3) 2, 2.45, 1.90, 1.81, 1.67 (dsp, 8H, 3jH = 7.5-7.9 Hz, 2jp = 2.0-4.0 Hz); CH(CH3)2, 1.30, 1.21, 1.19, 1.06, 1.05, 1.02, 0.92, 0.92 (overlapping dd, 48H, 2 J P = 12.0-14.0 Hz); P C H 2 C H 2 P , 1.1-1.2 (m, obscured by methyl groups); Rh-H-Rh, -8.20 (ttt, 21Vtrans = 60.0 Hz, 2j P c / j = 12.4 Hz, l J R h = 22.2 Hz). 13C{lH} NMR (C6D6, ppm): Cipso, 147.2; other aromatic carbons, 129.2, 127.2, 126.2; NCH(C6H5)2, 73.7; CH(CH3)2, 27.8-28.9; CH 2CH 2, CH(CH3)2, 17.1-25.2 (overlapping signals). 31p{lH} NMR (C6D6, ppm): Ptrans(N), 104.3 (dm, l J R h = 159 Hz); Pcis(N), 8 4 - 5 (dm. l jRh ~ 191 Hz). Anal. Calcd. for C 4iH 7 7P 4Rh 2N: C, 53.89; H, 8.49; N, 1.53%. Found: C, 54.07; H, 8.37; N, 1.60%. 181 Synthesis of [(dippe)Rh]2[MC9H8N)](p>H), 24a. To a stirred solution of [(dippe)Rh]2(u.-H)2, la, (0.140 g, 0.19 mmol) in hexanes (15 mL), was added dropwise a solution of isoquinoline (0.074 g, 0.57 mmol) in hexanes (5 mL). The green solution of la turned to a red-brown color within 10 minutes and the hexanes were removed in vacuo leaving a dark red-brown solid. Recrystallization from hexanes yielded a crop of red crystals (0.112 g, 68%) after washing with cold hexanes. *H NMR (C6D12, ppm): C-H4 or C-H5, 6.83 (m, 1H, 3 j H = 6.4 Hz); C-H4 or C-H5 and C-H3 or C-H6, 6.71 (m, 2H); C-Hi, 6.52 (d, 1H, 3 j H 2 = 7.6); C-H2, 4.98 (d, 1H); NCH2, 5.66 (br t, 2H, 4 J P = 2.2 Hz); CH(CH3)2, 2.07, 2.05, 1.97, 1.85 ( overlapping dsp, 8H, 3 j H = 7.5-8.0 Hz, 2 j p = 2.0-4.0 Hz); CH(CH3)2 and PCH2CH2P, 0.9-1.3 (overlapping signals); 4 Rh-H-Rh, -9.31 (ttt, 2Jptrans = 5 9 2 Hz> 2 jPc« = 1 3 6 H z » ljRh = 22.9 Hz). ^C{lR} NMR (C6D12, ppm): NCH, 148.7; C-C (bridgehead carbons), 136.1, 128.7; other aromatic carbons, 125.0, 123.2, 121.5, 119.8, 95.9; NCH2, 62.2; CH(CH3)2, 27.4, 27.3, 26.6, 25.5 (m); CH2CH2, 20.3 (m); CH(CH 3) 2, 21.0, 20.7, 20.6, 19.5, 19.4, 17.9, 17.5, 17.4. 31p{lH} N M R (C6Di2, ppm): Ptrans(N), H0.5 (dm, l J R h - 162 Hz); Pcis(N), 84.1 (dm, !jRh =185 Hz); minor regioisomer: Plrans(N), H5.7 (dm, URJ, ~ 175 Hz); Pcis(N)> 96.7 (dm, ijRh * 188 Hz). Anal. Calcd. for C 3 7H 7 3 P 4 Rh2N: C, 51.57; H , 8.54; N, 1.63%. Found: C, 51.53; H, 8.50; N, 1.62%. 182 Synthesis of [(dipope)Rh]2[p>(C6H5)N(CH2C6H5)Kp>H), 18d. To a stirred solution of [(dipope)Rh]2(|i-H)2, Id, (0.150 g, 0.17 mmol) in toluene (10 mL) was added solid N-benzylideneaniline (0.063 g, 0.35 mmol), in one portion. The reaction was stirred for 30-45 minutes, during which a slow red to yellow-brown color change was observed. The toluene was removed under reduced pressure, and the oily yellow-brown residue recrystallized from hexanes (<1 mL). After a quick wash with 0.5 mL of cold hexanes, 0.152 g (84%) of yellow crystals of [(dipope)Rh]2[|i-(C6H5)N(CH2C6H5)](n-H) was obtained. *H NMR (C7D8, ppm): C-Hi, 8.42 (dd, 2H, 3jH 2 = 7.4 Hz, 4 J H 3 = 1.9 Hz); C-H4, 7.61 (d, 2 H 3jH 5 = 8.0 Hz); C-H2 and C-H3 or C-H6, 7.01 (m, 3H); C-H5, 6.89 (m, 2H); C-H3 or C-H6, 6.58 (m, 1H); N C H 2 C 6 H 5 , 5.01 (br t, 2 H , 4 J P = 4.5 Hz); OCH(CH3)2, 5.07, 4.43, 4.40, 3.71 (dsp, 8 H , 3jH = 5.8-6.6 Hz, 3jP = 2.0-4.0 Hz); OCH(CH3)2, 1.39, 1.37, 1.30, 1.14, 1.05, 1.02, 1.00, 0.78 (d, 48H); P C H 2 C H 2 P , 1.5-1.8 (m, 8H); Rh-H-Rh, -8.32 (ttt, 2Jptrans = 84.1 Hz, 2 J P D J = 4.8 Hz, l j R h = 23.1 Hz). 13C{lH} NMR (C 7D 8, ppm): Cipso(N), 165.6 (t, 3jP = 3.3 Hz); £ipSo(C), 159.9; other aromatic carbons, 142.6, 131.3, 131.1, 126.7, 126.0, 125.9, 121.3, 117.7 (two signals obscured by solvent peaks); N C H 2 C 6 H 5 , 63.6; OCH(CH3)2, 67-70 (4 s); P C H 2 C H 2 P , 31.5-33 (m); OCH(CH3)2, 24.6-26 (8 s). 3lP{lH} NMR (C7D8, ppm): PC^N = A ; PtransN = B ; Rh = X : 8A = 73.9; 5B = 105.5. J A A ' = 44.8 Hz; J A B = 39.9 Hz; J A B 1 = 4.4 Hz; J A x = 191.3; J A X ' = 265.1 Hz; J B B ' = -3.8 Hz; J B x = 223.3 Hz; J B x ' = -2.4 Hz; J X x ' 3 H' 183 = 3.2 Hz. Anal. Calcd. for QiH/vC^Rl^N: C, 47.27; H, 7.45; N, 1.34%. Found: C, 47.52; H, 7.70; N, 1.40%. Synthesis of [(dipope)Rh]2[MCH3)N(CH2C6H5)](u.-H), 19d. To a stirred solution of [(dipope)Rh]2(p:-H)2, Id, (0.065 g, 0.08 mmol) in toluene (10 mL) was added neat N-benzylidenemethylamine (0.027 g, 0.24 mmol), in one portion. Stirring was continued for 10 minutes during which time a red to yellow-brown color change occurred. The toluene was then removed under reduced pressure and the oily residue remaining was redissolved in hexanes (0.5-1.0 mL). Upon cooling, a crop of oily yellow crystals was isolated in 61% yield (0.045 g). *H NMR (QD6, ppm): Eortho, 8.36 (d, 2H, ^Kmeta = 7-4 Hz); Umeta, 7-25 (m, 2H); Upara, 7.15 (m, 1H); N C H 2 C 6 H 5 , 5.04 (br t, 2H, 4 J P = 4.8 Hz); OCH(CH3)2, 5.12, 4.97, 4.84, 4.66 (dsp, 8H, 3 J H = 6.2-6.6 Hz, 3 J P = 2.0-4.0 Hz); NCH3, 3.99 (br t, 3H, 4 J P » 2-3 Hz); PCH2CH2P, 1.6-1.9 (m, 8H); OCH(CH3)2, 1.51, 1.50, 1.48, 1.36, 1.34, 1.32, 1.32, 1.22 (d, 48H); Rh-H-Rh, -7.33 (ttt, 2lptrans = 75.8 Hz, 2 j P c / 5 = 4.9 Hz, ijRh = 22.2 Hz). 3c{lH} NMR (C6D6, ppm): C i p s o , 143.4; other aromatic carbons, 129.9, 126.4, 125.9; NCH2, 72.7; N£H 3, 56.5 (t, J = 2.8 Hz); OCH(CH3)2, 69.7, 69.1(2), 68.7; PCH2CH2P, 32.0-32.8 (m); OCH(CH3)2, 24.9-25.8 (8 overlapping singlets). 31P{1H} NMR (C6D6, PPm): Ptrans(N), 202.2 (dm, ^ R h = 204 Hz); ?cis(Nh 177.8 (dm, ijRh » 272 Hz). Anal. Calcd. for C36H7408P4Rh2N: C, 44.18; H, 7.62; N, 1.43%. Found: C, 43.83; H, 7.93; N, 1.48%. Synthesis of [(dipope)Rh]2[p.-N(CH2C6H5)2](p:-H), 20d. To a stirred solution of [(dipope)Rh]2(p>H)2, Id, (0.038 g, 0.04 mmol) in toluene was added N-benzylidenebenzylamine (0.026 g, 0.13 mmol), in one portion. The reaction was stirred for two hours, during which a slow red to yellow-brown color change occurred. The toluene was completely remove in vacuo and the oily residue redissolved in hexanes (0.5 184 mL). Upon cooling, a crop of oily yellow crystals were obtained in 63% yield (0.029 g). *H NMR (C6D6, ppm): H^fo, 8.27 (d, 4H, ^ m e t a = 7.6 Hz); Umeta, 7.05 (m, 4H,); Upara, 7.00 (m, 2H); N C H 2 C 6 H 5 , 5.26 (br t, 4H, *J P = 4.8 Hz); OCH(CH3)2, 4.75, 4.28 (dsp, 8H, 3 j H = 6.2-6.6 Hz, 2j p = 2.0-3.0 Hz); PCH2CH2P, 1.65 (dm, 4H, J P » 30 Hz, J H - 8 Hz); PCH2CH2P, 1.58 (dm, 4H, J P - 30 Hz); OCH(CH3)2, 1.38, 1.27, 1.09, 0.98 (d, 48H); Rh-H-Rh, -7.44 (ttt, 2J?trans = 78.4 Hz, 2 j P r f j = 3.8 Hz, ijRh = 21.8 Hz). ™C{ lH} NMR (C6D6, ppm): Ci p s o , 144.2; other aromatic carbons, 129.5, 125.7, 125.3; NCH2, 71.6; OCH(CH3)2, 68.6, 68.5; PCH2CH2P, 31.2-32.6 (m); OCH(CH3)2, 25.3, 25.2, 24.7, 24.6. 31P{1H} NMR (C6D6, ppm): Ptrans(N), 200.3 (dm, l J R h ~ 204 Hz); Pcis(N), 176.7 (dm, l J R h - 268 Hz). Anal. Calcd. for C42H7908P4Rh2N: C, 47.78; H, 7.54; N, 1.33%. Found: C, 48.00; H, 7.53; N, 1.40%. Synthesis of [(dipope)Rh]2[p>HNCH(C6H5)2](p>H), 21d. To a stirred solution of [(dipope)Rh]2(p>H)2, Id, (0.045 g, 0.05 mmol) in toluene (10 mL) was added neat benzophenone imine (0.028 g, 0.15 mmol), in one portion. Stirring was continued for 5 minutes, during which time a red to yellow-orange color change occurred. After roto-evaporation of the toluene, the oily residue was redissolved in hexanes (0.5 mL). Upon cooling, a crop of yellow crystals was isolated yielding 0.025 g (46%) after washing with cold hexanes. *H NMR (C6D6, ppm): Hortho* 7.95 (d, 4H, ^ u m e t a = 7.4 Hz); Hmeta, 7.24 (m, 4H); Hp f lra, 7.11 (m, 2H); NCH(C6H5)2, 6.70 (br t, 4j P = 8.8 Hz); OCH(CH3)2, 5.06, 4.90, 4.60, 4.59 (dsp, 8H, 3 j H = 6.2-6.6 Hz, 2j p = 2.0-3.0 Hz); NH, 4.07 (br t, 1H, 3j P « 4-5 Hz); PCH2CH2P, 1.88, 1.60 (dm, 8H); OCH(CH3)2, 1.42, 1.37, 1.35, 1.26, 1.23, 1.21, 1.16 (d, 48H); Rh-H-Rh, -7.53 (ttt, 2Jptrans = 76.2 Hz, 2j P c i, = 4.0 Hz, !j Rh = 21.2 Hz). ^COH} NMR (C^D^, ppm): Ci p s o , 148.1; other aromatic carbons, 128.9, 127.6, 126.2; NCH(C6H5)2, 72.1; CH(CH3)2, 68.5-70.7; CH 2 CH 2 , 32.1 (m); OCH(CH3)2, 24.9-25.5 (overlapping singlets). 31p{lH} NMR (C6D6, ppm): Ptrans(N), 185 197.4 (dm, l j R h « 205 Hz); Pcis(N), 177.5 (dm, l J R h » 260 Hz). Anal. Calcd. for C41H77C-8P4RI12N: C, 46.27; H, 7.45; N, 1.34%. Found: C, 46.40; H, 7.53; N, 1.44%. Synthesis of [(dipope)Rh]2[MC9H8N)j(n-H), 24d. To a stirred solution of [(dipope)Rh]2(p>H)2, Id, (0.081 g, 0.09 mmol) was added a solution of isoquinoline (0.036 g, 0.30 mmol) in hexanes (5 mL), in one portion. After stirring for a further 30 minutes, a red to yellow color change was complete; concentration of the reaction mixture to ~1 mL in volume, followed by cooling led to a crop of yellow crystals (0.082 g, 88%). !H NMR (C 6 DI 2 , ppm): C-H4 or C-H5, 6.85 (m, 1H); C-H4 or C-H5, 6.76 (m, 1H); C-Hi, C-H3,and C-H6, 6.56 (m, 3H); C-H2, 5.0 (d, 1H, 3j H l = 7.4 Hz); NCH2, 4.75 (br t, 2H, 4 J P » 4 Hz); OCH(CH3)2, 4.95, 4.89, 4.81, 4.62 (overlapping dsp, 8H, 3 j H = 6.4-6.6 Hz, 2 J P = 2.0-3.0 Hz); PCH2CH2P, 1.35-1.65 (m, 8H); OCH(CH3)2, 1.23-1.32 (5 overlapping d, 30H); OCH(CH3)2, (3 overlapping d, 18H); Rh-H-Rh, -8.10 (ttt, 2]j>trans = 76.0 Hz, 2 J F c i s = 5.4 Hz, !jRh = 23.3 Hz). ^CpH} NMR (C6Di2, ppm): £-H, 143.2, 120.6, 119.2, 118.3, 116.4, 93.7 (the two quaternary bridgehead carbons were not observed); N£H2, 57.2; OCH(CH3)2, 64.8, 64.1, 63.9, 63.3 (s); PCH2CH2P, 24.8-28; OCH(CH3)2, 19.0-21.5 (overlapping resonances). 31p{lH} NMR (C6Di2, ppm): Ptrans(N), 201.0 (dm, lJ R h » 211 Hz); Pcis(N), 177.2 (dm, lJ R h » 252 Hz). 4 H 186 Synthesis of [(dippp)Rh]2[n-HNCH(C6H5)2](P--H), 21b. To a solution of [(dippp)Rh]2(M--H)2, lb, (0.062 g, 0.08 mmol) in toluene (10 mL) was added benzophenone imine (0.150 g, =10 equivalents), in one portion. The reaction was stirred at 50°C for 16 hours, which brought about a green to red-orange color change. The toluene was removed in vacuo and the residues recrystallized from toluene/hexanes (1:2). The crystals were washed twice with cold hexanes (1 mL) to remove excess imine. Yield: 0.059 g (77%). 1H NMR (C6D6, ppm): Eortho, 7.83 (d, 4H, ^Rmeta = 7.3 Hz); Emeta, 7.15 (m, 4H); H^ a r a , 7.06 (m, 2H); NCH(C6H5)2, 6.05 (br t, 1H, 4 J P « 6 Hz); NH, 3.87 (br s); CH(CH3)2, 2.31, 1.6-1.8 (overlapping dsp, 8H, 3jH = 7.6-8.0 Hz, 2 J P = 3.0-4.0 Hz); P C H 2 C H 2 C H 2 P , CH(CH3)2, 0.8-1.4 (overlapping multiplets, 60H); Rh-H-Rh, -10.00 (ttt, 2Jptrans = 57.9 Hz, 2 j P d j = 11.6 Hz, !jRh = 22.0 Hz). 3lp{lH} NMR (C6D6, ppm): Ptrans(N), 41.4 (dm, l J R n - 161 Hz); Pcis(N), 27.2 (dm, l J R h - 190 Hz). Anal. Calcd. for C43H8iP4Rh2N: C, 54.84; H, 8.67; N, 1.49%. Found: C, 54.59; H, 8.72; N, 1.35%. 4.6.3 Kinetic and Equilibrium Measurements. Equilibrium Measurements for the Equilibrium: la + 23 <==> 24a. Sample Preparation: [(dippe)Rh]2[p>(C9H8N)](p.-H) (0.120 g, 0.139 mmol) was dissolved in d/2-cyclohexane (0.30 mL). The resulting deep red solution was distributed evenly amongst three sealable 5 mm NMR tubes. Similarly, a solution of isoquinoline (0.020 g, 0.16 mmol) in d&-toluene (0.30 mL) was added in even portions to each of the three NMR tubes. To the first tube was added a further 0.40 mL of d/2-cyclohexane (C6D12 : C7D8 = 5:1); the second sample was augmented with 0.20 mL of ^ -cyclohexane and 0.20 mL of dg-toluene (C6D12 : C7D8 = 1:1), while the third sample was made up to volume with 0.40 mL of fife-toluene (C6D12: C7D8 = 1:5). The tubes were fitted with needle valves, attached to a vacuum line, degassed, and sealed under about 0.9 atmospheres of argon at -78°C. Each 187 sample contained 4.65 x 10"5 moles of [(dippe)Rh]2[|i-(C9H8N)](|j.-H) and 5.2 x 10"5 moles of isoquinoline in a total sample volume of 0.75 mL. Equilibrium Measurements: Measurement of the equilibrium constant at various temperatures was carried out by integration of the appropriate peaks in the 31P{ lH} NMR spectra. The samples were placed in a thermostatted probe and allowed to equilibrate at each temperature for 10 minutes before pulsing began. The 31P{1H} spectra were collected at constant temperature by using a 73° pulse, a 4 second relaxation delay, and broad band decoupling. The ratio of phosphorus containing constituents was determined via integration of the appropriate peaks; no correction for nOe effect were made. The possibility of intensity anomalies due to relaxation time differences was ruled out when a separate experiment yielded Ti values of between 1.5-2.5 seconds for each phosphorus nucleus in the spectrum. Equilibrium Measurements for the Equilibrium: Id + 23 <==> 24d. Sample Preparation: [(dipope)Rh]2[|i-(C9HgN)](n-H) (0.036 g, 0.036 mmol) was dissolved in pure n-hexane (0.30 mL) and the solution loaded into a sealable 5 mm NMR tube. A solution of isoquinoline (0.006 g, 0.05 mmol) in dg-toluene (0.30 mL) was added to the tube and a 180° needle valve attached. The assemblage was connected to a vacuum line and the sample degassed via 2 freeze-pump-thaw cycles. The tube was sealed under =0.9 atmospheres of argon. Equilibrium measurements were carried out in an identical procedure to that described above for the dippe system. Kinetic Measurements: [(dippe)Rh]2(|i-H)2 + N-Benzylideneaniline. Kinetic measurements of the reaction of [(dippe)Rh]2(|i-H)2, la, with N-benzylideneaniline were made by following the reaction by 31P{ lH} NMR. Samples were prepared by loading a sealable 5 mm NMR tube with 0.25 mL of a stock solution of la in dg-toluene (0.494 g in 5.00 mL C7Dg, 1.35 x 10'1 M) and a further 0.25 mL of ^ -toluene. The 188 tube was then fitted with a 45° needle valve and attached to a vacuum line. The sample was degassed via 2 freeze-pump-thaw cycles and cooled to -78°C. N-benzylideneaniline (0.10 mL of a stock solution containing 1.257 g in 5.00 mL C7D8, 1.39 M) was added through the needle valve via syringe under a strong flow of argon. The tube was then sealed without delay under about 0.85 atmospheres of argon and placed in a thermostatted probe set at either 204.0, 209.0, 214.0, or 218.0 K. While thermally equilibrating, a spectrum was collected; if observable amounts of the final amido hydride product 18a were present, the run was aborted. After an equilibration period of 10 minutes, spectra were collected at convenient intervals. Time points were taken as the mid-point of FID collection periods and calculated based on the number of scans collected, time at the start of data collection, and an acquisition time of 0.4096 seconds. The ratio of components in solution was determined from the ratios of appropriate integrals; anomalies due to differences in Ti values were ruled out when a separate experiment revealed similar Ti values for all phosphorus containing components in solution. When the first step in the reaction was complete, the probe was warmed rapidly to either 244.0, 248.5, 254.0, or 258.0 K. After a 10 minute equilibration period, the second step was followed in a manner analogous to that described above for the first step. Kinetic Measurements: rf2-[(dippe)Rh]2(|i-D)2 + N-Benzylideneaniline. This experiment was carried out in an identical procedure to that described above. Samples were prepared by dissolving 6?2_la (0.026 g, 0.035 mmol) in dg-toluene (0.4 mL) and loading into a sealable 5 mm NMR tube equipped with a 180° needle valve. After attachment to a vacuum line and degassing, a solution of N-benzylideneaniline (0.025 g, 0.14 mmol) in cfg-toluene (0.2 mL) was added via syringe under a strong flow of argon. The tube was sealed under 0.9 atmospheres of argon and the reaction's two steps were followed at 209 K and 254 K, respectively. 189 4.6.4 H/D Exchange Reactions. H/D Exchange Between C 7 D 8 and [(dippe)Rh]2[MC6H5)N(CH2C6H5)](|i-H). Method A. Aliquots (=2.5 mL) of a solution of [(dippe)Rh]2[p> (C6H5)N(CH2C6H5)]0J.-H) (=0.15 g) in cfo-toluene (=25 mL) were loaded into small reactor bombs and degassed on a vacuum line with two freeze-pump-thaw cycles. The solutions were then heated at 122-126° C for varying amounts of time. The deuterated solvent was vacuum transferred away from the solute and the 2 H NMR spectrum obtained. Method B. The exchange process was also monitored via *H NMR spectroscopy. A sample of [(dippe)Rh]2[p-(C6H5)N(CH2C6H5)](|i-H).(=0.03 g) was dissolved in dg-toluene (=0.7 mL) and loaded into a sealable 5 mm NMR tube. The sample was degassed and the tube sealed under about 0.9 atmospheres of argon. The sample was then heated at 116-119°C for about 3 days and the *H NMR spectrum taken at regular intervals. Intermolecular H/D Exchange in rf2-[(dippe)Rh]2[p>(C6H5)N(CHDC6H5)](p> D). A sample of d2-18a (=0.03 g) was dissolved in toluene, and the solution loaded into a sealable 5 mm NMR tube. After degassing the sample, the tube was sealed under about 0.9 atmospheres of argon. The sample was heated at 120°C for 12 hours, with periodic monitoring by 2 H NMR. H/D Exchange Between [(dippe)Rh]2[MC6H5)N(CH2C6H5)](p>H) a n d D 2 A sample of 18a (=0.03 g) was dissolved in toluene and the solution loaded into a 5 mm NMR tube. The sample was degassed and the tube sealed under 4 atmospheres of deuterium gas. The solution was heated at 50°C and monitored periodically by 2 H NMR. 190 4.6.5 Reactions of [(dippe)Rh]2(p>H)2, la, with the Iminium Salts 30 and Synthesis of {[(dippe)Rh]2[MC6H5)N=CH(C6H5)](p-H)}BF4, 32. The dihydride [(dippe)Rh]2(M--H)2, la, (0.200 g, 0.27 mmol) and the iminium salt generated from N-benzylideneaniline and HBF4 (0.073 g, 0.27 mmol) were loaded into a small reactor bomb equipped with a magnetic stir bar. The bomb was attached to a vacuum line, evacuated, and cooled to -78"C. THF (=10 mL) was vacuum transferred into the vessel. An immediate orange-brown colour obtained, different from the characteristic deep green of la; while warming to room temperature the solution became a deep red-purple in colour. Although gas evolution was not visually observed, a slight build up of pressure in the reaction vessel was apparent. The THF was removed in vacuo once the reaction had warmed to room temperature and the residues recrystallized from minimum THF/toluene. *H NMR ((CD3)2CO, ppm): C-Hi, 8.69 (d, 2H, 3 J H 2 = 8.0 Hz); C-H4 or C-H5, 7.94 (br s, 1H); CH=N, 7.78 (br t, 1H, J P = 1.5 Hz); C-H3, 7.40 (t, 1H, 3 J H 2 = 7.4 Hz); C-H4 or C-H5, 7.34 (br s, 1H); C-H6, C-H7, and C-Hg, 7.23 (br m, 3H); C-H2, 7.14 (m, 2H); CH(CH3)2 and P C H 2 C H 2 P , 1-8-2.5 (overlapping broad m, 16H); CHrCH^, 1.41, 1.29, 1.07 (3), 31. Ph R = Pri R' = Ph, 32 R' = CH3, 33 191 8 3 0.75 (3) (overlapping dd, 48H, 3 J H = 6.4-7.2 Hz, 2 J P = 11.6-16.4 Hz); Rh-H-Rh, -9.01 (ttt, 2 l p t r a n s = 50.1 Hz, 2 J P c i s = 13.6 Hz, URh = 23.6 Hz). 31P{!H} NMR ((CD3)2CO, ppm, +30°C): Ptrans(N), 107.7 (br dm, lJRh - 163 Hz); Pcis(N), 91.3 (br dm, l j R h » 157 Hz). (-45°C): Ptrans(N), 110.0 (dd, URK = 166.3 Hz, 2 J P = 31.2 Hz); P , ™ ^ , 106.4 (dd, llRh = 159.8 Hz, 2JP = 29.6 Hz); P c i s ( N ) , 92.8 (dm, l j R h = 182.5 Hz); P c i s ( N ) , 89.7 (dm, l j R h = 140.9 Hz). ^CfiH} NMR ((CD3)2CO, ppm): CH=N, 93.6 (s, = 168.4 Hz). Anal. Calcd. for C4iH76P4Rh2NBF4: C, 49.26; H, 7.66; N, 1.40%. Found: C, 49.46; H, 7.85; N, 1.30%. Synthesis of {[(dippe)Rh]2[MCH3)N=CH(C6H5)](M>H)}BF4, 3 3 -The dippe dihydride, la, (0.153 g, 0.21 mmol) and the tetrafluoroborate iminium salt of 19b (0.043 g, 0.21 mmol) were loaded into a small reactor bomb equipped with a magnetic stir bar. The vessel was evacuated and cooled to -78°C. Pure THF (10 mL) was vacuum transferred into the vessel. Although reaction was immediate, as evidenced by a green to red-brown color change, the mixture was stirred at -78°C until all the solid was dissolved prior to warming to room temperature (30 minutes). After warming, the THF was removed in vacuo and the remaining oily dark red solid was recrystallized from THF/toluene, 192 yielding 0.150 g (77%) of red purple crystals. lH NMR ((CD3)2CO, ppm): Eortho* 8.41 (d, 2 H , SjHmeta = 7 - 4 H z ) 5 CH=N> 7 - 7 0 ( * • 1H> *P ~ 2 -° Hz)^ *Wa, 7.34 (t, 1H ^ R m e t a = 7.2 Hz); Hmeta> 7.09 (m, 2H) ; N C H 3 , 3.65 (s, 3H) ; dippe ligand resonances are broad, CH(CH3)2, 2.4, 1.9 (8H); P C H J C H J P , 2.2, 1.8 (8H); CH(CH3)2, 0.9-1.4 (48H); Rh-H-Rh, -9.45 (m). 3lp{lH} NMR (THF, ppm, +30°C): Ptrans(N), 107.1 (br dm, l J R h » 163 Hz, 2jp » 29 Hz); Pcis(N), 92.4 (br dm, lJ R h - 169 Hz). (-50°C): PtransfN), 109.4 (dd, lJ R h = 159.6 Hz, 2 J P = 29.1 Hz); Ptrans(N), 105.5 (br dd, l J R h = 166.2 Hz, 2 J P = 26.6 Hz); Pcis(N), 95.3 (dm, l J R h = 183.0 Hz); Pcis(N), 90.1 (dm, l J R h = 143.9 Hz). 13 C{1H} NMR ((CD3)2CO, ppm): C i p s o , 139.6; other aromatic carbons, 129.5, 128.9, 128.1; CH=N, 102.8; NCH3, 57.9; ligand resonances, 18-30. The Reaction of {[(dippe)Rh]2[p-(C6H5)N = CH(C6H5)](p..H)}BF4, 32, and LiAlH(OBut)3. The pv-imine hydride complex, 32, (0.052 g, 0.052 mmol) was dissolved in THF (10 mL). To this stirred solution was added solid LiAlH(OBut)3 (0.015 g, 0.057 mmol) in small portions over a period of 5 minutes. The solution slowly turned from deep red to a green-orange colour. The THF was removed under reduced pressure and the residues extracted with hexanes (3x10 mL portions). The extracts were filtered through a pad of Celite and concentrated to =3 mL. Cooling led to the precipitation of an orange powder consisting of >95% pure [(dippe)Rh]2[u.-(C6H5)N(CH2C6H5)](p.-H), 18a, (0.37 g, 78%). This reaction was also carried out in an NMR tube at low temperature. Complex 32 (0.046 g, 0.046 mmol) was dissolved in THF and loaded into a sealable 5 mm NMR tube equipped with a 180° needle valve. The assemblage was connected to a vacuum line and degassed; the sample was then cooled to -78°C and a THF solution containing two equivalents of LiAlH(OBul)3 was syringed into the tube under a strong flow of argon. The tube was sealed immediately and the reaction monitored by 31p{ lH} NMR spectroscopy. 193 4.6.6 Hydrogenation Procedures for the Attempted Hydrogenation of N-Benzylideneaniline. The Use of [(dippe)Rh]2(|i-H)2, la, as Catalyst Precursor. The dihydride [(dippe)Rh]2(|i-H)2, la, (0.020 g, 0.027 mmol) and N-benzylideneaniline (0.098 g, 20 equivalents) were dissolved in THF (5.0 mL) and the solution loaded into a small reactor bomb. The solution was degassed on a vacuum line, placed under 4 atmospheres of dihydrogen and stirred at room temperature for 14 hours. The solvent and excess dihydrogen were removed in vacuo and the residues analyzed via *H NMR spectroscopy. Only traces of N-phenylbenzylamine were detected, corresponding to <3% hydrogenation. The Use of [(dippe)Rh]2(H)4, 5a, as Catalyst Precursor. The dihydride [(dippe)Rh]2(p-H)2, la, (0.020 g, 0.027 mmol) was dissolved in THF (4.0 mL) and the solution loaded into a small reactor bomb. The solution was degassed and placed under one atmosphere of dihydrogen, converting the dihydride into the tetrahydride 5a instantly. Under a strong flow of dihydrogen, a solution of N-benzylideneaniline (0.098 g, 20 equivalents) in THF (1.0 mL) was syringed into the reaction vessel. The solution was stirred for 12 hours and worked up in an analogous procedure to that described above. *H NMR spectroscopy of the residues indicated =35% conversion of the imine to the amine, corresponding to ~1 turnovers. Again, a similar procedure using [(dipope)Rh]2(p>H)2, Id, as catalyst precursor was performed; no hydrogenation was detected. 194 4.6.7 Miscellaneous Procedures and Reactions. The Reaction of [(dippe)Rh]2[MC6H5)N(CH2C6H5)](li-H), 18a, with Dihydrogen. The amido hydride 18a (0.032 g; 0.035 mmol) was dissolved in dg-toluene (=0.5 mL) and the solution loaded into a sealable 5 mm NMR tube fitted with an 180° needle valve. The assemblage was attached to a vacuum line with access to pure dihydrogen and the sample degassed with two freeze-pump-thaw cycles. The sample tube was cooled to -196°C and dihydrogen admitted to =0.9 atmospheres; the tube was sealed, thawed, and monitored periodically via *H NMR spectroscopy. Signals due to N-phenylbenzylamine slowly appeared and were identified through comparison with a spectrum of an authentic sample (Aldrich). After =3 days, the peaks due to 18a had disappeared, leaving only those of the free amine and the tetrahydride complex 5a. Spectroscopic Monitoring of Reactions between la and Imines. The following method was used to monitor the reactions of the dihydrides with various imines at low temperatures. A solution of la or Id was loaded into a sealable 5 mm NMR tubes equipped with a 180° needle valve. The assemblage was attached to a vacuum line, cooled to -78°C and degassed. A solution of the imine was then added via syringe through the needle valve port under a strong flow of argon. The tube was sealed under a partial vacuum (=0.9 atmosphere of argon) and, without agitating, taken to the NMR spectrometer. The sample was shaken vigorously and transferred into the pre-cooled probe as quickly as possible (< 5 seconds). After suitable equilibration periods, depending on the nature of the experiment, spectra were collected. 195 Attempted "Trapping" of the Intermediate 26. A toluene solution of N-benzylideneaniline (2 equivalents, 1 mL) was added via syringe to a solution of la (=0.050 g) in toluene (5 mL) cooled to -78°C. The reaction was stirred for 2 hours at temperatures never exceeding -60°C to ensure complete formation of the brown-green intermediate. At this point, the trapping agents, CCI4 (1 mL) or ethylene (1 atmosphere), were introduced into the reaction vessel. After stirring for a further 30 minutes at low temperature in the presence of these reagents, the solutions were allowed to slowly warm to room temperature. The reaction containing the CCI4 turned a black-brown colour upon warming, with significant deposits of metal observed. The solution to which ethylene had been admitted was worked up by removal of the solvent and excess ethylene in vacuo followed by recrystallization of the residues from toluene/hexanes. 31P{1H} NMR spectroscopy revealed the presence of only 18a and the p-vinyl hydride complex [(dippe)Rh]2(p-ri2-a-CH=CH2)(p-H). 4.7 References. 1. a) Lappert, M. F.; Power, P. P.; Sanger, A. P.; Srivastava, R. C. Metal and Metalloid Amides. Wiley: New York, 1979. b) ibid., p 468. 2. Fryzuk, M. D.; Montgomery, C. D. Coord. Chem. Rev. in press. 3. Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles and Applications of Organotransition Metal Chemistry, 2nd ed. University Science Books: Mill Valley, Ca., 1987, p 64. The chemistry of a bonded diimines is well developed. 4. Van Koten, G.; Vrieze, K. Adv. Organomet. Chem. 1982,27, 151. 5. Fryzuk, M. D.; Piers, W. E. Organometallics 1988, 7, 2062. 6. McKenna, S. T.; Andersen, R. A.; Muetterties, E. L. Organometallics 1986,5, 2233. 7. Fryzuk, M. D.; Piers, W. E. Polyhedron 1988, 7, 1001. 196 8. a) Parshall, G. W. Acc Chem. Res. 1970, 3, 139. b) Parshall, G. W. Acc. Chem. Res. 1975, 5, 113 and references therein. 9. a) Masters, C. J. / . Chem. Soc. Chem. Commun. 1973, 191. Recent examples: b) Zeilher, E. H. K.; DeWitt, D. G.; Caulton, K. G. / . Am. Chem. Soc. 1984,106, 7006. c) Chaudret, B. / . Organomet. Chem. 1984,265, C33. 10. Streirwieser, A.; Heathcock, C. H. Introduction to Organic Chemistry. Macmillan: New York, 1976. 11. Fryzuk, M. D.; Einstein, F. W. B.; Jones, T. Organometallics 1984,3, 185. 12. a) Burch, R. R.; Muetterties, E. L.; Shultz, A. J.; Gebert, E. G.; Williams, J. M. / . Am. Chem. Soc. 1981,103, 5517. b) Burch, R. R.; Shusterman, A. J.; Muetterties, E. L.; Teller, R. G.; Williams, J. M. /. Am. Chem. Soc. 1983,105, 3546. c) McKenna, S. T.; Muetterties, E. L. Inorg. Chem. 1987,26, 1296. 13. Doherty, N. M.; Bercaw, J. E. /. Am. Chem. Soc. 1985,107, 2670. 14. a) Longley, C. J.; Goodwin, T.; Wilkinson, G. Polyhedron 1986,5, 1625. b) Grigg, R.; Mitchell, T. R. B.; Tongpenyai, N. Synthesis 1981, 442. 15. a) Shapley, J. R.; Richter, S. I.; Tachikawa, M.; Keister, J. R. / . Organomet. Chem. 1975, 94, C43-C46. b) Gerlach, R. F.; Duffy, D. N.; Curtis, M. D. Organometallics 1983, 2, 1172. c) Iggo, J. A.; Mays, M. J.; Raithby, P. R.; Hendrick, K. J. Chem. Soc, Dalton Trans. 1983, 205. d) King, R. B.; Treichel, P. M.; Stone, F. G. A. /. Am. Chem. Soc. 1961, 83, 3600. e) Al-Obaidi, Y. N.; Baker, P. K.; Green, M.; White, N. D.; Taylor, G. E. / . Chem. Soc, Dalton Trans. 1981, 2321. f) Caddy, P.; Green, M.; Smart, L. E.; White, N. /. Chem. Soc, Chem. Commun. 1978, 839. g) Keister, J. R.; Shapley, J. R. / . Organomet. Chem. 1975, 55, C29. h) Deeming, A. J.; Hasso, S.; Underhill, M. / . Organomet. Chem. 1974, 80, C53. i) Guy, J. J.; Reichert, B. E.; Sheldrick, G. M. Acta. Cryst. 1976, B32, 3319. j) Clauss, A. D.; Tachikawa, M.; Shapley, J. R.; Pierpont, C. G. Inorg. Chem. 1981,20, 1528. 197 16. a) Vastag, S.; Bakos, J.; Toros, S.; Takach, N. E.; King, R. B.; Heil, B.; Marko, L. / . Mol. Catal. 1984, 22, 283. b) Vastag, S.; Heil, B.; Toros, S.; Marko, L. Transition Met. Chem. 1977,2, 58. c) Levi, A.; Modena, G.; Scorrano, G. / . Chem. Soc. Chem. Commun. 1975, 6. 17. a) Radhi, M. A.; Palyi, G.; Marko, LJ. Mol. Catal. 1983,22, 195. b) Radhi, M. A.; Marko, L. /. Organomet. Chem. 1984,262, 359. 18. Baranyai, A.; Ungvary, R.; Marko, L. /. Mol. Catal. 1985,32, 343. 19. Palagyi, J.; Nagy-Magos, Z.; Marko, L. Transition Met. Chem. 1985,10, 336. 20. a) Yoshida, T.; Harada, K. Bull. Chem. Soc. Jpn. 1972, 45, 3706. b) Roe, A.; Montgomery, J. A. / . Am. Chem. Soc. 1953, 75, 910. 21. "Organic Syntheses" Collective Vol 1, 2nd Ed. Gilman, H.; Blatt, A. H., Eds. Wiley: New York, 1941, 80. 198 CHAPTER 5 The Reactions of Binuclear Rhodium Hydrides with Nitriles: The Stepwise Reduction of C=N at Two Metal Centres 5.1 Introduction Several examples of transition metal complexes containing the alkylideneimido ligand, •N=CRR', are known, but this ligand is far less common than the amide ligands discussed in the previous chapter. Examples of compounds in which the alkylideneimido ligand binds to one,1-6 two,7-12 and three12-16 metal centres are documented. In binding to one metal centre, this ligand may act as either a two or a four electron donor, as shown in I and II in Figure 5-1. In type I (bent) the nitrogen is sp2 hybridized, whereas in the linear type n , in which the lone pair is also donated to the metal centre, it is sp hybridized. The available lone pair in type I allows the binding of another metal centre, creating a bridging mode of bonding, type III. Figure 5-1. Bonding modes for alkylideneimido ligands to one (I, II), two (HI) and three (TV) metal centres. 199 If a third metal centre is proximal, 7C-donation from the C=N bond of the p^ -alkylideneimido ligand may occur, producing the p 3 bonding mode depicted in IV. Most of the alkylideneimido complexes known were synthesized via one of two general routes. Reaction of anionic _N=CRR' with a metal halide produced many mononuclear alkylideneimido complexes of molybdenum and tungsten.1-8 Cleavage of the N-N bond in RR'C=N-N=CRR' upon reaction with metal carbonyls yielded several binuclear complexes with p.2-alkylideneimido ligands, i. e., type III.7 A relatively small number of examples exist in which the imido ligand originates from a nitrile reagent. Insertions of nitriles into the Zr-H bond of Schwartz's reagent16 (Cp2ZrClH) yield alkylideneimido complexes with the bent type I bonding.17 The binuclear rhenium dihydride Re2H2(dppm)(CO)6 reacts slowly with acetonitrile at reflux to produce the | i 2 -ethylideneimido complex HRe2[p>N=C(H)CH3](dppm)(CO)6.10 Reaction of Os2(H-H)2(CO)io with the activated nitrile C F 3 C N affords the complex HG*S3[u,-N=C(H)CF3](CO)io.n'12 Finally, the trinuclear complex HFe3[p:-N=C(H)CH3](CO)9, containing a p.3-ethylideneimido ligand was produced as one of the products of a complicated reaction involving Fe2(CO)82", WCCGOsI1", and refluxing, moist acetonitrile.13 In this chapter, the reactions of the dihydrides la and Id with nitriles to form binuclear p:2-alkylideneimido complexes are described. Along with a description of the synthesis and properties of these complexes, a brief description of the reactivity of select members of the family with dihydrogen is also presented. 200 5.2 The Synthesis and Properties of Binuclear Rhodium Alkylideneimido Hydrides. 5.2.1 Synthetic Procedures. The observed reactivity between the dihydrides [(dippe)Rh]2(|i-H)2, la, and [(dipope)Rh]2((X-H)2, Id, and carbon-nitrogen double bonds18 led us to explore the behaviour of these dimers towards the carbon-nitrogen triple bonds of simple nitrile substrates. The dihydrides were found to react smoothly with one equivalent of R'CN (R* = CH3, C6H5,0-CH3-C6H4) at room temperature to produce the |i-alkylideneimido hydride complexes 34-36a and 34-36d (equation 5-1). The reactions were spectroscopically clean, \ R h . . R h / (5-1) / X / \ CH3 CgHs 0-CH3-C6K4 Pr1, 34a 35a 36a OPr', 34d 35d 36d producing the products in >95% yield; isolated yields for 34-36a were 82-88%, whereas the more soluble compounds of the dipope series were isolated in 72-78% yield. Similar to the trends in the reactions with imines, the dippe dimer la reacted instantly with the nitriles employed, while the dipope dihydride required about 30 minutes to undergo the same transformations. The six membered chelate dippp dihydride lb undergoes a slow reaction with acetonitrile at elevated temperatures, but fails to react with either benzonitrile or o-tolylnitrile. The presence of the |i-ethylideneimido dippp product (34b) was noted spectroscopically, but ^ \ y \ / ^ Rh Rh ^ • P ^ H x p — + R — c = N toluene 25'C R = Pr'( la R = OPr', Id R' = R = R = 201 the prohibitively slow rate of this transformation discouraged complete characterization of this complex. All of the red or red-purple (i-alkylideneimido hydride complexes were found to be extremely air-sensitive. The lH, ^CpH}, and 31P{1H} NMR spectral data collected for these complexes were consistent with the proposed structure for these molecules; in addition, a single crystal X-ray analysis on 34a confirmed the structural assignment (vide infra). In the !H NMR spectra of these compounds, the signals of interest are those for the alkylideneimido proton and the bridging hydride ligand. The spectrum of 34a is given in Figure 5-2. Downfield of the usual jungle of signals associated with the dippe ligand, at 8.75 ppm, is a complex multiplet due to the ethylideneimido proton. Upon broadband 3 1P decoupling and homodecoupling of the methyl group (2.06 ppm) this multiplet collapses to a singlet, indicating no observable coupling to rhodium. The multiplicity arises from coupling to each of the four inequivalent phosphorus nuclei in the molecule. Couplings of 15.6, 8.8 and 4.8 Hz were observed for 34a, while, in addition to couplings of 20.5, 13.0 and 6.5 Hz, a fourth coupling of 1.8 Hz was observed for the same signal in the spectrum of 34d. The larger couplings are probably due to the phosphorus nuclei trans to the p.-alkylideneimido ligand, but precise assignment is speculative. The u,-hydrido ligand also gives rise to complex signals with couplings to four inequivalent phosphorus nuclei in evidence; one-bond rhodium coupling constants of =20-22 Hz are also observed. In the 13C{!H} NMR spectra, the p,-alkylideneimido carbon atom resonates at =159 ppm for the dippe compounds and =168 ppm in the complexes of the dipope d series. Little 1 3 C data has been reported for p> alkyhdeneimido complexes, but these chemical shifts do fall in the range in which the imine carbon in free imines usually resonate (=160 ppm).19 As expected, the 31P{ !H} spectra for these compounds are complex. Four phosphine environments and long range P-P and Rh-P couplings between opposite sides of the dimer combine to provide the spectrum of 34d, shown in Figure 5-3 along with a calculated spectrum. The now familiar large four-bond coupling between the two phosphorus nuclei cis Figure 5-2. 400 MHz *H NMR spectrum of the ethylideneimido hydride complex 34a. Upon broadband 3 1P decoupling, the hydride resonance collapses to a broad triplet (upper trace). 203 Figure 5-3. a) 121.4 MHz 3lp {lH} NMR spectrum of the ethylideneimido hydride 34d, incorporating the dipope ligand (scale is marked by 100 Hz gradations), b) Calculated spectrum based on an ABCDXY spin system (see experimental for parameters). 204 to the ethylideneimido ligand is again observed (4Jp.p = 68.3 Hz). Substantial three-bond couplings of 11.3 and 12.2 Hz from the cis phosphorus nuclei to the distal rhodiums are also present. The spectra of the dippe compounds 34-36a are broader with less well defined couplings. 5.2.2 The Solid State Structure of [(dippe)Rh]2[p>N=C(H)CH3](p>H), 34a. To firmly establish the bridging bonding mode of the |i-alkylideneimido ligands, a single crystal X-ray analysis of the dippe ethylideneimido hydride complex 34a was performed. ORTEP diagrams and a stereoview of the binuclear complex are shown in Figure 5-4a-b and 5-4c, respectively; Table 5-1 gives selected bond distances and angles. Each rhodium centre in the dimer is of a distorted square planar geometry. As seen from the view in Figure 5-4b, looking down the Rh-Rh vector, the two square planes are slightly bent with respect to each other. The angle between the mean square planes containing a rhodium atom and the four ligand atoms is 25.5°. The parameters associated with the chelating phosphine ligands are unremarkable; Rh-P bond lengths and the P-Rh-P angles of 86.44° and 85.89° are nearly identical to analogous values found in other structurally characterized complexes containing the dippe ligand.20'22 The angles P(2)-Rh(l)-Rh(2) and P(4)-Rh(2)-Rh(l) are 145.1° and 148.6° respectively and partially account for the large four-bond coupling between P(2) and P(4) (vide supra). The bridging hydride ligand was located and refined isotropically and was found to symmetrically bridge the two metal centres. The Rh-H bond lengths of 1.74(5)A are about 0.1 A shorter than those in other [(cbp,pe)Rh]2(|J.-X)(|±-H)20>21 complexes, which average =1.84A in length, but are comparable to the 1.70A distances found in the parent dihydride la 2 2 a The "tighter" molecular core is also reflected in a shorter than normal (»2.86A) Rh-Rh separation of 2.7992(4)A. Bond and angle parameters in the |i.-ethylideneimido ligand are typical of a (1-T)2 bonding mode with no re-donation from the C=N bond to one of the metal centres. The C=N 205 C(8) C ( 1 9 ) C(18) Figure 5-4. a) ORTEP drawing of the ethylideneimido hydride complex 34a. b) Another view looking dow the Rh(l)-Rh(2) vector, c) Stereoview of the molecule. 206 Table 5-1. Selected Bond Parameters for C3pH69P4Rh2N (34a). Distances (A) Rh(l)-P(l) 2.2186(10) P(2)-C(2) 1.853(4) Rh(l)-P(2) 2.2297(10) P(2)-C(7) 1.862(5) Rh(l)-N 2.030(3) P(2)-C(8) 1.858(5) Rh(l)-H(Rh) 1.74(5) C(5)-C(13) 1.517(7) Rh(2)-P(3) 2.2106(11) C(5)-C(14) 1.526(6) Rh(2)-P(4) 2.2327(10) C(6)-C(15) 1.513(7) Rh(2)-N 2.044(3) C(6)-C(16) 1.508(7) Rh(2)-H(Rh) 1.84(4) C(7)-C(17) 1.523(8) P(l)-C(l) 1.860(4) C(7)-C(18) 1.537(7) P(D-C(5) 1.858(4) C(8)-C(19) 1.511(8) P(l)-C(6) 1.856(4) C(8)-C(20) 1.507(7) N-C(29) 1.255(5) C(29)-C(30) 1.490(6) gles (deg) P(l)-Rh(l)-P(2) 86.96(4) P(3)-Rh(2)-P(4) 85.60(4) P(l)-Rh(l)-N 170.39(10) P(3)-Rh(2)-N 166.45(9) P(l)-Rh(l)-H(Rh) 86.8(14) P(3)-Rh(2)-H(Rh) 84(1) P(2)-Rh(l)-N 102.43(9) P(4)-Rh(2)-N 107.13(9) P(2)-Rh(l)-H(Rh) 170.1(14) P(4)-Rh(2)-H(Rh) 169(1) N-Rh(l)-H(Rh) 85.4(14) N-Rh(2)-H(Rh) 83(1) Rh(l)-N-Rh(2) 86.78(12) Rh(l)-H(Rh)-Rh(2) 103(2) Rh(l)-P(l)-C(l) 111.33(14) Rh(l)-P(2)-C(2) 110.9(1) Rh(l)-P(l)-C(5) 120.4(1) Rh(l)-P(2)-C(7) 121.6(2) Rh(l)-P(l)-C(6) 117.44(14) Rh(l)-P(2)-C(8) 116.1(2) C(l)-P(l)-C(5) 101.3(2) C(2)-P(2)-C(7) 101.5(2) C(l)-P(l)-C(6) 101.9(2) C(2)-P(2)-C(8) 103.2(2) C(5)-P(l)-C(6) 101.9(2) C(7)-P(2)-C(8) 101.1(2) P(l)-C(l)-C(2) 110.4(3) P(2)-C(2)-C(l) 112.0(3) P(l)-C(5)-C(13) 111.0(3) P(2)-C(7)-C(17) 111.2(3) P(l)-C(5)-C(14) 116.4(3) P(2)-C(7)-C(18) 115.4(4) C(13)-C(5)-C(14) 111.0(4) C(17)-C(7)-C(18) 110.0(5) P(l)-C(6)-C(15) 111.0(3) P(2)-C(8)-C(19) 111.1(4) P(l)-C(6)-C(16) 111.6(3) P(2)-C(8)-C(20) 111.4(4) C(15)-C(6)-C(16) 110.2(4) C(19)-C(8)-C(20) 109.5(5) Rh(l)-N-C(29) 128.8(3) Rh(2)-N-C(29) 141.3(3) N-C(29)-C(30) 128.3(4) 207 Table 5-II. Carbon-Nitrogen Bond Distances in |i-Alkylideneimido Complexes. Compound C=N Distance(s)a Reference Fe2(p:-N=CR2)2(CO)6 (R = p-tolyl) 1.24(2) 1.29(2) 7 Mn2(p:-N=CR2)2(CO)7 (R = CF 3) 1.258(9) 1.259(9) 8 HRu3(p:-N=CR2)2(CO)io (R = CH 3) 1.279(5) 1.280(6) 9 HOs3[p:-N=C(H)CF3](CO)9(L) (L = PMe2(C6H5)) 1.247(15) 11 HOs3[p.-N=C(H)CF3](CO)i0 1.271(8) 24 HRe2[H-N=C(H)CH3] (CO)6(dppm) 1.385(59) 10 HFe3[p>N=C(H)CH3](CO)9 1.321(5) 14 [(mppe)Rh]2[M>N=C(H)CH3](p:-H) (34a) 1.255(5) — Angstroms (A). bond length of 1.255(5)A is slightly shorter than an average carbon-nitrogen double bond (1.28A),23 but coincides well with other known C=N distances in complexes incorporating the |i-ri2-alkylideneimido ligand (Table 5-II). Interestingly, the C=N length of 1.385(59)A reported for the complex HRe2[M--N=C(H)CH3](dppm)(CO)6,10 was longer even than that of 1.321(5)A found for the p>T|3-N=C(H)CH3 ligand in HFe3[p>N=C(H)CH 3](CO) 9 1 4 in which the C=N bond 7t-donates electron density to a third metal centre. This apparently anomalous C=N length in the rhenium dimer was not commented on by the authors of the report. 208 The N-C(29)-C(30) bond angle of 128.3° suggests sp2 hybridization about C(29) (the other two angles, i. e., N-C(29)-H and C(30)-C(29)-H, are idealized since the ethylideneimido proton was not refined). This is also supported by large l T c - H coupling constants of «168 Hz in the alkylideneimido C-H bonds 2 5 The discrepancy between Rh(l)-N-C(29) (128.8°) and Rh(2)-N-C(29) (141.3°) indicates that the p.-ethylideneimido ligand is slightly tilted such that the methyl group, C(30), is moved away from the iso-propyl groups on P(4). Presumably, with the larger phenyl and o-tolyl substituents on C(29), this steric interaction would be exacerbated. It also provides a possible rationale for why the bulkier nitriles BuHDN and PriCN do not react with la or Id to produce the corresponding u.-alkyhdeneimido hydride complexes. 5.2.3 Fluxional Properties of the u,-AlkyIideneimido Hydride Complexes. The 31p{lH} NMR spectrum of [(dippe)Rh]2[u.-N=C(H)o-CH3-C6H4](p:-H), 36a, exhibits a peculiar temperature dependency, suggesting that at higher temperatures a fluxional process is operative. This is illustrated in Figure 5-5, which shows a series of 3lp{lH} NMR spectra of complex 36a. At lower temperatures (below 10°C), the chemical shifts of the signals for the four phosphorus nuclei are markedly temperature dependent. Particularly, the two downfield signals (each a doublet of doublets) drift further downfield as the temperature is raised from -50°C, until at 10°C they are nearly coincident and appear as one doublet of doublets pattern. Further warming of the sample to temperatures of 70°C or higher results in a steady broadening of all the signals in the spectrum. It appears as though the upfield set (two doublets of triplets) and the downfield set are coalescing as the temperature is raised, but the high temperature limit spectrum, in which a broad doublet in both the upfield and downfield region would be expected, was not reached. These spectra may be interpreted as indicating a fluxional process occurring at high temperatures which exchanges the two 209 Figure 5-5. 162.2 MHz 31p {lH} NMR spectra of the o-tolylideneimido hydride 36a recorded at various temperatures. 210 sides of the dimer. From the spectra obtained, a rninimum activation energy of =15 kcal/mol is calculated for this process. The limited information available renders any proposal to account for the observed 31P{1H} NMR spectra speculative. A plausible process is presented in Scheme 5-1. It begins with a rearrangement of the alkylideneimido hydride to a p>T]2-a species (O) via a Scheme 5-1 Rh* Rh* o Rh* = Rh(dippe) N .C H P Rh' R' - H C — H O' "Rh* 211 dissociation of the Rh-N bond trans to the o-tolyl substituent with concomitant 7t-donation to rhodium from the C=N double bond. This intermediate O is structurally analogous to the cationic i i - T ^ - a imine hydride complexes discussed in the previous chapter. Such a bonding mode for an alkylideneimido ligand has not been previously observed directly, but the possibility for |i-rj2-o bonding has been acknowledged.11 A windshield wiper process analogous to that proposed in Chapter 4 then accounts for the observed fluxionality in 36a. Note that in order to exchange the phosphorus environments the nitrogen lone pair must not be involved in the windshield wiper process. If the flipping movement was occurring through the |i-alkylideneimido hydride 36a, as depicted in Scheme 5-2, the phosphorus environments would not be equilibrated. In addition, dissociation of the Rh-N bond cis to the o-tolyl substituent results in the syn-O' species, an isomer which by analogy to the |i-alkenyl-hydride chemistry should be disfavoured on steric grounds.20 Thus, to account for the exchanging environments it is necessary to invoke P, in which a plane of symmetry through the bridging ligands exists and renders the cis and trans phosphorus atoms equivalent. A possible cause for this fluxionality resides in the steric interactions present between the aryl moiety on the |i-imido ligand and the adjacent dippe wopropyl groups (vide supra). Broadening at high temperatures was not observed in the 31P{ lK} NMR spectra of the p> Scheme 5-2 Rh* Rh* = Rh(dippe) 212 ethylideneimido hydride complex 34a, where R' = CH3. Replacement of CH3 with the bulkier aryl groups likely induces the observed behaviour. 5.3 Mechanistic Considerations. The reaction of d2-la with acetonitrile or o-tolylnitrile yielded c?2-34a and d2-36a, respectively (equation 5-2), in which the deuterium label appears in the p:-alkyhdeneimido and (5-2) d2-la R' = CH3, d2-34a R' = o-CH3-C6H4, d2-36a (i-hydrido positions. Unlike the reactions involving la and imines, however, intermediates in the reactions with nitriles were not observed when carried out at low temperatures. Only peaks for the dihydride and the final product 36a were observed in the 31P{1H} NMR spectrum of the proceeding reaction between la and o-tolylnitrile at -80°C. Evidently, the rearrangement that occurs subsequent to initial contact between the dimer and nitrile is too rapid to observe by NMR spectroscopy. The results of the labelling experiment are consistent with an insertion of the O N triple bond into a rhodium hydride bond, but the nature of any intermediate(s) en route to the ^ -alkylideneimido compounds (perhaps similar to those proposed in the imine reaction or the proposed |i-rj2-0" alkylideneimido species O and O') is unknown. 213 5.4 The Reactions of |i-AlkyIideneimido Complexes with Dihydrogen and Related Processes. The reaction chemistry of the |1-alkylideneimido hydrides 34-36 is largely unexplored. Preliminary studies have shown that complexes in the dippe series react readily with alkynes to give as yet unidentified binuclear products in which the p.-alkylideneimido ligand and the alkyne appear to have coupled. In a more thorough study, the complexes were shown to react with dihydrogen (equation 5-3) to produce, in near quantitative yield, amido \ / R » \ / R \ / R ^ \ /* \ R h R h 7 + Hg t o l u e n e , \ P h v R h 7 R R R R R R R R R'= C H 3 CeHg R'= C H 3 CsHg R = P r 1 , 34a 35a R = P r ' , 37a 38a hydride products akin to those produced in the reactions of the dihydrides with imines. The amido hydrides 37a and 38a (R* = C H 3 and 0 - C H 3 C 6 H 4 , respectively) were relatively inert to further reaction with dihydrogen and were thus isolated in 85-92% yield. The overall sequence from la and R'0=N to the amido hydrides represents a stoichiometric reduction of the nitrile to an amine mediated by two metal centres. Stepwise reduction of R'CN on trinuclear clusters has been observed previously ( M 3 = F e 3 , 1 3 < 1 5 O S 3 2 6 ) ; this is the first example of the transformation occurring on a binuclear cluster. These amido hydrides could also be produced directly from the tetrahydride 5a and nitriles. In addition, they could be synthesized from the dihydride and the appropriate amine, though by a much different mechanism involving N-H activation followed by the elimination 214 Scheme 5-3 215 of dihydrogen. Finally these particular amido hydrides (37a and 38a) would presumably also be accessible via reaction between la and the appropriate imine. These results are summarized in Scheme 5-3, with the dotted line for the imine reaction indicating that particular reaction as being predicted rather than observed. While the reactions of la with the amines were slowly reversible in the presence of excess dihydrogen, under no conditions explored could the reverse reaction from the primary amido hydrides to the alkyhdeneimido complexes be effected again illustrating the stability of these derivatives. Thermal or photochemical treatment of solutions of 37a produced no change in it's NMR spectra and chemical routes were not explored extensively. The production of the amido hydride 37a from the tetrahydride 5a and acetonitrile appeared to proceed mainly without initial formation of the alkyhdeneimido hydride, although detectable amounts of 34a were observed in the lH NMR spectrum of the proceeding reaction. Support for the existence of a pathway which does not include an alkylideneimido intermediate is found in the following experiment. When the tetrahydride 5c, incorporating the seven membered chelate ligand dippb, was treated with excess acetonitrile at 70°C for three days in the absence of excess dihydrogen (equation 5-4), the major product (>90% by *H NMR) was the p.-ethylamido hydride complex 37c. Detectable traces of the |i-ethylideneimido hydride 34c were observed, but were likely formed from the small amounts of thermally produced lc and acetonitrile. It is unlikely that, in the absence of excess toluene 70 -C 3 days c (5-4) R = Pr\ 5c R = Pr', 37c 216 dihydrogen, a sequence through 34c is feasible, because the partial pressure of dihydrogen present due to the elimination of dihydrogen from 5c would be too low to react with 34c at a reasonable rate. Even under 4 atmospheres of dihydrogen, 34a and 35a require 6-8 hours to undergo complete reaction. 217 5.5 Experimental. 5.5.1 General Procedures. Benzonitrile, orr/io-tolylnitrile and benzylamine were purchased from Aldrich Chemical Company and distilled prior to use. Spectrograde acetonitrile (BDH) was dried either by refluxing over calcium hydride under argon for 12 hours (large quantities) or by letting sit over activated 3A molecular sieves for 8 hours (small quantities). In the latter case, the acetonitrile was then degassed by vacuum transferring away from the molecular sieves and conducting two freeze-pump-thaw cycles. Ethylamine was obtained from Matheson Gas Products and condensed prior to use. It was then vacuum transferred from a -10°C bath to minimize transfer of any water present. The X-ray crystal structure analysis of 34a was performed by Dr. Steven J. Rettig of this department 5.5.2 The Syntheses of |i-Alkylideneimido Hydride Complexes. H v / C II R R R R R R R R R R' = C H 3 C 6 H 5 o-CH3-C6H4 R = Pr', R = OPr', 34a 35a 34d 35d 36a 36d 218 Synthesis of [(dippe)Rh]2[u.-N=C(H)CH3](p.-H), 34a To a stirred solution of [(clippe)Rh]2(p>H)2, la, (0.100 g, 0.14 mmol) in toluene (5 mL) was added a moderate excess of acetonitrile (2-3 drops). The deep green to orange-red colour change accompanying the reaction was instantaneous. The toluene and excess acetonitrile were removed in vacuo and the residues recrystallized from toluene/hexanes (1:1). After washing with cold hexanes, the yield of red crystals was 0.089 g (84%). *H NMR (C6D6, ppm): N=CHCH3, 8.75 (m, 1H, 3jH = 4.4 Hz, 4 J P couplings of 15.6, 8.8 and 4.8 Hz); CH(CH3)2, 2.23, 2.04-2.20 (overlapping dsp, 8H, 2 J P = 2.0-4.0 Hz, 3 J H = 6.4-7.6 Hz); N=CHCH3, 2.06 (d, 3H); CH(CH3)2, 1.43, 1.40, 1.32, 1.31, 1.20, 1.15, 1.14, 1.11 (overlapping dd, 48H, 3 J p = 12.4-13.4 Hz); PCH?CH?P, signals buried underneath methyl resonances; Rh-H-Rh, -7.54 (m, !jRh = 20.8-21.7 Hz). 31p{lH} NMR (C6D6, ppm): ?trans(Nh 104.0 (dd, lJ R h = 153.4 Hz, 2 J P = 26 Hz); Ptrans(Nh 102.6 (dd, lJ R h = 153.0 Hz, 2 J P = 24 Hz); Pcis(N), 87.0 ( br dm, l J R h = 195.6 Hz, 4 J P « 32 Hz); Pcis(N), 83.3 ( br dm, lJRh = 193.2 Hz). ^COH} NMR (C6D6, ppm): N=CHCH3, 159.1 (s); N=CHCH3, amongst ligand resonances; ligand resonances, 19-30 (overlapping signals). Anal. Calcd. for C30H69P4Rh2N: C, 46.58; H, 8.99; N, 1.81%. Found: C, 46.80; H, 9.11; N, 1.71%. Reaction of [(dippe)Rh]2(|i-D)2, d2-la, with CH 3CN. An identical procedure to that described above was followed using d2-la (0.033 g, 0.04 mmol). In the *H NMR spectrum of the product the signals due to the ethylideneimido proton and the bridging hydride ligand were absent; 2 H NMR spectroscopy confirmed the presence of the label in these positions. 219 Synthesis of [(dippe)Rh]2[u-N=C(H)C6Hs](u-H), 35a. To a stirred solution of [(dippe)Rh]2(p>H)2, la, (0.185 g, 0.25 mmol) in toluene (5 mL) was added a solution of distilled benzonitrile (0.029 g, 0.27 mmol) in toluene (1 mL). The solution changed colour from deep green to dark red instantly. The toluene was removed in vacuo and the residues recrystallized from toluene/hexanes (1:1) at -20° C. After washing with cold hexanes, 0.173 g (82%) of dark red-purple crystals was isolated. lH NMR (C7D8, ppm): N=CHC6H5, 9.00 (br m, 1H, coupled to four phosphorus nuclei); ^ ortho* 7.97 (d, 2H, 3 J u m e t a = 7.2 Hz); S™*,, 7.15 (m, 2H, 3jKpara = 7.2 Hz); Hp f l r a , 7.09 (m, 1H); CH(CH3)2, 2.05-2.25 (br overlapping dsp, 8H, 3jH = 6.6-7.6 Hz); CH(CH3)2, 1.34, 1.33, 1.30, 1.21, 1.05, 0.97, 0.96, 0.95 ( overlapping dd, 48H, 2 j p = 12.0-14.0 Hz); PCH2CH2P, 1.1 (br m, 8H); Rh-H-Rh, -7.82 (m, l J R h = 21.4-22.0 Hz). 31p{lH} NMR (C6D6, ppm): P ^ ; , 108.9 (dd, l J R h = 153.7 Hz, 2j p = 27 Hz); P ^ j , 108.3 (dd, ! j R h = 153.9 Hz, 2j p = 28 Hz); Pcis(N), 94.4 ( br dm, ! j R h = 195.9 Hz, 4 J P » 32 Hz); ?cis(N)> 8 7 - 5 ( b r dm. l j R h = 193.9 Hz). ^CpH} NMR (C6D6, ppm): N=CHC6H5, 159.9 (s, iJn = 163.6 Hz); Cj p s o , 143.9; other aromatic carbons, 129.1, 127.6, 127.1; ligand resonances, 19-28. Anal. Calcd. for C35H71P4RI12N: C, 50.31; H, 8.56; N, 1.68%. Found: C, 50.00; H, 8.40; N, 1.50%. Synthesis of [(dippe)Rh]2[p>N=C(H)0-CH3-C6H4](p>H), 36a. To a stirred solution of [(dippe)Rh]2(|i-H)2, la, (0.100 g, 0.14 mmol) in toluene (5 mL) was added a solution of distilled o-tolylnitrile (0.018 g, 0.15 mmol) in toluene (1 mL). The solution changed colour from deep green to dark red instantly. The toluene was removed in vacuo and the residues recrystallized from toluene/hexanes (1:1) at -20°C. After washing with cold hexanes, 0.105 g (88%) of dark red-purple crystals was isolated. *H NMR (C^Ds, ppm): N=CH, 10.09 (br m, coupled to 4 phosphorus nuclei); C-Hi. 8.18 (dd, 1H, 3 J H = 220 5.6 Hz, 4jH = 3.9 Hz); C-H2 and C-H3, 7.06 (m, 2H); C-H4, 6.93 (m, 1H); C-CH3, 2.41 (s, 3H); CH(CH3)2, 2.18 (8H, overlapping dsp); CH(CH3)2, P C H 2 C H 2 P , 0.9-1.4 (56H, br overlapping signals); Rh-H-Rh, -8.92 (ttt, 2 l p l r a n s = 60.4 Hz, 2 J P c i s =11.6 Hz, *JRh = 21.6 Hz). 31p{lH} NMR (C6D6, ppm): Ptrans(N), 107.9 (dd, !jRh = 152.6 Hz, 2 J P = 26.9 Hz); Ptrans(N), 107.8 (dd, l J R h = 152.5 Hz, 2JP = 26.8 Hz); P c i s ( N ) , 94.1 ( br dm, l j R h = 193.9 Hz, 4 J P » 32 Hz); Pcis(N), 82.9 ( br dm, l J R h = 193.8 Hz). ™C{ lH) NMR (C6D6, ppm): N=CH, 159.8 (s); CipSO(N), 143.5; Cipso(CH3> 135.3; other aromatic carbons, 130.2, 129.1, 127.1, 125.7; C-CH3, obscured in ligand region; ligand resonances, 19-29. Anal. Calcd. for C36H73P4Rh2N: C, 50.89; H, 8.66; N, 1.65%. Found: C, 51.22; H, 8.59; N, 1.74%. Reaction of [(dippe)Rh]2(p>D)2, with 0 - C H 3 - C 6 H 5 . An identical procedure to that described above was followed using <22-la (0.036 g, 0.05 mmol). In the *H NMR spectrum of the product the signals due to the o-tolylideneimido proton and the bridging hydride ligand were absent; 2 H NMR spectroscopy confirmed the presence of the label in these positions. 221 Synthesis of [(dipope)Rh]2[M>N=C(H)CH3](^-H), 34d. To a stirred solution of [(dipope)Rh]2(p>H)2, Id, (0.152 g, 0.18 mmol) in toluene (5 mL) was added two drops of acetonitrile. The reaction was allowed to stir for a further 30 minutes at which time the toluene was removed in vacuo. The orange residue was recrystallized from hexanes, yielding 0.122 g (77%) of 34d. *H NMR (C^D^, ppm): N=CHCH3, 9.24 (m, 1H, 3 J H = 4.9 Hz, and couplings of 20.0, 13.6, 6.4 and 1.8 Hz to phosphorus); OCH(CH3)2, 5.17, 5.01, 4.75, 4.66 (dsp, 2H each, 3 J H = 5.6-6.4 Hz, 2 J P = 2.0-3.0 Hz); N=CHCH3, 2.47 (d, 3H); P C H J C H J P , 1.73 (m, 8H); OCH(CH3)2, 1.42, 1.35, 1.34, 1.33, 1.21, 1.20,1.20, 1.19 (d, 48H); Rh-H-Rh, -6.42 (ttt, 2Jptrans = 76.0 Hz, Rh Rh M N 2 jPcw = 4-5 Hz, U R J , = 21.2 Hz). 3 1 P { lH} NMR (C6D6): Simulation of the spectrum of 34d was performed using the following parameters: 8A = 5220.0 Hz (actual chemical shift 189.3 ppm); 5B = 4945.4 Hz (actual chemical shift 186.9 ppm); 8M = 7640.0 Hz (actual chemical shift 209.1 ppm); 8N = 7500.8 Hz (actual chemical shift 207.8 ppm). J A B = 68.3 Hz; J A M = 30.2 Hz; J A N = 8.3 Hz; J A x = 250.3 Hz; J A Y = 12.2 Hz; J B M = 10.5 Hz; J B N = 29.8 Hz; J B x = H-3 Hz; J B y = 245.9 Hz; J M N = 0.0 Hz; J M x = 199.3 Hz; J M Y = 5.6 Hz; j N X = 4.4 Hz; J N Y = 200.1 Hz; J X y = 3.8 Hz. 13C{ !H} NMR (C6D6, ppm): N=CH, 168.9; 0£H(CH 3) 2, 4 signals 69.2-69.9; P C H 2 C H 2 P , 32.1-33.4 (m); N=CHCH3, 31.9; OCH(CH3)2, 8 signals 24.4-25.5. Anal. Calcd. for C3oH6908P4Rh2N: C, 39.97; H, 7.71; N, 1.55%. Found: C, 40.10; H, 7.81; N, 1.49%. 222 Synthesis of [(dipope)Rh]2[|i-N=C(H)C6H5](|i-H), 35d. To a stirred solution of [(dipope)Rh]2(p>H)2, Id, (0.167 g, 0.19 mmol) in toluene (5 mL) was added a solution of distilled benzonitrile (0.022 g, 0.21 mmol) in toluene (1 mL). The reaction was stirred for 30 minutes to insure completion; the colour change accompanying the reaction was slight. The toluene was removed in vacuo and the remaining red solid recrystallized from hexanes. After washing with cold hexanes, 0.137 g (73%) of crystals of 35d were obtained. *H NMR (C6D6, ppm): N=CHC6H5, 10.34 (br m, 1H, coupled to three phosphorus nuclei, Jp ~ 16, 16, and 4 Hz); Hortho, 8.17 (d, 2 H , ^Snmeta = 7-2 Hz); Hmeta, 7.29 (m, 2 H , ^ n p a r a = 7-3 Hz); Upara, 7-20 (m, 1H); OCH(CH3)2, 5.18, 5.17, 4.95, 4.50 (dsp, 8 H , 3 j H = 6.8-7.6 Hz, 2j p = 2.0-3.0 Hz); P C H 2 C H 2 P , 1.76, 1.70 (m, 8H); OCH(CH3)2, 0.95-1.5 (br signals, 48H); Rh-H-Rh, -6.80 (ttt, 2Jptrans = 75.2 Hz, 2 j P d , - 4 Hz, l J R h = 20.4 Hz). 31P{1H} NMR (QjDe, ppm): Ptrans(N), 204.7 (dd, lJ R h = 198.7 Hz, 2j P = 27.3 Hz); Ptrans(N), 201.0 (dd, l J R h = 203.6 Hz, 2j P = 36.1 Hz); Pcis(N), 187.0 (br ddd, ! j R h = 257.1 Hz, 2j p = 27.3 Hz, 4 J P = 59.3 Hz); PCis(N). 175.0 (br ddd, ! j R h = 263.3 Hz). ^CfiH) NMR (C6D6, ppm): N=CH, 168.0 (!j H = 169.3 Hz); Ci p s o , 140.7; other aromatic carbons, 129.3, 128.9, 127.7; OCH(CH3)2, 70.0 (2), 69.8, 69.1; P C H 2 C H 2 P , 31-34; OCH(CH3)2, 24.9-25.6. Anal. Calcd. for C 3 5 H 7 i0 8 P4Rh2N: C, 43.60; H, 7.47; N, 1.45%. Found: C, 43.43; H, 7.60; N, 1.40%. Synthesis of [(dipope)Rh]2[^-N=C(H)0-CH3-C6H4Kp>H), 36d. To a stirred solution of [(dipope)Rh]2(|i-H)2, Id, (0.102 g, 0.12 mmol) in toluene (5 mL) was added a solution of distilled benzonitrile (0.015 g, 0.13 mmol) in toluene (1 mL). The reaction was stirred for 45 minutes to insure completion. The toluene was removed in vacuo and the remaining red solid recrystallized from hexanes. After washing with cold hexanes, 0.088 g (76%) of crystals of 36d were isolated. *H NMR (C6D6, ppm): N=CH, 223 10.40 (br m, 1H, coupled to three phosphorus nuclei, Jp = 16, 16, and 4 Hz); C-Hi, 8.36 (dd, 1H, 3 J H = 7.2 Hz, 4 J H = 1.8 Hz); C-H2 and C-H3, 7.16 (m, 2H); C-H4, 7.01 (d, 1H, 3 J H = 6.8 Hz); OCH(CH3)2, 5.22, 5.16, 4.93, 4.46 (dsp, 8H, 3 J H = 6.4-7.6 Hz, 2 J P = 2.0-3.0 Hz); C-CH3, 2.55 (s, 3H); P C H 2 C H 2 P , 1.78, 1.73 (m, 8H); OCH(CH3)2, 1.45, 1.40, 1.36, 1.35, 1.27, 1.20, 0.97, 0.93 (d, 48H); Rh-H-Rh, -6.80 (br ttt, 2 Jp, r a / w = 73.6 Hz, 2jpdj = 4.8 Hz, l J R h = 20.8 Hz). 3lp{lH} NMR (CgDe, ppm): Ptrans(N), 207.8 (dd, l J R h = 200.0 Hz, 2 J P = 29.0 Hz); Ptrans(N), 201.2 (dd, l j R h = 203.6 Hz, 2 J P = 34.6 Hz); Pcis(N), 186.4 (br dm, lJRh = 256.2 Hz, 4 J P = 55 Hz); Pcis(N), 175.0 (br dm, !jRh = 262.7 Hz). 13C{1H} NMR (C6D6, ppm): N=CH, 167.4 (s); Cip S o ( N 0, 136.0; Cipso(CH3), 131.9; other aromatic carbons, 128.9, 128.7, 126.5, 126.4; OCH(CH3)2, 70.1(2), 69.8, 68.9; PCH2£H2P, 32.1-33.5 (m); OCH(CH3)2, 24.6-25.6; C-CH3, obscured in methyl region. Anal. Calcd. for C 36H 7iC>8P4Rh 2N: C, 44.23; H, 7.53; N, 1.43%. Found: C, 43.99; H, 7.56; N, 1.39%. 224 5.5.3 Reactions of ^ .-Alkylideneimido Hydrides with Dihydrogen. R i R^  „R H . C H 2 R . y R \ Rh Rh / R R R R R ' = CH3 C Q H 5 R = Pr1, 37a 38a Synthesis of [(dippe)Rh]2[MH)N(CH2CH3)](p-H), 37a. Solid [(dippe)Rh]2[|i-N=C(H)CH3](p>H), 34a, (0.072 g, 0.09 mmol) was loaded into a small reactor bomb along with toluene (5 mL) and a magnetic stir bar. The solution was degassed on a vacuum line and cooled to -196°C. Dihydrogen was admitted to one atmosphere of pressure and the solution was allowed to warm to room temperature. Vigorous stirring was continued for 6-7 hours during which time a red to yellow-orange colour change occurred. The toluene was removed in vacuo and the residues recrystallized from toluene/hexanes (1:1) yielding 0.059 g (78%) of oily yellow crystals. lH NMR (C6D6, ppm): N-H, 3.97 (br s, 1H); H-NCH ,^ 3.69 (br m, 2H); CH(CH3)2, 2.40, 2.28, 2.06, 1.92 (dsp, 8H, 2 J P = 2.0-4.0 Hz, 3 j H = 7.3-8.0 Hz); CH(CH3)2, 1.50, 1.33, 1.31, 1.20, 1.17, 1.13, 1.12, 1.01 (dd, 48H, 3 j p = 12.0-14.4 Hz); NCH2CH3, PCH^CH^P, obscured under methyl resonances; Rh-H-Rh, -7.30 (ttt, 2Jptrans = 61.2 Hz, 2 J P c i s = 13.2 Hz, l J R h = 21.2 Hz). 31P{lH} NMR (C6D6, ppm): Ptrans(N), 104.2 (dm, lJ R h » 154 Hz); Pcis(N), 87.1 (dm, l J R h » 193 Hz). 13c{lH} NMR (C6D6, ppm): H-NCH2, 52.7; NCH 2CH 3, ligand 225 resonances 17-28. Anal. Calcd. for C3oH 7 iP 4 Rh2N: C, 46.46; H, 9.23; N, 1.81%. Found: C, 46.09; H, 9.27; N, 1.81%. Synthesis of [(dippe)Rh]2[|i-(H)N(CH2C6H5)](p:-H), 38a. An identical procedure to that described above was employed using 35a (0.093 g, 0.11 mmol). Yellow crystals of the p>benzylamido hydride complex 38a were obtained (0.086 g, 81%). lH NMR (C6D6, ppm): Rortho, 8.01 (d, 2H, 3 J H m e f f l = 7.6 Hz); Umeta, 124 (m, 2H, 3JHPara = ?-5 Hz); Upara, 7.14 (t, 1H); H-NCH2, 5.17 (br m, 2H, J P « 3.0 Hz); N-H, 4.10 (br s, 1H); CH(CH3)2, 2.45, 2.16, 2.13, 1.95 (dsp, 8H, 2 J P = 2.0-4.0 Hz, 3 j H = 7.2-8.2 Hz); CH(CH3)2, 1.56, 1.45, 1.34, 1.18, 1.09, 1.07, 1.01, 0.78 (dd, 48H, 3 j P = 11.4-15.2 Hz); P C H 2 C H 2 P , 1.26 (m, 8H); Rh-H-Rh, -7.44 (ttt, 2Jptrans = 60.2 Hz, 2 J P c i s = 12.6 Hz, l j R h = 21.6 Hz). 31 P {1 H } NMR (C6D6, ppm): Ptrans(N), 103.7 (dm, ! j R h = 152 Hz); Pcis(N), 85.8 (dm, ijRh = 193 Hz). Anal. Calcd. for C35H73P4RI12N: C, 50.18; H, 8.78; N, 1.67%. Found: C, 49.96; H, 8.73; N, 1.80%. 5.5.4 Reactions of the Binuclear Tetrahydrides with Nitriles. Synthesis of [(dippe)Rh]2[MH)N(CH2CH3)](u.-H), 37a. A solution of [(dippe)Rh]2(p>H)2, la, (0.103 g, 0.14 mmol) in toluene (5 mL) was loaded into a small reactor bomb equipped with a magnetic stir bar. The solution was degassed on a vacuum line and dihydrogen admitted to a pressure of one atmosphere. To this solution of the tetrahydride 5a was added acetonitrile (0.1 mL) via syringe under a strong flow of dihydrogen. The reaction mixture was cooled to -196°C and the Kontes valve port sealed completely. Stirring was continued for 16 hours during which time a slow brown to yellow-orange colour change ensued. The toluene, excess dihydrogen, and excess acetonitrile were removed in vacuo and the residues recrystallized from toluene/hexanes (1:1) 226 yielding 0.070 g (64%) of crystals of 37a. The product was found to be spectroscopically identical to that obtained in the reaction of 34a and dihydrogen. Synthesis of [(dippe)Rh]2[MH)N(CH2C6H5)](p>H), 38a. An identical procedure to that described above was employed using la (0.103 g, 0.14 mmol) and benzonitrile (0.72 g in 0.5 mL toluene, 5 equivalents). The product, 38a, was isolated in 75% yield (0.088 g) and was spectroscopically identical to the product obtained in the reaction of 35a with dihydrogen. Synthesis of [(dippb)Rh]2[MH)N(CH2C6H5)](p>H), 38c. A solution of the tetrahydride [(chppb)Rh]2(H)(p>H)3,5c, (0.098 g, 0.12 mmol) in toluene (5 mL) was loaded into a small reactor bomb equipped with a magnetic stir bar. Acetonitrile (1 mL) was added and the reaction mixture degassed on a vacuum line via two freeze-pump-thaw cycles; the vessel was back-filled with pure dinitrogen. The reaction was stirred at 70°C for three days, at which time a colour change from brown-green to yellow was complete. Toluene and excess acetonitrile were removed in vacuo and the residues analyzed by !H and 31p{lH} NMR spectroscopy. !H NMR (C6D6, ppm): N-H, 3.41 (br s, 1H); H-N C H 2 , 3.27 (br m, 2H); CH(CH3)2, 2.35, 2.18, 2.15, 1.96 (dsp, 8H, 2jp = 2.0-4.0 Hz, 3jH = 7.4-8.0 Hz); NCH2CH3, 1.70 (t, 3H, 3jH = 7.6 Hz); CH(CH3)2, 1.58, 1.55, 1.46, 1.32, 1.30, 1.21, 1.19, 1.17 (dd, 48H, 3jp = 12.0-14.0 Hz); P(CH2)4P, resonances obscured by methyl signals; Rh-H-Rh, -10.87 (ttt, 2 Jp / n w j = 5 1 - 2 H z ' 2 j P c f a = 1 3 - 2 H z ' 1JRh = 21.6 Hz). 31P{1H} NMR (C6D6, ppm): Ptrans(N), 61.2 (dm, l J R h - 154 Hz); Pcis(N), 33.7 (dm, lJ R h - 195 Hz). 227 5.5.5 Reactions of [(dippe)Rh]2(|i-H)2 with Amines. Synthesis of [(dippe)Rh]2[n-(H)N(CH2CH3)](p:-H), 37a. A solution of [(dippe)Rh]2(p>H)2, la (0.092 g, 0.12 mmol) in toluene (5 mL) was loaded into a small reactor bomb equipped with a magnetic stir bar. The solution was degassed on a vacuum line. A moderate excess of ethylamine was then vacuum transferred into the vessel from a -10°C bath. The reaction was stirred until the green to yellow colour change was complete (8-12 hours). The toluene and excess ethylamine were removed under reduced pressure and the residues recrystallized from toluene/hexanes (1:1) yielding 0.082 g (84%) of crystals of 37a. The product was spectroscopically identical to samples of 37a obtained via the routes described above. Synthesis of [(dippe)Rh]2[p-(H)N(CH2C6H5)](p:-H), 38a. To a stirred solution of [(dippe)Rh]2(p>H)2, la, (0.087 g, 0.11 mmol) in toluene (5 mL) was added pure benzylamine (0.064 g, 5 equivalents). The reaction was stirred for two hours while a green to yellow-orange colour change occurred. Removal of the toluene, followed by recrystallization yielded 0.087 g (88%) of yellow-orange crystals of 38a. This product was spectroscopically identical to samples of 38a obtained via the routes describe above. 5.6 References. 1. Kilner, M. Adv. Organomet. Chem. 1972,10, 115. See especially p 160-164. 2. Kilner, M.; Midcalf, C; /. Chem. Soc. Chem. Commun. 1970, 552. 3. Kilner, M.; Midcalf, C; Farmery, K. J. Chem. Soc. (A) 1970, 2279. 4. Kilner, M.; Midcalf, C. / . Chem. Soc. (A) 1971, 292. 228 5. Kilner, M.; Pinkney, J. N. / . Chem. Soc. (A) 1971, 2887. 6. Erker, G.; Fromberg, W.; Kriiger, C.; Raabe, E. /. Am. Chem. Soc. 1988,110, 2400. 7. Bright, D.; Mills, O. S. J. Chem. Soc. Chem. Commun. 1967, 245. 8. Churchill, M. R.; Kuo-Kuang, G. L. Inorg. Chem. 1975,14, 1675. 9. Churchill, M. R.; de Boer, B. G.; Rotella, F. J. Inorg. Chem. 1976,15, 1843. 10. Prest, D. W.; Mays, M. J.; Raithby, P. R. /. Chem. Soc. Dalton Trans. 1982, 2021. 11. Adams, R. D.; Katahira, D. A.; Yang, L.-W. /. Organomet. Chem. 1981,219, 85. 12. Dawoodi, Z.; Mays, M. J.; Raithby, P. R. J. Organomet. Chem. 1981,219, 103. 13. Kaesz, H. D.; Andrews, M. A. / . Am. Chem. Soc. 1979,101, 7238. 14. Kaesz, H. D.; Knobler, C. B.; van Buskirk, G.; Andrews, M. A. /. Am. Chem. Soc. 1979, 101, 7245. 15. Kaesz, H. D.; Andrews, M. A. / . Am. Chem. Soc. 1979,101, 7255. 16. a) Kautzner, B.; Wailes, P. C; Weigold, H. /. Chem. Soc. Chem. Commun. 1969, 1105. b) Wailes, P. C; Weigold, H.; Bell, A. P. /. Organomet. Chem. 1972,42, C32. c) Wailes, P. C; Weigold, H.; Schwartz, J.; Jung, C. Inorg. Synth. 1979,19, 223. 17. Etievant, P.; Tainturier, G.; Gautheron, B. C. R. Acad. Sc. Paris, Ser. C. 1976, 283, 233. 18. a) Fryzuk, M. D. Piers, W. E. Organometallics, in press, b) See Chapter 4. 19. Levy, G. C; Lichter, R. L.; Nelson, G. L. Carbon-13 Nuclear Magnetic Resonance Spectroscopy, 2nd Ed.. Wiley: New York, 1980, p 32. 20. Fryzuk, M. D.; Einstein, F. W. B.; Jones, T. Organometallics 1984,3, 185. 21. Fryzuk, M. D; Jang, M.-L.; Jones, T.; Einstein, F. W. B. Can. J. Chem. 1986, 64, 174. 22. a) Table 2-1. b) Table 3-III. 23. a) "Interatomic Distances Supplement", Chem. Soc. Spec. Publ. 1965,18. b) Bohme, H.; Viehe, H. G., Eds. Iminium Salts in Organic Chemistry, Part I. Wiley: New York, 1976, 89. 229 24. Dawoodi, Z.; Mays, M. J.; Orpen, A. G. /. Organomet. Chem. 1981,219, 251. 25. a) Typical Uc-H values for sp2 hybridized carbon atoms fall in the range of 150-175 Hz. 2 5 b b) Levy, G. C.; Lichter, R. L.; Nelson, G. L. Carbon-13 Nuclear Magnetic Resonance Spectroscopy, 2nd Ed.. Wiley: New York, 1980, p 33-36. 26. Dawoodi, Z.; Mays, M. J.; Henrick, K /. Chem. Soc. Dalton Trans. 1984, 433. 230 CHAPTER 6 Concluding Remarks and Future Prospects 6.1 Conclusions. Until recently, the study and characterization of fundamental chemical transformations at more than one metal centre had largely been based on the thermally induced chemistry of organometallic clusters. Usually, intermediates in these reactions have been unobservable since the reactions often require elevated temperatures; much of the chemistry of trinuclear carbonyl clusters of the iron triad is of this type. However, the advent of stable, coordinatively unsaturated clusters has made the observation of intermediates in cluster reactions more common. The coordinatively unsaturated binuclear rhodium hydrides discussed in this thesis have been particularly useful in defining unprecedented reactions involving two transition metal centres, as well as new bonding modes for simple organic molecules. Throughout this chemistry, the effect of the ancillary diphosphine ligand on the reactions was significant and thought to be largely steric in origin. For example, the general reactivities of the dimers la and Id, containing the electronically very different dippe (electron rich) and dipope (electron poor) ligands, respectively, were quite similar, while a comparison of the chemistry of la vs lb, with varied chelate ring size in the dippe (five) and dippp (six) ligands, reveals significant differences in reaction rates, product geometries, and product distributions. This chelate ring size effect was demonstrated in the characterization of the fluxional tetrahydrides 5a-c, which result from the reaction of the dihydrides la-c with dihydrogen (Chapter 2). In these mixed valence binuclear tetrahydrides (formulated as containing one Rh(UI) centre and one Rh(I) centre) the geometry of both the ground and excited state 231 structures were found to be different in the dippe system than those of the dippp and dippb tetrahydrides. In the proposed hydrogenation cycle for the reduction of olefins by these tetrahydrides, this difference in geometry was thought to be a major factor leading to fragmentation and a cycle involving mononuclear intermediates for the dippp (5b) catalyzed reaction. The change in geometry at the unsaturated Rh(I) centre of the tetrahydrides from square planar (in 5a) to tetrahedral (in 5b) hampered olefin binding at this site, blocking the cycle involving binuclear intermediates and rendering the mononuclear cycle kinetically more important in the dippp system than in the dippe catalytic cycle. Chapter 3 described the reactions of la and lb with 1,3-dienes, and again, chelate ring size effects were significant. Two new bonding modes for 1,3-dienes to two metal centres were identified: the p:-r|4-a mode, in which the diene is 7t-bound to one centre and a-bonded to the other through a terminal carbon atom of the diene, and the p,-rj3-r|3 "partial sandwich" mode, in which the dienyl ligand is sandwiched between the P2R.f1 fragments and ti3-bonded to each of them. Studies on the fluxional behaviour in some of these systems led to a proposal which delineated some of the structural rearrangements possible for a dienyl ligand bonded to two metal centres; this proposal included interconversions between the two bonding modes mentioned above as well as other previously characterized diene bonding modes. Mechanistic studies revealed that dehydrogenation of the dihydride by one equivalent of butadiene was the first step in the reaction; this was supported via low temperature observation of the ri2-dienyl species present just prior to this dehydrogenation step. Subsequent reaction of the diene with the proposed binuclear d9-d9 rhodium(O) species [(P2)Rh]2 led to the binuclear products. The key step in this process was a proposed formal insertion of the diene into the Rh-Rh bond of [(P2)Rh]2 via a 4+2 cycloaddition. Rearrangement of the product of this process led to the |i-dienyl products. The accumulated evidence added to the plausibility of the proposal, which included several steps invoking transformations or rearrangements requiring the participation of both metal centres. 232 The reactions of la and Id, incorporating the 5 membered chelate ligands dippe and dipope, respectively, with simple imines was the subject of Chapter 4. Low temperature studies again allowed the characterization of intermediates in the reaction such that a clear picture of the pathway to the product amido hydrides was obtained. Initial coordination of the imine to one rhodium centre through the nitrogen lone pair was followed by 7t-donation from the C=N bond to the other rhodium atom with concomitant rupture of the Rh2(p>H)2 core of the molecule to form an intermediate which contained a U . - T | 2 - C T imine ligand, a new bonding mode for the imine ligand to two metal centres. A facile insertion of the 7t-bonded C=N bond into an Rh-H bond produced the observed amido hydride products. This last step showed that if re-coordination of C=N can be induced, insertion of the C=N bond into M-H bonds occurs rapidly, and also suggests that in order to effect reduction of imines, the lone pair on nitrogen must be occupied elsewhere to induce re-coordination over a-donation. This work demonstrates that another transition metal may perform this function, but the chemistry of this system suggests that the resulting bridging amido ligand is resistant to hydrogenolysis and may form a kinetically inert "intermediate" in a catalytic cycle. The mechanistic proposal was supported by kinetic and labelling studies as well as the specific synthesis of cationic derivatives of the |i-T|2-a imine intermediate in the reaction. Characterization of these cationic derivatives showed that the u.-T)2-0" imine ligand was fluxional, exhibiting the "windshield wiper" movement common to |i.-T|2-0" alkenyl ligands. The reactivity of unsaturated carbon-nitrogen functionalities with la and Id was extended to nitriles in the work described in Chapter 5. While labelling studies showed that the -^alkylideneimido products arise from a formal insertion of C=N into an Rh-H bond of the dihydrides, the inability to observe intermediates at low temperature precluded determination of a mechanistic pathway for this reaction. Reaction of the ji-alkylideneimido hydride products with dihydrogen produced amido hydrides, completing the stoichiometric reduction of the G=N bond at two metal centres. 233 6.2 Future Prospects. The work described in this thesis illustrates the value of these rhodium compounds in the study of fundamental processes involving two metal centres, but their potential for producing further interesting results is far from realized. The following discussion provides suggestions for future areas of development, some of which have already been explored in a preliminary fashion. One such area is a continued study of the insertions of unsaturated organic functional groups into the Rh2(fi-H)2 core of the dihydrides la and Id. Although ketone or aldehyde C=0 double bonds are unreactive towards the dihydrides,1 preliminary studies indicate that the heterocumulene bonds of isocyanates1 (RN=C=0) and thioisocyanates2 (RN=C=S) undergo facile reactions with both the dippe and dipope dimers to produce complexes with the organic reactant in a bridging mode of bonding. Indications are that the potential for observing intermediates in these reactions at low temperature is excellent; similar studies to those carried out in the imine reaction should be possible. Moreover, these reactivity studies could be extended to include other cumulenes, such as ketenes (R2C=C=0), allenes (R2C=C=CR2), carbodiimides (RN=C=NR) and possibly diazoalkanes (R2C=N+=N-). The observed reactivity of la and lb with Bronsted acids1'3 provides the impetus for another potential area of study. Kinetic and mechanistic studies should be straightforward and could provide valuable data on the activation of N-H and O-H bonds, given that it appears as though oxidative addition of these bonds is the first step in the dihydrides reactions with R2NH and ROH. In addition, reactions with other acids, such as RCOOH, have the potential for forming binuclear complexes with the conjugate bases in new bridging bonding modes. Reactivity studies on some of the complexes characterized in this thesis are also potentially interesting. The p>dienyl complexes of Chapter 3 activate butadiene in previously unobserved activation modes; a study attempting to delineate the implications of these activation modes with respect to the reactivity of the diene fragment towards various 234 nucleophiles and electrophiles might prove interesting. Direct attack by such reagents on the diene fragment may, however, be diverted by the coordinative unsaturation of these complexes. On the other hand, this feature of these complexes may give rise to coupling reactions in which the incoming reactant is joined to the diene fragment in some metal-mediated process; indeed, preliminary studies indicate that such a reaction type occurs when the dippe butadienyl hydride (8a, Chapter 3) is treated with terminal alkynes.1 Like the n-dienyl compounds, the cationic p>T|2-G imine hydride derivatives discussed in Chapter 4 form an interesting group of compounds. The reactions of cationic organometallic complexes with nucleophiles is a well defined reaction type;4 a study exploiting this strategy for the cationic p>rj2-a imine hydrides would be aimed at determining the site of nucleophilic attack and perhaps the formation of useful organic fragments. In addition, the thermal decomposition of these complexes must be studied in more detail. Finally, in a project with a slightly different theme, the dippe dimer la has been observed to react rapidly with the binuclear metal carbonyl complex Co2(CO)s to produce the heterobimetallic species [(dippe)Rh(CO)-Co(CO)4] along with Co4(CO)i2 and evolution of H2 and CO.1 The rhodium containing product has been structurally characterized5 and is structurally analogous to a fully carbonylated Rh-Co dimer which is the proposed active species in the catalytic hydroformylation of olefins by Rh/Co mixtures.6 Further studies on the potential of this compound as a hydroformylation catalyst as well as the reactions of la with other metal carbonyls (the reaction with Mn2(CO)io has also been studied in a preliminary fashion) are required. These are some of the primary areas for development in the chemistry of these dihydrides. There are undoubtedly many more, but one thing is clear: this family of binuclear rhodium hydrides will be providing fundamental information on reactivity at two metal centres for years to come. 235 6.3 References. 1. Fryzuk, M. D.; Piers, W. E. unpublished results. 2. Fryzuk, M. D.; Piers, W. E.; Rosenberg, L. unpublished results. 3. Fryzuk, M. D; Jang, M.-L.; Jones, T.; Einstein, F. W. B. Can. J. Chem. 1986, 64, 174. 4. Davies, S.G. Organotransition Metal Chemistry: Applications to Organic Synthesis; Pergarnmon Press: New York, 1982, p 116-154. 5. Fryzuk, M. D.; Piers, W. E.; Rettig, S. J. unpublished results. 6. Ojima, I.; Okabe, M.; Kato, K.; Kwon, H. B.; Horvath, I. T. /. Am. Chem. Soc. 1988, 110, 150. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0060285/manifest

Comment

Related Items