Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Cationic carbonyl derivatives of electron-rich metals-syntheses, structures, and spectroscopic properties Wang, Changqing 1996

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

Item Metadata

Download

Media
831-ubc_1996-148505.pdf [ 8.72MB ]
Metadata
JSON: 831-1.0059520.json
JSON-LD: 831-1.0059520-ld.json
RDF/XML (Pretty): 831-1.0059520-rdf.xml
RDF/JSON: 831-1.0059520-rdf.json
Turtle: 831-1.0059520-turtle.txt
N-Triples: 831-1.0059520-rdf-ntriples.txt
Original Record: 831-1.0059520-source.json
Full Text
831-1.0059520-fulltext.txt
Citation
831-1.0059520.ris

Full Text

CATIONIC CARBONYL DERIVATIVES OF ELECTRON-RICH METALS - SYNTHESES, STRUCTURES, AND SPECTROSCOPIC PROPERTIES by CHANGQING WANG B.Sc, The University of Henan, 1982 M.Sc., The University of Lanzhou, 1988 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Department of Chemistry) We accept this thesis as conforming to the required standard The University of British Columbia June 1996 ® Changqing Wang, 1996 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. I 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 of The University of British Columbia Vancouver, Canada Date DE-6 (2/88) ABSTRACT The range of thermally stable, predominantly a-bonded metal carbonyl derivatives has been expanded to include Pd(I), Pd(II), Ir(III), Os(II), and Ru(II). The new compounds, [Pd(CO)4][Sb2F„]2, m-Pd(CO)2(S03F)2, [Pd2(ix-CO)2](S03F)2, [Pd2(u-CO)2][Sb2F11]2, mer-Ir(CO)3(S03F)3, [Ir(CO)5Cl][Sb2Fn]2, [Ru(CO)6][Sb2Fu]2, [Os(CO)6][Sb2Fn]2, and cis-Pt(CO)2(Sb2Fn)2 have been synthesized. The compounds have been characterized using a variety of physico-chemical techniques: elemental analysis, FT-IR and Raman spectroscopy, and 1 3 C NMR spectroscopy. For cw-Pd(CO)2(S03F)2, [Pd2(M-CO)2](S03F)2, mer-Ir(CO)3(S03F)3, and [Ir(CO)5Cl][Sb2Fu]2, the molecular structures have been determined by single crystal X-ray diffraction. The complex cw-dicarbonyl palladium(II) fluorosulfate is the first example of a thermally stable, mononuclear bis(carbonyl) derivative of dipositive palladium, while in [Pd(CO)4][Sb2Fn]2 a square-planar carbonyl cation was found. The cyclic cation [Pd2(/i-CO) 2 ] 2 + is a rare example of the point group D2h and a complete vibrational analysis was attempted. Both isomers of Ir(CO)3(S03F)3 (mer and fac) have been shown to exist, with the /ner-isomer predominant in solution and the only one isolated as a pure compound, mer-Ir(CO)3(S03F)3 is the first thermally stable, structurally characterized, and predominantly o-bonded carbonyl derivative of a metal in the +3 oxidation state. Although the existence of octahedral dipositive metal carbonyl cations has been anticipated for a long time, the successful syntheses and spectroscopic characterization reported here for [Ru(CO)6][Sb2Fn]2 and [Os(CO)6][Sb2Fn]2 are the first well documented examples for such species. ii Two remarkable characteristics of the new carbonyl compounds reported in this thesis are the extremely high CO stretching frequencies and the relatively high thermal stabilities, usually well above 100 °C. The conventional synergetic bonding mechanism cannot be invoked, without modification, to explain their thermal robustness since the M-CO bond is expected to be weak in the near-absence of rc-back-donation, as indicated by the high v (CO) stretching frequencies observed for these compounds. The single crystal structural analysis of c/5-Pd(CO)2(S03F)2 sheds some light on this phenomenon. Secondary interactions, which involve the positively polarized electrophilic C atom of the CO group and the O or F atoms of the fluorosulfate groups, help to provide some degree of charge compensation in the absence of significant rc-back-donation and exert a stabilizing influence on the structure. Similar secondary interactions were observed in the structures of mer-Ir(CO)3(S03F)3 and Pr(CO)5Cl][Sb2F11]2. These interactions may well explain the formation of predominantly CT-bonded metal carbonyl derivatives. In addition, alternative, simplified synthetic routes to metal carbonyl salts of Au(I), Pd(II), and Pt(II) are reported. In these methods metal fluorosulfates and metal chlorides are used as precursors instead of metal carbonyl fluorosulfates, making the syntheses of the metal carbonyl compounds more convenient, less hazardous, and less time consuming. iii TABLE OF CONTENTS Abstract ii Table of Contents iv List of Tables xi List of Figures xiii List of Symbols and Abbreviations xjy Acknowledgements xvi CHAPTER 1 GENERAL INTRODUCTION 1 1.1 Metal-CO Bonding — a Descriptive Overview 1 1.2 Carbon Monoxide, CO, as a Potential Ligand 2 1.3 Synergic Bonding of CO in Typical Metal Carbonyl Compounds 6 1.4 Carbon Monoxide as a a-Donor Ligand in Inorganic Compounds 12 1.4.1 CO Adducts with Molecular Boron Lewis Acids 12 1.4.2 Carbonyl Cations 14 1.4.3 Metal Carbonyl Derivatives 15 1.5 Spectroscopic and Structural Studies on Noble-Metal Carbonyl Derivatives 18 1.6 Recent Developments of Cationic Metal Carbonyl Derivatives 22 1.7 Research Direction— Methods and Objectives 31 1.7.1 Research Direction 32 1.7.2 Synthetic Methodology and Cooperation 33 1.7.2.1 Synthetic Methodology 33 1.7.2.2 Cooperation with Professor H. Willner's Group 34 iv 1.8 Outline of This Thesis 35 References 36 CHAPTER 2 GENERAL EXPERIMENTAL 45 2.1 Introduction 45 2.2 Chemicals 45 2.2.1 Chemicals Used Without Further Purification 45 2.2.2 Purification of CO, HS0 3F, SbF5, and S0 2 46 2.2.3 Preparative Reaction of S 2 0 6 F 2 48 2.3 Apparatus and Equipment 52 2.3.1 Vacuum Line 52 2.3.2 Glove Box 52 2.3.3 Glass Vessels 54 2.4 Instrumentation and Methods 54 2.4.1 Infrared Spectroscopy 54 2.4.2 Raman Spectroscopy 56 2.4.3 X-ray Crystallography 56 2.4.4 NMR Spectroscopy 56 2.4.5 Microanalysis 57 2.4.6 Melting Point Determination 57 References 58 v CHAPTER 3 SYNTHESES, STRUCTURES, AND SPECTROSCOPIC STUDIES OF CATIONIC CARBONYLS OF PALLADIUM (TJ) AND PALLADIUM (I) 59 3.1 Introduction 59 3.2 Experimental 60 3.2.1 Synthesis of ris-Pd(CO)2(S03F)2 60 3.2.1.1 Gas-Solid Reaction 60 3.2.1.2 Reductive Carbonylation in HS0 3F 60 3.2.1.3 Preparation of Single Crystals of cw-Pd(CO)2(S03F)2 61 3.2.2 Synthesis of Single Crystals of [Pd2(p:-CO)2](S03F)2 62 3.2.3 Synthesis of [Pd(CO)4][Sb2Fn]2 63 3.3 Results and Discussion 64 3.3.1 Synthetic Aspects 64 3.3.1.1 c/5-Pd(CO)2(S03F)2 and [Pd2(/x-CO)2](S03F)2 64 3.3.1.2 [Pd(CO)4][Sb2Fn]2 70 3.3.2 Vibrational Spectra and Structural Aspects 70 3.3.2.1 [Pd2(/x-CO)2](S03F)2 70 3.3.2.2 cw-Pd(CO)2(S03F)2 88 3.3.2.3 [Pd(CO)4][Sb2Fn]2 102 3.4 Summary and Conclusions 107 References 109 CHAPTER 4 SYNTHESES, STRUCTURES, AND SPECTROSCOPIC STUDIES OF CATIONIC CARBONYLS OF IRIDIUM(III) 114 vi 4.1 Introduction 114 4.2 Experimental 115 4.2.1 The Synthesis of Ir(CO)3(S03F)3 115 4.2.2 The Synthesis of [Ir(CO)5Cl][Sb2F„]2 118 4.3 Results and Discussion 119 4.3.1 Ir(CO)3(S03F)3 119 4.3.1.1 Synthetic Aspects 119 4.3.1.2 Molecular Structure of mer-Ir(CO)3(S03F)3 124 4.3.1.3 Vibrational Spectra 129 4.3.2 [Ir(CO)5C13[Sb2Fn]2 136 4.3.2.1 Synthetic Aspects 136 4.3.2.2 Molecular Structure of [Ir(CO)5Cl][Sb2Fn]2 138 4.4 Summary and Conclusions 141 References 143 CHAPTER 5 SYNTHESES AND SPECTROSCOPIC CHARACTERIZATIONS OF CARBONYL CATIONS OF GROUP 8 METALS 145 5.1 Introduction 145 5.2 Experimental 148 5.2.1 Synthesis of [Ru(CO)6][Sb2F„]2 148 5.2.2 Synthesis of [Os(CO)6][Sb2Fn]2 149 5.2.3 Attempted Synthesis of Cationic Ruthenium and Osmium Carbonyl Fluorosulfates 151 vii 5.2.3.1 Ruthenium Carbonyl Fluorosulfate 151 5.2.3.2 Osmium Carbonyl Fluorosulfate 152 5.2.4 Attempted Synthesis of [Fe(CO)6][Sb2Fn]2 152 5.3 Results and Discussion 154 5.3.1 Synthetic Aspects 154 5.3.2 Vibrational Spectroscopy 156 5.3.2.1 [Ru(CO)6][Sb2F„]2 and [Os(CO)6][Sb2F„]2 156 5.3.2.2 Fe[SbF6]2 161 5.3.2.3 Ruthenium and Osmium Carbonyl Fluorosulfates 161 5.3.3 1 3 C MAS NMR Spectroscopy 163 5.4 Summary and Conclusions 163 References 166 CHAPTER 6 SOLVOLYSIS REACTIONS IN ANTIMONY PENTAFLUORIDE - Improved Synthetic Routes and Methods of Generating New Metal Carbonyl Derivatives 169 6.1 Introduction 169 6.2 Experimental 170 6.2.1 Solvolysis Reactions in SbF5 in the Presence of CO 170 6.2.1.1 The Synthesis of [Pd(CO)4][Sb2Fj,]2 from Pd[Pd(S03F)6] 170 6.2.1.2 The Synthesis of [Pd(CO)4][Sb2F, ,]2 from PdCl2 171 6.2.1.3 The Synthesis of [Pt(C0)4][Sb2Fn]2 from Pt(S03F)4 172 6.2.1.4 The Synthesis of [Au(CO)2][Sb2Fn] from Au(S03F)3 172 viii 6.2.1.5 The Synthesis of [Au(CO)2][Sb2F„] from AuCl 3 173 6.2.2 Solvolysis Reactions in the Absence of Additional Gaseous CO 174 6.2.2.1 The Synthesis of [Pd2(Li-CO)2][Sb2Fi ,]2 174 6.2.2.2 The Synthesis of d5-Pt(CO) 2(Sb 2F„) 2 175 6.2.2.3 The Attempted Synthesis of cw-Pd(CO)2(Sb2F„)2 175 6.3 Results and Discussion 177 6.3.1 Solvolysis Reactions in SbF5 in the Presence of CO - Improved Synthetic Routes to Metal Carbonyl Cations 177 6.3.2 Solvolysis Reactions in SbF5 in the Absence of Gaseous CO 183 6.4 Summary and Conclusions 189 References 191 CHAPTER 7 GENERAL CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK 194 7.1 Summary and General Conclusions 194 7.2 Suggestions for Future Work 205 7.2.1 Syntheses of New Cationic Metal Carbonyl Derivatives 205 7.2.2 Syntheses of New Cationic Metal Carbonyl Salts 206 7.2.3 Substitutions of CO by Other Weak Nucleophiles 206 7.2.4 Theoretical Studies and Practical Applications 206 References 208 APPENDICES ix Appendix A Crystallographic Data and Atomic Coordinates for [Pd2(p:-CO)2](S03F)2 209 Appendix B Crystallographic Data and Atomic Coordinates for cw-Pd(CO)2(S03F)2 211 Appendix C Crystallographic Data and Atomic Coordinates for mer-Ir(CO)3(S03F)3 213 Appendix D Crystallographic Data and Structural Parameters for [Ir(CO)5Cl][Sb2Fu]2 215 Appendix E Selected Structural Parameters and the Molecular Structure of [Hg(CO)2][Sb2F„]2 222 x LIST OF TABLES Table 1-1 Selected Physical, Structural and Spectroscopic Properties of Carbon Monoxide 3 Table 1-2 Spectroscopic Criteria for Synergic Metal-CO Bonding in Metal Carbonyls 8 Table 1-3 Highly Reduced Carbonylmetallates 11 Table 1-4 Examples of Mononuclear Homoleptic Metal Carbonyl Species 16 Table 1-5 Selected Carbonyl Derivatives of Group 10 and 11 Metals and Their Spectroscopic Properties 20 Table 2-1 Chemicals Used Without Further Purification 46 Table 3-1 Bond Lengths [A] For [Pd2(/i-CO)2](S03F)2 71 Table 3-2 Bond Angles [°] for [Pd2(/i-CO)2](S03F)2 72 Table 3-3 Relevant Bond Distances and Angles for Various Fluorosulfate Derivatives 75 Table 3-4 Selected Bond Parameters for the Cyclic [Pd2(Lt-CO)2]2 + Cation 76 Table 3-5 Vibrational Spectra Data for [Pd20i-CO)2](SO3F)n and (CH3)2Sn(S03F)2 81 Table 3-6 The Vibrational Data (cm1) for cw-Pd(CO)2(S03F)2, cw-Pt(CO)2(S03F)2, and Au(CO)S03F with Estimated Band Intensities and Approximate Assignments 89 Table 3-7 Selected Intramolecular Distances (A) and Angles (°) for cw-Pd(CO)2(S03F)2 at 220K 95 Table 3-8 CO Bond Distances and Average CO Stretching Frequencies for Cationic Metal Carbonyl Derivatives 102 Table 3-9 Vibrational Data for [Pd(CO)4][Sb2Fn]2 and [Pt(CO) 4][Sb 2F„] 2 103 Table 3-10 CO and CN Stretching Vibrations and Force Constants For Selected Noble Metal Carbonyl and Cyanide Complexes 106 Table 4-1 Selected Intramolecular Distances (A) and Angles (°) for mer-Ir(CO)3(S03F)3 at 200 K 125 Table 4-2 Vibrational Data for mer-Ir(CO)3(S03F)3 and cw-Pt(CO)2(S03F)2 130 Table 5-1 Binary Metal Carbonyls of Group 8 147 xi Table 5-2 Vibrational Data for Solid [M(12CO)6][Sb2Fn]2 and [M( 1 3CO) 6][Sb 2F„] 2, M = Ru or Os 158 Table 5-3 Vibrational Data (v(CO) region) and 1 3 C Chemical Shifts for Octahedral, Isoelectronic [M(CO)6] Species 160 Table 6-1 Vibrational Data for Solid c-[Pd2(u.-CO)2][Sb2F11]2 185 Table 7-1 Thermally Stable (Persistent) Cationic Carbonyl Derivatives 195 Table 7-2 Homoleptic Metal Carbonyl Cations 197 Table 7-3 Vibrational Data (v(CO) region), Stretching Force Constants ft, and 1 3 C Chemical Shifts for Homoleptic, Isoelectronic and Octahedral M(CO)6 Species of the 5d Block Metals 198 xii LIST OF FIGURES Figure 1-1 MO scheme for carbon monoxide 5 Figure 2-1 Storage ampoule for SbF5 47 Figure 2-2 Schematic diagram of the preparation of bis(fluorosulfuryl) peroxide (S206F2) 49 Figure 2-3 The reactor used in the preparation of bis(fluorosulfuryl) peroxide (S206F2) 50 Figure 2-4 Typical apparatus for the preparation of cationic metal carbonyl derivatives 53 Figure 2-5 One-body vacuum filtration apparatus 55 Figure 3-1 Stereoview of a single layer of the polymeric sheet structure of [Pd2(/x-CO)2](S03F)2 73 Figure 3-2 Perspective view of [Pd2(/i-CO)2](S03F)2 73 Figure 3-3 FT-Raman spectrum of [Pd2(/*-CO)2](S03F)2 82 Figure 3-4 Fundamental vibrations of [Pd2(/x-CO)2]2+ 86 Figure 3-5 The vibrational spectra of m-Pd(CO)2(S03F)2 91 Figure 3-6 The molecular structure of cw-Pd(CO)2(S03F)2 96 Figure 3-7 The stereo diagram of the crystal structure of d5-Pd(CO)2(S03F)2 96 Figure 4-1 Two-part reactor designed and used for the synthesis of Ir(S03F)3 and Ir(CO)3(S03F)3 116 Figure 4-2 The molecular structure of mer- Ir(CO)3(S03F)3 124 Figure 4-3 Stereoscopic view of four complete molecules of Ir(CO)3(S03F)3 with secondary interactions shown by thin lines 128 Figure 4-4 The FT-IR spectrum of mer-Ir(CO)3(S03F)3 132 Figure 4-5 The FT-Raman spectrum of Ir(CO)3(S03F)3 in the CO-stretching region for crystalline mer-Ir(CO)3(S03F)3 and a solution spectrum of the mother liquor 133 Figure 4-6 The molecular structure of [Ir(CO)5Cl][Sb2Fn]2 139 Figure 7-1 The distribution of metal carbonyl species in the periodic table 199 Figure 7-2 Three types of metal-CO bonding and approximate ranges of v (CO) 201 xiii LIST OF SYMBOLS AND ABBREVIATIONS A angstrom unit, 10"1 m as, asym asymmetric or antisymmetric av average b, br broad (in vibrational spectroscopy) b bridging (in molecular structure) c speed of light in vacuum, 2.997925 x 1010 cm s" c cyclo calc. calculated d day(s) dp depolarized fr stretching force constant, N m"1 h hour(s) int. intensity IR infra red m medium (intensity) NMR nuclear magnetic resonance p polarized R Raman Ref. reference(s) s strong (intensity) solv. solvated xiv sh shoulder sym symmetrical t terminal v very w weak (intensity) v vibrational frequency in wavenumbers (cm1) co vibrational frequency in wavenumbers corrected for unharmonicity (cm"1) 8 vibrational deformation mode chemical shift in ppm (NMR) \i bridging ligand (in molecular structure) reduced mass (in vibrational spectroscopy) v vibrational stretching mode A v Raman shift (cm"1) p vibrational rocking mode xv ACKNOWLEDGEMENTS First, I would like to express my appreciation and gratitude for Professor Friedhelm Aubke, my research supervisor, for a rewarding experience under his guidance. My gratitude and thanks also extend to the members of my Guidance Committee, Professors A. Bree, C. Orvig, and M. Tanner who provided many helpful suggestions and constructive criticism. Professors A. Storr and R. C. Thompson are thanked for providing many types of assistance during my years at UBC. Dr. S. J. Rettig and Professor J. Trotter of this Department and Dr. R. J. Batchelor and Professor F. W. B. Einstein at Simon Fraser University are thanked for their help in crystal structure determination. Professor H. Willner and his research group at der Universitat Hannover, Germany, are thanked for their assistance in recording solid state 1 3 C NMR spectra and some FT-Raman spectra. I am indebted to Mr. Andrew R. Lewis for reading the entire thesis and the many enlightening suggestions he made. Many thanks are also due to my present and former co-workers Dr. Dingliang Zhang, Sun C. Siu, Germaine Hwang, Dr. Fred Mistry, and Dr. Shah Roshan Cader for their pleasant friendship, close cooperation, and enlightening discussions during my years of graduate studies. Dr. Mike Logan, Jim Sawada, Chris Simpson, Dave Summers, Shihua Xia, and Dr. Guowei Wei are thanked for their assistance in proofreading parts of this thesis. The exceptional service staff of this Department are too numerous to list all, but a few deserve special mention: Liane Darge and Marietta Austria (NMR services); Tom Markus, Roily Chan, and James Adair (Electronics); Ron Marwick, Bill Henderson, and Brin Powell (Mech. Shop); E. Varty and S. Rollinson (Chern. Illustration); Tilly Schreinders, Irene Rodway, Lani Collins, Bev Evans, and Sheri Harbour (Main Office). Mr. Steve Rak, the glass blower, and Mr. Peter Borda, the microanalyst, are specially thanked for their excellent professional expertise. Ms. Carolyn Joyce is thanked for her great assistance in many ways. Finally I would like to express my gratitude and thankfulness to my family for all their support, understanding, and encouragement during my years of graduate study. xvi CHAPTER 1 GENERAL INTRODUCTION 1.1 Metal-CO Bonding — a Descriptive Overview The chemistry of transition metal carbonyl derivatives is of both academic and practical interest. Numerous reactions and processes involving transition metal carbonyls, either as reactants or as catalysts, are presently known and in use on the laboratory as well as industrial scales. Each year, new metal carbonyl compounds are synthesized and increased applications are developed. Recent development in the field of metal carbonyl chemistry reveals that some metal carbonyl complexes have bonding features hot previously recognized. Therefore, a brief overview of CO as a ligand in coordination chemistry seems appropriate at the beginning of this chapter. Depending on the nature and electronic structure of the transition metal involved, two principal processes, CO-to-metal a-donation and metal-to-CO rc-back-donation, may be involved. The overall "synergic" process results in a strengthening of the metal-carbon bond and appears to be responsible for the thermal stabilities of many transition-metal carbonyl complexes. The poor donor ability of CO and the resulting emphasis on the 7t-acceptor ability in these compounds has led to an interesting suggestion by Pearson1 that in these complexes CO acts predominantly as a Lewis acid. There have recently been two developments which appear to depart from the synergic 1 bonding model: (i) In basic media and with the aid of powerful reducing agents such as group 1 metals, highly reduced carbonyl anions form where CO binds predominantly via rc-back-donation (vide infra) to the metal center. The metal centers are found in Groups 4 (Ti, Zr, Hf) to 8 (Co, Rh, Ir) and include most of the early transition metals. The oxidation states of the metals and hence the anion charges range from -2 to -4. (ii) In stark contrast, in highly acidic media a number of carbonyl cations or cationic derivatives form, which involve electron-rich metals in Groups 10 (Pt) and 11 (Ag, Au) in oxidation states of +2 and +1. The known examples include the first metal carbonyls 2 3fl reported by Schiitzenberger. ' Evidence suggests that bonding modes in these compounds are predominantly, but not exclusively, o-donation from CO to the metal cation. The latter group will be the subject of this thesis. Hence, from the above brief description, CO emerges as a ligand of great versatility. The reason for this will now be discussed. 1.2 Carbon Monoxide, C O , as a Potential Ligand Carbon monoxide, as the physical properties in Table 1-1 indicate, is a gas at ambient temperature. Hence the reactions of metals or metal ions with CO are largely heterogeneous and frequently require high pressures and temperatures because they involve a reduction in volume and have high activation energies. 2 Table 1-1 Selected Physical, Structural and Spectroscopic Properties of Carbon Monoxide' Physical properties: m.p. (°C) -199, b.p. (°C) at 1 atm. -191.5 Dipole moment n (D): 0.112 Structural properties: Internuclear distance (A): CO(g) 1.12819, CO+(g) 1.11506 van der Waals radii according to Bondi (A):b C 1.70, O 1.52, Sum of C+O 3.22 Covalent single bond radii (A): C, 0.77; O, 0.66 13C-NMR chemical shift 8: 184 ppm Vibrational data: Stretching frequency (cm"1) v 1 2 C 1 6 0 2143.16 1 3 C i e O 2096.07 [ 1 2 C 1 6 0] + 2184 a 1 2 C 1 6 0 2170.21 , 3 C 1 6 0 2121.41 [ 1 2 C 1 6 0] + 2214.41 Point group C„v Electronic structure: Electronic groundstate: 1 Z + Electronic configuration: (la)2(2a)2(3a)2(4o)2(l7t)4(5cr)2 First ionization energy (eV): 14.014 Electron affinity (eV): -1.8 8The data are taken from standard references such as: Gmelin's Handbuch der Anorganischen Chemie; 8th ed.; Springer-Verlag: Berlin, Carbon and its Compounds, 1970; pp 115-125, and Herzberg, G. Spectra of Diatomic Molecules; 2nd ed.; Van Nostrand: Toronto, 1966; pp 520-522. b Bondi, A. J. Phys. Chern. 1964, 68, 441. 3 The dipole moment of CO is rather small as a consequence of its electronic structure (vide infra). The polarity C 8 *— O8" suggested by the electronegativities of carbon and oxygen is most probably not correct and should, based on the experimental results3b'c and theoretical calculations,3d rather be C8" — (f*. Formation of a coordinate bond using a-donation only from the lone pair on carbon would result in a reversal of charges for CO to C 8 *— O5". Therefore, a juxtaposition of two positively charged atoms such as M n + <- C 8 *— O8" is anticipated. This situation is relaxed in typical transition metal carbonyls involving zero- or low-valent metals by rc-back-donation, which provides for intra-molecular charge compensation. Judging by the short C — O distance of 1.12819 A, the CO bond is best viewed as a triple bond, the canonical resonance form Ic'^ O"1"1! making a major contribution to the overall bonding in free carbon monoxide. A simple molecular orbital approach suggests a bond order of three for CO. As seen in Figure 1-1, the HOMO orbital of CO is the 5c MO, a carbon-based orbital that is predominantly non-bonding although slightly antibonding. The antibonding nature of the 5a orbital is apparent from the decrease in the internuclear distance and the increase in the stretching frequency when going from CO to C 0 + . The LUMOs of CO are the 2K molecular orbitals which are antibonding and serve as acceptor orbitals for filled metal d-orbitals of suitable symmetry. The Jt-back-bonding will result in shortened M-C and lengthened C-0 bonds. 4 6d /'< 2TI 1 1 / \ \ \ ' '1 M L' 1 ,V\ LA At A V, / , ' — 2 s 26 1s-•1s \ 16" / Figure 1-1 MO scheme for carbon monoxide (taken from Ref. 5). Although changes in the vibrational frequency, v(CO), on bonding to a metal are large, the lengthening of the CO bond distance is very slight. It becomes apparent that in case of high C-O bond order (approximately 3), vibrational spectroscopy will provide a clearer picture of small changes in the CO bond order than does X-ray crystallography. In addition, limits in accuracy, expressed in terms of estimated standard deviations (ESDs), the thermal motion of oxygen, and weak secondary contacts in the solid state tend to obscure the CO bond lengths obtained via X-ray analysis. Another advantage of vibrational spectroscopy is that in most CO derivatives the CO stretching vibrations fall into an uncluttered spectral region, where overlap with most other bands except for C = N or C = C vibrations is unlikely. However, vibrational mixing (e.g. with CH or BH) and Fermi resonance can still present 5 problems that may often be overcome by a complete vibrational assignment, a normal coordinate analysis, force-constant calculations, and a comparison of observed and calculated frequencies. Isotopic substitutions of 1 3 C for 1 2 C and 1 80 for 1 60 are of enormous use in a 13 careful vibrational analysis, as are C-NMR measurements on isotopically enriched samples. In general, for terminal CO groups in transition metal carbonyls, the CO bond distances are approximately 1.15 A, with little variation, while v(CO) ranges from 2125 to 1850 cm"1. Finally, consistent with the overall view of CO as a very poor Lewis base, the first ionization energy of CO is, at 14.014 eV, much higher than that found for many molecular N-and O-donor systems that will, where present, compete with CO for coordination sites. In summary, the review of physical, spectroscopic, and structural parameters of CO allows the conclusion that chances for the use of CO as a Lewis base on a synthetic scale and at ambient temperatures do not seem to be promising. The observation that CO forms numerous thermally stable compounds with transition metals is mainly due to its ability to engage in synergic bonding, which will be discussed in the next section. 1.3 Synergic Bonding of CO in Typical Metal Carbonyl Compounds Carbon monoxide, CO, is the most important and extensively studied ligand in transition metal chemistry, and carbonyl derivatives of one type or another are now known for all transition metals.4"6 This may seem surprising, because CO is a very poor Lewis base or a-electron pair donor. However, the ability of carbon monoxide to function as a it-acceptor or rc-acid ligand results in synergic bonding, where 7t-back-donation from filled d-orbitals of the 6 metal to the 7c*-molecular orbitals of CO increases its basicity and the ability to function as a a-donor towards the metal. The weakening of the CO bond is most readily detected by vibrational spectroscopy,7"8 where v(CO), the stretching frequency, is found to be very sensitive to changes in the CO bond order caused by rc-back-donation. Shifts to lower wavenumbers from the value of v(CO)(g) at 2143 cm"1 9 are used not only to detect and estimate the extent of synergic bonding, but also to assign coordination modes of the carbonyl ligands: terminal monodentate CO groups are usually found to absorb between 2125 and 1850 cm"1, while bridging bidentate groups have v(CO) values between 1860 and 1700 cm"1,4'7'8 as is seen in Table 1-2. Metal-carbon vibrations are less easily detected and their shifts are not readily analyzed because they are usually found in a more "cluttered" region of the vibrational spectrum and the bands are occasionally of low intensity. Some attempts, however, have been made.10 1 3 C -NMR has become a valuable diagnostic tool,11 where chemical shifts are observed between 190 and 215 ppm, compared with 184 ppm for free CO. Some limited use has been made of X-ray photoelectron spectroscopy (XPS) where Cls and Ols binding energies shift to lower energies relative to free CO. These shifts in binding energy are related with the extent of metal-to-CO 7t-back-donation.12'13 In addition to these experimental criteria, which are summarized in Table 1-2, the conditions for the formation of synergic metal-carbon bonds have become well established and are summarized briefly as follows. (i) The metal should have electrons in orbitals of suitable symmetry (d-orbitals) 7 Table 1-2 Spectroscopic Criteria for Synergic Metal-CO Bonding in Metal Carbonyls" Vibrational spectroscopy fr (N m"1)" 18.8-14.5 x 102 /r (Nm'1) 14.6-13.0 x 102 JJC-NMR 8 (ppm) (i) Terminal CO 185-200 (ii) Bridging bidentate CO 239-275 (i) Terminal CO vC-0 (cm ) 2125-1850 vM-C (cm1) 360-560 (ii) Bridging bidentate CO vC-O (cm1) 1860-1700 XPS first ionization energy (eV) 14.013(3) Binding energies (eV) (i) Cls 295-300 (ii) Ols 541-538 a See Ref. 4, 5, 7, and 8. b / r = force constant, see Harris D. C ; Bertolucci, M. D. Symmetry and Spectroscopy; Oxford: UK, 1978; pp 104, 105, 108, and 109. available for rc-back-donation. Suitable metals are found in the middle section of the 3 d-series (Groups 6 to 9) metals, and should be neutral or in low oxidation states. Positively charged metals are much less likely to engage in rc-back-donation and very few metal carbonyl cations are known.14 On the other hand, synergic bonding is not very effective in metal carbonyl anions either because the metals are now poor a-acceptors. 8 (ii) The ready formation of supported and unsupported metal-metal bonds will widen the scope from mono-nuclear to polynuclear complexes or metal atom cluster carbonyls with the metal again in a low overall oxidation state. (iii) Of the additional ligands, better Jt-acceptor ligands will labilize the metal-CO bond and will often replace CO in substitution reactions whereas strong electron-donor ligands will strengthen the metal-carbon bond. Among the halides, a very common substitutent in metal-carbonyl derivatives, strongly electron-withdrawing halides decrease the ability of the metal center to engage in 7t-back-donation, while good rc-donors will enhance the scope of derivatives and their thermal stabilities. For metal-carbonyl halides the general order of stability is found to be I > Br > CI, with metal carbonyl fluorides being rather uncommon.4,5 (iv) The effective atomic number rule (EAN) is generally obeyed. This rule is used to judge chemical reactivity (e.g. oxidizing or reducing ability) and the thermal stability of transition-metal carbonyls. It also permits one to deduce the presence or absence of metal-metal bonds in polynuclear complexes, where nuclear structures are either not available or are difficult to interpret owing to bridging ligands. The best known exception among mono-nuclear carbonyls is V(CO)6, which has 17 electrons associated with the central metal. In addition to the mono- and poly-nuclear, and homo- and hetero-leptic metal carbonyl derivatives summarized above, the essential features of synergic metal-CO bonding appear also to apply to two other additional, interrelated groups: (i) CO adsorbates on transition-metal surfaces,15,16 and (ii) the products of the reaction of metal atoms with carbon monoxide, studied by matrix-isolation techniques17 in cases where no thermally stable metal carbonyls are 9 known. However, the synergic bonding model does not seem to apply well to the highly reduced carbonylate anions such as [M(CO)5]3",18 M = Nb or Ta, or [Ti(CO)6]2" 1 9 since the high negative charge on the metal will not accept a-donation easily. For metal-CO adsorbates, v (CO) for monodentate terminal groups are found between 2130 and 2000 cm' 1, 1 5 ' 1 6 with lower v(CO) values attributed to various bridging configurations.15 Of the thermally unstable binary carbonyls studied by matrix isolation, three examples of particular relevance to this thesis may be mentioned: tetrahedral Pd(CO)4,20 Pt(CO)4,21 and linear Au(CO) 2. 2 2 The tetrahedral tetracarbonyls of palladium and platinum extend the series in Group 10 from Ni(CO)4, the first binary carbonyl reported,23 downwards. All of these molecules are observed together with smaller fragments of the type M(CO)n, M = Pd or Pt and n = 1, 2 or 3, and the evidence of bonding of CO to gold via oxygen is also obtained from vibrational spectra. In all instances of matrix-isolated carbonyls the CO stretching frequencies are observed well below 2143 cm"1, the value found for gaseous C O . 9 Alkali metal reduction in liquid ammonia produces isolable, thermally stable, but occasionally shock-sensitive, carbonylate anions with the metal in very high negative formal oxidation states such as -3 or -2, 1 8 , 1 9 , 2 4 , 2 5 termed highly reduced carbonylmetallates.x%* High formal negative charge on the metal favors rc-back-donation and v (CO) may be found between 1 18 1800-1500 cm" . Conventional geometries (octahedral, trigonal bipyramidal, or trigonal planar) appear to be present. In all instances the EAN rule is obeyed. The known carbonylate anions are listed in Table 1-3. 10 Table 1-3 Highly Reduced Carbonylmetallates * [Ti(CO) 6f [V(CO) 5f [CrCCO)^4- [Mn(CO)4]3' [Co(CO)3f [Zr(CO)6f [Nb(CO)5f [Mo(CO)4]4" [Rh(CO)3f [Hf(CO)6]2" rTa(CO)5]3- [W(CO)4f [Re(CO)4f [Ir(CO)3]3" *SeeRef. 18, 19, 24, and 25. Of these three groups, the highly reduced carbonylmetallates are useful reagents in organometallic chemistry, while the metal CO adsorbates on metallic surfaces play an important role both in surface chemistry, the characterization of heterogeneous catalysts, and in the elucidation of heterogeneously catalyzed carbonylation processes.15'16 The matrix-isolated binary metal carbonyls, owing to their low thermal stabilities, have not progressed markedly beyond mere laboratory curiosities. The extensive studies17 have not yet revealed any new bonding modes or unexpected coordination geometries. In summary, metal-to-carbon rc-back-donation, as an important component of synergic bonding, plays an essential role in numerous transition metal carbonyl derivatives, including CO-adsorbates and matrix-isolated metal-carbonyl fragments. It is for this reason that CO is ranked on a par with or even superior to the isoelectronic cyanide ion, CN", in the spectrochemical series as a strong field ligand.26 However, thermally 11 stable binary carbonyls are largely confined to Groups 5-10 while binary metal-cyanide derivatives are known for all transition metals, including some of the lanthanides and 27 28 actinides. ' In addition, stable cyanide complexes form also in cases where conditions are not conducive to rc-back-donation (e.g. metals in high oxidation states or with d° configurations). Hence , it is apparent that the cyanide ion is an excellent a-donor ligand and does not have to rely on rc-back-bonding to form stable metal complexes with metals in relatively high oxidation states. These conclusions have also been reached by comparative MO 29 calculations. The question now is whether and under what conditions CO can function predominantly as a a-donor ligand and form thermally stable metal carbonyls where conditions to form synergic bonding are not favorable. 1.4 Carbon Monoxide as a a-Donor Ligand in Inorganic Compounds There are two typical examples where CO functions largely as a a-donor ligand: (i) compounds obtained by the interaction of CO with neutral, molecular Lewis acids where the elements do not have d-electrons for rc-back-donation and still form sufficiently stable adducts with the Lewis base CO; (ii) compounds derived from metal and non-metal cations that cannot or may not engage in rc-back-bonding. 1.4.1 CO Adducts with Molecular Boron Lewis Acids In the first group, three 1:1 CO adducts with BH 3 , B 4 H 8 , and BF 3 are known and merit some consideration. The gaseous adduct BH 3CO was first reported in 1937.30 Its molecular structure was determined by electron diffraction31 and microwave spectroscopy,32 and while 12 v(CO) is, at 2169 cm"1, 3 2 3 5 higher than 2143 cm"1 for gaseous CO, it appears from a subsequent force-field study36 using the Wilson FG-matrix method36" that the high band frequency is in part due to force-field mixing with BH 3 vibrations. The stretching force constant/, obtained is 18.05 x 102 N m"1, lower than that of free CO (18.6 x 102 N m"1) and the calculated stretching frequencies are between 2101 and 2107 cm"1 for the 1 0B, n B , ! H and 2 D isotopomers. The bonding in BH 3CO and other BH 3 adducts has been studied by photoelectron spectroscopy,37 and it is concluded that -^back-donation from a filled BH 3 molecular orbital (the HOMO) to the %* MO of CO (the LUMO) results in a significant stabilization of the 38 BH 3CO as previously suggested. In addition, configurational mixing of the 4a and 5a MOs results in a strengthening of the C-O bond. Hence it appears that 7C-bonding plays a role in BH 3CO. However, the B-C bond remains weak with an estimated bond energy of ca. 80 kJ ,-1 38 mol . The chemistry of BH 3CO provides examples of nucleophilic attack either on B or on C, e.g. trimethylamine will displace CO to give the adduct H 3BN(CH 3) 3 , 3 0 while NH 3 , CH 3 NH 2 , and (CH3)2NH all form 2:1 adducts with BH 3CO 3 9 where v(CO) is lowered to about 1640 cm"1. The adduct with methylamine is formulated as an ionic solid of the type [H 3NCH 3] +[H 3BC(0)NHCH 3]' 3 9 ' 4 0 and the crystal structure of this salt has been obtained.40 BH 3CO is relatively stable with respect to dissociation into CO and BH 3 or B 2 H 6 . This is not the case for BF 3 • CO, which according to a study41 by molecular beam resonance 13 spectroscopy is a van der Waals molecule with a B-C bond length of 2.886 A, compared with 1.540 A i n B H 3 C 0 . 3 2 1.4.2 Carbonyl Cations By far the simplest cation and the strongest Lewis acid is the free proton H + . Its adduct with CO, the formyl cation has been suggested as a primary ionic product in the combustion of hydrocarbons. This adduct has also been observed in interstellar molecular clouds,42 and its microwave spectrum is known.43 In addition, the formyl cation is postulated as a reactive intermediate in many acid- and superacid-catalyzed formylation reactions with C O . 4 4 Attempts to observe the cation by NMR spectroscopy under stable ion conditions using 13C-enriched CO in superacid solution have been unsuccessful, due to rapid proton exchange even at low temperatures.44 However matrix-isolation experiments have allowed detection of weakly bound molecular complexes of the type OC - H F and CO ••• HF, with the former more stable.45 Even though the microwave spectrum of H C O + was observed in 1975 for the first time in the laboratory,43b recording the complete vibrational spectrum became a protracted process.46-48 By now all three fundamentals are known, and force-field calculations have been reported recently.49 While v(CO) is, at 2184 cm - 1 , 4 6 well above the value for v(CO)(g) of 2143 cm"1,9 the stretching force constant of 21.3 x 102 N m"1 is the highest reported so far.49 The relatively low v(CO) value of 2184 cm"1 is seemingly caused by vibrational mixing between v (CO) and v (CH). It appears that the stretching force constant of v (CO) for H C O + represents the upper limit for a complex ion with solely cr-bonded CO in the complete absence of rc-back-donation, and the value obtained49 may be used as a benchmark in judging the extent 14 of rc-back-donation in other complexes. In addition, H C O + has played an important role in recent developments of carbonyl coordination chemistry. An attempt to protonate carbon monoxide in the superacid system HS03F-Au(S03F)350'51 resulted in the discovery of a thermally stable metal carbonyl cation [Au(CO) 2] +. 5 2 1.4.3 Metal Carbonyl Derivatives As already discussed in previous sections, carbon monoxide, CO, can function both as a a-donor and 7t-acceptor ligand. In favorable cases, as in low valent metal carbonyls, these two bonding modes reinforce each other, and hence lead to strong metal-carbon bond at the expense of the carbon-oxygen bond. Therefore, the CO stretching frequency is reduced relative to that of free CO at 2143 cm"1. The formation of most metal carbonyl compounds rely on this synergic bonding mechanism. In Table 1-4 the known mono-nuclear metal carbonyls are listed. These compounds manifest the synergic bonding mechanism. Their molecular geometries are the usual octahedral, trigonal bipyramidal, and tetrahedral configurations. The EAN rule is obeyed by these compounds with the exception of V(CO)6, which is a 17 e" species. Oxidation states of the metal are most often zero, sometimes -1 and more rarely +1. Of the latter group only [M(CO)6]+, M = Mn, Tc, or Re, are known with [A1C14]" or [FeCLJ as counteranions. 15 Table 1-4 Examples of Mononuclear Homoleptic Metal Carbonyl Species' n Structure 6 Octahedral 5 Trig, bipyramidal 4 Tetrahedral M(CO) n + M = Mn, Tc, Re M(CO)n M = V, Cr, Mo, W M = Fe, Ru, Os M = Ni (Pd, Pt)* M(CO)n" M = V, Nb, Ta M = Mn, Tc, Re M = Co ' Ref. 5. b Abel, E. W.; Tyfield, S. P. Adv. Organomet. Chern. 1970, 8, 117. * M = Pd or Pt are matrix isolated molecules. Only a small number of metal carbonyl derivatives belong to a second group, in which Jt-back-donation seems to be insignificant in the formation of metal-carbon bond. The known examples include metal carbonyl cations as well as metal carbonyl halides. While noble-metal carbonyl cations were only discovered very recently, the cationic halide complexes have a very long history. The discovery and separation of the three platinum(II) carbonyl chlorides, Pt(CO)2Cl2, Pt2(CO)2Cl4 and Pt2(CO)3Cl4 by P. Schutzenberger in 1868-18702'3" predate the first synthesis 23 of nickel tetracarbonyl, Ni(CO)4, by 22 years and mark the beginning of metal-carbonyl chemistry. A subsequent report of the corresponding three Pd(II) carbonyl chlorides53 in 1898 was later found to be erroneous by Manchot and Konig.54 A year later the same authors 16 (Manchot and Konig) reported the first synthesis of a palladiumfll) carbonyl derivative, Pd(CO)Cl2.55 The nearly simultaneous discovery of gold(I) carbonyl chloride, Au(CO)Cl, by Manchot and Gall, 5 6 and Kharash and Isbell57 occurred in 1925. It is interesting to note that Au(CO)Cl remained for about 65 years the only mononuclear carbonyl derivative of gold. The principal method of synthesis involved reductive carbonylation of gold(III) chloride. Manchot and Konig also pioneered the use of a strong protonic acid, concentrated 58 H 2 S0 4 , in an unsuccessful attempt to obtain a silver(I) carbonyl derivative. A material of the composition Ag 2S0 4- CO, produced in situ, was found to be thermally unstable. Early postulates regarding the existence of copper carbonyl species also go back to early times and concern two separate areas: (i) the transport of metallic copper in a CO atmosphere59 and (ii) the absorption of gaseous CO by solution of Cu(I) in various solvents60,61 at a 1:1 stoichiometric ratio (e.g. in aqueous hydrochloric acid). The isolation and structural 62 characterization of Cu(CO)Cl, however, is very recent. As is apparent from this brief summary, the roots of cationic metal carbonyls can be 63 traced well back into the last century. Bruce has reviewed early work involving Group 11 metals (Cu, Ag and Au), and reviews on carbonyls of platinum metals (Rh, Ir, Pd and Pt)64 and on palladium carbonyls65 are available. However, the unusual nature of the metal-carbon bond was not always recognized in all those early studies, and the next section discusses the use of spectroscopic and structural techniques. 17 1.5 Spectroscopic and Structural Studies on Noble-Metal Carbonyl Derivatives The unusual nature of the metal ion-CO bond in noble metal carbonyl derivatives was first revealed in the study of the vibrational and 1 3C-NMR spectra66 of Au, Pt, and Pd carbonyl halides. The implication that the metal ion to CO 7c-back-donation was substantially reduced was soon recognized by Calderazzo and coworkers.67"69 These researchers made extensive use of aprotic solvents such as SOCl2 both in the synthesis and in the spectroscopic solution studies of noble-metal carbonyl chlorides. In thionyl chloride, CO addition to metal halides at relatively mild conditions and low CO pressures is possible70'71 to produce poly- or mono-nuclear metal carbonyl chlorides: 2PdCl2 + 2CO > Pd2(CO)2Cl4 (1-1) or PtCl2 + 2CO > cw-Pt(CO)2Cl2 (1-2) The conversion of the thermodynamically more stable c«-Pt(CO)2Cl2 to the trans-isomer via monomer-dimer equilibria of types (1-3) and (1-4) in S0C1 2, 7 3' 7 4 is observed. 2 cw-Pt(CO)2X2 > Pt2(C0)2X4 + 2CO (1-3) Pt2(CO)2X4 + 2CO >2 mww-Pt(CO)2X2 X = Cl, Br or I (1-4) Also, the reductive carbonylations of higher valent metal halides such as AuCl 3 7 5 or PtCl 4 7 2 b are now easily accomplished, via e.g. Equation 1-575 18 AuCl 3 + 2C0 > Au(CO)Cl +C0C12 (1-5) Finally, addition of AuCl 3 to Au(CO)Cl according to76 AuCl 3 + Au(CO)Cl > Au2(CO)Cl4 (1-6) afforded the first binuclear carbonyl derivative of gold with v (CO) at 2180 cm'1. In addition to interesting synthetic procedures, the use of SOCl 2 and similar solvents allows various spectroscopic (NMR, vibrational) and calorimetric measurements in solution.67"69 The spectroscopic results are summarized in Table 1-5. It is noteworthy that a rather poor donor solvent like SOCl2 is so useful in these studies. However, the use of SOCl2 and similar aprotic solvents combined with the availability of metal chlorides as suitable starting materials may be seen as a limitation of research in this area to the study of metal-carbonyl chlorides. The extensive studies,67"69 nevertheless, resulted in the synthesis and characterization of a substantial number of mononuclear and binuclear, neutral or anionic metal-carbonyl derivatives of Pt, Pd, and Au. In addition to CO addition, an interesting new synthetic route to Pt(CO)2Cl2 via the reaction of Pt atoms with oxalyl chloride C 20 2C1 2 has been reported,77 but this approach does not widen the scope of known carbonyl derivatives of the noble metals. The general order of thermal stability of the M(CO)„Xm derivatives with M = Pt, Pd or Au and X = CI, Br or I, is found to be CI > Br > I,69'78 which is a reversal of the order observed for the typical transition-metal carbonyl halides, as discussed previously. The reversed stability order for noble-metal carbonyl halides 19 Table 1-5 Selected Carbonyl Derivatives of Group 10 and 11 Metals and Their Spectroscopic Properties Compound vCCO^Ccm 1) 8 81 3C-NMR (ppm) Referenceb c«-Pt(CO)2Cl2 2157.5 151.6/( ,95Pt-13C 1576 Hz) 66 (2164.5) c«-Pt(CO)2Br2 2143 152.0 7(195Pt-13C 1566 Hz) 66 (2153) Au(CO)Cl 2162 170.8 66 (2183) [Pt(CO)Cl5r 2191 80 Au2(CO)Cl4 2180 170.8 76b Pd2(CO)2Cl4 2166 70, 71 Pt2(CO)2Cl4 2146 80 [Cu(CO)][AsF6] 2180 82 CO on ZnO, 77 K 2169-2178 83 The wave numbers in brackets refers to Raman shifts of solid samples, while IR spectra are obtained in solutions of aprotic solvents. b The reference refers to the spectroscopic data not to the synthesis. 20 is best illustrated by the known gold(I) complexes. As discussed, Au(CO)Cl ' has been know for almost 70 years, while Au(CO)Br is thermally unstable and is only characterized in 76 cyclohexene solution, with Au(CO)I so far unknown. Other limitations are posed by the low ionizing ability of thionylchloride, SOCl 2, and similar aprotic solvents used in the past, and the relatively high nucleophilicity of the halide ligand. As a consequence, cations of the noble-metal carbonyls are virtually unknown. Where molecular structures have become known, as for Au(CO)Cl,7 9 which is a four-atomic linear monomer, or for Cu(CO)Cl,62 a chloride-bridged polymer, or Pt2(CO)2I4,80 carbonyl cations are not present. Only in two instances, both involving fluorine-containing anions ([Cu(CO)][CF3C02]81 and [Cu(CO)][AsF6]82), has the presence of [Cu(CO)]+ been suggested, based solely on spectroscopic evidence. For the hexafluoroarsenate a rather high v(CO) 1 82 — value of 2180 cm" is reported, and it is interesting to note that similarly high v(CO) values are reported for CO adsorbates on solids containing metal cations, as for example in ZnO, where v(CO) depends on the extent of surface coverage by CO; the value ranges between 2169 and 2177 cm" according to a study at 77 K. Similar vibrational spectra have been reported for CO adsorbates on cuprous fluoride and on Cu0. 8 4 a In a different approach, CO complexes with monomeric nickel(II) halides were studied by matrix isolation84b in an attempt to extend the field of Group 10 metal-carbonyl halides67"69 to nickel. For a complex formulated as NiF 2- CO, v(CO) was found at 2200 cm"1, the highest reported CO stretching frequency at that time. There is also a theoretical study of the bonding in CO adducts to nickel ions N i 2 + and N i 3 + . 8 4 c 21 These examples make a simple point: changing the metal surface ' to a metal salt reduces rc-back-bonding in CO adsorbates as well as in matrix-isolated metal-chloride CO 17 20 22 complexes, when compared with metal atoms. ' ' This is manifested by the shifts of v (CO) to higher wavenumbers with respect to free CO. In summary, a number of Group 10 and 11 metal-carbonyl derivatives including CO adsorbates on metal salts83'848 and matrix species84b are known, where vibrational spectra as well as 1 3C-NMR spectra indicate reduced rc-back-bonding. In addition, some of the descriptive features of these compounds show differences from those of the typical transition-metal carbonyl compounds. The oxidation state of the metal ranges from +1 for Cu, Au, Pd, and Pt, to +4 in the case of the anion [Pt(CO)X5]_, X = Cl or Br, obtained by the oxidative O f addition of Cl 2 or Br2 to [Pt(CO)X3]\ Oligomerization occurs either through halide bridges or, in the case of some Pd(I) compounds, through bridging C O 6 4 ' 6 5 rather than via direct metal-metal bonds, and the EAN rule is frequently not obeyed. 1.6 Recent Developments of Cationic Metal Carbonyl Derivatives Strong protonic acids such as H 2S0 4 , HS0 3F, HS0 3CF 3 , HF, and B F 3 H 2 0 have been used in the past to generate and stabilize metal-carbonyl cations.58'86"91 Even though it has not been possible to isolate solid metal-carbonyl salts from the acid solutions, possibly owing to their low thermal stability, two interesting consequences of these studies may be mentioned, (i) It is reportedly possible to coordinate up to 4 moles CO per mole Cu(I) and up to 2 moles CO per mole A g + , depending on the temperature, CO pressure, and the acid chosen, (ii) CO uptake will decrease with increasing solvating tendency or "basicity" of the acid in the 22 following order: H 2 S0 4 > HS0 3F > HS0 3CF 3 > HF > B F 3 H 2 0 . Possibly on account of the thermal instability and concomitant high reactivity of the metal ion-CO solutions, extensive use has been made of them in carbonylation reactions of olefins, alcohols, and saturated hydrocarbons.91"94 The low thermal stabilities of the solvated silver(I) carbonyl cations must be seen as a principal reason for the failure to isolate solid silver(I) carbonyl compounds from their solutions in strong protonic acid. Recently, Strauss and coworkers95 succeeded in the isolation and characterization by X-ray diffraction of the first Ag(I) CO complex of the composition of Ag(CO)B(OTeF5)4. The formation reaction AgOTeF5 + B(OTeF5)3 + CO > Ag(CO)B(OTeF5)4 (1-7) is reversible. Low temperatures and high pressures favor CO addition, while at room temperature and low CO pressure the compound releases CO within a few hours to produce the starting materials AgOTeF5 and B(OTeF5)3. In spite of the reversible interaction of CO with A g + , the molecular structures of Ag(CO)B(OTeF5)495 and more recently of [Ag(CO)2][B(OTeF5)4]96 have been obtained at -125 and -100°C respectively, by single-crystal X-ray diffraction. Ag(CO)B(OTeF5)4 has a nearly linear AgCO group (<AgCO = 176(1)°). Silver is bonded to two oxygen atoms of the -OTeF5 groups. The reported bond distances, a long Ag-C bond of 2.10A and a short C-0 bond of 1.077(16)A, as well as the high v(CO) value of 2204 cm"1, are indicative of substantially reduced 7t-back-bonding. In [Ag(CO)2][B(OTeF5)4],96 a nearly linear [Ag(CO)J+ cation (initially identified in 23 acid solution ' ) is present, but the large estimated standard deviations do not permit a meaningful discussion of the reported C-0 bond distances. High v(CO) values of 2207 cm'1 for the IR-active mode for this compound, and similar high CO stretching frequencies observed in the IR spectra of other silver mono-carbonyl and di-carbonyl complexes with [Zn(OTeF5)4]2" as counteranions, are consistent with the presence of coordinated CO ligands with very little metal-to-CO rc-back-donation for these compounds. However a complete vibrational analysis is not provided by the authors. It appears that the use of large and weakly coordinating anions made up of highly electronegative -OTeF5 (teflate) groups allows the isolation and the structural characterization of various silver(I) carbonyl compounds. The previously noted58'86'87 thermal instability of silver(I) carbonyls still remains a problem. A different approach is chosen for the generation of the dicarbonyl gold(I) cation [Au(CO)J+ because suitable gold(I) compounds comparable to AgOTeF 5 9 7 are not 98 99 available ' and have to be generated in situ. An interesting approach is the oxidation of gold by UF 6 in anhydrous HF as solvent in the presence of carbon monoxide100 according to: Au +2CO + UF 6 > [Au(CO)2][UF6] (1-8) The low thermal stability of the reaction product permits the characterization of the [Au(CO)2]+ cation by IR spectroscopy only. The reduction of gold(III) compounds provides another route to [Au(CO)2]+.5 2 Again a protonic acid, HS0 3F, is employed as a solvent with CO as a reducing agent as well as a ligand: 24 Au(S03F)3 + 3C0 H S O i F ) [Au(CO)J+ + S03F" + C 0 2 + S 2 0 5 F 2 (1-9) The reductive carbonylation of Au(S03F)3 in HS0 3F proceeds similarly to the reduction of AuCl 3 in S0C1275 with two exceptions: (i) the phosgene analogue CO(S03F)2 expected during the reaction is thermally unstable and decomposes to C 0 2 and S 2 0 5 F 2 , 1 0 1 and (ii) the use of a strongly ionizing solvent such as HS0 3F allows the formation of the electrophilic [Au(CO)2]+ cation and its detection by IR, Raman, and 1 3C-NMR spectroscopy.49 Attempts to isolate an [Au(CO)2]+ fluorosulfate by removing the solvent and the volatile reaction products in vacuo were unsuccessful. It appears that the S03F" ion is sufficiently basic to compete successfully with CO for a coordination site on gold, and Au(CO)S03F is isolated in quantitative yield52 according to: [AufCO);,]-^ + S0 3F ~ H S O i F ) Au(CO)S03F + CO(g) (1-10) when the solvent is removed in vacuo. Gold(I) carbonyl fluorosulfate is thermally stable up to about 190°C, has a v(CO) of 2197 cm"1, and will ionize to give the [Au(CO)]+ ion in HS0 3F and magic acid (HS03F-SbF5).4 9 According to the vibrational spectra the OS02F group is weakly bonded to gold as a monodentate ligand. Interestingly, Au(CO)S03F is, in addition to Au(CO)Cl (reported about 65 years earlier), the only thermally stable mononuclear Au(CO) derivative. In contrast to the Ag(I)CO derivatives discussed above, the CO uptake is not reversible and the decomposition mode is still unclear. The unknown compound AuS03F is not obtained by the thermal decomposition of Au(CO)S03F. 25 When low CO pressure is employed, partial reduction of gold(III) fluorosulfate is 102 observed according to 2Au(S03F)3 + CO > 2Au(S03F)2 + C 0 2 + S 2 0 5 F 2 (1-11) This reduction by CO has a parallel in the formation of (AuCl2)4 from AuCl 3 . 1 0 3 The diamagnetic solid of composition Au(S03F)2 is best formulated as a mixed valence gold(I)-gold(III) fluorosulfate. The compound is sparingly soluble in HS0 3F. The resulting paramagnetic solutions contain the solvated A u 2 + cation according to a recent electron spin resonance study.1 0 2 , 1 0 4 When sufficiently high CO pressure is used, a near-quantitative yield of Au(CO)S03F is easily obtained as described. The subsequent conversion of Au(CO)S03F to [Au(CO)2][Sb2Fn] is motivated by the intent to replace the fluorosufate anion by an even less basic antimony(V)fluoro anion in the quest for a carbonyl cation.105 The method chosen, the solvolysis of Au(CO)S03F in liquid SbF5 with CO present according to Au(CO)S03F + CO + 4SbF5 > [Au(CO)2][Sb2F„] + Sb 2F 9S0 3F (1-12) follows published precedents, where main-group, organometallic or transition-metal fluorosulfates are solvolysed in an excess of antimony(V) fluoride to permit the syntheses of fluoroantimonate derivatives such as Br2[Sb3Fi6],1 0 6 [(CH3)2Sn][Sb2Fu]2,107 or Pd[SbF6]2.108 This route has become the preferred synthetic approach where fluorosulfates are more readily available than the corresponding fluorides. 26 Bis(carbonyl)gold(I) undecafluoroantimonate(V), [Au(CO)2][Sb2Fn], is the first reported example of a thermally stable compound with a linear metal carbonyl cation. The compound is stable up to 130°C, in contrast to the corresponding dicarbonyl silver(I) cation. 13 18 The use of C and O isotopic substitution has allowed a complete vibrational assignment, a normal coordinate analysis, and general valence field calculations. The results of the force field calculations permit a comparison with previously reported force constants for the 109 isoelectronic species Hg(CN>2 and [Au(CN)2]\ It is found that the strength of the metal-carbon bond decreases in going from [AuCCN)^ to Hg(CN)2 to [Au(CO)2]+, while the intra-ligand bond increases in strength in the same order, which reflects decreasing 7t-back-donation. The stretching force constant ft for [Au(CO)2]+ (20.0 x 102 N m"1) is below the value of 21.3 x 102 N m"1 for H C O + ; 4 9 however the average v(CO) is, with 2235.5 cm"1, over 90 cm"1 higher than that in CO itself. Within the group of known gold(I) carbonyl derivatives, 7t-back-donation, as judged by v(CO), decreases in the order Br" > Cl" > S03F" > UF 6" > Sb2F„", which appears to coincide with the order of decreasing ion basicity. Within the group of [Au(CO)2]+ derivatives inclusive of the solvated cation,52 the thermal stability increases again with decreasing anion basicity: S03F" < UF 6 < Sb 2Fn". These trends suggest that positive charges on gold, and consequently polar contributions to the metal-carbon bond play an important role in the formation of thermally stable gold(I) carbonyl compounds. In addition to polar contributions, the noted ability of 98 univalent gold to form strong covalent bonds in compounds mostly with coordination number 27 of 2 and a linear geometry becomes important as well. Covalent bond strength is possibly aided by relativistic effects.110 These are said to be at a maximum for gold according to recent calculations for gold(I) halide anions,111 and may explain the striking difference in thermal stability between gold(I) and silver(I) carbonyl derivatives. The reversibility of CO uptake by silver(I) complexes, but not by gold compounds, has a simple explanation: dissociation of CO leads in the case of Ag(I) compounds to existing thermodynamically stable compounds such as 97 AgOTeF5. The corresponding gold(I) compounds such as AuS03F, AuUF 6, AuSb 2F n or 99 AuSbF6 are, just like AuF, unknown and evidently not preparable by pyrolysis of the carbonyl precursors discussed. In all instances a more complex decomposition reaction is observed, which results essentially in the formation of elemental gold and various volatile decomposition products.49'52 While the clean elimination of CO on heating is not observed, the reactivity of the gold-carbon bond is evident from CO cleavage and ligand substitution reactions. Fast CO exchange between [Au(13CO)2]+ and [Au(12CO)2]+ is observed at 25°C in HS0 3F using vibrational spectroscopy.49 Slow exchange between [Au(1 3CO)2]+ and [Au(12CO)]+ in HS0 3F or HS03F/SbF5, magic acid, is observable by 1 3C-NMR, where the two single-line resonances 174 and 158-162 ppm respectively coalesce on warming from 17 to 52°C. The formation of a binuclear intermediate in a binuclear reaction via a bridging CO group or possibly an isocarbonyl linkage ( Au • • • CO • • • Au) is likely. Ligand exchange was observed inadvertently during an attempt to recrystallize [Au(CO)2][Sb2Fn] from acetonitrile, CH 3 CN. The reaction proceeds cleanly according to 28 [Au(CO)2][Sb2Fn] + 3CH 3CN > [Au(NCCH3)2][SbF6] + S b F 5 N C C H 3 + 2C0 (1-13) and the resulting product, [Au(NCCH3)2][SbF6] , crystallizes in a cubic NaCl structure with a linear [Au(NCCH3)2]+ cation on the space diagonal of the unit cell. The reaction suggests a wide potential use of [Au(CO)2]+ in the synthesis of various [AuL ]^"1" cations, involving ligands with better a donor ability than CO. Not many examples of this type are known.98 The facile formation of [Au(CO)2]+ ( s o l v ) and Au(CO)S03F by reductive carbonylation of Au(S03F)3 in the previously investigated superacid system HS0 3F/Au(S0 3F) 3 3 6 ' 3 7 suggests the possible extension of this synthetic approach to the superacid system HS03F/Pt(S03F)4.1 1 2 The expected reductive carbonylation occurs with similar ease, but depending on the reaction conditions two different solid materials are obtained: partial reduction of Pt(S03F)4 produces a yellow diamagnetic solid of the composition Pt(C0)2(S03F)3,113 and a complete reduction to Pt(II) leads to colorless Pt(CO)2(S03F)2 according to Pt(S03F)4 + 3CO > Pt(C0)2(S03F)2 + C 0 2 + S 2 0 5 F 2 (1-14) Vibrational spectroscopy shows that the solid isolated from solution is exclusively cis-Pt(CO)2(S03F)2114 in analogy to the formation of cw-Pt(C0)2Cl2 in the reductive carbonylation ofPtCl 4 inSOCl 2 . 6 7 , 6 9 Yellow Pt(CO)2(S03F)3 has no precedent among previous reported carbonyl derivatives of platinum.64 Vibrational analysis suggests formulation as [Pt(CO)4][Pt(S03F)6]113 by comparison with reported spectra for the anion [Pt(S03F)6]2" with Ba 2 + or Cs + as cations.112 The formation reaction according to 29 2Pt(S03F)4 + 5C0 > [Pt(CO)4][Pt(S03F)6] + C 0 2 + S 2 0 5 F 2 (1-15) requires low CO pressure and reaction temperature. The isolation of solid [Pt(CO)4]-[Pt(S03F)6] from solution and separation from cw-Pt(CO)2(S03F)2 are facilitated by the limited solubility of the former in HS0 3F. In addition, cw-Pt(CO)2(S03F)2 is soluble in S0 2, while [Pt(CO)4][Pt(S03F)6] is not. The [Pt(CO)4]2+ cation is also observed in solution of HS0 3F by 1 3C-NMR. The cation is generated either by redissolving [Pt(CO)4][Pt(S03F)6] in HS0 3F or during the reductive carbonylation of Pt(S03F)4 in solution. Finally, solvolysis of c«-Pt(CO) 2(S0 3F) 2 under CO pressure in liquid SbF5 proceeds in complete analogy to the solvolysis of Au(CO)S03F according to: Pt(CO)2(S03F)2 + 2CO + 8SbF5 > [Pt(CO)4][Sb2F„]2 + 2Sb2F9(S03F) (1-16) and results in the formation of a white solid of composition [Pt(CO)4][Sb2Fn]2, which again contains the square-planar cation [Pt(CO)4]2+ and is thermally stable up to 120°C. 1 1 5 It represents the first example of a dipositive, homoleptic carbonyl cation. Its square planar geometry evident from vibrational spectra is unprecedented in metal carbonyl chemistry. In summary, carbon monoxide is a very versatile ligand in terms of both the number of the metal carbonyl compounds it forms and the bonding types found in these compounds. However, most of the metal carbonyl complexes so far known have much lower average v (CO) stretching frequency than free CO. The synergic bonding mechanism seems applicable to explain the majority of these metal carbonyl complexes. The bonding model is especially 30 valid for d-block metals in low oxidation states. However, the accepted synergic bonding mechanism appears to be unsatisfactory when it is applied to the cationic metal carbonyl complexes with average v (CO) frequencies well above 2143 cm"1. 1.7 Research Direction— Methods and Objectives From the preceding review of the history and recent development of cationic noble metal carbonyl complexes, it is obvious that a very promising start has been made towards the synthesis and understanding of this unusual group of metal complexes. However, examples of this type of compounds are still rare. In addition, very little structural information has been reported, and it is unclear why some of these compounds should have relatively high thermal stabilities since the metal-carbon bond is expected to be weak in the absence or the near-absence of rc-back-bonding. In order to have a better understanding of the bonding nature and to investigate further the chemistry of cationic metal carbonyl derivatives, it is highly desirable to expand the scope of known cationic metal carbonyls. Palladium(II) becomes the metal of choice because of the known similarity of its chemistry to that of platinum(II). However, no thermally stable mono-nuclear palladium(II) carbonyl chlorides have so far been reported in spite of the long existence of cw-Pt(CO)2Cl2, and this deserves further study and presents a synthetic challenge. Another target is to generate octahedral metal carbonyl cations such as [M(CO)6]2 + or [M(CO) 6] 3 +. For this purpose, osmium, ruthenium, iridium, and rhodium seem to be promising candidates. 31 Therefore, the direction of this research will now be formulated, the synthetic approaches will be outlined, and the questions which this thesis attempts to answer will be posed. 1.7.1 Research Direction (i) Expansion to other electron-rich metals: An expansion of cationic metal carbonyl derivatives from Pt(II) to Ir(III) and Os(II) seems logical and feasible. Of particular interest is the synthesis of Ir(III) carbonyl derivatives because iridium forms both synergically bonded carbonyls (e.g. Ir4(CO)i2) and highly reduced carbonylates like [Ir(CO)3]3". In addition, an expansion to the 4d transition metals should subsequently lead to cationic Pd(II), Rh(III), and Ru(II) carbonyl derivatives. The obvious synthetic targets would be [Pd(CO)4]2+ and cis-Pd(CO)2(S03F)2 because of the strong similarity between the chemistry of Pt(II) and Pd(II). The synthetic efforts will also be extended to obtain dipositive octahedral carbonyl cations, such as [Ru(C0)6]2+ or [Os(CO)6]2+. (ii) In order to have a better understanding of metal-CO bonding and the thermal stabilities of these new metal carbonyl compounds, precise molecular structural information is needed, and hence efforts will also be directed to obtain single crystals for X-ray diffraction analysis. Except for the low temperature structures of Ag(CO)B(OTeF5)495 and [Ag(CO)2]-[B(OTeF5)4]96 which are of low accuracy, and of Au(CO)Cl,7 9 no molecular structures of noble metal carbonyl derivatives were known at the outset of this thesis. Hence precise molecular structure determination is required in order to understand why cationic carbonyl derivatives have, in the near-absence of rc-back-donation, relatively high thermal stabilities. 32 (iii) The development of new synthetic routes to cationic carbonyl derivatives. The cationic carbonyl derivatives of electron-rich metals appear to have a very promising and interesting chemistry. It is necessary to develop a unique, highly productive procedure which hopefully will not require the use of hazardous, corrosive, and commercially unavailable reagents. 1.7.2 Synthetic Methodology and Cooperation 1.7.2.1 Synthetic Methodology There are at present three synthetic approaches to chose from: (i) The oxidative carbonylation of metals in HF employed in the synthesis of [Au(CO)2][UF6].100 (ii) Addition of CO to known metal salts applied in the synthesis of [Ag(CO)][B(OTeF5)4]95,96'116 and the related compounds, (iii) Reductive carbonylation in strong protonic acids and superacids, and the subsequent solvolysis in SbF5 in the presence of gaseous CO, exemplified by the synthesis of [Au(CO) 2][Sb 2F„] 2 4 9 and [Pt(CO)4][Sb2Fn]2.115 The first two methods have failed to meet two simple objectives: (a) to produce thermally stable compounds and (b) to generate multiply-charged carbonyl cations. It appears doubtful that UF 6 or other metal hexafluorides will readily effect 2e~ or 3e~ oxidations and whether the resulting [MF6]"" anions will be capable of stabilizing highly electrophilic cations. In addition, the second method (the approach of CO-addition) suffers from the unavailability of suitable metal salts and also from the reversibility of such reactions. It does not seem to be promising to produce M(II) or M(III) carbonyls since this would be opposed by 33 the lattice energies of the salts used. The reductive carbonylation in HS0 3F or related superacids and the solvolysis in SbF5 becomes the method of choice. However, the use of corrosive (HS03F and SbF5), hazardous (S206F2), and commercially unavailable (noble metal fluorosulfates) chemicals presents a limitation to a wider use of this synthetic approach. In addition, the synthetic process involves the synthesis of a metal carbonyl fluorosulfate as an intermediate. In the case of dealing with extremely moisture-sensitive and hazardous chemicals, this should be avoided whenever possible. Therefore, the development of a one-step procedure using only SbF5 instead of both SbF5 and HS0 3F is another goal of this research. 1.7.2.2 Cooperation with Professor H . Winner's Group Previous research work in our group, in particular on gold(I) and platinum(II) carbonyls, was based on a close cooperation between our group at the University of British Columbia (UBC) and Professor H. Willner's group at der Universitat Hannover, Germany. This cooperation has continued to the present and a very useful arrangement has been established. Synthetic work on Pd, Ir, Rh, Os, and Ru as well as vibrational spectroscopy and X-ray diffraction studies was carried at UBC and in part at Simon Fraser University (SFU). Professor Willner's group repeated our synthetic procedures, provided solid state NMR spectra and performed, where appropriate, vibrational coordinate analyses and CO stretching force constant calculations. At the same time, Professor Willner's group extended our synthetic approach to post-transition metals, in particular to Hg and Tl, with the UBC group now involved in repeating 34 and checking the results obtained. 1.8 Outline of This Thesis Following this introductory chapter, general experimental procedures used in this research will be described in Chapter 2. In Chapter 3, 4, and 5, the syntheses and characteri-zations of cationic metal carbonyls of palladium® and palladium(II), iridium(III), and ruthenium(II) and osmium(II) will be discussed. Solvolysis reactions in SbF5 with and without additional gaseous CO will be described in Chapter 6. Finally, Chapter 7 will summarize and conclude this thesis with proposed suggestions for future work. In addition, answers to the pursuant three simple but challenging questions will be attempted in the final chapter. (i) Why and under what conditions are cationic carbonyl derivatives thermally stable? (ii) How is metal-CO bonding best described? Will the synergic bonding model have to be replaced or modified? (iii) How to decide upon the terminology of the cationic metal carbonyl derivatives with very high CO stretching frequencies? 35 References 1 Pearson, R. G. Inorg. Chem. 1984, 23, 4675. 2 Schutzenberger, P. Bull. Soc. Chim. Fr. 1868, 10, 188. 3 (a) Schutzenberger, P. C.R. Acad. Sci. 1870, 70, 1134, 1287. (b) Rosenblum, B.; Nethercot, A. H. Jr.; Towns, C. H. Phys. Rev. 1958, 109, 400. (c) Meerts, W. L. ; De Leeuw, F. H. ; Dymanus, A. Chem. Phys. 1977, 22, 319. (d) Wiberg, K. B.; Hadad, C. M . ; LePage T. J.; Breneman, C. M. J. Phys. Chem. 1992, 96, 671. 4 Cotton, F. A. and Wilkinson, G. Advanced Inorganic Chemistry; 5th ed.; Wiley: New York, 1988; pp58 and 1021. 5 Pruchnik, F. P. Organometallic Chemistry of the Transition Elements; Plenum: New York, 1990;p 23. 6 Werner, H. Angew. Chem., Int. Ed. Engl. 1990, 29, 1077. 7 Braterman, P. S. Metal Carbonyl Spectra; Academic: New York, 1975. 8 Kettle, S. F.A. Top. Curr. Chem. 1977, 71, 111. 9 Nakamoto, K. Infrared Spectra of Inorganic Compounds and Coordination Compounds; 1st ed.; Wiley: New York, 1963; p 72. 1 0 Adams, D. M. Metal Ligand and Related Vibrations; Edward Arnold: London, 1967; p 84. 1 1 Mann, B. E. Adv. Organomet. Chem. 1974, 12, 133. 1 2 Barber, M . ; Connor, J. A. ; Guest, M. F.; Hall, M. B.; Hillier, I. H . ; Meredith, W. N. E. J. Chem. Soc, Faraday Discuss. 1972, 54, 219. 1 3 (a) Jolly, W. L. Top. Curr. Chem. 1977, 71, 149. 36 (b) Avazino, S. C ; Bakke, A. A. ; Chen, H. -W.; Donahue, C.G. ; Jolly, W. L. ; Lee, T. H . ; Ricco, R. J. Inorg. Chern. 1980, 19, 1931. 1 4 Abel, E. W.; Tyfield, S. P. Adv. Organomet. Chern. 1974, 12, 133. 1 5 Ford, R. R. Adv. Catal. 1970, 21, 52. 1 6 (a) Little, L. H. Infrared Spectra of Adsorbed Species; Academic: London, 1966; p 67. (b) Sheppard, N. Vibrational Spectroscopy of Adsorbates; Willis, R. F., Ed.; Springer: Berlin, 1979; p 169. 1 7 (a) Moskovits, M . ; Ozin. G. A. Cryochemistry; 1st ed.; Wiley: New York, 1976; p 263. (b) Kiindig, E. P.; Moskovits, M . ; Ozin. G. A. Angew. Chern. Int. ed. Engl. 1975, 14, 292. 1 8 (a) Ellis, J. E. Adv. Organomet. Chern. 1990, 31,1. (b) Warnock, G. F. P.; Sprague, J.; Fjare, K. L. ; Ellis, J. E. J. Am. Chern. Soc. 1983, 105, 672. 1 9 Chi, K. M . ; Frerichs, S. R.; Philson, S. B.; Ellis, J. E. J. Am. Chern. Soc. 1988, 110, 303. 2 0 (a) Darling, J. H . ; Ogden, J. S. J. Chern. Soc, Dalton Trans. 1973, 1079. (b) Huber, H . ; Kundig, E. P.; Moskovits, M . ; Ozin. G. A. Nature (Phys. Sci), 1972, 98, 235. 2 1 Kundig, E. P.; Mcintosh, D.; Moskovits, M . ; Ozin. G. A. J. Am. Chern. Soc. 1973, 95, 7234. 2 2 Mcintosh, D. ; Ozin. G. A. Inorg. Chern. 1977, 16, 51. 37 u Mond, L. ; Langer, C ; Quincke, F. J. Chem. Soc. 1890, 749. 2 4 Ellis, J. E.; Fjare, K. L. ; Hayes, K. L. J. Am. Chem. Soc. 1981, 103, 6100. 2 5 Dewey, C. G. ; Ellis, J. E.; Fjare, K. L. ; Pfahl, K. M . ; Warnock, G. F. P. Organometallics, 1983, 2, 388. 26 Huhey, J. E.; Keiter, E. A. ; Keiter, R. L. Inorganic Chemistry, 4th ed.; Harper-Collins: New York, 1993; p 405. 2 7 Chadwick, B.M.; Sharpe, A. G. Adv. Inorg. Chem. Radiochem. 1966, 8, 83. 28 Sharpe, A. G. The Chemistry ofCyano Complexes of the Transition metals; Academic: New York, 1976. 2 9 DeKock, R. L. ; Sarapu, A. C ; Fenske, R. F. Inorg. Chem. 1971, 10, 38. 3 0 Burg, A. B.; Schlesinger, H. J. J. Am. Chem. Soc. 1937, 59, 780. 3 1 Bauer, S. H. J. Am. Chem. Soc. 1937, 59, 1804. 3 2 Gordy, W.; Ring, H; Burg, A.B. Phys. Rev. 1948, 74, 1191. 3 3 Cowan, R. D. J. Chem. Phys. 1949, 77, 218. 3 4 (a) Bethke, G. W.; Wilson, M . K. J. Chem. Phys. 1957, 26, 1118. (b) Bethke, G. W.; Wilson, M. K. J. Chem. Phys. 1957, 27, 978. 3 5 (a) Taylor, R. C. J. Chem. Phys. 1957, 26, 1131. (b) Taylor, R. C. J. Chem. Phys. 1957, 27, 979. 3 6 (a) Sundaram, S.; Cleveland, F. F. J. Chem. Phys. 1960, 32, 166. (b) Jones, L. H . ; Taylor, R. C ; Paine, R. T. J. Chem. Phys. 1979, 70, 749. 3 7 Beach, D. B.; Jolly, W. L. Inorg. Chem. 1985, 24, 567. 38 DeKock, R. L. ; Gray, H. B. Chemical Structure and Bonding; Benjamin/Cummings: Menlo Park, CA, 1980; p 323 ( and references cited therein). 3 9 Carter, J. C ; Parry, R. W. J. Am. Chern. Soc. 1965, 87, 2354. 4 0 Parry, R. W.; Nordman, C. E.; Carter, J. C ; Terhaar, G. Advances in Chemistry Series No. 42; American Chemical Society: Washington, DC, 1964; p 302. 4 1 Janda, K. C ; Berntein, L. S.; Steed, J. M . ; Novick, S. E.; Klemperer, W. J. Am. Chern. Soc. 1978, 100, 8074. 4 2 Talbi, D. ; Pauzat, F.; Ellinger, J. Chern. Phys. 1988, 126, 291(and references cited therein). 4 3 (a) Hitora, E.; Endo, J. Mol. Spectrosc. 1988, 727, 527. (b) Woods, R. C ; Dixon, T. A. ; Saykally, R. J.; Szanto, P. G. Phys. Rev. Lett. 1975, 35, 1269. 4 4 Olah, G. A. ; Prakash, G. K. S.; Sommer, J. Superacids; Wiley: New York, 1985. 4 5 Schatte, G. ; Willner, H . ; Hoge, D. ; Knozinger, E.; Schrems, O. J. Phys. Chern. 1989, 93, 6025. 4 6 Foster, S. C ; Mckellar, A. R. W.; Sears, T. J. J. Chern. Phys. 1984, 81, 578. 4 7 Kawaguchi, K.; Yamada, C ; Saito, S.; Hirota, E. J. Chern. Phys. 1985, 82, 1750. 4 8 Amano, T. J. Chern. Phys. 1983, 79, 3595. 4 9 Willner, H . ; Schaebs, J.; Hwang, G.; Mistry, F.; Jones, R.; Trotter, J.; Aubke, F. J. Am. Chern. Soc. 1992, 114, 8972. 5 0 Lee, K. C ; Aubke, F. Inorg. Chern. 1979, 18, 389. 39 3 1 Lee, K. C ; Aubke, F. Inorg. Chem. 1980, 19, 119. 5 2 Willner, H . ; Aubke, F. Inorg. Chem. 1990, 29, 2195. 5 3 Fink, E. C. R. Acad. Sci. 1898, 126, 646. 5 4 Manchot, W.; Konig, J. Chem. Ber. 1925, 58, 2518. 5 5 Manchot, W.; Konig, J. Chem. Ber. 1926, 59, 883. 5 6 Manchot, W.; Gall, H. Chem. Ber. 1925, 58, 2175. 5 7 Kharash, M . S.; Isbell, H. S. J. Am. Chem. Soc. 1930, 52, 2919. 5 8 Manchot, W.; Konig, J.; Gall, H. Chem. Ber. 1924, 57, 1157. 5 9 Kunz-Krause, H. Apoth. Ztg. 1916, 31, 66. 6 0 Leblanc, F. C. /?. Acad. Sci. 1850, 50, 483. 6 1 Berthelot, M. Ann. Chim. Phys. 1856, 346, All. 6 2 Hakansson, M . ; Jagner, S. Inorg. Chem. 1990, 29, 5241. 6 3 Bruce, M . I. J. Organomet. Chem. 1972, 44, 209. 6 4 Thripathy, S. C ; Srivastava, S. C ; Mani, R. P.; Shrimal, A. K. Inorg. Chim. Acta. 1976, 17, 257. 6 5 Vargaftic, M . N. ; Stromnova, T. A. ; Moiseev, I. I. Russ. J. Inorg. Chem. (Engl, trans.), 1980, 25, 127. 6 6 Browning, J.; Goggin, P. L. ; Goodfellow, R. J.; Norton M. J.; Rattray, A. J. M . ; Taylor, B. F.; Mink, J. J. Chem. Soc, Dalton Trans. 1977, 2061. 6 7 Calderazzo, F. Pure Appl. Chem. 1978, 50, 49. 6 8 Calderazzo, F.; Belli Dell'Amico, D. Pure Appl. Chem. 1986, 58, 561. 40 Calderazzo, F. J. Organomet. Chern. 1990, 400, 303. 7 0 Calderazzo, F.; Belli Dell'Amico, D. Inorg. Chern. 1981, 20, 1310. 7 1 Belli Dell'Amico, D. ; Calderazzo, F.; Zandona, N. Inorg. Chern. 1984, 23, 137. 7 2 (a) Belli Dell'Amico, D. ; Calderazzo, F.; Dell'Amico, G. Gazz. Chim. Ital. 1977, 707, 101. (b) Belli Dell'Amico, D. ; Calderazzo, F. Gazz. Chim. Ital. 1979, 109, 99. 7 3 Calderazzo, F.; Cotton, F. A. Inorg. Chern. 1962, 1, 30. 7 4 Calderazzo, F.; Fachinetti, G. ; Floriani. C. J. Am. Chern. Soc. 1974, 46, 3695. 7 5 Belli Dell'Amico, D. ; Calderazzo, F. Gazz. Chim. Ital. 1973, 103, 1099. 7 6 (a) Belli Dell'Amico, D. ; Calderazzo, F.; Robino, P.; Segre\ A. Gazz. Chim. Ital. 1991, 121, 51. (b) Belli Dell'Amico, D. ; Calderazzo, F.; Robino, P.; Segrd, A. J. Chern. Soc, Dalton Trans. 1991, 3017. 7 7 Brown, P. R.; Geoffrey, F.; Cloke, N. ; Green, M . L. H. Polyhedron. 1985, 4, 869. 7 8 Malatesta, L. ; Naldini, L. Gazz. Chim. Ital. 1960, 90, 1505. 7 9 Jones, P. G. Z. Naturforsch., TeilB 1982, 37, 823. on Andreini, B. P.; Belli Dell'Amico, D. ; Calderazzo, F.; Venturi, M. G.; Segr6, A. J. Organomet. Chern. 1988, 354, 357. 8 1 Scott, A. F.; Wilkening, L. L.; Rubin, B. Inorg. Chern. 1969, 8, 2533. 8 2 Desjardin, C. D. ; Edwards, D. B.; Passmore, J. Can. J. Chern. 1979, 57, 2714. 8 3 Ghiotti, G. ; Boccuci, F.; Scala, R. J. Catal. 1985, 92, 79. 41 (a) Gardner, R. A. ; Petrucci, R. H. J. Phys. Chem. 1963, 67, 1376. (b) DeKock, C. W.; Van Leirsburg, D. A. J. Am. Chem. Soc. 1972, 94, 3235. (c) Pollitzer, P.; Kasten, S. D. Surf. Sci. 1973, 36, 186. (a) Crocker, A. ; Goggin, P. L. ; Goodfellow, R. J. J. Chem. Soc, Chem. Commun. 1978, 1056. (b) Belli Dell'Amico, D. ; Calderazzo, F.; Marchetti, F.; Merlino, S. J. Chem. Soc, Dalton Trans. 1982, 2257. Souma, Y.; Iyoda, J.; Sano, H. Inorg. Chem. 1976, 75, 968. Neppel, A. ; Hickey, J. P.; Butler, I. S. J. Raman. Spectrosc. 1979, 8, 57. Backen, W.; Ericsson, A. Acta. Chem. Scand., Ser. A 1981, 35, 1. Souma, Y.; Sano, H. Nippon Kagaku Zasshi, 1970, 91, 625. Souma, Y.; Sano, H. Chem. Lett. 1973, 1059. Souma, Y.; Sano, H. J. Org. Chem. 1973, 38, 2016. Souma, Y.; Sano, H. Bull. Chem. Soc. Jpn. 1973, 46, 3237. Souma, Y.; Sano, H. J. Org. Chem. 1973, 38, 3633. Souma, Y.; Sano, H. Bull. Chem. Soc. Jpn. 1974, 47, 1717. Ffurlburt, P. K.; Anderson, O. P.; Strauss, S. H. / . Am. Chem. Soc. 1991, 773, 6277. Hurlburt, P. K.; Rack, J. J.; Dec, S. F.; Anderson, O. P.; Strauss, S. H. Inorg. Chem. 1993, 32, 373. Strauss, S. H . ; Noirot, M. D. ; Anderson, O. P. Inorg. Chem. 1985, 24, 4307. (a) Puddephatt, R. J. In The Chemistry of Gold; 1st ed. Elsevier: Amsterdam, 1978. 42 (b) Puddephatt, R. J. In Comprehensive Coordination Chemistry; Wilkinson, G., Ed.; Pergamon: Oxford, U. K., 1987; Vol. 7, p 861. 9 9 Muller, B. J. Angew. Chern., Int. Ed. Engl. 1987, 26, 861. 1 0 0 Adelhelm, M . ; Bacher, W.; Hohn, E. G. Jacob, E. Chern. Ber. 1991, 124, 1559. 1 0 1 Lustig, M. Inorg. Chern. 1965, 4, 1828. 1 0 2 Willner, H . ; Mistry, F.; Hwang, G.; Herring, F. G.; Cader, M.S. R.; Aubke, F. J. Fluorine Chern. 1991, 52, 13. 1 0 3 Belli Dell'Amico, D. ; Calderazzo, F.; Marchetti, F.; Merlino, S.; Perego, G. J. Chern Soc, Chern. Commun. 1977, 31. 1 0 4 Herring, F. G.; Hwang, G.; Lee, K. C ; Mistry, F.; Phillips, P. S.; Willner, H . ; Aubke, F. J. Am. Chern. Soc. 1992, 114, 1271. 1 0 5 Mallela, S. P.; Yap, S.; Sams, J. R.; Aubke, F. Inorg. Chern. 1986, 25, 4327. 1 0 6 Wilson, W. W.; Thompson, R. C ; Aubke, F. Inorg. Chern. 1980, 19, 1484. 1 0 7 Mallela, S. P.; Yap, S.; Sams, J. R.; Aubke, F. Rev. Chim. Miner. 1986, 23, 572. 1 0 8 Cader, M.S. R.; Thompson, R. C ; Aubke, F. Can. J. Chern. 1989, 67, 1942. 1 0 9 (a) Chadwick, B. M . ; Frankiss, S. G. J. Mol. Struct. 1916, 31, 1. (b) Jones, L. H. J. Chern. Phys. 1957, 27, 468. 1 1 0 Pyykko, P. Chern. Rev. 1988, 88, 563. 1 1 1 Schwerdtfeger, P. J. Am. Chern. Soc. 1989, 777, 7261. 1 1 2 Lee, K. C ; Aubke, F. Inorg. Chern. 1984, 23, 2124. 1 1 3 Hwang, G.; Bodenbinder, M . ; Willner, H. ; Aubke, F. Inorg. Chern. 1993, 32, 4667. 43 1 1 4 Hwang, G.; Wang, C ; Bodenbinder, M . ; Willner, H . ; Aubke, F. J. Fluorine Chem. 1994, 66, 159. 1 1 5 Hwang, G.; Wang, C ; Aubke, F.; Willner, H . ; Bodenbinder, M. Can. J. Chem. 1993, 71, 1532. 1 1 6 Hurlburt, P. K.; Rack, J. J.; Luck, J. S.; Dec, S. F.; Webb, J. D. ; Anderson, O. P.; Strauss, S. H. J. Am. Chem. Soc. 1994, 116, 10003. 44 CHAPTER 2 GENERAL EXPERIMENTAL 2.1 Introduction General experimental techniques as well as the sources, purification and, where necessary, preparation of starting materials used in this study will be described in this chapter. Specific syntheses and procedures will be discussed in the appropriate chapters. 2.2 Chemicals 2.2.1 Chemicals Used Without Further Purification Many chemicals used in this study were obtained from a supplier in a user-ready form. They are listed in Table 2-1, along with their source and percentage purity. Palladium(II) chloride and platinum(II) chloride were supplied courtesy of Professor B. James of this Department. Fluorolube grease Series 25-10M, CF2C1(CF2-CFC1)„CF2C1, obtained from Halocarbon Products Corporation, was sparingly used to seal and lubricate ground glass connections. Its low volatility and relative inertness toward halogen-containing compounds made it suitable in most cases. Where elevated temperatures were necessary and hence possible by-reaction was suspected, Teflon joint sleeves (Nalge Company, New York, USA) were used instead. 45 Table 2-1 Chemicals Used Without Further Purification Chemicals Source Purity (%) F 2 Air Products Inc. Technical Grade S0 3 Du Pont de Nemours Co. 98 (stabilized) Wilmington, Delaware, USA P 2 0 5 BDH 98 CaCl2 Fisher 97.1 N 2 Canadian Liquid Air Dry K-grade Pt, 60 mesh Johnson Matthey Catalog Co. 99.9 Pd, 60 mesh Johnson Matthey Catalog Co. 99.95 Ir, 60 mesh Alfa 99.5 Os, 60 mesh Alfa 99.5 Ru, 80 mesh Ventre Alfa Corp. 99.5 NiCl 2 Aldrich Chern. Co. 99 RhCl 3 , anhydrous Ventre Alfa Corp. not given FeCl 3 , anhydrous Aldrich Chern. Co. 98 IrCl3, anhydrous Aldrich Chern. Co. not given TIBr BDH >99.5 2.2.2 Purification of CO, HSO3F, SbF5, and S0 2 (i) Carbon monoxide, CO (C. P. grade, 99.5% purity) was obtained from Linde Gases and dried by passing it through a trap at -196 °C to remove trace amounts of C0 2 , H 2 0 and other impurities. 46 (ii) Fluorosulfuric acid, HS0 3F (technical grade), was obtained from Orange County The constant boiling fraction at 162-163 °C was either collected directly into a reactor for synthetic use or distilled into a 100 mL Pyrex storage vessel for later vacuum transfer. (iii) Antimony pentafluoride, SbF5, was obtained from Ozark-Mahoning and purified into a special storage vessel using the following procedure. To keep contamination by atmospheric moisture to a minimum, a special one-part distillation apparatus was used to purify antimony pentafluoride. Under a flow of dry nitro-gen at atmospheric pressure, crude antimony pentafluoride was distilled r -Chemicals and purified by double distillation under dry nitrogen as previously described. vacuum line via a B19-socket/B14-vacuum at 0 °C. The purified SbF5 designed storage ampoule (Figure 2-1) amount of air and possibly HF, in the cone adapter. was then transferred into a specially distillation, the flask was attached to a with a B19 ground glass cone. After into a 500 mL round bottom flask flask was removed under a dynamic N 2 , with a trace for future use. Figure 2-1 Storage ampoule for SbF5. 47 (iv) Sulfur dioxide, S0 2 (99.5% purity), was obtained from Matheson Gases and dried by storing over P 20 5 . 2.2.3 Preparative Reaction of S 2 0 6 F 2 Bis(fluorosulfuryl)peroxide, S 20 6F 2 , was prepared by the direct reaction between F 2 and S0 3 with AgF 2 acting as a catalyst. Large quantities of S 2 0 6 F 2 could be produced using a system adapted from the original reports.2'3 Over the years, some improvements were made to the system in this laboratory. The apparatus and reaction conditions used most recently are briefly outlined below. The system (Figure 2-2) comprised three parts: supply of gaseous reactants, the reactor and the collection of products. Fluorine gas was of technical grade (Air Products, Long Beach, California) contained variable amounts of 0 2, OF 2 , HF and fluorocarbons totaling ca. 1%. Of these HF should be removed, and a NaF tower was used for this purpose. Sulfur trioxide (Du Pont de Nemours Co., Wilmington, Delaware) was carried by a stream of nitrogen gas into the flow reactor. A heating mantle was placed under the round bottom flask to increase the vapor pressure of S0 3 and to facilitate its transport by the stream of nitrogen. Four Pyrex glass traps were used to collect the crude product. Trap A was kept at room temperature and used for pre-cooling and monitoring the products. A soda lime tower was used to dispose of any unreacted F 2 and the possible by-product FOS0 2F. The tower reactor (Figure 2-3) was made in the Mechanical Engineering Shop of this Department. It was constructed of a 4" o/d Monel-K tube with a flange and a removable top 48 \Whitey #SS-1KS4) Crosby Monei Pressure Gauge 2 . Figure 2-2 Schematic diagram of the preparation of bis(fluorosulfuryl) peroxide (S206F2). to allow the loading and removal of the catalyst. The Monel-K tube was wrapped in two Chromel heating coils which were used to control the temperatures of the upper and lower halves of the reactor separately. Thermocouple wires were also wrapped inside in order to monitor the temperatures. This assembly was surrounded by Fiberfax insulating paper and Pyrex glass wool, and then encased in 0.02" thick stainless steel plating held tight by hose clamps. Two additional thermocouple wires were inserted into the top and bottom of the reactor to monitor the internal temperature. 49 Thermocouple (3) VIRGIN TEFLON SEAL COPPER FLANGES SILVER SOLDER (Johnson Matthey easi fto 45) SO3/N2 INLET (3/8" o.d. Monel K) Cotran/csUltra-Temp 3300 bonded with Eisenglass 16/£i HEATERS Chromel-A(2-20ga (Upper & Lower) Thermocouple (2) COPPER GAUZE 6-1/4" N.C. Stainless Steel Screws Thermocouple Wells a I (1/4" o.d. x 0.035" Monel) F2 INLET rar; ^-^ (3/8"o.d. Monel K) To 110VVARIAC MONEL-K TUBE 4"o.d.x0.065", 47" overall length CATALYST (AgF2on Copper Turning) To 110 VVARIAC - f f ^ — - - • GAS OUTLET I {316" o.d. Monel K) 1/4" Monel K Thermocouple (4) Figure 2-3 The reactor used in the preparation of bis(fluorosulfuryl) peroxide (S206F2). 50 The catalyst, AgF 2, was made by fluorination of silver supported on copper turnings. Silver was plated onto the copper turnings using an aqueous solution of AgN0 3 and KCN. The silver-plated copper turnings were washed, dried and loaded into the reactor. Fluorination of the silver plating was carried out in a slow, undiluted stream of F 2 . The flow rate of F 2 was controlled so that the fluorination process proceeded smoothly. After completion of the fluorination, the reactor was cooled down to room temperature by flushing with nitrogen-diluted fluorine, followed by purging with dry nitrogen. The inlets and outlet of the reactor were sealed until the reactor was used. The whole process of preparing S 2 0 6 F 2 needed constant monitoring and adjusting of the reaction temperature because the reaction is exothermic and the reaction proceeds extremely fast at temperatures above 180 °C. The ideal temperature was about 170 °C based on experience. Keeping the stoichiometric ratio at approximately 1:1 was also crucial to the success of the preparation. A higher ratio produced more FOS0 2F, which is known to be a hazardous and unpredictable compound.4'5 Lower ratios left some unreacted S0 3, which was very difficult to remove from the crude product. The crude product collected in the cooling traps at -78 °C still contained small amounts of S0 3 and FOS0 3F. The actual amounts of the two by-products depended upon the conditions used. Although FOS0 2F is by far the more hazardous compound, it is easier to remove from the mixture by trap-to-trap distillation than S0 3. The traps with crude product were allowed to warm to room temperature on the collection line for a brief period of time to reduce the content of FOS0 2F. The final purification was accomplished by trap-to-trap distillation with three traps kept at -10 °C, -55 °C and -78 °C respectively. The trap at -55 51 °C contained pure S 20 6F 2 . The purity of S 2 0 6 F 2 was checked conveniently by gas phase pressure, and then more accurately by infrared and 1 9 F NMR spectroscopy. Pure S 2 0 6 F 2 was stored in sealed glass ampoules at room temperature. So far no entirely satisfactory approach has been found to remove S0 3 from the crude product. However, an alternative approach for obtaining pure S 2 0 6 F 2 from a mixture with a significant concentration of S0 3 proved to be very useful. The crude product with a high concentration of S0 3 was allowed to react with fluorine gas once more under slightly different conditions. To avoid formation of FOS0 2F, the reactor temperature was lowered to 110-120 °C and the F 2 flow rate was drastically reduced. About one pound of crude product with a high content of S0 3 was converted to pure S 2 0 6 F 2 in this way. 2.3 Apparatus and Equipment 2.3.1 Vacuum Line A Pyrex vacuum line was employed for synthetic reactions. A typical apparatus for the preparation of cationic metal carbonyl derivatives is shown in Figure 2-4. The glass vacuum line had a 60 cm-long manifold with three BIO sockets, one BIO cone, and one B24 socket equipped with Kontes Teflon stem stopcocks. A Setra 280 E pressure transducer with a digital output was attached to the BIO cone to measure (or monitor) the gas phase pressure. 2.3.2 Glove Box A "DRI-LIB" Model DL-001-S-G glove box (Vacuum Atmosphere Corp., Hawthorne, 52 53 CA, USA) was used in the handling and storing of hygroscopic solids and liquids with low volatility. The glove box was filled with dry nitrogen, and the removal of trace moisture inside the glove box was accomplished by an automatic circulating system over molecular sieves located within the "DRI-TRAIN" Model HE-493 (Vacuum/Atmosphere Corp., CA, USA). The molecular sieves were regenerated periodically to maintain efficient moisture removal. 2.3.3 Glass Vessels Most synthetic reactions were carried out either in Pyrex round-bottom flasks (50 or 100 mL) fitted with 4 mm stopcocks and BIO ground glass cones, or in a thick-wall glass tube reactor if relatively high pressures were anticipated. To keep the contamination from grease to a minimum, a one-body filtration apparatus (Figure 2-5) was designed and used to carry out vacuum filtration. 2.4 Instrumentation and Methods 2.4.1 Infrared Spectroscopy Infrared spectra down to 400 cm"1 were recorded at room temperature on a Bomem MB 102 Fourier-transform infrared (FT-IR) spectrometer. For solid samples AgBr or AgCl windows were used with the edges sealed with vinyl insulating tape. For liquid samples a tiny drop of solution was put between silicon windows. A thin film of solution sample was formed by pressing the windows tightly together. To prevent air from coming into contact with the 54 solution sample, the, silicon windows were mounted in a specially designed air-tight Teflon holder inside a glove box. 55 Recording of gaseous spectra was accomplished using a glass gas cell fitted with AgBr windows and a 4 mm Kontes Teflon stopcock. A cold finger was employed to facilitate sample collection if an extremely low pressure was expected. 2.4.2 Raman Spectroscopy Raman spectra were obtained on a Bruker FRA 106 Raman accessory mounted on a IFS-66v FT-IR optical bench. Samples were loaded into melting point capillary tubes, which were temporarily sealed with halocarbon grease inside a glove box. The tubes were flame-sealed under nitrogen upon taking them out of the glove box. Some spectra were recorded by Professor H. Willner, der Universitat Hannover, Germany. 2.4.3 X-ray Crystallography X-ray diffraction data were collected on two machines: A Rigaku AFC6S diffractometer and an Enraf Nonius CAD4F diffractometer equipped with an in-house modified low-temperature attachment using graphite monochromatized Mo K a radiation. The structures were solved either by Dr. S. J. Rettig and Professor J. Trotter of this Department or by Dr. R. J. Batchelor and Professor F. W. B. Einstein at Simon Fraser University. 2.4.4 NMR Spectroscopy 1 9 F NMR spectra were recorded on a Varian XL-300 FT multinuclear spectrometer operating at 282.231 MHz. CFC13 was used as external reference. The spectra were acquired without a lock. 56 1 3C-MAS-NMR spectra were recorded on a Bruker MSL-200 FT spectrometer operating at 50.322 MHz courtesy of Professor H. Willner, der Universitat Hannover, Germany. 2.4.5 Microanalysis Elemental analyses were performed by Mr. Peter Borda of this Department. Carbon and hydrogen contents of the samples were determined on a Carlo Erba Elemental Analyzer Model 1106. Sulfur contents of the samples were titrimetrically determined. 2.4.6 Melting Point Determination Melting points, or decomposition temperatures, were obtained on a Gallenkamp melting point apparatus. Solid samples were finely ground and loaded into 2.0 mm o/d capillary tubes. The capillary tubes were first sealed with halocarbon grease in the glove box, and then flame-sealed upon taking them out of the glove box. 57 References 1 Barr, J.; Gillespie, R. J.; Thompson, R. C. Inorg. Chem. 1964, 3, 1149. 2 Dudley, F. B.; Cady, G. H. / . Am. Chem. Soc. 1957, 79, 513. 3 Cady, G. H . ; Shreeve, J. M. Inorg. Synth. 1963, 7, 124. 4 Cady, G. H. Chem. Eng. News, 1966, 40. 5 Cady, G. H. Inorg. Synth. 1968, 11, 155. 58 CHAPTER 3 SYNTHESES, STRUCTURES, AND SPECTROSCOPIC STUDIES OF CATIONIC CARBONYLS OF PALLADIUM(II) AND PALLADIUM (I) 3.1 Introduction Palladium and platinum find extensive chemical uses as catalysts, including the new and increasing use in reducing pollution from the exhaust gases of cars.1 Many reactions involve metal carbonyl derivatives as precursors or intermediates. Platinum carbonyl chloride derivatives reported by Schiizenberger2 mark the beginning of transition metal carbonyl chemistry. Subsequently, nickel tetracarbonyl, Ni(CO)4, was prepared from nickel and carbon monoxide in 1890,3 which provides the first example of a homoleptic binary metal carbonyl compound. The corresponding homoleptic carbonyls of palladium and platinum exist only as "matrix molecules" by co-condensation of the metal atoms with CO in an inert matrix at temperatures between 4 and 10 K. 4 ' 5 Despite the fact that both cis- and fraAw-Pt(CO)2Cl26 have been known for a long time, the palladium analogues remain elusive. In addition, no thermally stable mononuclear bis(carbonyl) derivatives of dipositive palladium have ever been reported. The recent generation and characterization of platinum carbonyls, [Pt(CO)4][Pt(S03F)6],7 cis-Pt(CO)2(S03F)2,8 and [Pt(CO) 4][Sb 2F„] 2, 9 provides an incentive to extend the synthetic methodology to palladium. 59 3.2 Experimental 3.2.1 Synthesis of cw-Pd(CO)2(S03F)2 3.2.1.1 Gas-Solid Reaction Initially 109 mg (1.024 mmol) of palladium powder was converted quantitatively to 415.5 mg (0.514 mmol) of Pd[Pd(S03F)6] by reaction with S 2 0 6 F 2 in a solution of HS0 3F as previously reported.10 After removal of all volatiles (HS03F and excess of S206F2) and checking the course of the reaction by weighing the reactor, purified CO at an initial pressure of 580 mbar was admitted to the reactor. The dark-brown color of the starting material, distributed as a thin film on the reactor wall, changed to pale yellow within several minutes. To ensure complete reaction, the reactor was kept under CO pressure for another two days. After removal of all volatiles, 370.1 mg (1.027 mmol) of pale yellow solid product was obtained. Elemental analysis of the product gave the composition of Pd(CO)2(S03F)2 (Analysis: Calculated for C2F208PdS2: C, 6.66; S, 17.78%. Found: C, 6.24; S, 17.1%). The compound was established as cw-Pd(C0)2(S03F)2 by vibrational spectroscopy. When cw-Pd(CO)2(S03F)2 was heated to 117 °C in a sealed capillary tube, there was a slight color change from yellow to orange, and the compound decomposed with evolution of gas at 120 °C. 3.2.1.2 Reductive Carbonylation in HSO3F Palladium powder (236 mg) was used to prepare Pd[Pd(S03F)6] as previously described (vide supra). Approximately 5 mL of HS0 3F was added to the 100-mL one-part 60 reactor that contained the freshly prepared Pd[Pd(S03F)6]. A CO pressure of about 500 mbar was introduced to the system afterwards. The rapid CO uptake was noticed both from the pressure decrease and the physical change of the suspension. The original red-brown suspension changed to a clear bright yellow solution within half an hour. The solvent and the volatile reaction products were removed in vacuo first at room temperature and then at 60 °C for 2 hours. An orange^yellow crystalline product was obtained. Elemental analysis of the product gave the composition of Pd(CO)2(S03F)2 (Analysis: Calculated for C 2F 20 8PdS 2: C, 6.66; S, 17.78%. Found: C, 6.67; S, 17.97%). However, vibrational spectra showed greater complexity than that expected for cw-Pd(C0)2(S03F)2, which suggested the presence of a mixture. The product was allowed to stand at room temperature for two months, at which point a complete conversion to cw-Pd(CO)2(S03F)2 occurred. 3.2.1.3 Preparation of Single Crystals of cw-Pd(CO)2(S03F)2 About 150 mg of pale yellow, freshly prepared c«-Pd(CO) 2(S0 3F) 2 from the gas-solid reaction was dissolved in approximately 5 mL of fluorosulfuric acid. The color of the solution changed gradually to orange red, and after about one week, a small amount of yellow-orange crystalline material formed. In order to obtain larger crystals, the mixture was heated in a water bath at 60 °C for 3 hours, whereupon the orange red solid redissolved. On cooling of the solution to room temperature and reducing its volume by removing some of the acid in vacuo, a single yellow crystal plate, together with small amount of powdery precipitate, formed at the bottom of the reactor. The yellow-colored crystal grew over a period of 3 weeks into a relatively large crystal with dimensions of about 2.2 x 1.5 x 1.5 mm. 61 The remaining orange liquid contained cw-Pd(CO)2(S03F)2 and small amount of [Pd2(p> CO)2](S03F)2 as detected from an IR spectrum. The crystal was isolated inside a glove box by removing the remaining solution by pipette and washing the crystal 3 times with small quantities of fluorosulfuric acid. The crystal was dried by removing all volatiles in vacuo at room temperature. 3.2.2 Synthesis of Single Crystals of [Pd2(p>CO)2](S03F)2 Approximately 150 mg of solid pale yellow Pd(CO)2(S03F)2, obtained by reductive carbonylation of Pd[Pd(S03F)6] in HS0 3F, was placed in a round-bottom flask and redissolved in about 5 mL of HS0 3F. The color of the solution changed immediately from pale yellow to light orange. After the solution stood for four days at room temperature, small needle-like orange-red crystals were observed. An additional two weeks was needed to allow the crystals to grow to a size suitable for single crystal X-ray diffraction analysis. The crystals were isolated by pipetting the orange red solutions out of the flask and washing the crystals repeatedly with small amounts of HS0 3F. The fluorosulfuric acid was removed at room temperature first by pipette and then by vacuum distillation. The needles, which had grown to a length of 10-12 mm, were cut into small pieces inside a glove box and fitted into Lindermann glass capillaries. Samples used for vibrational spectroscopy were ground into a fine powder, again, inside a glove box. [Pd2(p>CO)2](S03F)2 is an orange red, extremely hygroscopic crystalline material. Elemental analysis of the product gave the composition of Pd(CO)(S03F) (Analysis: Calculated for CF04PdS: C, 5.14; S, 13.73%. Found: C, 5.09; S, 13.97%). The identity of [Pd2(u;-CO)2](S03F)2 was established by vibrational spectroscopy 62 and single crystal X-ray analysis. [Pd 2(Li-CO) 2](S0 3F)2 melted at 157 °C with decomposition to give a black residue. 3.2.3 Synthesis of [Pd(CO)4][Sb2Fn]2 In a typical reaction, 0.987 mmol of cw-Pd(CO) 2(S03F)2 prepared by the solid-gas reaction was used to synthesize [Pd(CO)4][Sb2Fn]2. The pale yellow powdery cis-Pd(CO) 2 (S0 3 F) 2 was finely ground and transferred into a 100-mL one-part reactor. Approximately 5 mL of antimony pentafluoride was admitted into the reactor by vacuum distillation. A CO pressure of about 1 atmosphere at room temperature was obtained by immersing the reactor in liquid nitrogen and filling the reactor with CO to 250 mbar. A hot water bath and magnetic stirring bar were used to facilitate the reaction. The original viscous yellowish suspension gradually changed color and became less viscous. Within two hours an almost white suspension was obtained, but some yellowish particles suspended in SbF5 were still noticeable. The reaction was allowed to continue for another 2 days until the white suspension became homogeneous. Careful trap-to-trap distillation of the volatile materials from the solvolysis was used to find any evidence of reduction of palladium(II) to palladium(I) because a number of palladium(I) carbonyl derivatives are known.11 The absence of C 0 2 and S 2 0 5 F 2 ruled out this possibility. Removal of all volatiles in vacuo left a white, hygroscopic powdery product. Elemental analysis of the product gave the composition of [Pd(CO)4][Sb2Fn]2 (Analysis: Calculated for C4F2204PdSb4: C, 4.27%. Found: C, 4.41%). In a sealed capillary tube [Pd(CO)4][Sb2F,,]2 began to decompose at 140 °C with a color change from white to orange. At 155 °C the compound melted with gas evolution. 63 3.3 Results and Discussion 3.3.1 Synthetic Aspects 3.3.1.1 ris-Pd(CO)2(S03F)2 and [Pd2(ji-CO)2](S03F)2 The reductive carbonylation of the mixed-valence oxidation state compound PdnPdIV(S03F)6 in fluorosulfuric acid at room temperature, followed by removing the volatiles first at room temperature and then at 60 °C, produced a yellow product, which had a composition of Pd(CO)2(S03F)2. However, the vibrational spectrum was complicated in the S03F" stretching and deformation regions. There were also four bands in the CO-stretching region. In addition to the two bands of medium intensity at 2208 and 2228 cm"1, attributed to cw-Pd(CO)2(S03F)2, two strong bands at 2179 and 1967 cm"1 were also observed. The band positions suggest the presence of both bridging and terminal CO groups, which is evidence for an oligomeric, CO-bridged isomeric species. In order to separate the apparent mixture, two different approaches were employed: (i) recrystallization from fluorosulfuric acid and (ii) an alteration of synthetic conditions - the reductive carbonylation of solid PdIIPdIV(S03F)6 in the absence of the solvent HS0 3F. While the recrystallization in fluorosulfuric acid did not result in the expected product cis-Pd(CO)2(S03F)2, the compound was obtained quantitatively by the gas-solid reaction between PdnPdIV(S03F)6 and CO. During the process of recrystallization of the mixture in HS0 3F, a needle-like crystalline material formed in the solution over a period of 3 weeks when the solution was kept at room temperature. However, there were indications of substantial changes: The color 64 of the solution changed from pale yellow to orange, and the infrared spectrum of the mother liquor showed three bands in the CO-stretching region. Bands at 2213 and 2233 cm"1 belong to cw-Pd(CO)2(S03F)2. The small shift of CO-stretching frequencies to higher wavenumbers is possibly due to some partial ionization of m-Pd(CO)2(S03F)2 in fluorosulfuric acid. In addition to these two bands, a third band appeared at 1982 cm"1, which is slightly higher than that subsequently found in the infrared spectrum of crystalline [Pd2(/x-CO)2](S03F)2. The bands at 2179 and 1967 cm"1, attributed to the oligmeric isomer, were no longer observed, which may suggest that only this form is involved in the formation of [Pd2(^i-CO)2](S03F)2. However, the mother liquor of [Pd2(/x-C0)2](S03F)2, which should only contain two compounds: [Pd2(/x-C0)2](S03F)2 and c/.y-Pd(CO)2(S03F)2 according to IR, produced more crystalline [Pd2(/i-CO)2](S03F)2 when it was allowed to stand at room temperature for another two weeks. The oligomeric isomer, thought to be [Pd2(CO)20/-CO)2(SO3F)2](SO3F)2 (vide infra), could be an intermediate for the formation of [Pd2(/*-CO)2](S03F)2 from cis-Pd(CO)2(S03F)2 in HS0 3F. The failure to detect the oligomeric form in the mother liquor by IR was probably due to its low concentration. In the solid state, however, m-Pd(CO)2-(S03F)2 is more stable and [Pd2(CO)2(/t-CO)2(S03F)2](S03F)2 slowly converts to cw-Pd(CO)2. (S03F)2. This also explains why the gas-solid reaction produces pure ris-Pd(CO)2(S03F)2 and in HS0 3F a mixture of cw-Pd(CO)2(S03F)2 and [Pd2(CO)2(/*-CO)2(S03F)2](S03F)2 forms. The orange-red needle-like crystals of [Pd2(/x-CO)2](S03F)2 obtained in the recrystallization process appear to be a decomposition product of the original mixture. The compound has a strong CO stretching band at 1977 cm"1 in the infrared and 2027 cm"1 in the Raman spectrum. In addition, the decomposition point of [Pd2(/*-CO)2](S03F)2 (157 °C) was 65 higher than that of cw-Pd(CO)2(S03F)2 and the composition, determined by microanalysis, was found to correspond closely to the summary formula of Pd(CO)(S03F). It appears that reductive elimination of S0 3F radicals, their subsequent dimerization to S 20 6F 2 , and the simultaneous loss of 1 mol of carbon monoxide could have occurred according to 2Pd(CO)2(S03F)2 -> [Pd2(/x-CO)2](S03F)2 + 2CO + S 2 0 6 F 2 (3-1) which is followed by oxidation of CO to C0 2 : CO + S 2 0 6 F 2 C 0 2 + S 2 0 5 F 2 (3-2) as observed recently.12 Even though the small quantity of the material used does not permit an unambiguous identification of the gaseous byproducts, the 1 9F-NMR spectrum of the residual solution clearly shows the resonance of S 2 0 5 F 2 at 47.5 ppm.13 As a relevant precedent, the rather uncommon reductive elimination of S 2 0 6 F 2 on heating is observed for palladium fluorosulfate:10 PdIIPdIV(S03F)6 1 2 0 ° c )2Pd(S03F)2 + S 2 0 6 F 2 (3-3) The process involves reduction of Pd(IV) to Pd(H) and requires elevated temperatures. In addition, bis(carbonyl)derivatives of palladium(II) are rare and are considered to be thermally unstable.11 The only other Pd(CO)2-derivatives previously known, organometallic derivatives of the type cw-Pd(CO)2R2 (R = C 6 F 5 , C6C15),1 4 are stable only below -30 °C under CO 66 pressure, but lose CO readily under normal conditions. In addition, the displacement of CO by S03F" in HS0 3F according to [Au(CO)2]+ + SO3F" -» Au(CO)S03F + CO (3-4) is known.12 It appears that the formation of [Pd2(/*-CO)2](S03F)2 from Pd(CO)2(S03F)2 in HSO3F is a rather slow process but ultimately produces single crystals suitable for X-ray diffraction studies (to be discussed later). There are a number of seemingly polymeric or oligomeric precedents of diamagnetic palladium(I) carbonyl derivatives of the type [Pd(CO)X]„, with X = CI, Br, or 0 2 C R . n The first member of this group, [Pd(CO)Cl]„, was reported by Manchot and Konig in 192615 and the molecular structure of an acetate derivative of the composition [Pd(CO)CH3C02]4 • 2CH 3 C0 2 H is known16'17 and will be discussed later in this chapter. The most common synthetic route to these compounds is the reaction of CO with PdX2 (X = CI, Br, or 02CR) in various solvents.11 The conversion of dimeric [Pd(CO)Cl2]2, with bridging CI and terminal CO, to [Pd(CO)Cl]„ in the presence of CO and with acetic anhydride as solvent was also reported.18 But the route discovered here is novel and confirms again that HS0 3F is a good medium for the generation of noble-metal carbonyl cations12'19 and their derivatives. Although recrystallization attempts result in a reduction of Pdn to Pd1, in which Pd(CO)2(S03F)2 slowly converts to [Pd2(/x-CO)2](S03F)2, single crystals of cw-Pd(CO)2-(S03F)2 was obtained from fluorosulfuric acid under slightly different conditions. When a solution of Pd(CO)2(S03F)2 in HS0 3F was allowed to stand for 2 to 3 days, tiny orange crystals formed. The newly formed crystals redissolved when the solution was heated 67 intermittently for 3 hours at 60 °C. According to the IR spectrum of the solution, primarily c«-Pd(CO) 2(S0 3F) 2 was present with [Pd2(/x-CO)2](S03F)2 only as a minor constituent. At room temperature and after gradual, slow removal of H S 0 3 F in vacuo, crystallization of cis-Pd(CO)2(S03F)2 began. The identity of the crystalline material was ascertained by vibrational spectroscopy. Hence solutions of the yellow isomeric mixture in HS0 3F with the composition of Pd(CO)2(S03F)2 will produce, depending on the experimental conditions, two different products: orange-red needles of [Pd 2(/i-CO) 2](S0 3F) 2 or pale yellow plates of cis-Pd(CO)2(S03F)2, both as single crystals. It is fortuitous that the two different compounds both obtained as single crystals from nearly identical solutions allow the determination of two highly unusual structures. One features the first example of a cyclic [Pd2(/*-C0)2]2+ cation and the other represents the first thermally stable bis(carbonyl)palladium(II) derivative, m-Pd(CO) 2 (S0 3 F) 2 . Both compounds are examples of predominantly o-bonded metal-carbonyl derivatives. A third compound, oligomeric Pd(CO)2(S03F)2, still eludes complete characterization since it has never been obtained as a pure compound. Based on the vibrational spectrum, which suggests both bridging and terminal CO ligands, and microanalysis of the mixture, the oligomeric Pd(CO)2(S03F)2 may be tentatively formulated as [Pd2(CO)2(/i-CO)2(S03F)2]-(S03F)2. The v (CO) of the oligomer at 1967 cm"1 is close but measurably different from the v(CO) of [Pd20i-CO)2](SO3F)2 at 1976 cm"1. 68 However, it is found that the infrared spectrum of Pd(CO)2(S03F)2 prepared from fluorosulfuric acid undergoes continuous change in terms of relative intensities of the CO stretching bands. This indicates that a chemical reaction is still proceeding even in the solid state. Therefore, a systematic monitoring of this change was performed by taking the infrared spectrum of the solid sample periodically. Initially, the bands at 2179 and 1967 cm"1 were quite strong, whereas the bands due to cw-Pd(CO)2(S03F)2 ( 2228 and 2208 cm"1) were very weak. As time went by, the intensities of the two originally strong bands (2179 and 1967 cm"1) diminished simultaneously, while the two weak bands due to cw-Pd(CO)2(S03F)2 increased concomitantly in intensity. After over two months, the infrared spectrum did not show further obvious change, and the final infrared spectrum only shows two strong peaks at 2228 and 2208 cm"1. It is concluded that both in solution and in the solid state oligmeric Pd(CO)2(S03F)2 is unstable, and that it will convert either to ds-Pd(CO)2(S03F)2 in the solid state or decompose to [Pd2(/i-CO)2](S03F)2 in HS0 3F. Therefore, the monitoring of the mixture conversion at room temperature via infrared spectroscopy not only lends support to the tentative assignment of the two CO stretching bands at 2179 and 1967 cm"1, but also provides another synthetic route to the preparation of pure c«-Pd(CO) 2(S0 3F) 2 . Compared with the solid-gas reaction mentioned above, this method necessitates relatively longer reaction time (more than two months). Nevertheless, the alternative synthetic route is preferred if a relatively large amount of product is required. 69 3.3.1.2 [Pd(CO)4][Sb2Fn]2 In the synthesis of [Au(CO)2][Sb2Fn], solvolysis in antimony pentafluoride in the presence of carbon monoxide was used to convert Au(CO)S03F to [Au(CO)2][Sb2Fn].19 The application of this procedure to the synthesis of [Pt(CO)4][Sb2Fu]2 was equally successful.9 Hence this approach was adopted to the synthesis of [Pd(CO)4][Sb2Fn]2. Both the pure compound cw-Pd(CO)2(S03F)2 obtained from the gas-solid reaction and the isomeric mixture of Pd(CO)2(S03F)2 prepared from fluorosulfuric acid can be used as a starting material in the preparation of [Pd(CO)4][Sb2Fn]2. The microanalysis results of the samples prepared from the two precursors show no significant difference, and this provides further evidence to support the tentative assignment of the two bands at 2179 and 1967 cm'1 to [Pd2(CO)2(/*-CO)2(S03F)2](S03F)2. The overall solvolysis reaction occurs according to Pd(CO)2(S03F)2 + 8SbF5 + 2CO 6 0 C-> [Pd(CO)4][Sb2Fn]2 + 2Sb2F9S03F SoFs (3-5) As previously noted,19 the reaction mixture is initially very viscous and gentle heating at 60-80 °C is usually required to allow efficient magnetic stirring. 3.3.2 Vibrational Spectra and Structural Aspects 3.3.2.1 [Pd2(At-CO)2](S03F)2 A. X-ray Crystallographic Analysis The crystal structure of [Pd2(/i-CO)2](S03F)2 was determined by Dr. S. J. Rettig and Professor J. Trotter of this Department. Details can be found in Appendix A. Selected bond lengths and bond angles appear in Tables 3-1 and 3-2 respectively. 70 The packing within a polymeric sheet is shown in Figure 3-1, and the coordination environment of palladium, including selected interlayer contacts, is depicted in Figure 3-2. The molecular structure consists of cyclic, planar four-membered [Pd2(jt-CO)2] units (point group £>2/l) with symmetrical carbonyl bridges and a cross-ring Pd-Pd distance of 2.6929(6) A . Individual metallocycles are connected by symmetrically bridging, O-bidentate fluorosulfate groups, giving rise to polymeric sheets. The coordination environment of palladium, ignoring the Pd-Pd crossing interaction, may be described as approximately square planar, with bond angles between 94 and 86.1° and average Pd-C and Pd-0 distances of 1.975 and 2.156 A , respectively. Table 3-1. Bond Lengths [A] a For [Pd2(/i-CO)2](S03F)2 Atom Atom Distance Atom Atom Distance Pd(l) Pd(l)1 3.4777(7) Pd(l) Pd(l)2 2.6939(6) Pd(l) 0(1) 2.159(3) Pd(l) 0(2)3 2.153(4) Pd(l) 0(3)4 2.653(4) Pd(l) C(l) 1.984(4) Pd(l) C(l)2 1.966(4) S(l) F(l) 1.545(4) S(l) 0(1) 1.449(4) S(l) 0(2) 1.444(4) S(D 0(3) 1.420(5) 0(4) C(l) 1.133(6) a Superscripts refer to symmetry operations: (1) 1 - x, y, 1/2 - z; (2) 1 - x, - y, - z; (3) 1/2 - x, y - 1/2, 1/2 - z; (4) 1/2 - x, 1/2 - y, - z; (5) 1/2 - x, 1/2 + y, 1/2 - z. 71 Table 3-2 Bond Angles [°] a for [Pd2i>-GO)2](S03F)2 Atom Atom Atom Angle Atom Atom Atom Angle Pd(l)1 Pd(l) Pd(l)2 101.25(2) Pd(l)1 Pd(l) O(l) 72.3(1) Pd(l)1 Pd(l) 0(2)3 79.4(2) Pd(l)1 Pd(l) 0(3)4 153.94(10) Pd(l)1 Pd(l) C(l) 96.9(2) Pd(l)1 Pd(l) C(l) 2 98.4(2) Pd(l)2 Pd(l) O(l) 138.4(1) Pd(l)2 Pd(l) 0(2)3 87.4(2) Pd(l)2 Pd(l) 0(3)4 104.18(10) Pd(l)2 Pd(l) C(l) 46.7(1) Pd(l)2 Pd(l) C(l) 2 47.3(1) O(l) Pd(l) 0(2)3 87.4(2) 0(1) Pd(l) 0(3)4 84.8(2) O(l) Pd(l) C(l) 168.3(2) 0(1) Pd(l) C(l) 2 92.1(2) 0(2)3 Pd(l) 0(3)4 87.4(2) 0(2)3 Pd(l) C(l) 86.1(2) 0(2)3 Pd(l) C(l) 2 177.8(2) 0(3)4 Pd(l) C(l) 104.6(2) 0(3)4 Pd(l) C(l) 2 94.7(2) C(l) Pd(l) C(l) 2 94.0(2) F(l) S(l) 0(1) 102.8(2) F(l) S(l) 0(2) 104.2(3) F(l) S(D 0(3) 105.4(3) 0(1) S(l) 0(2) 111.1(3) 0(1) S(l) 0(3) 115.5(3) 0(2) S(l) 0(3) 116.1(3) Pd(l) 0(1) S(l) 137.3(3) Pd(l)5 0(2) S(l) 138.8(3) Pd(l)4 0(3) S(l) 130.0(3) Pd(l) C(l) Pd(l)2 86.0(2) Pd(l) C(l) 0(4) 135.8(4) Pd(l)2 C(l) 0(4) 138.2(4) a Superscripts refer to symmetry operations: (1) 1 - x, y, 1/2 - z; (2) 1 - x, - y, - z; (3) 1/2 - x, y - 1/2, 1/2 - z; (4) 1/2 - x, 1/2 - y, - z; (5) 1/2 - x, 1/2 + y, 1/2 - z. 72 Figure 3-1 Stereoview of a single layer of the polymeric sheet structure of [Pd20*-CO)2](SO3F)2. Pdl* Pdl* Figure 3-2 Perspective view of [Pd2(/i-CO)2](S03F)2. 73 There are, in addition, interlayer contacts: a Pd • • • Pd contact of 3.4777(7) A and a Pd • • • 0(3) interaction of 2.653(4) A involving the third oxygen atom of the fluorosulfate group, which results in a rather distorted octahedral environment for each palladium. The observed 4:2 coordination is commonly observed in primarily square planar solid-state structures. A very similar molecular structure type is found for dimethyltin(IV)bis(fluorosulfate), 20 (CH3)2Sn(S03F)2. Again, a sheet-like polymer is present with bidentate, symmetrically bridging fluorosulfate groups and a regular square planar environment for the tin atom consisting of four weakly coordinating oxygen atoms. The octahedral coordination for tin is completed, however, by the two methyl groups, and the third oxygen atom of the fluorosulfate group does not appear to be involved in any significant interaction with the tin atoms. Even though there are additional obvious differences (the [Pd2GK-CO)2] ring is co-planar with the polymeric sheet, while the linear dimethyltin(IV) moiety is aligned perpendicular to it), there are two strong similarities, as can be seen in Table 3-3, where the essential bond distances and angles are listed and compared to the structural data for the bridging fluorosulfate group of dimeric [Au(S03F)3]221 and for NH 4 S0 3 F, 2 2 where an ionic S03F" group is present: (i) Metal-oxygen bonds are long and weak for both (CH3)2Sn(S03F)220 and rPd2(/i-CO)2](S03F)2. The longer Sn-0 distance of 2.24 A , compared to 2.156 A for Pd-O, can be attributed to a slightly larger covalent radius for tin. (ii) The sulfur-oxygen bond lengths for the two fluorosulfate groups in the polymers, labeled S-O(l) and S-0(2), differ only slightly from that of S-0(3), where 0(3) indicates the third oxygen not involved in 74 direct coordination to Sn or Pd. The bond length difference for two types of S-0 groups is only ca. 0.025 A, compared to 0.085 A for [Au(S03F)3]2.21 It is concluded that departure from C3v symmetry is very slight in both cases. In addition, all S-0 and S-F bond lengths are identical within error limits to those reported for NH 4 S0 3 F. 2 2 The assumption that dipositive cations, linear (CH 3) 2Sn 2 + 2 0 and cyclic [Pd2(Li-2+ CO)2] , are linked by S03F" ions packed in bidendate fashion to give layered structures is, in the dimethyltin(IV) compound, consistent with the reported 1 1 9Sn Mossbauer data23'24 for this and other dimethyltin(IV) salts of very strong protonic acids and superacids.25 Table 3-3 Relevant Bond Distances and Angles for Various Fluorosulfate Derivatives Bond Length3 Me2Sn(S03F)2b [Pd2(li-C0)2](S03F)2 [Au(S03F)3]2c NH 4 S0 3 F d and Angles d SO(l) [A] 1.472(8) 1.449(4) 1.467(6) 1.45(2) d SO(2) [A] 1.473(11) 1.444(4) 1.471(7) 1.45(2) d SO(3) [A] 1.454(7) 1.420(5) 1.384(7) 1.45(2) d SF [A] 1.561(8) 1.545(4) 1.507(7) 1.55(2) <0(l)SO(2) [°] 110.5(5) 111.1(3) 111.3 113 < M O S a v [ ° ] 148 138 130.5 -d M O a v [A] 2.24(1) 2.156(4) 2.018 -a 0(1) and 0(2) denote the two coordinated oxygen atoms. b Ref. 20. c Ref. 21, only bond parameters for the bidentate bridging fluorosulfate group are listed. d Ref. 22. 75 Table 3-4 Selected Bond Parameters for the Cyclic [Pd2(p:-CO)2]2+ Cation Bond Distances and Angles [ P d ^ L t - C O ^ ^ F ^ [Pd4(Li-CO)4(Li-CH3C02)4] • 2CH 3 C0 2 H* d Pd-Pd [A] 2.6929(6) 2.663(1) d Pd Pd [A]b 3.4777(7) 2.909(1) d Pd(l)-C(l) [A] 1.966(4) 1.97 dPd(2)-C(ll) [A] 1.984(4) 2.01 dPdO a v[A] 2.156(4) 2.12 dCO[A] 1.133(6) 1.15 <Pd(l)C(l)Pd(2) [°] 86.0 84.2 a Ref. 16, ESDs are only quoted where they could be found in the reference. b Denotes in the case of [Pd2(p>CO)2](S03F)2 the interlayer Pd-Pd contact, and for Pd4(u.-CO)4(u.-CH3C02)4] • C H 3 C 0 2 H the Pd-Pd bonds with two acetate bridges. The discussion now turns to the cyclic [Pd2(p>CO)2] cation. While several molecular structures of palladium(I) carbonyl derivatives with bridging CO groups are known, one of these, that of a tetranuclear cluster found in a complex of the composition [Pd4(p> CO)4(|i-CH3C02)4] • 2CH 3 C0 2 H, 1 6 ' 1 7 is of particular relevancy. The cluster consists of two [Pd2(p>CO)2]2+ cations with symmetrically bridging CO groups and bond distances and angles very similar to those of the [Pd2(p:-CO)2]2+ cation in [Pd2(/x-CO)2](S03F)2, as the summary in Table 3-4 indicates. However, in the case of the tetranuclear cluster, two [Pd2(p>CO)2]2+ cations are linked by symmetrically bridging acetate ions into dimers, giving rise to a planar, nearly rectangular arrangement of four Pd(I) centers alternately bridged by two CO and two 76 CH 3 C0 2 groups. In addition, hydrogen-bonded acetic acid dimers are found in lattice spaces between the metal clusters. As can be seen from Table 3-4, there is a very good correlation between the two sets of data, with only minor exceptions. The rather long Pd • • • Pd contact of 2.909(1) A for [Pd4(p>CO)4(Li-CH3C02)4] • 2CH 3C0 2H, supported by two bridging acetate groups, is compared in the case of [Pd2(/x-CO)2](S03F)2 to an interlayer contact of 3.4777(7) A. The shorter Pd-O bond length of 2.12 A are a reflection of greater basicity of the acetate group compared to that of the fluorosulfate group. On the other hand, the short Pd-Pd bond distances, supported by two bridging CO groups in each case, are almost identical to the previously reported26'27'28'29 Pd-Pd single bonds for Pd(I) complexes, which are commonly found in the range of 2.53-2.69 A. The two Pd-C distances are in each case not completely identical, but the differences are almost within the estimated standard divisions. The C-0 bond distance in the [Pd2(u.-CO)2]2+ cation is rather short for a bridging carbonyl ligand. The distance is 1.133(6) A, only slightly longer than that of free CO, 1.1281 A. 3 0 For the acetate derivative a value of 1.15 A is reported.16 It is anticipated that the CO-stretching frequencies for both compounds, to be discussed subsequently, will allow a better comparison. There is less agreement with structural data reported for other Pd(I)-carbonyl derivatives where other rc-acceptor ligands are also present in the molecule. Frequently, as in Pd2(Li-CO)Cl2 • 3PEt2Ph29 or [Pd(dppm)02CCF3]2(Li-CO)31 (dppm = bis(diphenylphosphino)-77 methane), asymmetric carbonyl bridges are present with C-0 distances of 1.156(4) and 1.20(3) A, 3 1 respectively. In the case of (Li-carbonyl)(Li-(bis(diphenylarsino)-methane)-dichlorodipalladium(I), a symmetric carbonyl bridge is claimed;32 however, the Pd-C distances of 1.84(5) and 1.95(6) A are not sufficiently accurate to be conclusive evidence for this claim. In almost all of these structurally characterized palladium© carbonyl deriva-tives, 1 5 , 1 6 ' 2 9 , 3 1 ' 3 2 only bridging CO ligands are encountered. There is, however, one exception - bis[(u.-chloro)dicarbonylpalladium(I)], [Pd(CO)2(U--Cl)]2,33 where bridging chlorides and only terminal CO ligands are found. While there is no precedent for both composition and structure among other palladium carbonyl derivatives,11 a ' b close inspection of the reported unit-cell parameters, internuclear distances, and the IR stretching frequencies of [Pd(CO)2(u,-Cl)]2 3 3 results in the surprising observation that these data are identical within error limits to the previously reported parameters for the well-known complex [Rh(CO)2(Li-Cl)]2.34'35 2 + In summary, the cyclic [Pd2(Li-CO)2] discussed here departs from previously characterized palladium(I) carbonyl derivatives in three aspects: (i) It is the first structurally characterized binary carbonyl cation formed by a group-10 metal; (ii) the molecular cation is a rare case of a thermally stable, inorganic cyclic molecule of D2h symmetry; and (iii) the CO bond distance is very short for a symmetrically bridging carbonyl ligand. B. Vibrational Spectra Vibrational spectroscopy has been very useful in studies of metal carbonyl 36 37 compounds ' for three reasons: (i) CO-stretching vibrations are usually observed between 78 2100 and 1600 cm"1, where they are easily identified and where vibrational coupling to other fundamentals is usually not important, (ii) The band position of v (CO) is very sensitive to the subtle changes in the coordination and bonding of the CO ligand to the metal center. Hence information regarding the molecular geometry of metal carbonyl compounds as well as 36 37 the bonding characteristics of the CO ligand has become readily available in the past ' from the study of the CO-stretching vibration, (iii) In areas of high C-0 bond order (between 2 and 3), small changes in bond order are more clearly manifested in shifts of v (CO) (cm1) than in the internuclear distance of C-0 (A), where estimated standard deviations and thermal motions cause uncertainties. In spite of these advantages, a detailed vibrational analysis, except for the CO-stretching region, of transition-metal carbonyl derivatives is rarely attempted. This is largely due to experimental limitations. Equipment to record far-IR spectra is commonly not available. Raman spectra are difficult to obtain because many metal-carbonyl derivatives are intensely colored and photolytically labile. In addition, necessary structural information is often either not available, or the molecular structures are far too complex for a meaningful vibrational analysis of the molecules. All the aforesaid limitations do not apply in this case. The molecular structure of [Pd2(/x-CO)2](S03F)2 reveals two weakly interacting ions, and the molecular structure of 20 23 24 (CH3)2Sn(S03F)2 (where the complete vibrational spectrum is known ' ) provides a useful precedent. The cyclic [Pd2(/i-CO)2]2+ cation has D2h symmetry. Since there is no preferred rotational axis, arbitrary choices have to be made. The 2-fold axis with both CO groups 79 placed on it is designated as the z axis, and the Pd-Pd bond axis becomes the y axis. The irreducible representations of normal vibrations for the cation [Pd2(/t-CO)2]2+ are r v i b = 3A g[v„ v2, v3(R, p] + B2g[v4(R, dp)] + 2B3g[v5, v6(R,dp)] + 2Blu[v7, v8(IR)] + 2B2JVO, v10(IR)l + 2B3U[v„, v12(IR)] (3-6) for a total of twelve nondegenerate fundamental. Six should be observable only in the Raman, and the other six, only in the IR spectrum. Even though it has been argued in the preceding section that the fluorosulfate ion, on the basis of bond distances and angles, approaches C3v symmetry, the interaction with cyclic [Pd2(/x-CO)2]2+ cation results in a symmetry reduction to Cs. The irreducible representations for this point group applied to a five-atom molecule are r v i b = 6A'[v,-v6 (R, p, IR)] + 3A" [v7-v9(R, dp, IR)] (3-7) With all 9 fundamental due to the anion being both Raman and IR active, while the 12 fundamentals due to the cation are subject to the mutual exclusion rule, a differentiation between two sets of fundamentals and hence an unambiguous identification of the bands due to the cation should in principle be possible. The vibrational bands for [Pd2(/*-CO)2](S03F)2 down to 50 cm"1 are listed in Table 3-5 and compared to the bands due to fluorosulfate group in (CH3)2Sn(S03F)223'24 down to 300 cm"1. Of the Raman bands of the latter compound, a weak S0 3 stretch at 1231 cm"1 had previously not been observed but was detected in the FT-Raman spectrum obtained in this study. The FT-Raman spectrum of [Pd2(^-CO)2](S03F)2 is shown in Figure 3-3. 80 Table 3-5 Vibrational Spectra Data* for \Pd2(M>-C0hKS03p)n and (CH3)2Sn(S03F)2 IR Spectra Raman Spectra rPd,(M-co),](SOO09 (CIL,)9Sn(SChF)7 rPd9(M-CO)9l(SO,F)9 (CH1)9Sn(S03F)9 vtcnr1] Int. vfcnr1] Int. A vfcnr1] Int. Avrcnr1] Int. 1977 s ~1940 vw, sh 1311 s 1351 vs, br 2027 vs 1330 m 1303 1354 s 1202 vs 1189 vs, br 1215 vw 1189 w 1231 w 1083 s 1088 m 1072 s, br 1086 vs 1088 s 827 ms 827 m, sh 823 m 826 ms 682 ms 638 m,sh 603 m 600 m, sh 600 w 610 w 585 m, s 581 ms 586 m 584 m 567 s 556 ms 572 551 mw 435 m 417 s 435 m 420 w 424 m 426 m 416 m s 406 m 409 s 395 ms 360 w 304 w 265 ms 256 ms 367 mw 320 mw 243 vs 112 ms ' 76 m 118 w 97 vw 85 w * Vibrational data refer to solid samples, and the data listed in italics refer to [Pd2(/*-CO)2] 81 9801 82 The bands attributed in both compounds to the fluorosulfate group show reasonable correspondence. The S03-stretching vibrations of [Pd2(/ti-CO)2](S03F)2 found at ca. 1310 and 1200 cm"1 are both split by about 25 cm"1 in the Raman spectrum only. This splitting may be caused by factor group effects. The two S0 3 bands arise from the removal of the degeneracy of the asymmetric S03-stretching mode (E) for ionic fluorosulfates, usually found at about 1280 cm"1,37 due to a symmetry reduction from C3v to Cs. The extent of this splitting may be taken as a measure of the covalent anion-cation interaction and the departure from C3v symmetry for the fluorosulfate group. The actual band separations are ca. 110 cm"1 for [Pd2(/*-CO)2](S03F)2 and ca. 170 cm"1 for (CH3)2Sn(S03F)2, consistent with observations in regard to S-O bond lengths for both compounds, as summarized in Table 3-3 and discussed in the preceding section. Finally it is noted that the S-F stretching vibrations are intermediate between v(SF) values for ionic alkali metal fluorosulfates at 750-780 cm"1 3 8 and covalent fluorosulfates at about 880 cm" . In summary, the internal vibrations of the fluorosulfate groups in [Pd20*-CO)2](SO3F)2 are consistent with weakly interacting fluorosulfate ions, arranged in bidentate fashion. It hence appears justified to recognize some bands as being characteristic of the anion with the remaining bands, which will follow the mutual exclusion rule, as arising from the cation. In view of the rather weak covalent anion-cation interaction and interlayer contacts revealed by the X-ray molecular structure determination, vibrations involving both the cation and anion should be observable only at very low wavenumbers. Consistent with this view of weak covalent anion-cation interaction, only single bands due to the CO-stretching vibrations are observed in the Raman at 2027 cm"1, and the IR 83 spectrum at 1977 cm"1, for [Pd2(/i-C0)2](S03F)2. A very weak shoulder at about 1940 cm"1 in the IR spectrum is assigned to v( 1 3CO). For [Pd4(/i-CO)4(/i-02CCH3)4] two intense IR bands are found at 1940 and 1975 cm"1 1 6 and for the solvate with acetic acid, where the molecular structure is known,16'17 the IR bands are shifted to 1934 and 1968 cm"1. Unfortunately, Raman spectra do not appear to have been reported for either cluster compounds. For seemingly polymeric [Pd(CO)X] (X = Cl, Br),4 0 intense IR bands at 1978 and 1953 cm"1, respectively, are reported, but for both compounds several bands of low intensity are observed in the CO-stretching region, at both slightly lower and slightly higher wavenumbers. The bands at lower wavenumbers, two for each compound, may be interpreted either as 13C-satellite bands or as combination bands. The bands at higher wavenumbers, at 2023 and 2002 cm"1 for the chloro and at 2008 and 1996 cm"1 for the bromo derivative, may be symmetric CO-stretching vibrations, which are split by vibrational coupling and should be only Raman active for point group D2h. Their observation in the IR spectrum implies that the symmetry in the suggested [Pd2(Li-CO)2]2+ units40'41 is lower than D2h. Due to thermal instability of the two Pd(I)-carbonyl derivatives, a Raman spectrum, which would confirm this interpretation, is again not reported. A common feature of all Pd(Li-CO) derivatives discussed here16'40 and of [Pd2(Ai-CO)2](S03F)2 is the unusual band position of v(CO), which is about 200 cm"1 higher than for bidentate bridging groups in typical transition metal carbonyl compounds.36'37 For neutral, polynuclear Pd(I) derivatives, the situation is different. In addition to the derivatives cited above, which are of the general type [Pd(CO)X]n with X = Cl, Br, CH 3 C0 2 , or 84 S O 3 F , 1 6 ' 1 7 ' 4 0 ' 4 1 where rc-back-donation appears to be reduced, there are examples among the complexes with reported crystal structures where the CO-stretching vibrations are considerably lower and fall into a more normal range for bridging carbonyls.36'37'41 However, in these cases other rc-acid ligands are present in the molecule in addition to CO. In [Pd(dppm)(02CCF3)]2(p>CO)31 (dppm = bis(diphenylphosphino)methane), an asymmetric CO bridge is observed and v(CO) is reported at 1720 cm"1. For the analogous dpam derivative (dpam = bis(diphenylarsino)methane), [Pd(dpam)Cl]2(p>CO),32 the CO bridge is symmetric within experimental error (Pd-C = 1.84(5) and 1.95(6) A), and v(CO) is again observed at 1720 cm"1, which is within the range 1700-1860 cm"1 quoted for bridging CO ligands in transition-metal carbonyl derivatives.40 The latter examples allow the comment that addition of a bidentate it-acceptor ligand such as dpam to polymeric [Pd(CO)Cl]n with the principal v(CO) band at 1978 cm"1,3 8'3 9 reduces v(CO) by 258 cm"1. Besides structural changes, the CO-Pd bond type is altered by increasing the metal to CO rc-back-donation in the resulting complex. For the cation [Pd2(p>CO)2]2+ the identification and assignment of v(CO) and the identification of the remaining fundamentals is easily accomplished. As seen from Table 3-5 a total of 10 vibrational bands (6 IR active, 4 Raman active), listed in italics, are attributed to the cation. The 12 fundamental vibrations of the [Pd2(p>CO)2]2+ ring are depicted in Figure 3-4 together with tentative assignments. These assignments are, in the absence of relevant precedents, based largely on the relative masses of the atoms involved in various vibrational motions. 85 ° 1 I* i + R,P p < > * p \ , P i * P d v T p d ?A % v [cm"1] o+ Vj 2027 ° v 2 416 ° V 3 118 -o B 2 g . . A Pd Pd R, dp vtcm'1] + ° V4~600(n.o.) -<-o o R,dP P d w p d PdVf vtcrrT1] ° ^ V 5 ~600(n.o.) 0 V 6-260 O f |0 B l u t / 5 N * IR P d w P d P \ V d ?! I? v [cm'1] 0 t V 7 1977 +° V 8 395 o-+- o-* IR P d w P d ^ P d N ^ P d v [cm1] 0 V 9 682 0 V 1 0 638 +o + o ra " P d Pd" Pd Pd v [cm-1] +° V U 243 +° V 1 2 112 (n.o. denotes not observed) Figure 3-4 Fundamental vibrations of [Pd2(At-CO)2]2+'. 86 It appears that two of the six Raman-active vibrations are not found. The two missing bands are v4(B2g), a symmetric, out-of-plane CO deformation, and v5(B3g), a symmetric in-plane CO deformation mode. Both vibrations are similar and may even coincide, due to accidental degeneracy. They are expected in the region of about 600 cm'1, where overlap with anion bands can occur. In addition, as in the case of [Au(CO) 2] 2 +, 1 9 the band due to deformation modes may be of low intensity. Therefore an attempt has been made to synthesize similar compounds with same cation but different anion in order to find the two missing vibrational modes. The target compound [Pd2(|i-CO)2][Sb2Fn]2 (see Chapter 6) was obtained successfully via the solvolysis in antimony pentafluoride without additional CO. However, the assignment of the two missing bands was not achieved due to a similar problem. In summary, the cyclic Pd2(|i-CO)2 unit appears to be present in all Pd(I)-monocarbonyl derivatives of the type [Pd(u.-CO)X]„ (X = CI, Br, CH 3 C0 2 , S03F) known so far. Structural differences appear to be caused by the bridging anion X. Strong, effective bridging by the halide anions appears to lead to, alternately, CO and X (X = CI or Br) bridged polymers according to their vibrational spectra.40'41 Less effective bridging by bidentate acetate groups results in the pairing of two [Pd2(H-CO)2] units according to the molecular structure,16'17 consistent with the vibrational spectra. The weakly nucleophilic, poorly bridging fluorosulfate anions, arranged in a bidentate mode, isolate individual [Pd2(p> CO) 2 ] 2 + cations sufficiently to permit a vibrational treatment in terms of point group D2h. Common to all palladium(I) carbonyl derivatives of the [Pd(p>CO)X]n type are the unusually high CO stretching frequencies, which for [Pd2(/t-CO)2](S03F)2 average out to 2002 1 1 Qi» >n cm" , about 200-300 cm" higher than observed for typical transition-metal carbonyls. ' ' 87 The bonding situation hence parallels other cationic Pd(II)-carbonyl derivatives with terminal CO groups, to be discussed later in this chapter. 3.3.2.2 cw-Pd(CO)2(S03F)2 A. Vibrational Spectra Although the synthesis of d.y-Pd(CO)2(S03F)2 is quite different from that of cis-Pt(CO)2(S03F)2, the IR and Raman spectra for both compounds show many similarities. Their infrared and Raman frequencies, together with that of Au(CO)S03F as reported previously,12 are listed in Table 3-6. Also listed in Table 3-6 are the approximate assignments for the Pd(II) and Pt(II) derivatives. The IR spectrum of cw-Pd(CO)2(S03F)2 and the v(CO) region of the Raman spectrum are shown in Figure 3-5. The similarity between cis-Pd(CO)2(S03F)2 and cw-Pt(CO)2(S03F)2 in terms of band positions and intensities becomes obvious from the listing in Table 3-6. The greater complexity of the two sets of spectra for both compounds compared to the rather simple spectra of Au(CO)S03F1 2 is not unexpected on account of vibrational coupling primarily between the two fluorosulfates which, from the small band separations, appear to be weakly coordinated to the two respective metal centers. A similar band distribution and coordination mode has been observed previously for M(CO)5S03F (M = Mn or Re).43 The low frequency of v(SF) at about 800 cm"1 and the position of v (SO • • • M) at 1000-1040 cm"1 are consistent with a high ionic character for the metal-OS02F bonds in all instances. 88 Table 3-6 The Vibrational Data* (cnr1) for ri.y-Pd(CO)2(S03F)2, cw-Pt(CO)2(S03F)2, and Au(CO)S03F with Estimated Band Intensities and Approximate Assignments a>Pd(CO) 2 (S0 3 F) 2 m-Pt(CO) 2(S0 3 F) 2 a A u ( C O )S0 3 F b Approx. Assignm. IR Raman IR Raman IR Raman 2228 ms 2208 s 2166 w,sh 2228 vs 2212 w, sh 2207 ms 2219 s 2185 vs 2145 vw 2218 vs 2191 m 2181 s 2195 s 2198 vs v s y m ( 1 2 CO) v s y m ( 1 3 CO) Vas (1 2CO) v a s (1 3CO) 1380 s 1360 s 1382 s 1358 ms 1397 s 1389 s, sh 1378 vs 1362 w, sh 1395 s 1376 m 1360 vs 1368 w Vas S 0 2 1229 s, sh 1206 vs 1208 s 1230 m,sh 1209 vs 1212 s 1198 vs 1200 m v s S 0 2 1038 s 1019 vs 935 vw 1031 vs 1003 s 1034 m, sh 1026 s 1009 vs 1019 s 993 m 1020 vs 1020 s VSO...M 794 vs 808 m 789 mw 799 s 815 w 792 w 810 s 768 s 792 w, br vSF 648 s 638 m, s 649 vs 657 ms 648 msh 656 s 641 m 640 m vMO + 5S02 586 vs 586 m 579 m 589 s, sh 584 s 588 mw 580 m 577 s 578 m 5as S 0 3 554 s 559 m 552 m, w 557 m s 551 s 565 w 554 w 554 s 553 w 5sym S O 3 514 ms 476, 472 ms 475 w, sh 8M-CO? 89 Table 3-6 (Continued) 440 ms 455 mw 436 w 462 ms 445 w 456 w VMO + 8S03 417 vw 405 w 411 vw 412 w 411 w 400 w pS0 3 F 406 ms 388 s 390 w 290 vs 291 m 276 vw 251 m 268 m unassigned torsion 167 m, sh 175 w and 147 ms 151 m deformation 134 m 138 vw 115 w modes * Vibrational dada refer to solid samples. a See reference 8. b See reference 12. Abbreviations: IR = Infrared, s = strong, m = medium, w = weak, v = very, br = broad, sh = shoulder, v = stretching mode, 8 = deformation mode p = rocking, sym = symmetric, as = asymmetrical. It is also apparent that the spectra are dominated by the fluorosulfate bands in the frequency range below 600 cm"1. Hence it is very difficult to assign metal-carbon stretching and C-0 deformation modes unambiguously without help from labeled (1 3C and 180) isotopomers. Based on the previous report for [Au(CO)2]+,1 9 these bands are very weak and difficult to recognize. They also coincide partly with deformation modes due to the fluorosulfate group. On the other hand, the CO-stretching modes are clearly recognizable and allow the following conclusions: 90 3 D M V 1 1 I W S M V U L % 91 (i) The principal species present appears to be ri.y-Pd(CO)2(S03F)2 as determined from the coincident bands in the Raman and IR spectra. This is also true in the case of cis-Pt(CO)2(S03F)2. However, there are weak bands present in the Raman and IR spectra of both Pd(CO)2(S03F)2 and Pt(CO)2(S03F)2, 2212 and 2166 cm"1 for the former and 2191 and 2145 cm"1 for the latter. According to the mutual exclusion rule, this may lead to a tentative assignment of the vibrational modes to the corresponding trans isomers. The crystal structure analysis of a.y-Pd(CO)2(S03F)2 (vide infra) and the IR spectrum taken of the single-crystal sample argues against this tentative assignment. Therefore, it is more likely that the weak peaks that appear in the vibrational spectra of both compounds come from isotopic satellites rather than isomeric impurities. (ii) The average stretching frequency is higher for cw-Pd(CO)2(S03F)2 (2218 cm"1 IR and 2217.5 cm"1 Raman) than for cw-Pt(CO)2(S03F)2 (2202 cm"1 IR and 2198.5 cm"1 Raman). This has been noted previously for the binuclear pair M2C14(C0)2 (M = Pd or Pt),44 for Pd(CO)2R2 (R = C 6 F 5 or C6C15),1 4 and for the anion pairs [MX3(CO)]" (M = Pd or Pt, and X = Cl or Br).45 In all these instances terminal rather than bridging CO groups are present. (iii) As the anionic group becomes less nucleophilic, the average CO stretching frequency will increase. This conclusion is also based on a series of platinum(II) carbonyl derivatives because d.s-Pd(CO)2(S03F)2 remains the only thermally stable mononuclear Pd(II) carbonyl derivative. However, attempts have been made to synthesize the corresponding compound dj-Pd(CO)2(Sb2Fn)2 ( see Chapter 6 ), which provides further evidence to support the conclusion reached here. 92 The CO average stretching frequency for d.s-Pt(CO)2(S03F)2, observed in the Raman spectrum of a solid sample at 2198.5 cm"1, is higher than that reported for cw-Pt(CO)2Cl2 (2164.5 cm"1) or cw-Pt(CO)2Br2 (2153 cm"1).45 Similar small shifts are found for Au(CO)S03F (2197 cm"1)12 and Au(CO)Cl (2083 cm"1).45 As observed previously for cis-Pt(CO)2Cl245 and cw-M(CO)2R214 (M = Pt or Pd, R = C 6 F 5 or C6C15), the symmetric CO stretching frequencies are observed at higher wavenumbers than their asymmetric counterparts. The assignment here is based on the intensities of the IR and Raman bands in this region. Finally, an interesting observation deserves some comment. When the IR spectra were recorded for solutions of the two carbonyl derivatives in HS0 3F, the two v(CO) bands observed were shifted to slightly higher wavenumbers. For Pd(CO)2(S03F)2, both v(CO) bands were shifted from 2208 and 2228 cm"1 to 2214 and 2233 cm"1, respectively. For Pt(CO)2(S03F)2, v(CO) shifts from 2185 and 2219 cm"1 to 2191 and 2224 cm"1 were noted on dissolution in fluorosulfuric acid. It is suggested that partial ionization according to M(CO)2(S03F)2 - * [M(CO) 2SO 3F] + ( s o l v 0 + S0 3F' (3-8) may be responsible for the observed upward shifts by about 6 wavenumbers. Another possibility is that the neutral molecules may be protonated in HS0 3F according to M(CO)2(S03F)2 + HS0 3F -> [M(CO) 2(S0 3F) 2H] + ( s o l v ) + S03F" (3-9) 93 B. X-ray Crystal Structure Analysis The molecular structure of cw-Pd(CO)2(S03F)2 was solved by Dr. R. J. Batchelor and Professor F. W. B. Einstein at Simon Fraser University. Crystallographic details and the atomic coordinates are listed in Appendix B. Rigid body analysis46 of the anisotropic thermal parameters of the molecule yielded R = 0.27 for the agreement between observed and calculated and, as expected, underestimated the thermal motion of many of the outer atoms while overestimating that for the palladium atom. Analysis of the internal motion of the molecule in terms of the segmented rigid body47 in which each -OS02F group was allowed internal libration about the S(l)-0(11) and S(2)-0(21) bond axes gave R = 0.17 but showed similar discrepancies between calculated and observed thermal parameters as did the simple rigid body model. Models in which different internal librational axes for the OS0 2F groups (such as the Pd-0 bond axes) were specified either produced no significant improvement over this model or were ill-conditioned. Each OS02F group, when independently modeled as a rigid body, gave R = 0.015, 0.031 for the groups containing S(l) and S(2), respectively, and no significant residual motion. These analyses, along with independently determined riding motion corrections for selected pairs of atoms, were used to estimate the probable ranges of corrections (always positive) for the bond lengths given in the footnote to Table 3-7. The molecular structure, shown in Figure 3-6, shows that the coordination about the palladium atom is square planar with ris-stereochemistry. Bond distances and angles are given in Table 3-7. The molecule has no crystallographic symmetry and deviates from 94 Table 3-7 Selected Intramolecular Distances (A) and Angles (°) for c/s-Pd(CO)2(SC>3F)2 at 2 2 0 K . * Bond Distances Pd-O(ll) 2.016(3)a S(l)-F(l) 1.557(3) Pd-0(21) 2.006(3) a S(2)-0(21) 1.480(3) Pd-C(2) 1.945(5) b S(2)-0(22) 1.402(4) Pd-C(l) 1.919(5)b S(2)-0(23) 1.394(4) S(l)-0(11) 1.479(3)c S(2)-F(2) 1.565(4) S(l)-0(12) 1.422(4) d C(2)-0(2) 1.102(6) S(l)-0(13) 1.413(4)e C(l)-0(1) 1.114(6) Bond Angles 0(11)-Pd-0(21) 86.00(13) F(l)-S(l)-0(13) 105.43(22) 0(11)-Pd-C(2) 91.71(16) 0(21)-S(2)-0(22) 113.70(22) 0(11)-Pd-C(l) 176.23(17) 0(21)-S(2)-F(2) 99.62(21) 0(21)-Pd-C(2) 176.25(16) 0(21)-S(2)-0(23) 111.30(23) 0(21)-Pd-C(l) 90.39(17) 0(22)-S(2)-F(2) 103.19(24) C(2)-Pd-C(l) 91.83(19) 0(22)-S(2)-0(23) 121.4(3) 0(11)-S(1)-0(12) 113.58(19) F(2)-S(2)-0(23) 104.2(3) 0(11)-S(1)-F(1) 101.68(19) Pd-0(11)-S(l) 123.48(19) 0(11)-S(1)-0(13) 111.35(22) Pd-0(21)-S(2) 124.68(19) 0(12)-S(1)-F(1) 104.50(20) Pd-C(2)-0(2) 178.9(4) 0(12)-S(1)-0(13) 118.24(23) Pd-C(l)-0(1) 177.5(4) * Thermal motion corrections using rigid body, segmented rigid body or riding models indicate that the actual bond lengths should be greater by: a 0.004 to 0.009 A; b 0.001 to 0.004 A; c 0.001 to 0.019 A; d 0.008 to 0.014 A; e 0.009 to 0.022 A; f 0.010 to 0.024 A; & 0.018 to 0.020 A; h 0.020 to 0.049 A; 10.024 to 0.033 A; •> 0.002 to 0.023 A. 95 96 molecular mirror symmetry in the torsion angles about the Pd-0 bonds (cf. 0(21)-Pd-0(11)-S(l), -122.0(2)°; 0(11)-Pd-0(21)-S(2), 106.5(2)°) and, more markedly, in the torsion angles about the O-S bonds (cf. F(l)-S(l)-0(11)-Pd, 73.8(2)°; F(2)-S(2)-0(21)-Pd, 83.0(2)°. Note that mirror symmetry would require the opposite sense of rotation, which is not the case, for the latter pair of torsion angles. The lengths of the chemically equivalent bonds are generally not significantly different with the notable exceptions of the terminal S-0 bonds and also Pd-C bonds. The variations in both the S-O and Pd-C bond lengths appear to reflect the degree to which the oxygen and carbon atoms concerned are involved in weakly attractive intermolecular and intramolecular interactions (see below). The bond lengths fall within the expected ranges. The lengths of the C-0 bonds lie at the lower end of the range exhibited for transition metal carbonyls, in keeping with the spectroscopic results, as discussed above. There is a very small displacement (0.036(3) A) of the palladium atom out of the least-squares plane of the coordinating atoms toward an oxygen atom of a fluorosulfate group on a neighboring molecule (Pd ••• 0(22)', 3.007(4) A, where ' indicates 1-x, -y, -z). The sum of There are no other intermolecular distances to palladium; which are less than the sums of the corresponding pairs of van der Waals radii. On the opposite side of the coordination plane the nearest interatomic distances from a neighboring molecule to the palladium atom are as follows: Pd • • • 0(23)", 3.250(3)A, and Pd-"0(13)", 3.301(4) A (where" indicates 1/2+x, 1/2-y, 1/2+z). There are a number of intermolecular C - 0 contacts (four for each carbon atom) to the terminal oxygen atoms of the fluorosulfate groups in the range of 2.839(6)-3.172(6) A which may provide a stabilizing influence on the structure. In a complementary fashion, each 97 of the terminal oxygen atoms of the fluorosulfate groups has two intermolecular contacts to carbon atoms. There is also an intermolecular contact between 0(2) and 0(13)"' of 2.861(5) A (where " ' indicates 14-x, y, z). In light of numerous C - 0 contacts it is of interest that this latter distance represents the only significant nonbonded contact involving a carbonyl oxygen atom. The aforementioned Pd • • • 0(22)', C - 0 , and O • • • O distances are the only inter-molecular separations significantly less than the sums of the appropriate pairs of van der Waals radii. These nonbonded contacts as well as the "unimposed" intramolecular contacts C(l) • • • F(2) (3.051(6) A) and C(2) • • • 0(12) (2.843(6) A) are shown in Figure 3-7. This figure includes the four complete molecules whose palladium atoms lie within the unit cell plus additional atoms to show the significant intermolecular contacts to one molecule, the atomic coordinates of which are given in Appendix B. The unit cell is shown for reference. For clarity, not all of the molecular fragments which protrude into the cell are depicted. The weak C • O (and C ••• F) interactions discussed above allow another comment. In previous discussion of thermally stable noble metal carbonyl salts and cationic derivatives such as Au(CO)S03F and [Au(CO)2][Sb 2Fn], 1 2 , 1 9 the thermal stability as well as the extent to which v (CO) was raised above the value of v (CO) in gaseous CO was explained in terms of polar contributions to the metal-CO bond by the weak nucleophilicity of the counteranion or the anionic moiety. It appears now that the molecular nature of the anions may allow for additional stabilization of the coordinated carbonyl ligands via C ••• O or C ••• F contacts. Secondary interactions involving the carbon atom of the CO ligand have seemingly not been noticed for other noble metal carbonyl derivatives.49'50'51'52 It is interesting however to note that, in the structure of linear Au(CO)Cl,49 individual molecules are packed in such a 98 manner that, in addition to Au ••• Au contacts of 3.38 A, each carbon has four contacts to Cl of adjacent molecules of about 3.35 A; slightly less than the van der Waals radii of 3.45 A. 4 8 C. Structural Comparison to Related Compounds As stated earlier in this chapter, cw-Pd(CO)2(S03F)2 is the first example of a thermally stable, mononuclear carbonyl derivative of palladium(II). Its molecular structure is here solved by single crystal X-ray diffraction analysis. There is a report on the synthesis and the molecular structure determination of a dinuclear complex with the formula [Pd((x-Cl)(CO)2]2,33 where palladium has the formal oxidation state of +1. The central atoms appears to be in a distorted square planar environment, the CO ligands are terminal, and asymmetric chloride bridges are claimed.33 Both the composition and the structure are uncharacteristic and unprecedented for a palladium© carbonyl derivative.11 In all reported Pd(I) structures,16'17'29'31'32 including [Pd2(/i-C0)2](S03F)2 obtained in this research, bridging 16 32 CO ligands are observed while chloride, where present, ' is a monodentate, terminal ligand. Since the reported unit cell parameters, internuclear distances, bond angles, and the IR band positions and relative intensities reported for [Pd(Li-Cl)(CO)2]2 are, within accepted error limits, identical to those published previously for [Rh(u.-Cl)(CO)2]2,34'35 a reinvestigation into the Pd system is suggested. The two anions [Pd(CO)Cl3]~ and [Pd(CO)Br3]"53 have the square planar structure in common with cw-Pd(CO)2(S03F)2. While the CO distances reported are not significantly different, the Pd-C distance for [Pd(CO)Cl3]" is 1.87(1) A, significantly shorter than 1.945(5) and 1.919(5) A found for cw-Pd(CO)2(S03F)2, which would suggest more 7t-back-donation 99 for the Pd-CO moiety in this anion. This is consistent with v(CO) of 2132 cm"1 for [Pd(CO)Cl3]" compared with v(CO)a v of 2218 cm"1 for cw-Pd(CO)2(S03F)2. The structure determination of [Pd(CO)Br3]~ is of too low a precision (Pd-C = 1.87(3) A ) to make a useful comparison; however, the v(CO) value of 2120 cm"1 allows a similar conclusion as reached for the chloro anion. Structural features for [Pd2(/i-CO)2](S03F)2, where the CO groups and the fluoro-sulfate anions are both bridging, may be contrasted with those of cw-Pd(CO)2(S03F)2. Pd-C (1.984(4), 1.966(4) A ) and C-0 (1.133(6) A ) distances within the cyclic [Pd2(/i-CO)2]2+ cation are respectively longer than those in c«-Pd(CO) 2(S0 3F) 2 . For the anion in [Pd2(/x-CO)2](S03F)2, the difference in the S-0 bond length, d(S-0), between coordinated and noncoordinated O atoms is about 0.025 A , consistent with weakly coordinated fluorosulfate ions. For the monodentate covalent fluorosulfate groups in cw-Pd(CO)2(S03F)2 this difference is about 0.072 A . Finally, stronger Pd-0 bonds are found for cw-Pd(CO)2(S03F)2 (d(Pd-0)av = 2.011 A ) compared to 2.156 A for [Pd20*-CO)2](SO3F)2. Even though both molecular structures differ appreciably, the long Pd-C bonds found and the vibrational spectra in the v(CO) region suggest that in both palladium carbonyl fluorosulfates the so-called synergetic bonding mechanism does not play an important role in forming these metal carbonyl compounds. Covalent monodentate fluorosulfate groups bound to a noble metal are also found in dimeric [Au(S03F)3]2.21 The higher positive formal charge on the central metal and the noted54'55 ability of gold to form strong covalent bonds results in a longer S-Ob bond of 100 1.501(8) A and correspondingly shorter S-F and S-Ot bonds, where the subscripts t and b denote terminal and bridging oxygen atoms respectively. A more detailed comparison between both OS0 2F groups is perhaps inconclusive on account of the thermal motion of the terminal O and F atoms observed in the structure of [Au(S03F)3]2.21 Consistent with vibrational data for cw-Pd(CO)2(S03F)2, Au(CO)S03F,1 2 and previously reported M(CO)5S03F (M = Mn or Re),43 the fluorosulfate group in these compounds may be termed semi-ionic. This is best reflected in the difference in S-0 bond lengths between coordinating and terminal oxygen atoms. For cw-Pd(C0)2(S03F)2, the average differences are 0.063 and 0.082 A for the slightly nonequivalent fluorosulfate groups. For [Au(S03F)3]221 the difference is about 0.11 A. Both CO bonds appear to be slightly shorter than in CO itself (1.1281 A),3 0 but when thermal motion corrections are employed using rigid body, or riding models, the situation becomes less clear, as seen in Table 3-7. While the vibrational spectra with v (CO)av at 2218 cm"1 vs. 2143 cm"1 3 0 ' 5 5 for gaseous C!0 strongly suggest that the CO bond lengths should be shorter in cw-Pd(CO)2(S03F)2 than in free CO, this is not unambiguously established by the structural data obtained here when these data are carefully analyzed. The same conclusions are reached when CO bond lengths and vibrational CO stretching frequencies for cationic CO derivatives of group 10, and 11 elements,49"52'56 summarized in Table 3-8, are examined. In all instances the sometimes substantial estimated standard deviations should allow no firm conclusions based on C-0 bond length alone. Furthermore it is unclear whether thermal motion corrections have been applied, since all previous reports are either short communications49'50 or preliminary accounts with insufficient details.51'52 101 Table 3-8 CO Bond Distances and Average CO Stretching Frequencies for Cationic Metal Carbonyl Derivatives Compound d(c-o) [A] v(CO)av[cm-l] Reference Cu(CO)Cl 1.11(2) 2120 50 Ag(CO)B(OTeF5)4 1.077 (16) 2204 51 [AgCCOJtBCOTeFs)^ 1.07 (5), 1.09 (5) 2198a 52 1.09 (6), 1.08 (4) Au(CO)Cl 1.11(3) 2170 43, 49 [Pd2(Ll-CO)2](S03F)2b 1.133 (6) 2002 m-Pd(CO)2(S03F)2 1.102 (6) 2218 1.114 (6) CO 1.12822 (7) 2143 30, 55 a IR data only are reported. b Molecule has a symmetrically bridging CO ligand. 3.3.2.3 [Pd(CO)4][Sb2Fn]2 A. Vibrational Spectra of [Pd(CO)4][Sb2F11]2 Since all attempts to prepare single crystals of [Pd(CO)4][Sb2Fu]2 have failed so far, conclusions regarding the coordination geometry of the molecule are based largely on vibrational spectroscopy. The vibrational (infrared and Raman) data obtained from a solid sample of [Pd(CO)4][Sb2Fn]2 are listed in Table 3-9. 102 Table 3-9 Vibrational Data* for [Pd(CO)4][Sb2Fn]2 and [Pt(CO)4][Sb2Fn]2 [Pd(CO)4][Sb2F„]2 [Pt(CO) 4][Sb 2F„] 2 a Approx. Assignment IR Raman IR Raman 2279 vs 2289 vs v(CO) A l g 2263 vs 2267 vs v(CO) B l g 2248 s 2244 s v(CO) Eu 2205. w 2204 w v(13CO) 708 vs 707 vs v(Sb 2F„) 689 vs 688 vs v(Sb2Fn) 686 m v(Sb 2F„) 675 s 675 s v(Sb 2F„) 662 s 666 s v(Sb 2F„) 686 m v(Sb 2F„) 669 m v(Sb 2F„) 656 s 657 s v(Sb 2F„) 650 vs v(Sb 2F„) 604 w v(Sb2Fn) 595 w 596 w v(Sb 2F„) 517 m 504 w 503 w 484 m 473 m 383 w 306 w 305 w v(Sb 2F„) 272 w v(Sb 2F„) 231 w 231 w v(Sb 2F„) 136 w 140 w Vibrational data refer to solid samples. a Ref. 9. 103 The CO stretching frequencies of [Pd(CO)4][Sb2Fn]2 are found at 2248 cm"1 (EJ in the infrared and 2263 (Blg) and 2279 cm"1 (A,g) in the Raman spectrum. There are two characteristic features in the CO stretching region: (i) Vibrational bands are either infrared or Raman active; and (ii) CO stretching modes are at extremely high frequencies (the v (CO)av is 2259.5 cm"1 and compares well with that found for [Pt(CO)4][Sb2F11]2). The non-coincidence of the infrared and Raman CO vibrations suggests the presence of a symmetric center in the molecule, and the extremely high CO stretching frequencies usually 9 -4-indicates a carbonyl cation. Therefore, a centrosymmetric square planar [Pd(CO)4] cation with point group is postulated. Further support to this assignment comes from the following observations: (i) The compound is diamagnetic. (ii) Vibrations attributed to [Sb2Fn]~ follow those observed in the published precedent of [Au(CO)2][Sb2Fn].19 (iii) The bands in the CO stretching range display similar band positions and intensity patterns to those observed for CN vibrations of the isoelectronic anion [Pd(CN)4]2" although v(CN) are found at lower wavenumbers. B. Comparison with Related Compounds As has been mentioned above, CO stretching frequencies in [Pd(CO)4][Sb2Fn]2 are exceptionally high. With v(CO)a v = 2259.5 cm"1, it is only marginally lower than that found for [Pt(CO)4][Sb2Fn]2 with a v (CO)av value of 2261 cm"1. Higher CO stretching frequencies are found for [Hg(CO)2][Sb2F11]2, which has an infrared band at 2278 cm"1 and a Raman band at 2281 cm"1. Therefore, with a v(CO)a v value of 2279.5 cm'1, [Hg(CO)2][Sb2F„]2 104 represents the highest average CO stretching frequency so far reported for a metal carbonyl derivative. The vibrational data of several carbonyl cations or cationic compounds, together with its cyanide counterparts, are listed in Table 3-10. Also listed are the stretching force constants 57 58 59 which are derived either according to Jones or using the Cotton-Kraihanzel approach. ' From the data for the metal carbonyl derivatives collected in Table 3-10 it appears that the synergetic bonding model does not apply here. In these compounds C O stretching frequencies are generally observed between 2200 and 2290 cm"1, well above the typical values between 2125 and 1850 cm"1 for terminal CO in metal carbonyls. CO stretching force constants, usually between 15 and 18 x 102 N m"1, are now in excess of 20 x 102 N m"1 for binary carbonyl cations. Judging from the force constant of ft = 21.3 x 102 N m"1 calculated for H C O + , 1 9 a molecule where rc-back-donation is impossible, rc-back-donation can only make an insignificant contribution to the M - C O bonding in noble metal carbonyl cations. As discussed for [Au(CO)2] + , 1 9 the ability of gold, and also of platinum and palladium to form strong covalent bonds, possibly aided by relativistic effects60 and polar interactions, is seen as an important contribution to the bonding in this group of cationic coordination complexes of CO with the noble metal cations. 105 Table 3-10 CO and CN Stretching Vibrations and Force Constants For Selected Noble Metal Carbonyl and Cyanide Complexes Molecule or ion Point group vCX sym rem-il V C X syma [cm-1] vCX as. [cm-1] vCX av. [cm-1] / r x l 0 2 N n r 1 Ref. [Pt(CO)4]2+ b D4h 2289 2267 2244 2261 20.64e 9 [Pt(CO)4]2 + c D4h 2281 2257 2233 2251 20.47e 9, n [Pd(CO)4]2+ £>4h 2278 2263 2248 2259 20.63e 1 [Pt(CN)4]2- D4h 2168 2149 2133 2146 17.38f 8 [Pd(CN)4]2- D4h 2160 2146 2136 2144.5 17.35f g [Au(CO)2] + 2254 2217 2235.5 20.18e 20. l f 19 9 [Au(CN)2J- 2162 2142 2152 17.60f 57 d5-Pt(CO)2(S03F)2^ ~C2v 2219 2185 2202 19.56e 8,9 cw-Pd(CO)2(S03F)2d ~ C 2 v 2228 2208 2217.5 19.87e Au(CO)S03Fd ~c, 2196 2196 19.5f 19 CO Cooy 2143 2143 18.6 CO+ c 2184 2184 19.3f 19 HCO+ c 2184 2184 21.3f 19 a v" refers to the B l g mode, b S b 2 F n - as anion, c [Pt(S03F)6]2" as anion, d IR bands, e Cotton-Kraihanzel values refs. 58,59, f according to Jones: Ref. 57, gKubus, G. J; Jones, L. H. Inorg. Chem. 1974, 13, 2816. Another contributing factor to the thermal stabilities of noble metal carbonyl cations may come from secondary interionic or intermolecular contacts mainly involving a positively 106 charged C atom of the mainly a-bonded carbonyl group and the F atoms of the fluoro-antimonate anions Sb2Fu", which result in extended three dimensional networks. These interactions play an important role in stabilizing cationic metal carbonyl compounds such as cw-Pd(CO)2(S03F)2 and [IIg(CO)2][Sb2F11]2.61 3.4 Summary and Conclusions Using superacids as reaction media, three new cationic metal carbonyl compounds, cis-Pd(CO)2(S03F)2, [Pd2(/i-CO)2](S03F)2, and [Pd(CO)4][Sb2F„]2, have been prepared. All the three compounds are thermally stable to well over 100 °C. Their compositions and structures were characterized by microanalysis, vibrational spectroscopy and, in the case of cis-Pd(CO)2(S03F)2 and [Pd20*-CO)2](SO3F)2, by single crystal X-ray analysis. One of the common features of these three compounds is the extremely high CO stretching frequencies. The average CO stretching frequency of c«-Pd(CO) 2(S0 3F) 2 , v(CO)a v = 2218 cm"1, is the highest so far reported for a neutral noble-metal carbonyl complexes. The v(CO)a v for [Pd(CO)4][Sb2Fn]2 is 2259.5 cm"1, only less than [Pt(CO)4]-[Sb 2Fu] 2 with v(CO)a v = 2261 cm"1 and [Hg(CO)2][Sb2F„]2 with v(CO)a v = 2279.5 cm"1. It therefore appears that the synergetic bonding mechanism cannot be responsible for the observed high thermal stabilities of the compounds discussed here. According to the single crystal X-ray structure of cw-Pd(CO)2(S03F)2, secondary interactions provide a charge compensation for a positively charged carbon atom of the carbonyl group and stabilize the compound in the solid state by forming an extended 3-dimensional network. In HS0 3F, it seems that solvation may well stabilize the cationic species by similar weak O • • • C 107 interactions. Furthermore, the ability of gold, platinum, and apparently also palladium to form strong covalent metal-carbon bonds may also make an important contribution to the thermal stabilities of these cationic metal carbonyls in the near-absence of rc-back-bonding. It is interesting to note that all the palladium carbonyl derivatives discussed here do not obey the EAN rule. As in the case of the corresponding platinum carbonyl compounds, the metal center is surrounded by 16 electrons instead of the normal 18 electrons expected of a metal carbonyl compound. 108 References 1 Lee, J. D. Concise Inorganic Chemistry, 4th ed.; Chapman & Hall: London, 1991; pp 802 and 803. 2 (a) Schutzenberger, P. Bull. Soc. Chim. Fr. 1868, 10, 188. (b) Schutzenberger, P. C. R. Acad. Sci. 1870, 70, 1134, 1287. 3 Mond, L. ; Langer, C ; Quincke, F. J. Chern. Soc. 1890, 749. 4 (a) Darling, J. H . ; Ogden, J. S. J. Chern. Soc, Dalton Trans. 1973, 1079. (b) Huber, H . ; Kundig, E. P.; Moskovits, M . ; Ozin. G. A. Nature (Phys. Sci), 1972, 98, 235. 5 Kundig, E. P.; Mcintosh, D. ; Moskovits, M . ; Ozin. G. A. J. Am. Chern. Soc. 1973, 95, 7234. 6 (a)Calderazzo, F.; Cotton, F. A. Inorg. Chern. 1962, 1, 30. (b) Calderazzo, F.; Fachinetti, G. ; Floriani. C. J. Am. Chern. Soc. 1974, 46, 3695. 7 Hwang, G. ; Bodenbinder, M . ; Willner, H . ; Aubke, F. Inorg. Chern. 1993, 32, 4667. 8 Hwang, G. ; Wang, C ; Bodenbinder, M . ; Willner, H . ; Aubke, F. / . Fluorine Chern. 1994, 66, 159. 9 Hwang, G. ; Wang, C ; Aubke, F.; Bodenbinder, M . ; Willner, H. Can. J. Chern. 1993, 77, 1532. 1 0 (a) Lee, K. C ; Aubke, F. Can. J. Chern. 1979, 57, 2058. (b) Lee, K. C ; Aubke, F. Can. J. Chern. 1977, 55, 2473. 1 1 (a) Thripathi, S. C ; Srivastava, S. C ; Mani, R. P.; Shrimal, A. K. Inorg. Chim. Acta 1976, 77, 257. 109 (b) Vargaftic, M. N. ; Stromnova, T. A. ; Moiseev, I. I. Russ. J. Inorg. Chern. (Engl, trans.), 1980, 25, 121. 1 2 Willner, H . ; Aubke, F. Inorg. Chern. 1990, 29, 2195. 1 3 Walter, V. C. Ph. D. Thesis, The University of British Columbia, 1989. 1 4 Uson, R.; Forni£s, J.; Thomas, M . ; Menjdn, B. Organometallics 1985, 4, 1912. 1 5 Manchot, W.; Konig, J. Chern. Ber. 1926, 59, 883. 1 6 Stromnova, T. A. ; Kuz'mina, L. G.; Vargaftik, M . N. ; Mazo, G. Ya.; Struchkov, Yu. T.; Moiseev, I. I. Izv. Akad. Nauk SSSR, Ser. Khim. 1978, 720. 1 7 Moiseev, I. I.; Stromnova, T. A. ; Vargaftik, M. N. ; Mazo, G. Ya.; Kuz'mina, L. G.; Struchkov, Yu. T. J. Chern. Soc, Chern. Commun. 1978, 27. 1 8 Belli Dell'Amico, D. ; Calderazzo, F.; Zandona, N. Inorg. Chern. 1984, 23, 137. 1 9 Willner , H . ; Schaebs, J.; Hwang, G.; Mistry, F.; Jones, R.; Trotter, J.; Aubke, F. J. Am. Chern. Soc. 1992, 114, 8972. 2 0 Allen, F. H . ; Lerbscher, J. A. ; Trotter, J. J. Chern. Soc. A 1971, 2507. 2 1 Willner, H . ; Rettig, S. J.; Trotter, J.; Aubke, F. Can. J. Chern. 1991, 69, 391. 2 2 O'Sullivan, K.; Thompson, R. C ; Trotter, J. J. Chern. Soc. A 1970, 1814. 2 3 Yeats, P. A. ; Sams, J. R.; Aubke, F. Inorg. Chern. 1972, 11, 2634. 2 4 Yeats, P. A. Ph. D. Thesis, University of British Columbia, 1973. 2 5 Mallela, S. P.; Yap, S.; Sams, J. R.; Aubke, F. Inorg. Chern. 1986, 25, 4327. 2 6 Holloway, R. G.; Penfold, B. R.; Colton, R.; McCormick, H. J. J. Chern. Soc, Chern. Commun. 1976, 485. 2 7 Goodfellow, R. J.; Herbert, I. R.; Orpen, A. G. J. Chern. Soc, Chern. Commun. 1983, 1386. 110 Bender, R. Braunstein, P.; Tiripicchio, A. ; Tiripicchio-Camellini, M. J. Chem. Soc, Chem. Commun. 1984, 42. 2 9 Feltham, R. D. ; Elbaze, G. ; Ortega, R.; Eck, C ; Dubrawski, J. Inorg. Chem. 1985, 24, 1503. 30 Herzberg, G. Molecular Spectra and Molecular Structure; I Spectra of Diatomic Molecules; D. Van Nostrand: Toronto, 1950; p 250, 522. 3 1 Wink, D. J.; Creagan, T. ; Lee, S. Inorg. Chim. Acta 1991, 180, 183. 3 2 Colton, R.; McCormick, M. J.; Parman, D. C. J. Chem. Soc, Chem. Commun. 1977, 823. 3 3 Soriano-Garcia, M . ; Rosas, N. ; Gomez-Lara, G. ; Toscano, R. A. Acta Crystallagr. 1987, C43, 1679. 3 4 Dahl, L. F.; Martell, C.; Wampler, D. L. J. Am. Chem. Soc. 1961, 83, 1761. 3 5 Yang, A. C.; Garland, C. W. J. Phys. Chem. 1957, 61, 1504. 36 Braterman, P. S. Metal Carbonyl Spectra; Academic Press: New York, 1975. 3 7 Kettle, S. F. A. Top. Curr. Chem. 1977, 71, 111. 3 8 Ruoff, A. ; Milne, J. B.; Kaufmann, G.; Leroy, M. Z. Anorg. Allg. Chem. 1970, 372, 119. 3 9 Qureshi, A. M . ; Levchuk, L. E.; Aubke, F. Can. J. Chem. 1971, 49, 2544. 4 0 Goggin, P. L. ; Mink, J. J. Chem. Soc, Dalton Trans. 1974, 534. 4 1 Colton, R.; Farthing, R. H . ; McCormick, M. J. Aust. J. Chem. 1973, 26, 2607. 4 2 Cotton, F. A. ; Wilkinson, G. Advanced Inorganic Chemistry, 5th ed.; John wiley & Sons: New York, 1988; p 1034. 4 3 Mallela, S. P.; Aubke, F. Inorg. Chem. 1985, 24, 2969. 4 4 Calderazzo, F.; Belli-Dell'Amico, D. Inorg. Chem. 1981, 20, 1310. Ill 4 5 Browning, J.; Goggin, P. L. ; Goodfellow, R. J.; Norton, M. G.; Rattray, A. J. M . ; Taylor, B. F.; Mink, J. J. Chern. Soc, Dalton Trans. 1977, 2061. 4 6 Shomaker, V, ; Trueblood, K. N. Acta Crystallogr. 1968, B24, 63. 4 7 Trueblood, K. N. Acta Crystallogr. 1978, A34, 950. 4 8 Bondi, A. J. Phys. Chern. 1964, 68, 441. 4 9 Jones, P. G. Z. Naturforsch. 1982, 37B, 823. 5 0 Hakansson, M . ; Jagner, S. Inorg. Chern. 1990, 29, 5241. 5 1 Hurlburt, P. K.; Anderson, O. P.; Strauss, S. H. J. Am. Chern. Soc. 1991, 113, 6277. 5 2 Hurlburt, P. K.; Rack, J. J.; Dec, S. F.; Anderson, O. P.; Strauss, S. H. Inorg. Chern. 1993, 32, 373. 5 3 Andreini, B. P.; Belli Dell'Amico, D. ; Calderazzo, F.; Venturi, M. G. ; Pelizzi, G. J. Organomet. Chern. 1988, 354, 369. 5 4 Puddephatt, R. J. The Chemistry of Gold; Elsevier: Amsterdam, 1978. 5 5 Puddephatt, R. J. In Comprehensive Coordination Chemitry; Wilkinson, G., Ed.; Pergamon Press: Oxford, U. K., 1987; Vol. 7, p 861. 5 6 Wang, C ; Bodenbinder, M . ; Willner, H. ; Rettig, S. J.; Trotter, J.; Aubke, F. Inorg. Chern. 1994, 33, 779. 5 7 Jones, L. H. Inorganic Vibrational Spectroscopy; Mercel Dekker: New york, 1071; Vol. 1, p 122. 5 8 Cotton, F. A. ; Kraihanzel, C. S. J. Am. Chern. Soc. 1962, 84, 4432. 5 9 Kraihanzel, C. S.; Cotton, F. A. Inorg. Chern. 1963, 2, 533. 6 0 Pyykko, P. Chern. Rev. 1988, 88, 563. 112 Bodenbinder, M . ; Balzer-Jollenbeck, G.; Willner, H . ; Batchlor, R. J.; Einstein, F. W. B.; Wang, C ; Aubke, F. Inorg. Chem. 1996, 35, 82. 113 CHAPTER 4 SYNTHESES, STRUCTURES, AND SPECTROSCOPIC STUDIES OF CATIONIC CARBONYLS OF IRIDIUM(III) 4.1 Introduction Iridium was discovered in 1803 by Tennant in the residue left when crude platinum was dissolved by aqua regia.1 The metal is white, similar in appearance to platinum, but with a slight yellowish cast. Iridium is very hard and brittle, making it very difficult to machine, form, or work. It is the most corrosion-resistant metal known and is not attacked by any of the acids nor by aqua regia. However, iridium metal reacts with molten salts, such as NaCl and NaCN. The most important Ir(I) complexes are the so-called Vaska's compound, trans-IrCl(CO)(PPh3)2, and its analogues with other phosphines.2 IridiumfJII) complexes can be obtained by oxidative-addition using Vaska's compound as a precursor (4-1): fra/w-IrCl(CO)(PPh3)2 + AB <-» Ir(III)ClAB(CO)(PR3)2 (4-1) where AB represents oxidative diatomic molecules. Both anionic and cationic iridium carbonyls are known, for example, fra/w-IrCl(CO)(PPh3)2 + CO N a l H g > Na[Ir(CO)3(PR3)] (4-2) 114 and mz/w-IrCl(CO)(PPh3)2 + CO N a C ' ° 4 > [Ir(CO)3(PR3)2]C104 (4-3) These iridium carbonyl compounds, however, involve other ligands, such as PR3, which usually enhance rc-back-donation to CO and thus strengthen the metal carbon bonds. Some recent research work on iridium carbonyls includes the synthesis and characterization of Ir(CO)3F34 and the highly reduced carbonyl anion [Ir(CO)3]3".5 In both cases, the synergic bonding model does not seem to apply for different reasons, as discussed in the introduction of this thesis. Following the success with cationic palladium carbonyls, the investigation is now extended to iridium. Compared with the previous studies on gold, platinum, and palladium, in which the oxidation states of the metals are usually +1 or +2, the study on iridium carbonyls is anticipated to generate cationic metal carbonyls with higher oxidation states. It is expected that the a-character in Ir(III) carbonyl derivatives should be more pronounced since the expected higher oxidation state (+3) of the metal will make the jc-back-donation even less likely. 4 .2 Experimental 4.2.1 The Synthesis of Ir(CO)3(S03F)3 Iridium metal powder was oxidized to Ir(S03F)4 by reaction with a large excess of S 2 0 6 F 2 in HS0 3F at 140 °C as described earlier.6 The reaction was carried out by removing 0 2 and S 2 0 5 F 2 (formed by the decomposition of S206F2) and adding S 2 0 6 F 2 repeatedly. As a 115 consequence a large amount of S 2 0 6 F 2 was consumed over a prolonged time to generate Ir(S03F)4 or Ir(S03F)3, the latter being produced via thermal decomposition of Ir(S03F)4. To reduce the consumption of S 2 0 6 F 2 and make the oxidization more efficient, a two-part reactor shown in Figure 4-1 was employed. In a typical reaction 0.0976 g of iridium powder was added through the stopcock via a funnel to Part A of the reactor. Approximately 2.5 mL of HS0 3F and 1 mL of S 2 0 6 F 2 were added to Part A by distillation in vacuo. The reaction mixture was heated to 120 °C for 16 hours with vigorous magnetic stirring. Some of the metal was oxidized at this stage and a dark brown solution formed. The stirring was Kontes Stopcock w. Teflon Stem BIO Ground Glass Joint Figure 4-1 Two-part reactor designed and used for the synthesis of Ir(S03F)3 and Ir(CO)3(S03F)3. 116 stopped and the mixture was allowed to cool to room temperature. The unreacted metal settled out over a period of 30 minutes. The supernatant brown solution was decanted into Part B of the reactor. The volatiles (HS03F and S206F2) were distilled back onto the unreacted metal in Part A by heating Part B to 60 °C while cooling Part A in liquid N 2 . Some more (ca. 1 mL ) S 2 0 6 F 2 was added to part A at this point by vacuum transfer and the oxidation reaction was continued for another 24 hours. The procedure was repeated four more times without further addition of S 2 0 6 F 2 until all of the metal powder appeared to have been consumed. The dark brown solution in Part B was heated at 140 °C for 12 more hours to ensure complete oxidation. Finally, the color of the solution changed to a deep purple due to a decomposition of Ir(S03F)4 to Ir(S03F)3. The solution was then transferred to Part A of the reactor for the carbonylation reaction. Prior to carbonylation, all volatiles (02, S 2 0 5 F 2 , S 2 O e F 2 , HS0 3F, and a trace amount of SiF4) were distilled off in vacuo. About 3 mL of HS0 3F were added to Part A of the reactor, producing a deep blue-purple solution. CO gas (about 2 bars) was then admitted to the reactor and the solution was heated at 60 °C for 4 days. The color of the solution changed gradually from blue to violet, then to brown, and to black. Finally a bright yellow solution resulted and no further color change was noted. Trace amounts of precipitate were observed occasionally during the carbonylation reaction. The bright yellow solution was filtered and slow removal of most volatiles from the filtrate by vacuum distillation at room temperature produced a pale yellow crystalline product. Recrystallization from HS0 3F produced single crystals suitable for X-ray diffraction analysis. 117 The total yield of crystalline material isolated and identified by vibrational spectroscopy as mer-Ir(CO)3(S03F)3 was estimated to be ca. 70%. wer-Ir(CO)3(S03F)3 is a pale yellow (almost colorless) material. Heating at the melting point of 150 °C in a sealed capillary tube resulted in gas evolution and a color change to light brown. Elemental analysis of the crystalline pale yellow product was as follows: Calculated for C 3F 3Ir0 1 2S 3: C, 6.28; H, 0; S, 16.77. Found C, 5.7; H, 0.1; S, 18.25. The presence of trace amounts of hydrogen (0.1%) detected during microanalysis and the higher sulfur content, 18.25% compared to 16.77%, indicate some trapped solvent, HS0 3F, in the sample. If a composition of Ir(CO)3(SO3F)3.0.7 HS0 3F is assumed, the calculated values for C, H, and S (5.5, 0.11, and 18.16% respectively) agree well with the microanalysis values. There is no evidence for the presence of HS0 3F in samples used for vibrational spectroscopy, which were obtained using a different preparation. The difficulties in obtaining reasonable quantities of Ir(S03F)3 and subsequently Ir(CO)3(S03F)3 made it necessary to use small quantities of the sample for product identification. Since the molecular structure was solved by X-ray diffraction, no further attempts were made to get better microanalysis data. 4.2.2 The Synthesis of [IrtCCOsCIHSbjFnk Crystalline mer-Ir(CO)3(S03F)3 (40 mg) was added to a 50-mL one-part reactor, and less than 1 mL of HS0 3F was condensed onto the crystals by vacuum distillation. The reactor was then warmed to room temperature with gentle swirling and an almost colorless solution 118 was obtained. Approximately 2 mL of SbF5 was admitted to the reactor by vacuum distillation. A slightly viscous solution formed after gently warming the reaction mixture to 40 °C. The solution, although slightly viscous, was much more mobile than antimony pentafluoride. Approximately 2 bars of CO pressure was added to the reactor at this stage. After six weeks of standing at room temperature, a crystal formed at the bottom of the reactor. The solution was allowed to remain at the same conditions for another two weeks,, but no more crystals were observed, and it seemed that the original crystal did not grow larger either. The crystal was carefully picked out from the solution and cut into several pieces in a glove box. The single crystals suitable for X-ray analysis were mounted in capillary tubes and sealed. The X-ray analysis results indicate the compound to be prCCO^CrjfS^Fnk. The introduction of the chlorine atom into the molecule will be discussed later in this chapter. The experiment was repeated but pr(CO)5Cl][Sb2Fu]2 crystals were not obtained. Therefore, microanalysis and vibrational spectroscopic studies were not possible. 4.3 Results and Discussion 4.3.1 Ir(CO)3(S03F)3 4.3.1.1 Synthetic Aspects Like Au(CO)S03F7 and cw-M(CO)2(S03F)2,8 M = Pd or Pt, tris(carbonyl)-iridium(III) fluorosulfate, Ir(CO)3(S03F)3, is formed from a binary fluorosulfate precursor in fluorosulfuric acid by a reaction with gaseous CO according to 119 (4-4) There are three differences in the formation of iridium carbonyls compared to the earlier synthesis of gold, palladium, and platinum carbonyl fluorosulfates: (i) The oxidation state of iridium remains +3 while Au(CO)S03F,7 c/j,-Pt(CO)2-(S03F)2,8 and cis-Pd(CO)2(S03F)2 form in reductive carbonylation reactions from [Au(S03F)3]2,9 Pt(S03F)4,10 or Pd(n)Pd(IV)(S03F)6n respectively. This is confirmed by the absence of the by-products, C 0 2 and S 20 5F 2 , in the carbonylation reaction of the iridium(III) fluorosulfate. (ii) Reductive carbonylation reactions are found to be very fast for noble metal fluorosulfates and proceed easily at 25 °C within a few hours, in particular in the case of the reduction of the gold(III) and the mixed valency palladium precursors. Formation of Ir(CO)3(S03F)3 from Ir(S03F)3 is very slow and requires heating at 60 °C over 4 days. The reaction is easily followed by the color changes: The deep blue purple color of Ir(S03F)36 dissolved in HS0 3F gradually changes to dark brown and finally to pale yellow-orange. While there is no structural information available for either solid Ir(S03F)3 or its solution in HS0 3F, the observed weak temperature independent paramagnetization of the former6 is consistent with a d6 low spin configuration in an approximately octahedral environment. The compound is hence expected to be inert to the substitution reaction by CO. (iii) Evidence from vibrational and 1 3C-NMR spectra suggests that binary metal carbonyl cations, such as linear [Au(CO)2]+ 7 and square planar [M(CO)4]2 + (M = Pt or Pd),8 are formed initially. These three cations have subsequently been obtained as [Sb2F,,]" 120 salts.12,13 Substitution of CO by S03F" leads to Au(CO)S03F,7 cw-Pt(CO)2(S03F)2,8 and cis-Pd(CO)2(S03F)2 respectively. Consistent with the greater fra/w-directing ability of C O , 1 4 the cw-isomers forms exclusively. In contrast, a fra/w-isomer, /ner-Ir(CO)3(S03F)3, is predominant in the carbonylation reaction of Ir(S03F)3. Judging by this observation, a similar binary metal carbonyl cation such as [Ir(CO)6]3+ is unlikely to be a precursor in the formation of Ir(CO)3(S03F)3. Observation of a fra/w-Ir(CO)2 moiety is consistent with the gradual addition of CO to a fluorosulfate species, rather than with CO substitution by S03F" as observed during the formation of Au(CO)S03F,7 c«-Pt(CO) 2 (S0 3 F) 2 , 8 and cis-Pd(CO)2(S03F)2. There is, however, no conclusive evidence regarding the nature of solvated Ir(S03F)3 in HS0 3F to permit a more detailed discussion of the formation of mer-Ir(CO)3(S03F)3. It is noted that Ir(CO)3I3 was also formed by the CO addition to Irl3 at 100 °C and 250 atm of C O . 1 5 For this compound two bands in the CO stretching region of the IR spectrum have been reported, which is consistent with a ,/aoisomer. The compound had low thermal stability and lost CO readily. Ir(CO)3F3,4a the other precedent for Ir(CO)3(S03F)3, was formed by fluoronation of Ir4(CO)12 with a 6-fold excess of XeF2 in HF at -40 °C. 1 9 F and 1 3C-NMR spectra suggested initial formation of a mixture of mer- and ^ oc-isomers. The pale yellow solid isolated from solution was the yac-isomer.4a The proposed jac-configuration is confirmed by a molecular structure, obtained by extended X-ray absorption fine structure spectroscopy.415 In the case of Ir(CO)3(S03F)3 the /ner-isomer is formed predominantly. Slow evaporation of the solvent produced crystals of mer-Ir(CO)3(S03F)3, which were isolated by 121 removing the mother liquor by pipette. The total yield was about 70%. The mer-geometry is retained on re-dissolution in H S 0 3 F , as is evident from the 1 9F-NMR spectrum, and isomerization in HS0 3F is not observed. The mother liquor has, after separation from the crystals, a measurable concentration of the y^c-Ir(CO)3(S03F)3, best detected by its Raman spectrum (vide infra). While the carbonylation of iridium(III) fluorosulfate to give Ir(CO) 3(S0 3F) 3 is slow but straightforward, the synthesis of Ir(S03F)3 via the oxidation of iridium metal by S 2 0 6 F 2 in HS0 3F is troublesome and time consuming for three main reasons. (i) The system is chemically complex. At temperatures below 100 °C the reaction is very slow. At 130 °C the reaction proceeds to give dark brown Ir (S0 3 F) 4 , which then decomposes to blue purple Ir(S03F)3, S 2 0 5 F 2 , and 0 2 . Ir(S03F)3 is then re-oxidized and the reaction sequence constitutes a catalytic decomposition of S 2 0 6 F 2 into S 2 0 5 F 2 and 0 2 . At 160 °C the decomposition of S 2 0 6 F 2 becomes very rapid, and the oxygen thus produced has to be removed frequently in order to avoid the risk of explosion. Consequently, a fresh portion of S 2 0 6 F 2 was added from time to time because of the rapid conversion of S 2 0 6 F 2 to S 2 0 5 F 2 and 0 2 in the presence of iridium fluorosulfate according to 2 S 2 0 6 F 2 Ir^SOf^ _> 2S 2 0 5 F 2 + 0 2 (4-5) A It is, therefore, easy to understand why it is difficult to force the trace amount of metal powder remaining to react. As the oxidization reaction proceeds, the amount of Ir(S03F)4, whose presence is considered to be a catalyst for reaction (4-5), increases. At the later stage of the reaction, the high concentration of Ir(S03F)4 rapidly converts S 2 0 6 F 2 to S 2 0 5 F 2 and 0 2 . 122 It is estimated that a 20- to 40-fold molar excess of S 2 0 6 F 2 is required and the reaction takes several weeks to go to completion. (ii) Another difficulty encountered is at temperatures of ca. 140 °C fluorosulfuric acid, which is used as a solvent, will undergo self-dissociation according to: HS0 3F -> S0 3 + HF. This results in extensive glass etching and formation of SiF4. It is thus not possible to follow the course of the reaction by weight balance. Metal reactors, such as Monel-A or nickel reactors, are not suitable substitutes to overcome glass etching since contamination by other metal fluorosulfates is possible, and this is highly undesirable in the synthesis and characterization of new compounds. (iii) The deep blue purple color of the product Ir(S03F)3 makes it difficult to judge, qualitatively, whether all of the metal is consumed during the oxidation reaction. In this study, a reactor with a sidearm (see Figure 4-1), was used which permitted decanting of the supernatant solution after the remaining metal had settled to the bottom, and allowed for subsequent condensation of the oxidizer and solvent back into the main reactor. In this manner the reaction products, Ir(S03F)4 and Ir(S03F)3, were largely isolated from the reaction system. Hence, the catalytic cycle (which produces large amount of S 2 0 5 F 2 and 02) is hindered, and the S 2 0 6 F 2 consumption is reduced greatly. In order to further reduce the hazards associated with S 20 6F 2 , only small amount of Ir (less than 100 mg) was used per reaction in this research. 123 4.3.1.2 Molecular Structure of mer-Ir(CO)3(S03F)3 The crystal structure of /ner-Ir(CO)3(S03F)3 was solved by Dr. R. J. Batchelor and Professor F. W. B. Einstein, Department of Chemistry, Simon Fraser University. Details are found in Appendix C. The molecular structure of mer-Ir(CO)3(S03F)3, depicted in Figure 4-2, shows that the coordination about the iridium atom is approximately octahedral with a meridional stereochemistry. The lengths of chemically equivalent bonds are not significantly different (see Table 4-1). Variations in chemically inequivalent Ir-C, Ir-O and S-0(Ir) bond lengths are all consistent with the relative trans influences of the ligands. As expected from the 0(33)1 0(31) C(3) 0(22) 0(3) 0(11) 0(13) 0(12) F(l) Figure 4-2 The molecular structure of mer- Ir(CO)3(S03F)3. 124 Table 4-1 Selected Intramolecular Distances (A) and Angles (°) for /ner-Ir(CO)3(S03F)3 at 200 K Bond Distances Ir-O(ll) Ir-0(21) Ir-0(31) Ir-C(l) Ir-C(2) Ir-C(3) S(l)-0(11) S(l)-0(12) S(l)-0(13) S(l)-F(l) 0(3)-C(3) Bond Angles 0(11)-Ir-0(21) 0(ll)-IrO(31) 0(11)-Ir-C(l) 0(11)-Ir-C(2) 0(11)-Ir-C(3) 0(21)-Ir-0(31) 0(21)-Ir-C(l) 0(21)-Ir-C(2) 0(21)-Ir-C(3) 0(31)-Ir-C(l) 0(31)-Ir-C(2) 0(31)-Ir-C(3) 2.055(4)" 2.038(4)" 2.038(4)" 1.937(7)" 2.006(6)* 1.999(6)" 1.487(5)c 1.411(5/ 1.402(5/ 1.544(4/ 1.108(8)° 88.25(17) 88.49(17) 174.41(23) 93.46(23) 84.67(23) 176.74(17) 95.74(22) 84.30(20) 93.79(21) 87.52(22) 95.80(20) 86.00(21) S(2)-0(21) S(2)-0(22) S(2)-0(23) S(2)-F(2) S(3)-0(31) S(3)-0(32) S(3)-0(33) S(3)-F(3) 0(1)-C(1) 0(2)-C(2) 0(12)-S(1)-0(13) F(2)-S(2)-0(21) F(2)-S(2)-0(22) F(2)-S(2)-0(23) 0(21)-S(2)-0(22) 0(21)-S(2)-0(23) 0(22)-S(2)-0(23) F(3)-S(3)-0(31) F(3)-S(3)-0(32) F(3)-S(3)-0(33) 0(31)-S(3)-0(32) 0(31)-S(3)-0(33) 1.511(4) 1.410(5) 1.402(6) 1.545(5) 1.512(4) 1.413(5y 1.406(5/ 1.545(5) 1.094(8) 1.114(8) 119.6(3) 100.8(3) 105.9(3) 105.3(3) 109.7(3) 112.6(3) 120.4(3) 101.1(3) 105.9(3) 105.2(3) 113.4(3) 109.0(3) 125 Table 4-1 (continued) C(l)-Ir-C(2) C(l)-Ir-C(3) C(2)-Ir-C(3) 90.8(3) 91.2(3) 177.4(3) 101.5(3) 105.1(3) 105.2(3) 113.2(3) 110.2(3) Ir-0(11)-S(l) Ir-0(21)-S(2) Ir-0(31)-S(3) Ir-C(l)-0(1) Ir-C(2)-0(2) Ir-C(3)-0(3) 0(32)-S(3)-0(33) 120.0(3) 128.1(3) 123.07(24) 123.68(25) 177.8(6) 176.1(5) 176.1(6) F(l)-S(l)-0(11) F(l)-S(l)-0(12) F(l)-S(l)-0(13) 0(11)-S(1)-0(12) 0(11)-S(1)-0(13) a~° Thermal motion corrections using rigid body, segmented rigid body or riding models indicate that the actual bond lengths could be greater by from: a 0.003 to 0.008 A; * 0.003 to 0.004 A ; c 0.003 to 0.011 A; d 0.009 to 0.013 A; e0.011 to 0.020 A;'0.014 to 0.016 A; 8 0.010 to 0.010 A; * 0.014 to 0.022 A; ' 0.019 to 0.024 A; J 0.009 to 0.010 A; * 0.009 to 0.016 A;10.015 to 0.020 A; m 0.005 to 0.018 A; " 0.003 to 0.010 A; ° 0.008 to 0.026 A. vibrational spectroscopic results (very high v (CO) values), the C-0 distances lie at the lower end of the range exhibited by transition metal carbonyls.16 The relatively large uncertainties in the C-0 bond lengths from X-ray analysis preclude a detailed correlation with spectroscopic parameters. For /ac-Ir(CO)3F34b similar long Ir-C bond distances (2.030(8) A) are found. The Ir-F distances are with 1.920(6) A shorter than Ir-0 distances in /ner-Ir(CO)3(S03F)3. As found for cw-Pd(CO)2(S03F)2 (see Chapter 3) there are a number of intermolecular and unimposed intramolecular C ••• O contacts (three and one respectively for C(l), two of each for C(2), and one of each for C(3)) to terminal oxygen atoms of the fluorosulfate groups in the range of 2.722(8) A to 3.069(8) A in the crystal structure of mer-Ir(CO)3(S03F)3. C(3) has, in addition, a weaker unimposed intramolecular contact of 3.056(7) A with F(2). The 126 respective sums of accepted van der Waals radii are: C+0 3.22 A and C+F 3.17 A. 1 7 Thus, these interactions are thought to be attractive in nature and provide a stabilizing influence analogous to those in rij-Pd(CO)2(S03F)2. In the square planar palladium structure each carbon atom has five such contacts, however the C • • O distances are somewhat longer ( C - 0 range: 2.869(6)-3.172(6) A and one C ••• F 3.051(6) A). This is thought to be a consequence of the lower coordination number and square planar geometry at Pd, which allows for more such secondary contacts to the CO groups from neighboring molecules. The intermolecular secondary interactions involving C atoms of CO and O atoms of S03F" in the structure of mer-Ir(CO)3(S03F)3 link the molecules into an extended three-dimensional structure. There are also a few short non-bonded O • • • O contacts (in the range of 2.917(7) - 3.010(8) A) and F • • O contacts (2.880(7) - 2.970(6) A). These distances, however, are only marginally shorter than the respective sums of accepted van der Waals radii (O+O 3.04 A and O+F 2.99 A)17 and are judged to be caused mainly as a result of the attractive C • • • O and C • • • F interactions. There are no further intermolecular separations significantly less than the sums of appropriate pairs of van der Waals radii. Figure 4-3 depicts four complete molecules whose iridium atoms lie within a common unit cell. Additional atoms are shown to illustrate all the significant attractive C • • • O and C • • F contacts to one molecule. The atomic coordinates for wer-Ir(CO)3(S03F)3 at 200 K are given in Appendix C. In spite of the differences in stereochemistry and oxidation state of the two respective metals, the molecular structures of wer-Ir(CO)3(S03F)3 and of cw-Pd(CO)2(S03F)2 show strong similarities with respect to bond lengths and bond angles for the M(CO)n-moiety, 127 Figure 4-3 Stereoscopic view of four complete molecules of Ir(CO)3(S03F)3 with secondary interactions shown by thin lines. where M = Ir or Pd and n = 3 or 2. The same holds true for the fluorosulfate groups. Even the metal-oxygen bond distances are comparable in mer-Ir(CO)3(S03F)3 and cis-Pd(CO)2(S03F)2. The iridium-carbon bond lengths in the trans-lv(CO)2 group are 1.999(6) and 2.006(6) A. These bond distances are measurably longer than the third Ir-C distance of 1.937 (6) A where CO is trans to a fluorosulfate group. The difference is thought to be caused by a strong trans influence of S03F" (instead of competing for rc-electron-density, S03F" is a weak n-donor ligand). All three metal-carbon bond distances are relatively long when compared to those for other transition metal carbonyls.16 These results, combined with the evidence from 128 vibrational spectroscopic studies (vide infra), suggest a strongly reduced rc-back-donation from Ir to CO, particularly in the trans-h(CO)2 moiety. 4.3.1.3 Vibrational Spectra The FT-IR and FT-Raman band positions observed for crystalline mer-rr(CO)3(S03F)3 are listed in Table 4-2 together with estimated relative intensities. The corresponding vibrational data for c/5-Pt(CO)2(S03F)2 (which are very similar to those of c«-Pd(CO)2-(S03F)2) are also shown along with approximate assignments for the bands. The FT-IR spectrum of mer-Ir(CO)3(S03F)3 is shown in Figure 4-4. Figure 4-5 compares the CO-stretching regions of the Raman spectrum of the crystalline mer-Ir(CO)3(S03F)3 and the solution Raman spectrum of the mother liquor (separated from the crystals by pipette). Except for some more extensive band splitting for wer-Ir(CO)3(S03F)3 (particularly in the 1000 to 800 cm"1 region) assigned to v(SO ••• M) and v(SF), the vibrational spectra for wer-Ir(CO)3(S03F)3 and ris-Pt(CO)2(S03F)2 in the S0 3F vibration region are remarkably similar. Some small variations in band positions in the region of 400 to 512 cm"1 are due to the metal carbonyl moieties of the two compounds. While these bands are difficult to assign and in some cases even more difficult to detect, the identification of CO stretching vibrations is relatively simple and straightforward. For /ner-Ir(CO)3(S03F)3 (C2v point group) three CO stretching modes are expected, and they should both be infrared and Raman active. However, one of these modes is extremely weak and poorly resolved in the Raman spectrum though a 129 Table 4-2 Vibrational Data* for mer-Ir(CO)3(S03F)3 and cw-Pt(CO)2(S03F)2 mer-Ir(CO)3(S03F)3 cw-Pt(CO),(S03F)2 IR Raman IR Raman Approx. Assignment 2249 w 2208 s 2198 s 2168, 2150 vw 2249 vs 2206 w, sh 2196 s 2219 (s) 2185 (vs) 2218 (vs) 2181 (s) v ( , 2CO) v (12CO) v (12CO) v (13CO) 1409 s, sh 1394 vs -1380 m,sh 1397 (s) 1389 (s, sh) 1378 (vs) 1362 (w, sh) 1230 (m, sh) 1395 (s) 1376 (m) Vasym(S03) vsym(S03) 1213 vs 1215 s 1209 (vs) 1212 (s) 1030 s 1011 s 987 vs 932 w, sh 1056 m 1024 m 1007 1034 (m, sh) 1026 (s) 1009 (vs) 1019 (s) 993 (m) v S O - M 829 s 818 s 802 s 825 mw 815 mw 799 (s) 815 (w) 792 (w) vSF 658 m, sh 643 m, sh 640 s 657 (ms) 648 (m, sh) 656 (s) v(MO) + 5(S02) 595 s 585 m-w 589 (s, sh) 584 (s) 588 (mw) 580 (m) SaSym(S03) 548 m, sh 530 s 554 m 543 m,sh 557 (ms) 551 (s) 565 (w) 554 (w) ssym(so3) 451 vs 512 m 450 ms 403 w 276 s 137 s 476, 472 (ms) 436 (w) 411 (vw) 475 (w, sh) 462 (ms) 412 (w) 291 (m) 276 (vw) 175 (w) 151 (m) 138 (vw) 8(M-CO)? v(MO) + 5(S03) p(S03F) unassigned torsion and deformation modes 130 * Vibrational data refer to solid state samples. Abbreviations: s = strong, m = medium, w = weak, v = very, br = broad, sh = shoulder, v = stretching mode, sym = symmetric, asym = asymmetrical, 8 = deformation mode, p = rocking. shoulder band does appear at 2206 cm"1 in the spectrum of the solid and at 2208 cm"1 in the spectrum of the solution in HS0 3F. This band is assigned to the B2 mode since it is expected to be weaker in the Raman spectrum than the two Ai modes, which lie at 2249 and 2196 cm"1 respectively. In the IR spectrum one of the A, modes is weak. The relative intensities observed in the spectra are all consistent with the expected changes in the dipole moments or polarisabilities of the corresponding vibrational modes of the mer isomer. The advantage of using a combination of infrared and Raman spectroscopy is manifested in this case. If either IR or Raman spectrum were used alone to characterize the compound, it would seem very likely that a C3v symmetry (/aoisomer) might be assigned (incorrectly) to Ir(CO)3(S03F)3 since only two relatively strong bands were observed in either the IR or the Raman spectrum. The two fundamental bands (Ax and E) expected of fac-Ir(CO)3(S03F)3 should be both Raman and IR active and occur at the same positions in both spectra. The observation that this coincidence did not occur contradicts the assignment of the molecule to a C3v point symmetry. The average CO stretching vibration, va v(CO), is found to have a value of 2218 cm"1, which is about 75 cm"1 wavenumbers higher than that found for gaseous CO (2143 cm"1).18 From this, a stretching force constant (ft) of 19.9 ± 0.1 x 102 N m"1 is obtained using the two mass model. The CO stretching force constant is usually more accurate and reliable than its 131 FT-Raman of crystalline roer-Ir(CO)3(S03F)3 v (CO) region <3\ cs CS — i 1 1 1 1 r 2300 2200 2100 Av (cm") FT-Raman of Ir(CO)3(S03F)3 in HS0 3F solution v (CO) region 8 r f a c ( A , ) fac(E) i r 2300 T 1 2200 2100 Av (cm"1) Figure 4-5 The FT-Raman spectrum of Ir(CO)3(S03F)3 in the CO-stretching region for crystalline mer-Ir(CO)3(S03F)3 and a solution spectrum of the mother liquor. stretching frequency in deducing the relative importance of 7t-back-donation. Compared to Au(CO)S03F7 (v(CO) = 2196 cm"1, fr = 19.5 x 102 N m"1) and m-Pt(CO)2(S03F)28 (vav(CO) = 2202 cm"1, ft = 19.6 x 102 N m"1), both the weighted average CO stretching vibration and the stretching force constant for /ner-Ir(CO)3(S03F)3 are the highest among the three carbonyl fluorosulfates. From this it appears that for the noble metal carbonyl fluorosulfates in the 5d-block, the strength of the CO bond gradually increases with increasing 133 formal charge of the metal. This in turn suggests gradually decreasing 7t-back-donation from metal to CO in the order Au(CO)S03F > cw-Pt(CO)2(S03F)2 > Ir(CO)3(S03F)3. Vibrational studies of iridium(III) carbonyl derivatives19 and complexes of the type IrCl3(CO)(PR3)2, R = C2F£5 or n-C 4H 9 , have been reported where v (CO) is usually found between 2010-2090 cm"1. From these results, it is evident that the phosphines are better c-donor ligands than CO and enhance rc-back-donation from the metal. The carbonyl stretching frequencies for these compounds fall in the normal range for terminal CO ligands.19 As seen in Figure 4-5, two additional v(CO) stretching modes are observed in the Raman spectrum of the mother liquor in HS0 3F solution. The two bands at 2233 cm"1 (A,) and 2157 cm"1 (E) with IR counterparts at 2232 and 2156 cm"1 are assigned, based on band intensities, to ^ 2C-Ir(CO)3(S03F)3. These band positions agree with those reported for fac-Ir(CO)3F3 (at 2213 and 2165 cm"1). The average v(CO) values for/ac-Ir(CO)3(S03F)3 and jac-Ir(CO)3F3 of 2283 and 2181 cm"1 respectively suggest CO bonds of comparable strength, and stretching force constants of 19.2 ± 0 . 1 x 102 N m"1 are obtained for both of the compounds. Finally for Ir(CO)3I315 (assuming a facial configuration and C3v symmetry), a weighted average v(CO)a v of 2135 cm"1 and/r of 18.41 x 102 N m"1 are calculated based on the two reported bands. The CO stretching force constant for Ir(CO)3I3 is, therefore, slightly lower than that for gaseous CO (18.6 x 102 N m"1). When mer-Ir(CO)3(S03F)3 is redissolved in HS0 3F two bands at 2253 and 2197 cm"1 are found in the Raman spectrum, with a shoulder band at ca. 2208 cm"1 (weak and poorly resolved). For the same solution two 1 9F-NMR resonances due to the complex are observed 134 at 43.35 and 43.57 ppm with a peak area ratio of ca. 1:2, as expected of the two types of coordinated S03F" groups in mer-Ir(CO)3(S03F)3. It thus appears that isomerization in solution does not occur. It appears that CO bonding is slightly stronger in wer-Ir(CO)3(S03F)3 than in the fac-isomer as reflected by a difference of 35 cm'1 in v(CO)a v. It has been argued in the preceding section that the trans influence exerted by the fluorosulfate group is responsible for the difference in Ir-C bond lengths. By the same token, it can also be said that in the trans-Ir(CO)2-moiety there is more competition for 7t-electron density from the metal while the oxygen of the fluorosulfate group functions as a weak rc-donor towards the iridium(III) center. For compounds of the type IrCl3(CO)(PR3)2 (R = CH 3 or C 4 H Q ) , 2 0 several geometrical isomers have been obtained and studied in solution by IR using CHC13 as the solvent. It has been observed that v (CO) is always higher by 30 to 40 wavenumbers when the CO group is trans to a phosphine rather than trans to a chlorine. The CO stretching frequencies for these compounds are observed in the region from 2050 to 2100 cm"1 because of the aforementioned effect of the phosphine group. In compounds with v(CO) above 2200 cm"1, a complete absence of 7t-back-bonding is suggested.21,22 However, the calculated CO stretching force constants of these metal carbonyl compounds are still below the/r value for H C O + (21.3 x 102 N m"1).12 Thus it seems that a small amount of rc-back-donation still exists in these metal carbonyl complexes. The iridium(III) carbonyl derivatives discussed here clearly show that the metal-carbon bonding changes gradually from synergetic bonding to mainly a-bonding as phosphines or 135 other 7t-acid ligands are replaced by CO to give tris(carbonyl) iridium(III)-derivatives. The same trend is observed as the anionic ligands become harder and more electron withdrawing. 4.3.2 rjr(CO)sCl][Sb2Fn]2 4.3.2.1 Synthetic Aspects [Ir(CO)5Cl][Sb2Fn]2 was obtained unexpectedly in an attempt to synthesize [Ir(CO)6]-[Sb2FH]3. As already discussed in Chapter 3, metal carbonyl fluorosulfate can be used as a precursor for the synthesis of the corresponding metal carbonyl cation via a solvolysis reaction in antimony pentafluoride in the presence of CO. Initially, the solvolysis reaction was carried out in liquid antimony pentafluoride in the presence of CO at 60-80 °C. The reaction seemed to go to completion within 3 to 4 days. However, both the IR and Raman spectra showed more complicated patterns than would have been expected for an octahedral metal carbonyl cation. It was thought that the reaction might not have been complete. Therefore, SbF5 and CO was once again added to the solid product in the reactor and the reaction was allowed to continue for another 4 days. CO gas phase determination indicated no further CO uptake by the prolonged reaction, and the IR and Raman spectra gave the same results as before. The two relatively strong peaks at 2295 and 2275 cm"1 in the Raman spectrum have no counterparts in the IR spectrum. The peak at 2275 cm"1 is the stronger of the two Raman bands. It has been observed in the Raman spectrum of other M(CO)6 species 2 3 ' 2 4 that the A , g stretching vibrations are of lower intensity and found at higher wavenumbers than the E g vibrations. The extremely high band positions (2295 and 2275 cm"1), the intensity pattern, 136 and the non-coincidence of IR and Raman bands (the mutual exclusion rule) suggest the presence of [Ir(CO)6]3+ with a point group symmetry of Oh in the reaction mixture, even though it could not be isolated. The complexity of both the Raman and IR spectra is caused by the presence of at least one other iridium carbonyl species. In order to isolate the unidentified Ir carbonyl species a slightly different synthetic approach was employed. Instead of using antimony pentafluoride, "magic" acid (a mixture of fluorosulfuric acid and antimony pentafluoride) was used. Crystalline mer-Ir(CO)3(S03F)3 was dissolved in the magic acid and the reaction was carried out at room temperature without being stirred in the hope of obtaining single crystals. The addition of fluorosulfuric acid and antimony pentafluoride was accomplished by vacuum transfer via a distillation bridge. Fluorolube grease Series 25-10M, CF 2Cl(CF 2-CFCl)nCF 2Cl, was used to lubricate all ground glass connections. This may be a source of the chlorine found in the molecule of crystalline product since none of the chemicals used during the reaction contained chlorine. The grease used during vacuum transfer usually does not cause problems as long as the duration of exposure to the grease is brief and the transfer is carried out at room temperature. The prolonged time, slight warming, and the use of magic acid could have caused a partial decomposition (fluorination by SbF5) of the grease, producing Cl" as a by-product. The extremely small amount of the reaction product indicates that only trace amounts of chlorine were present. The small quantity of [Ir(CO)5Cl][Sb2Fn]2 obtained as single crystals prevents the recording of IR and Raman spectra. This appears unnecessary since the crystal structure was successfully determined by single crystal X-ray analysis, to be discussed in the next section. 137 4.3.2.2 Molecular Structure of [Ir(CO)5Cl][Sb2Fn]2 The crystal structure of [Ir(CO)5Cl][Sb2Fu]2 was solved by Dr. S. J. Rettig and Professor J. Trotter of this Department. Bond lengths, bond angles, and non-bonded contacts extending to 3.60 A are found in Appendix D. The molecular structure, depicted in Figure 4-6, shows that the complex consists of a ternary iridium carbonyl cation, [Ir(CO)5Cl]2+, and two fluoroantimonate(V) anions, [Sb 2F n]\ The cation adopts a slightly distorted octahedral configuration with the iridium atom at the center of the octahedron. The bond distances of the five iridium-carbon bonds do not vary significantly (see Appendix D), ranging from 2.00(2) to 2.05(1) A. These metal-carbon bond distances are comparable with those of mer-Ir(CO)3(S03F)3 (vide supra), but are significantly longer than the reported values for terminal CO groups in iridium carbonyls.16 Obviously, the metal-carbon bond is more sensitive (compared to C-O bond) to a reduction in Tt-back-donation from Ir(III). Hence, a much reduced rc-back-bonding (or predominant o-bonding) is encountered in Pr(CO)5Cl]2 +. Likewise, the C-0 bond distances are at the lower end of the values typically displayed by metal carbonyls. However, a more detailed analysis of the bond distances is difficult since a thermal motion correction has not been made for the C-0 bond lengths in [Ir(CO)5Cl][Sb2Fn]2. The bond angles of 174.0(6)° and 174.2(6)° for C(4)-Ir(l)-C(5) and C(2)-Ir(l)-C(3) respectively indicate that the four carbonyls which are cis to chlorine are not coplanar but are bent slightly towards the chlorine, constituting an overall umbrella-shape for the cation. However, the C4v symmetry of the ternary carbonyl cation is basically retained since all the CO groups except the one trans to the chlorine deviate by a similar degree. 138 139 The two [Sb2Fn] anions in the molecule are not identical and have slightly different bond angles. Both are severely distorted from symmetry in several ways. First, the Sb(l)-F(l)-Sb(2) and Sb(3)-F(12)-Sb(4) bridges are bent (153.2(6)° and 156.5(6)° respectively) and are slightly asymmetric. Secondly, the equivalent Sb-F bonds are no longer the same length and are no longer coplanar. Similar observations have been made for the same anion in the [Hg(CO)2][Sb2F,1]225 species where the Sb-F^- Sb angle was found to be 147.6(3)° (see Appendix E). The reason suggested for these distortions is that they arise as a consequence of secondary interactions with the corresponding cations. In the near absence of 7t-back-donation, carbon atoms in the metal carbonyls are expected to be partially electrically positive and unstable. The observed secondary contacts provide stabilizing effect by compensating for the partially positive charge via inter-ionic contacts to electronegative atoms such as oxygen and fluorine. There are altogether 15 C • • • F close contacts in the range of 2.69(1) to 3.00(2) A, which are noticeably less than the sum of the respective van der Waals radii (1.70 + 1.47 = 3.17 A).1 7 However, these close contacts are not evenly distributed among the 5 carbon atoms of [Ir(CO)5Cl][Sb2F„]2 in the crystal structure. C(2), C(3), and C(5) have 3 such contacts each in the ranges of: 2.72(2) to 3.00(2) A; 2.84(2) to 2.91(2) A; and 2.69(1) to 2.98(1) A respectively. C(4) has 4 such contacts in the range of 2.77(2) to 2.96(2) A. C(l), which is trans to the chlorine, only has 2 such contacts of 2.84(2) and 2.90(2) A. These close contacts between carbon and fluorine are attractive in nature, and their presence seems to be a common feature of cationic metal carbonyl complexes. 140 There are also a number of F • • • F (in the range of 2.79(2) to 2.90(1) A) and F • • • O contacts (in the range of 2.81(2) to 2.92(1) A) which are slightly less than the sums of the van der Waals radii of these elements (2.94 and 2.99 A respectively). Similar contacts have been observed in the crystal structure of [Hg(CO)2][Sb2FH]2.25 In contrast to the interaction between fluorine and carbon, the F • • F and F • O interactions are rather weaker and probably repulsive in nature. It may be considered that their presence is caused or imposed by the numerous relatively strong interactions between fluorine and carbon atoms in the crystal structure. The chlorine atom does not seem to have been involved in such secondary interactions with other atoms in the molecule. 4.4 Summary and Conclusions The synthesis of mer-Ir(CO)3(S03F)3 was accomplished by CO addition to Ir(S03F)3. Single crystals of /ner-Ir(CO)3(S03F)3 were obtained and the molecular structure was determined by single-crystal X-ray diffraction studies. The vibrational spectra suggest a greatly reduced iridium to CO rc-back-bonding. The crystal structure reveals intra- and inter-molecular contacts between the electropositive C atom of the CO and the O or F atoms of the fluorosulfate groups. Hence mer-tris(carbonyl)iridium(III) fluorosulfate becomes the first thermally stable, structurally characterized, and predominantly c-bonded carbonyl derivative of a metal in the +3 oxidation state. As observed previously for cw-Pd(CO)2(S03F)2 (see Chapter 3) significant inter- and intra-molecular SO • • • CO contacts appear to exert a stabilizing influence on the crystal structure of mer-Ir(CO)3(S03F)3. The two compounds, /ner-Ir(CO)3(S03F)3 and cis-Pd(CO)2(S03F)2, also show similarities in terms of bond lengths and angles. The two metal 141 carbonyl fluorosulfates are, however, very different in their stereochemistry, which is cis for Pd(CO)2(S03F)2 and mer (or trans) for Ir(CO)3(S03F)3. This appears to be a consequence of how each complex forms: cw-Pd(CO)2(S03F)2 forms when two CO ligands in square planar [Pd(CO)4]2+ are replaced by two S03F" ions; in solvated Ir(S03F)3 gradual addition of CO takes place. The resulting geometries are thus a consequence of the strong trans directing ability of the CO group in each case. The attempts to obtain [Ir(CO)6][Sb2Fn]3 by solvolysis reaction in SbF5 via mer-Ir(CO)3(S03F)3 produces a mixture of products. Vibrational spectroscopy provides strong evidence for the presence of the target compound, [Ir(CO)6][Sb2Fn]3. The isolation of the pure compound from the mixture has, however, been unsuccessful. The use of magic acid as a reaction media produces the unexpected compound, [Ir(CO)5Cl][Sb2Fn]2, for which the molecular structure has been obtained by single crystal X-ray diffraction studies. [Ir(CO)sCl][Sb2Fn]2and wer-Ir(CO)3(S03F)3 are the only two structurally characterized Ir(III) carbonyl derivatives which show a significant reduction in rc-back-donation. One additional point remains to be made to conclude this chapter. The pure form of Pr(CO)6][Sb2Fu]3 has been obtained recently by Professor Winner's group in Hannover.26 The compound is formed by reductive carbonylation of IrF6 in SbF5 in a CO atmosphere. The vibrational spectra of pure [Ir(CO)6][Sb2Fn]3 confirms that carbonylation of Ir(CO)3-(S03F)3 in SbF5 in a CO atmosphere, discussed earlier in this chapter, does produce [Ir(CO)6][Sb2FM]3 as a major product. 142 References 1 Lide, D. R. CRC Handbook of Chemistry and Physics, 73rd ed.; CRC: Boca Raton, 1992-1993; p 4-15. 2 Burk, M . J.; Crabtree, R. H. Inorg. Chern. 1986, 25, 931. 3 Cotton, F. A. ; Wilkinson, G. Advanced Inorganic Chemistry, 5th ed.;Wiley: New York, 1988;p 907. 4 (a) Brewer, S. A. ; Holloway, J. H . ; Hope, E. G.; Watson, P. G. J. Chern. Soc, Chern. Commun. 1992, 1577. (b) Brewer, S. A. ; Brisdon, A. K.; Holloway, J. H . ; Hope, E. G. ; Peck, L. A. ; Watson, P. G. J. Chern. Soc, Dalton Trans. 1995, 2945 5 Ellis, J. E. Adv. Organomet. Chern. 1990, 31, 1 and references therein. 6 Lee, K. C ; Aubke, F. J. Fluorine Chern. 1982, 19, 501. 7 Willner, H . ; Aubke, F. Inorg. Chern. 1990, 29, 2195. 8 Hwang, G.; Wang, C ; Bodenbinder, M , ; Willner, H . ; Aubke, F. J. Fluorine Chern. 1994, 66, 159. 9 Lee, K. C ; Aubke, F. Inorg. Chern. 1979, 18, 389. 1 0 Lee, K. C ; Aubke, F. Inorg. Chern. 1984, 23, 2124. 1 1 (a) Lee, K. C ; Aubke, F. Can. J. Chern. 1977, 55, 2473. (b) Lee, K. C ; Aubke, F. Can. J. Chern. 1979, 57, 2058. 1 2 Willner, H . ; Schaebs, J.; Hwang, G.; Mistry, F.; Jones, R.; Trotter, J.; Aubke, F. J. Am. Chern. Soc. 1992, 114, 8972. 143 1 3 Hwang, G.; Wang, C ; Aubke, F.; Willner, H . ; Bodenbinder, M. Can. J. Chern. 1993, 71, 1532. 1 4 Wilkins, R. G. Kinetics and Mechanism of Reactions of Transition Metal Complexes, VCH: Weiheim, Germany and references therein. 1 5 Malatesta, L. ; Naldini, L. ; Cariati, F. J. Chern. Soc. 1964, 961. 1 6 Orpen, A. G. ; Brammer, L. ; Allen, F. H . ; Kennard, O.; Watson, D. G.; Taylor, R. J. Chern. Soc, Dalton Trans. 1989, 511. 1 7 Bondi, A. J. Phys. Chern. 1964, 68, 441. 1 8 Nakamoto, K. Infrared Spectra of Inorganic and Coordination Compounds, 4th ed.; Wiley: New York, 1986. 1 9 Adams, D. M . Metal-Ligand and Related Vibrations; Edward Arnold: London, UK, 1967. 2 0 Chart, J.; Johnson, N. P.; Shaw, B. L. J. Chern. Soc. 1964, 1625. 2 1 Calderazzo, F. J. Organomet. Chern. 1990, 400, 303. 2 2 Hurlburt, P. K.; Rack, J. J.; Luck, J. S.; Dec, S. F.; Webb, J. D. ; Anderson, O. P.; Strauss, S. H. J. Am. Chern. Soc. 1994, 116, 10003. 2 3 Abel, E. W.; Mclean, R. A. N. ; Tyfield, S. P.; Braterman, P. S.; Walker, A. P.; Hendra, P. J. J. Mol. Spectrosc. 1969, 30, 29. 2 4 Jones, L. H . ; McDowell, R. S.; Goldblatt, M. Inorg. Chern. 1969, 8, 2349. 2 5 Bodenbinder, M . ; Balzer-Jollenbeck, G.; Willner. H . ; Batchelor, R.; Einstein, F. W. B.; Wang, Aubke, F. Inorg. Chern. 1996, 35, 82. 26 Bach, C. Diploma Thesis, der Universitat Hannover, 1995. 144 CHAPTER 5 SYNTHESES AND SPECTROSCOPIC CHARACTERIZATIONS OF CARBONYL CATIONS OF GROUP 8 METALS 5.1 Introduction The major thrust of this chapter is the synthesis and characterization of some ruthenium and osmium carbonyl cations. The attempted synthesis of the corresponding iron carbonyl will also be discussed for the simple reason that the three elements of the Group 8 metals (Fe, Ru, and Os) show much similarity in their carbonyl chemistry. Like other heavier Group VIII (8-10) metals, ruthenium and osmium are rare elements with an abundance of ca. 10"7% of the Earth's crust1. They occur as metals, often as alloys such as "osmiridium," and in sulfide, arsenide, and other ores. They are commonly associated with Cu, Ag, and Au. The main sources of these metals are South Africa, Russia, and Canada. In contrast to the rarity of ruthenium and osmium, iron is the second most abundant metal (by weight) after aluminum and the fourth most abundant element in the Earth's crust (after O, Si, and Al). The Earth's core is believed to consist mainly of iron and nickel, and the occurrence of iron meteorites suggests that it is abundant throughout the solar system. Ruthenium metal carbonyls were prepared as early as 1910 when Mond et al.2'2 prepared an orange crystalline solid by the action of carbon monoxide on metallic ruthenium 145 at 350-450 atm/300 °C. Of course, the product could not be fully characterized at that time. Since then a number of ruthenium carbonyls have been prepared and identified (vide infra). The binary carbonyls known so far include mononuclear and multinuclear species. The interesting aspect about the chemistry of Group 8 metal carbonyls is that the group homology is very much pronounced in terms of the varieties and structures of the metal carbonyls formed by these metals. As can be seen from Table 5-1, all Group 8 metals form thermally stable mononuclear, binuclear, and trinuclear metal carbonyls.4 Among the many applications of Group 8 metal carbonyls, their catalytic activity has long been noticed. They can be used in various reactions, for example,5 2HC=CH + 3CO + H 2 0 F e ( c o h )HO-C 6H 4-OH + C 0 2 (5-1) and C 6 H 5 N0 2 + 2CO + H 2 * " 3 ( C Q ) l 2 ) C 6 H 5 N H 2 + 2C0 2 (5-2) Although pentacarbonyls of ruthenium and osmium have been known for several decades, these compounds are very volatile and difficult to obtain as pure compounds. Their preparation and characterization were accomplished recently by Pomeroy and co-workers6 using Ru3(CO)12 and Os3(CO)12 as precursors respectively. Vibrational spectroscopy has played a major role in the identification of these metal carbonyl complexes. In addition to binary metal carbonyl species of Group 8 metals, there have also been extensive studies of the metal carbonyl halides.7 Carbonyl halides are very important 146 Table 5-1 Binary Metal Carbonyls of Group 8* Compound Comments Fe(CO)5 Yellow liquid; tbp; Fe-C = 1.810(3)ax, 1.832(2)eq Mononuclear Ru(CO)5 Colorless liquid; tbp (by IR) Os(CO)5 Colorless liquid; tbp (by IR) Fe2(CO)9 Shiny golden plates, confacial bioctahedron Binuclear Ru2(CO)9 Dark orange; loses CO to give Ru 3(CO)i 2 Os2(CO)9 Orange yellow; decomposes in a few days at -20 °C Highly unsymmetrical, terminal and bridging CO D3h symmetry, only terminal CO D3h symmetry, only terminal CO * Reference 4. tbp = trigonal bipyramidal Fe3(CO)12 Trinuclear Ru3(CO)12 Os3(CO)12 precursors in the preparation of metal carbonyl cations. For example, pentacarbonyl chlorides of Group 7 metals form stable, diamagnetic homoleptic metal carbonyl cations according to8 M(CO)5X + A1C13 + CO -> [M(C0)6]+[A1XC13]" M = Mn, Tc, or Re, and X = Br or Cl. (5-3) The formation of similar compounds, the homoleptic metal carbonyl dications of the type [M(CO) 6] 2 +, were claimed for Group 8 metals of Fe and Os. 9 , 1 0 However, these claims 147 have been subsequently repudiated,11 and it is accepted that none of these cations exist in isolable, thermally stable compounds.12'13 Therefore, it appears that the Lewis acids used so far in these reactions, such as A1C13 and FeCl3, are still not strong enough, and the resulting conjugate anions, ([A1C14]~, [FeCl4]~), are incapable of stabilizing the dipositive cations. The successful isolation and characterization of homoleptic cations, [M(CO)4]2 + (M = Pd or Pt),14 [Hg(CO) 2] 2 +, 1 5 , 1 6 and [Au(CO)2]+ 1 7 all as [Sb2Fn]" salts, show that solvolysis reactions in antimony pentafluoride in the presence of CO are very promising synthetic methods. Hence, an attempt to prepare the long anticipated dications [M(CO)6]2 + (M = Fe, Ru, or Os) by this means seems appropriate. 5.2 Experimental 5.2.1 Synthesis of [Ru(CO)6][Sb2Fn]2 In a typical reaction, 0.098 g of ruthenium powder was allowed to react at 60 °C for 24 hours with a mixture of approximately 1.5 mL of S 2 0 6 F 2 and 3 mL of HS0 3F in a one-1 8 part reactor with magnetic stirring according to 2Ru + 3S 20 6F 2 H S O i F ) 2Ru(S03F)3 (5-4) Completion of the reaction was indicated by the consumption of all of the metal powder and the formation of a dark red-brown solution. After removal of all volatiles first at room temperature and then at 60 °C, a solid dark red-brown product was obtained. This solid product, Ru(S03F)3 (identified by mass balance and IR spectroscopy), was used as a precursor for the synthesis of [Ru(CO)6][Sb2F„]2. 148 Approximately 5 mL of SbF5 was added by vacuum distillation to the reactor which contained Ru(S03F)3, and a CO pressure of about 2 bars was introduced afterwards. The reaction was carried out at 60-80 °C with magnetic stirring. A color change from the initial dark red-brown to purple was observed within one or two hours. Further gradual color changes were also noticed during the first two days of the reaction. The reaction was allowed to continue for another two days, and finally a homogenous white suspension was obtained. The CO uptake was determined, and the reactor was recharged with CO. However no further CO uptake was observed. Distillation of all the volatiles in vacuo left behind a white powdery solid product. The carbon content of the product (weight %) was analyzed to be 5.76%, compared with the calculated value of 6.14% for [Ru(CO)6][Sb2F,,]2. A melting point determination was carried out with the sample in a sealed capillary tube, but the compound did not show a distinct melting point. At 165 °C, the sample started shrinking but without obvious color change. At 220 °C, the sample turned into a floating powdery product inside the capillary tube. Black spots were observed when the temperature reached 304 °C, and at 310 °C the sample became uniformly black. 5.2.2 Synthesis of [Os(CO)6][Sb2Fn]2 A procedure similar to that used to obtain Ru(S03F)3 was adopted for the preparation of the starting material Os(S03F)3,19 but no HS0 3F was added. In a typical preparation, excess S 2 0 6 F 2 (about 2 mL) was vacuum distilled onto 38.6 mg (0.2029 mmol) of osmium powder in a 50-mL one-part reactor. The reaction was carried out at 60 °C with magnetic stirring. After three days a bright, clear green solution was obtained, and no metal powder 149 was detectable. The excess S 2 0 6 F 2 was removed by distillation in vacuo with the reactor first at room temperature and then at 60 °C for a short period of time. About 3-4 mL of SbF5 was added to the reactor by vacuum distillation, and a CO pressure of ca. 500 mbar was introduced to the system with the reactor immersed in a liquid nitrogen bath. The reactor was allowed to warm to room temperature, and a 60 °C water bath was used to facilitate magnetic stirring. Shortly after the mixing of antimony pentafluoride and osmium fluorosulfate, the green color of osmium fluorosulfate began to fade, and the suspension turned purple. The color of the suspension gradually changed to off-white within 24 hours. The reaction was allowed to continue for another 3 days, and a homogeneous off-white suspension was obtained. The volume and pressure of the unreacted CO were measured, and about 1.4 mmol of CO was consumed during the reaction. No further CO uptake was observed even though the reactor was recharged with a fresh portion of CO. The volatiles were removed in vacuo at room temperature overnight, and then a 60 °C water bath was used to ensure a complete removal of the remaining volatiles. The final product was a off-white powdery solid with a purple cast. Carbon content (weight %) was analyzed to be 5.12%, compared with the calculated value of 5.70% for [Os(CO)6][Sb2Fu]2. The slightly low carbon content than the expected value may arise from some trapped SbF5. Another possible reason is that trace amounts of Os(III) in Os(S03F)3 could be reduced to osmium metal, as indicated by the slightly purplish cast of the solid product. Nevertheless, the dication [Os(CO)6]2+ was unambiguously established by vibrational spectroscopy (vide infra). 150 The melting point determination of [Os(CO)6][Sb2Fj,]2 gave no distinct melting point. The white fine powder began to shrink at 172 °C and a pale orange color was observed. At 185 °C, the sample turned into a white powder again. At 230 °C, the white powder started to float inside the capillary tube. Some yellow and black spots appeared when the temperature reached 310 °C. The decomposition was complete at 320 °C when the sample turned to a uniformly gray black color. 5.2.3 Attempted Synthesis of Cationic Ruthenium and Osmium Carbonyl Fluorosulfates Preliminary attempts were made to investigate the generation of cationic metal carbonyl complexes in fluorosulfuric acid, HS0 3F. Further research in this direction needs to be carried out in order to reach conclusive results. 5.2.3.1 Ruthenium Carbonyl Fluorosulfate Ruthenium powder, 52.8 mg (0.5224 mmol), was used to prepare Ru(S03F)3 as 1 Q previously described. About 2 mL of HS0 3F was added after removing all excess S 20 6F 2 in vacuo. CO was subsequently introduced to the system. Under constant magnetic stirring at 60 °C, the original black brown suspension changed to an almost clear orange solution overnight. A slight color change was noticed thereafter, and the reaction was allowed to continue until a final pale greenish yellow color was reached. The CO uptake was determined after 4 days of reaction, and approximately 2 mmol CO was consumed by the reaction. The reactor was recharged with CO, but no further CO uptake was observed. 151 The infrared spectrum of the solution showed three CO stretching bands. All three bands were relatively strong, with a band at 2154 cm"1 being the most intense, and the other two bands at 2199 and 2185 cm"1 of medium intensity. 5.2.3.2 Osmium Carbonyl Fluorosulfate Approximately 3 mL of S 2 0 6 F 2 and HS0 3F (1:1 mixture) was condensed over 40.7 mg (0.2140 mmol) of osmium powder in a 50-mL one-part reactor via vacuum distillation. After all the metal powder was consumed and a clear green solution was obtained, the excess S 2 0 6 F 2 was removed, and another 2 mL of HS0 3F was added. Upon introducing CO to the reactor, the original clear solution turned cloudy, and the color changed from green to red brown. The deep dark color of the system made it difficult to tell whether a precipitate had formed. The magnetic stirring was stopped, but allowing the mixture to stand overnight did not produce any noticeable precipitate. Gas phase determination indicated that ca. 1.0 mmol of CO had been consumed and approximately 0.2 mmol of C 0 2 had been produced during the reaction. The reactor was recharged with CO once more, but no further CO uptake was observed. The infrared spectrum of the solution was complicated. In the CO stretching region, there were two sharp bands at 1880 and 2037 cm"1, a broad band at 2102 cm"1, and two shoulders at 2122 and 2150 cm"1 respectively. 5.2.4 Attempted Synthesis of [Fe(CO)6][Sb2Fn]2 In a glove box, 137.9 mg (0.8502 mmol) of anhydrous iron(III) chloride (FeCl3) was transferred into a 100-mL one-part reactor. Approximately 6 mL of SbF5 was introduced into 152 the reactor by vacuum distillation. Carbon monoxide was subsequently added. The amount of CO added to the reactor was determined by the volume of the vacuum line and the CO pressure measurement. A water bath at 60 to 80 °C was used to facilitate magnetic stirring and the mixing of the reactants. The reaction proceeded relatively fast. Within 15 minutes, most of the black brown crystalline FeCl3 appeared to have been dissolved in SbF5, and the color changed to brick-red. The color of the suspension faded as the reaction continued. A uniformly pale purplish red suspension was reached after 4 days of reaction, and the physical appearance did not change significantly even though the reaction was allowed to continue for another 3 days. The amount of CO consumption was determined by measurement of gas pressures; first with the reactor immersed in a liquid nitrogen bath, and again after the reactor was allowed to warm to room temperature. The results indicated that approximately 0.45 mmol CO of had been consumed during the reaction. All the volatiles were removed in vacuo first at room temperature overnight, and then at 60 °C for 3 hours. An off-white crystalline product was obtained, and the weight of the product was determined by the weight difference of the reactor. The actual weight of 455.9 mg corresponded closely to that expected for Fe[SbF6]2 (448.0 mg). 153 5.3 Results and Discussion 5.3.1 Synthetic Aspects The synthesis and characterization of [M(C06][Sb2F11]2, M = Ru or Os, provides the first evidence that dipositive metal carbonyl cations do exist. Once again, SbF5 is found to be a very useful reagent in stabilizing metal carbonyl cations. Compared to [A1C14]" or [FeCl4]", [Sb2Fn] is less nucleophilic, and weakly coordinating. It is suspected that secondary interactions, first noticed in the cationic palladium carbonyls of m-Pd(CO)2(S03F)220 and later seen in the crystal structures of /ner-Ir(CO)3(S03F)3, [Hg(CO)2][Sb2Fn]2,16 and Pr(CO)5Cl][Sb2Fn]2, are partly responsible for the formation of the cations. Single crystals are difficult to obtain from SbF5 solutions. Both [Ru(CO)6][Sb2Fn]2 and [Os(CO)6][Sb2Fn]2 are insoluble in S0 2, and the attempted recrystallization from HS0 3F or HS03F-SbF5 has resulted in a partial decomposition with CO evolution. The resulting products have not been fully characterized. The synthetic procedure adopted here is a one-step solvolysis in SbF5 in the presence of CO. This follows the success in preparing [Pd(CO)4][SbFn]2 (see Chapter 6) from the binary fluorosulfate. Previously, metal carbonyl fluorosulfates, such as Au(CO)(S03F),21 and cw-M(CO)2(S03F)2 (M = Pd or Pt),22 have been used as precursors for the synthesis of the metal carbonyl cations. The direct solvolysis reaction discussed here has been employed in the synthesis of [Hg(CO)2][Sb2Fn]2.15'16 The new homoleptic octahedral metal carbonyl cations are produced according to 154 2M(S03F)3 + 16SbF5 + 13C0 -> 2[M(CO)6][Sb2Fn]2 + C 0 2 + S 2 0 5 F 2 + 4Sb2F9(S03F) (M = Ru or Os) (5-5) Both [Ru(CO)6][Sb2Fn]2 and [Os(CO)6][Sb2Fn]2 are white, moisture-sensitive amorphous solids. They were identified by microanalysis, the mass balance of the reaction, and CO uptake measurements. Under similar conditions, however, the solvolysis reaction of FeCl3 in SbF5 does not lead to the generation of the expected iron hexacarbonyl cation, [Fe(CO)6]2+. Instead, the known compound, Fe[SbF6]2,23 is formed according to 2FeCl3 + 8SbF5 + CO -> 2Fe[SbF6]2 + COCl 2 + 4SbF4Cl (5-6) The reaction may be seen as a new method for the synthesis of Fe[SbF6]2. FeCl3 is more easily available than FeF2, which is a usual starting material for Fe[SbF6]2. In addition, the method discovered here does not involve the use of any additional solvents, such as HF or S0 2. The reductive carbonylation of Ru(S03F)3 and Ru(S03F)3 in fluorosulfuric acid was only partially successful. The preliminary results indicate that cationic metal carbonyls can be generated in HS0 3F. However, further investigation is needed in order to fully characterize the complexes formed in HS0 3F. The thermal stabilities of [Ru(CO)6][Sb2Fn]2 and [Os(CO)6][Sb2Fn]2 are comparable. On heating in a capillary tube, both solids shrink at approximately same temperature, 165 °C for [Ru(CO)6][Sb2F„]2, and 172 °C for [Os(CO)6][Sb2F„]2. Both compounds undergo a 155 series of decomposition processes with color change and gas evolution. Complete decomposition occurs when the temperature rises above 300 °C; 310 °C for [Ru(CO)6]-[Sb 2Fu] 2 and 320 °C for [Os(CO)6][Sb2F„]2. Solutions of [Ru(CO)6][Sb2Fn]2 and [Os(CO)6][Sb2F„]2 in HS0 3F or in HS03F-SbF5, the magic acid, are unstable. In addition, both compounds are insoluble in sulfur dioxide so attempts to obtain single crystals were unsuccessful. The characterization of the cations then rests on vibrational and 1 3 C MAS NMR spectroscopy. 5.3.2 Vibrational Spectroscopy 5.3.2.1 [Ru(CO)6][Sb2Fn]2 and [Os(CO)6][Sb2Fn]2 As discussed previously, vibrational spectroscopy is very useful in characterizations of metal carbonyl compounds.24'25 The vibrational analysis is also facilitated by the analogous 13 13 syntheses of the C isotopomers by using CO. The high symmetry of some cations necessitates the use of both IR and Raman spectroscopy. CO stretching vibrations for terminal CO groups generally give rise to bands between 2100 and 1850 cm"1, where they are easily identified and where vibrational coupling to other fundamentals is usually not important. However, bending and deformation modes of CO are at much lower frequencies, and are often difficult to identify. This difficulty can largely be overcome by isotopic substitution. Equation (5-7) describes the relationship between a vibrational frequency and some related properties for a diatomic molecule: 1 f v = -=- ,— (5-7) 2%c \ \i 156 where v = vibrational frequency, c = speed of light, ft = stretching force constant, and \i — reduced mass (1//J = l/nii + l/mj, nij and m2 are the masses of the two atoms joined by a chemical bond). Because the force constant, / „ is the same for both 1 2 CO and 1 3 CO, the bond stretching frequency of the CO entity is inversely proportional to the square root of / i , which is different for 1 2 CO and I 3 CO. Therefore, the CO vibration will be shifted to lower wavenumbers when 1 2 CO is substituted by 1 3 CO. In the case of [Ru(CO)6][Sb2Fn]2 and [Os(CO)6][Sb2F„]2, 1 3 CO substitution will only affect metal carbonyl cations, while the vibrations of the anions will in principle remain unchanged. Hence, vibrations of cations and anions should be distinguishable via isotopic substitution. As can been seen from the infrared and Raman data for [Ru(CO)6][Sb2Fu]2 and [Os(CO)6][Sb2Fn]2 listed in Tables 5-2, no vibrational band is both infrared and Raman active in the CO stretching region for either compound. It is apparent that the mutual exclusion rule applies, and the metal carbonyl cations, [Ru(CO)6]2+ and [Os(CO)6]2+, should have Oh symmetry. The irreducible representations of normal vibrations for the octahedral cations [M(CO)6]2 +, (M = Ru or Os) are as follows. F v ib = 2A l g [v„ v2(R, p)] + 2Eg[v3, v4(R, dp)] + lFlg[v5(inactive)] + 4Flu[v6, v7, v8, v9(IR)] + 2F2g[v10, vn(R, dp)] + 2F2u[v12, v13(inactive)] (5-8) Among a total of thirteen fundamental frequencies, six are Raman active, and four are infrared active modes. In principle, none of them should appear in both infrared and Raman spectra simultaneously. 157 co o t-l o rt5 X ! oo O U 00 O U • i - H 00 3 Q 13 c o •X3 2 x> > Xi 9 H c *- .2 § 3 a u C <0 OJQ -C co , co fa < & PL, <s X> 00 o U CO o <s X) o u CO o PL, X) o u PL, x> 00 O 3 rt rt rt rt c e rt 5b „ s •7" $t> — < P3 PL, NO v© O NO CN CN CN > PL, wo oo o ON oo r*» NO NO NO ON OO rf ON NO wo \ > > < W PL, o o rt PL, O 00 ON CN co «—i CO CO CN rt CN H N H M h h H H ( f | ON f» wo NO NO NO OO 1/0 ^4 ON rt ON wo wo rt o rt CO NO -H NO OO O O f - CO CO CO CN CN ON OO wo rtI CN CN CN CN 1/0 WO O ON OO O NO NO NO ON O NS NO O O rt o oo ON CN ON CO CN r - co co CN CN o ON i-, CN H M H M h h rt, CO ON t wo NO NO NO ON CN —< ON NO ON WO WO rt NO rt co NO I NO o o r-CO CO CN 00 CO CN CO CN CN t~-CN CN CN rt O ON oo r-NO NO NO ON ON rt ON NO wo rti ON ON r -co co oo ON CN wo o CO CO CN ^ ON rt O CO H M H (S h h O O rt ON r- wo NO NO NO r - CN ON rt w0 w0 CN ON rt CO co NO - H o o co co NO t--CN 00 CO CN co CO > £ £ £ £ £ rt CN W0 CN CN CN CN CN rt r- © ON oo r-NO NO NO O ON w0 ON NO w0 rt CN ON OO co co 00 ON CN > co w v / w0 ~ * CO CO CN co £ B ON ON CN CN rti cN rt, CN r- r-O O rt ON t wo NO NO NO 6 ON r - CN ON w0 ON w0 w0 rt / — \ CO c c CO > —' 1 ' s - ^ NO NO ,_| NO OO CO o o cn CO cn cn CN CN 158 Five out of six Raman active and three out of four IR active fundamentals of [M(CO) 6] 2 +, (M = Ru or Os) are identified by their 1 2 C - 1 3 C shifts. However, two bands, v9(F lu) and v1 0(F2 g), are missing. They are of low energy, and probably of low intensity. These bands could be buried beneath the many anion bands at low wavenumbers, some of which have relatively high intensity. 26 27 As has been observed in the Raman spectra of other octahedral M(CO)6 species, ' the Raman active A l g CO stretching vibrations are found at higher wavenumbers, but are of lower intensity, than the E g stretching fundamentals. As summarized in Table 5-3, the CO stretching vibrations of the two homoleptic carbonyl cations [M(CO) 6] 2 +, (M = Ru or Os) occur at considerably higher wavenumbers than for other isoelectronic M(CO)6 species, and the stretching force constants are 19.82 and 19.74 x 102 N m"1 respectively, compared to 17.15 x 102 N irf1 for Mo(CO)6and 17.02 x 102 N m"1 forW(CO)6 respectively.27 The vibrational spectrum of [Sb2Fn] anion has been reported previously for [XeF][Sb2Fn],28 Cs[Sb 2F„], 2 9 and [C 3F 3][Sb 2Fn]. 3 0 However, the reported data are not completely consistent; measurements in the lower spectral region are limited, and the vibrational assignments are incomplete. In contrast, a much more detailed study has been made of [Au(CO)2][Sb2Fn].17 With the assistance of 1 3 C isotopic substitution, the anion [Sb2Fu] has been found to have symmetry. Of the expected nine IR-active and eleven Raman-active modes, seven modes in each group are observed. The very recent studies on 159 a cs o X B o O U a o a Ml CO e o on < N O N O oo o\ CN 00 O N O N CN O N O N CN o u 00 O N N O O N + CM o u c/j o r--00 cn N O O o CN CN CN CN — •w O N r-m O N CN CN CN CN CN + o u "of T f CN O O N CN -H 3 o CN, cn CN o CN O N O o O N CN N O N O N O N O Y-H *-H O N O N CN, CN T—H ' N O _ l N O t-» o CN CN O o CN CN CN CN CN cn • — V cn cn O N O N CN I i * — S ' o m sO P» O N 00 oo P-o O N O N CN CN CN /•—V N O y—\ cn t- N O •—1 O N CN CN T—1 CN 00 CN 00 CN i - H CN o O N CN CN o O N CN CN CN CN »—I N O N O o CN in CN O U co % O H S o co 6 U 5 -<u Q 8 a, 2 B Vi O Vi s « si •o 13 c o > * CO Vi I cx c/j S3 x> a e (U I N O CN f -CN "9 CN O N 3 CN cn 3 CN o CN P O N pq pd cf 160 [Hg(CO)2][Sb2Fn]2 1 5'1 6 and [Hg2(CO)2][Sb2Fn]2 1 5 , 1 6 show much more complicated infrared and Raman spectra for both compounds, indicating the symmetry of the anions in these compounds is lower than D^,. As can be seen in Table 5-2, the mutual exclusion rule is, in principle, still obeyed by the anion, [Sb2Fn]~, even though one infrared band does have a very weak counterpart in the Raman spectrum, and vice versa. Therefore, it can be concluded with reasonable certainty that the symmetry of the anion is essentially retained in [Ru(CO)6][Sb2Fn]2 and [Os(CO)6][Sb2F„]2. 5.3.2.2 Fe[SbF6]2 Both the infrared and Raman spectra of Fe[SbF6]2 have been reported by Holloway and co-workers.23 The infrared bands and Raman bands of Fe[SbF6]2 prepared in this study are at quite similar positions to those reported previously.23. Compared with Cs[SbF6],31 Fe[SbF6]2 exhibits more bands in the infrared and Raman spectra than expected for an exact octahedral anion (Oh point group). The spectra are closer to those of [XeF][SbF6] in which the [SbF6]" anions are found to have C4v symmetry.32 5.3.2.3 Ruthenium and Osmium Carbonyl Fluorosulfates The infrared spectra of ruthenium and osmium carbonyl fluorosulfates in the CO stretching region show no resemblance to each other; the IR spectrum of osmium carbonyl fluorosulfate is much more complicated than that of ruthenium carbonyl fluorosulfate. 161 In the CO stretching region of osmium carbonyl fluorosulfate, the band at 1880 cm"1 is very probably due to a bridging CO, since a bridging CO stretching vibration at such a high frequency has been noticed in the case of c-[Pd2(u.-CO)2](S03F)2. Other CO stretching vibrations (2150, 2122, 2102, and 2037 cm"1) are considered to arise from terminal CO groups. Compared to ruthenium carbonyl fluorosulfate and other dipositive metal carbonyls, all the CO stretching frequencies for the osmium compound are low. It is suspected that Os(III) in Os(S03F)3 may be reduced to an oxidation state of lower than +2 since a lower oxidation state would make rc-back-donation more favorable. Therefore, the enhanced n-back-donation would strengthen metal-carbon bond at the expense of CO bond. In contrast to the rather complicated CO stretching pattern for the osmium carbonyl fluorosulfate, there are only three sharp peaks at 2199, 2185, and 2154 cm 1 in the CO stretching region of ruthenium carbonyl. The CO stretching frequencies are higher than those of osmium compound, but significantly lower than those observed for /ner-Ir(CO)3(S03F)3 (see Chapter 4). It seems reasonable to assume that ruthenium(II) is produced by reductive carbonylation. The isolation of solid product and subsequent microanalysis should confirm this. However, attempts to isolate the solid compound have so far been unsuccessful. Based on the CO consumption of the reaction and the infrared spectrum of ruthenium carbonyl fluorosulfate, the ruthenium carbonyl fluorosulfate obtained here is tentatively assigned to a dimer species which has a local symmetry of C2v for each ruthenium atom. Further work on this complex system is needed. 162 5.3.3 1 3 C MAS NMR Spectroscopy 1 3 C MAS NMR spectra of isotopically enriched samples of both [Ru(CO)6][Sb2Fn]2 and [Os(CO)6][Sb2Fn]2 were recorded in order to investigate the bonding nature of the new cationic metal carbonyl compounds. The results are summarized in Table 5-3. As the oxidation state of the metal increases, rc-back-bonding appears to diminish. This is reflected in an increase of both v (CO)av and the stretching force constants ft, as well as a shift of the 1 3 C resonance to higher fields. As seen in Table 5-3, 1 3 C NMR chemical shifts are found at 166.1 and 147.3 ppm for ruthenium(II) and osmium(II) hexacarbonyl cations respectively. Compared with the value of free CO at 184 ppm, the chemical shifts of 1 3 C obtained for both compounds are very low, indicating a much reduced rc-back-bonding. This is consistent with the vibrational spectroscopic results for [Ru(CO)6][Sb2Fn]2 and [Os(CO)6][Sb2Fu]2. Similar observations have been made on a study of protonated metal carbonyls of ruthenium and osmium.338 The reported 1 3 C chemical shifts of 179 and 176.9 ppm for [Ru(CO)5H]+ and 159 ppm for [Os(CO)5H]+ correlate well with the results obtained here. In contrast, chemical shifts of 217 and 211 ppm are reported for the carbonyl anions [Nb(CO)6]" and [Ta(CO)6] , where rc-back-bonding is strengthened by the negative charge on the metal center. 5.4 Summary and Conclusions The synthesis of [Ru(CO)6][Sb2Fn]2 and [Os(CO)6][Sb2F„]2 was accomplished by the solvolysis of Ru(S03F)318 and Os(S03F)319 respectively in liquid antimonyfV) fluoride at 60-80 °C in a CO atmosphere. The existence of the dications [M(CO)6]2 + (M = Fe, Ru, or Os), 163 has long been anticipated, and there have been early claims of the identification of [Fe(CO)6]2+ and [Os(CO)6]2+ as [A1C14]" salts or as products of an amine-catalyzed disproportionation of Fe(CO)5.10 However, these early claims have been proven to be erroneous,11 and the counteranions ([A1C14]~ or [A1C14]") used are not capable of stabilizing the highly electrophilic carbonyl cations. Solvolysis reactions in SbF5 have been used initially to synthesize fluoroantimonatefV) salts containing main group cations, such as C10 2 + , 3 4 I 2 + 3 5 and B r 2 + , 3 5 and (CH 3 ) 2 Sn 2 + . 3 6 The new application of this synthetic approach in the presence of gaseous CO has also been adopted to the preparation of other metal carbonyl cations of late- or post-transition metals, for example, [Au(CO)2][Sb2F,,],17 [M(CO) 4][Sb 2F„] 2 (M = Pd or Pt),14 and [Hg(CO)2][Sb2Fn]2.16 The characterization of the new compounds is based on microanalysis, vibrational 13 spectroscopy and C MAS NMR spectroscopy. The non-coincidence of infrared and Raman bands in the CO stretching region points to the assignment of Oh symmetry to the two cations obtained in this study. The weighted average CO stretching vibrations (2216, and 2211 cm"1) found for the two cations are very high, and so are the calculated force constants for both of them (see Table 5-3). The/r(CO) values of 19.82 x 102 N m"1 for [Ru(CO)6][Sb2Fn]2 and 19.76 x 102 N m"1 for [Os(CO)6][Sb2Fn]2 are quite close to that of H C O + (21.3 x 102 N m"1), for which 7t-back-bonding is impossible. 1 3 C MAS NMR spectroscopic studies of the 1 3 C isotopomers of the two cations give the chemical shifts of 166.1 and 147.3 ppm for [Ru(CO)6][Sb2Fn]2 and [Os(CO)6][Sb2F,,]2 respectively. All of these spectroscopic features indicate that a much reduced rc-back-bonding is encountered in these cations. 164 The reductive carbonylation of Ru(S03F)3 and Os(S03F)3 in fluorosulfuric acid generates the corresponding ruthenium and osmium carbonyl fluorosulfates respectively. Further investigations, however, need to be carried out in order to isolate the new compounds and to fully characterize them. The attempted reductive carbonylation of FeCl3 in the presence of SbF5 did not produce the target compound [Fe(CO)6][Sb2Fu]2. The failure to obtain the expected iron carbonyl cation has recently resulted in the development of a different synthetic approach: Fe(CO)5 is oxidized by AsF5 in liquid SbF5 in a CO atmosphere to produce [Fe(CO)6]-[Sb 2F n] 2 . 3 7 The vibrational spectrum of [Fe(CO)6][Sb2Fn]2 is remarkably similar to that of [Ru(CO)6][Sb2Fn]2. Details will be published soon.37 165 References 1 Cotton, F. A. ; Wilkinson, G. Advanced Inorganic Chemistry; 5th ed.; Wiley: New York, 1988;p 868. 2 Mond, L. ; Hirtz, N. ; Cowap, M. D. Proc. Chem. Soc. 1910, 26, 27. 3 Mond, L. ; Hirtz, N. ; Cowap, M. D. J. Chem. Soc. 1910, 97, 798. 4 Cotton, F. A. and Wilkinson, G. Advanced Inorganic Chemistry; 5th ed.; Wiley: New York, 1988; pp 1023-1027. 5 Lee, J. D. Concise Inorganic Chemistry, 4th ed.; Chapman & Hall: London, 1991; p 764. 6 Rushman, P.; Buuren, G. N. ; Shirallan, M . ; Pomeroy, R. K. Organometallics, 1983, 2, 693. 7 Tripathi, S. C ; Srivastava, S. C ; Mani, R. P.; Shrimal, A. K. Inorg. Chim. Acta, 1975, 15, 249 8 Abel, E. W.; Tyfield, S. P. Adv. Organomet. Chem. 1970, 8, 137. 9 Hieber, W.; Kruck, T. Angew. Chem. 1961, 73, 580. 1 0 Sternberg, H. W.; Friedel, R. A. ; Shufler S. L. ; Wender, I. J. Am. Chem. Soc. 1955, 77, 2675. 1 1 Hieber, W.; Frey, V. ; John, P. Chem. Ber. 1967, 100, 1961. 1 2 Aubke, F.; Wang, C. Coord. Chem. Rev. 1994, 137, 483. 1 3 Abel, E. W.; Tyfield, S. P. Adv. Organomet. Chem. 1970, 8, 143. 1 4 Hwang, G. ; Wang, C ; Aubke, F.; Willner, H . ; Bodenbinder, M . Can. J. Chem. 1993, 71, 1532. 166 1 5 Willner, H . ; Bodenbinder, M . ; Wang, C ; Aubke, F. J. Chern. Soc, Chern. Commun. 1994, 1189. 1 6 Bodendinder, M . ; Balzer-Jollenbeck, G.; Willner. H . ; Batchelor, R.; Einstein, F. W. B.; Wang, C ; Aubke, F. Inorg. Chern. 1996, 35, 82. 1 7 Willner, H . ; Schaebs, J.; Hwang, G. ; Mistry, F.; Jones, R.; Trotter, J.; Aubke, F. J. Am. Chern. Soc. 1992, 114, 8972. 1 8 Leung, P. C ; Aubke, F. Can. J. Chern. 1984, 62, 2892. 1 9 Leung, P. C ; Wong, G. B.; Aubke, F. / . Fluorine Chern. 1987, 35, 607. 2 0 Wang, C ; Willner. H . ; Bodendinder, M . ; Batchelor, R.; Einstein, F. W. B.; Aubke, F. Inorg. Chern. 1994, 33, 3521. 2 1 Willner, H . ; Aubke, F. Inorg. Chern. 1990, 29, 2195. 2 2 Hwang, G.; Wang, C ; Bodenbinder, M . ; Willner, H . ; Aubke, F. J. Fluorine Chern. 1994, 66, 159. 2 3 Gantar, D. ; Leban, I.; Frlec. B.; Holloway, J. H. J. Chern. Soc, Dalton Trans. 1987, 2379. 2 4 Braterman, P. S. Metal Carbonyl Spectra; Academic Press: New York, 1975. 2 5 Kettle, S. F. A. Top. Curr. Chern. 1977, 77, 111. 2 6 Abel, E. W.; Mclean, R. A. N. ; Tyfield, S. P.; Braterman, P. S.; Walker, A. P.; Hendra, P. J. J. Mol. Spectrosc. 1969, 30, 29. 2 7 Jones, L. H . ; McDowell, R. S.; Goldblatt, M. Inorg. Chern. 1969, 8, 2349. 2 8 Gillespie, R. J.; Landa, B.; Inorg. Chern. 1973, 12, 1383. 2 9 Bonnet, B.; Mascherpa, G.; Inorg. Chern. 1980, 19, 785. 3 0 Craig, N. C ; Fleming, G. F.; Pranata, J. J. Am. Chern. Soc. 1985, 107, 7324. 167 3 1 Birchall, T. ; Dean, P. A.W.; Gillespie, R. J. J. Chem. Soc A, 1971, 1777. 3 2 Frlec, B.; Holloway, J. H. J. Chem. Soc, Dalton Trans. 1975, 535. 3 3 (a) Brewer, S. A. ; Holloway, J. H . ; Hope, E. G. J. Fluorine Chem. 1995, 70, 167. (b) Ellis, J. E. Adv. Organomet. Chem. 1990, 31, 1. 3 4 Yeats, P. A. ; Aubke, F. J. Fluorine Chem. 1974, 4, 343. 3 5 Wilson, W. W.; Thompson, R. C.; Aubke, F. Inorg. Chem. 1980, 19, 1489. 3 6 Mallela, S. P.; Yap, S.; Sams, J. R.; Aubke, F. Rev. Chim. Miner. 1986, 23, 572. 3 7 Bley, B.; Willner, H . ; Aubke, F. To be published. 168 CHAPTER 6 SOLVOLYSIS REACTIONS IN ANTIMONY PENTAFLUORIDE - Improved Synthetic Routes and Methods of Generating New Metal Carbonyl Derivatives 6.1 Introduction Since its discovery by Ruff and Plato1 antimony pentafluoride, SbF5, has found wide use as a chemical reagent. It is regarded as the strongest molecular Lewis acid,2'3 and has become a component of two important conjugate superacid systems,3'4 magic acid, HS0 3F-SbF 5 3' 4 and HF-SbF 5, 3 ' 4 which appear to be the strongest protonic superacids.3 A large 5 3 4 number of carbocations and a wide range of inorganic cations ' have been generated in these highly ionizing media. Many of these cations are obtained in conjunction with fluoroantimonate(V) anions such as [SbF6]", [Sb2Fn]~, and [Sb3Fi6]".4 SbF5 found its application in metal carbonyl chemistry when a linear metal carbonyl cation, [Au(CO)2]+, was first isolated as a thermally stable compound, [Au(CO)2][Sb2Fu].6a However, as has already been mentioned in the introduction of this thesis, the synthesis of cationic metal carbonyl derivatives often involves other corrosive and hazardous chemicals such as HS0 3F and S 20 6F 2 , and some of these chemicals are not commercially available. In addition, the preparation of metal carbonyl cations requires a two-step synthesis: (i) synthesis of a metal carbonyl fluorosulfate, and (ii) a solvolysis reaction in SbF5 in the presence of additional gaseous CO. Therefore, the development of a simplified synthetic approach to the 169 generation of cationic metal carbonyl derivatives is one of the goals of this research. As a substitute for metal carbonyl fluorosulfate, metal fluorosulfates or metal halides were chosen as precursors in a one-step solvolysis reaction in antimony pentafluoride, SbF5, in the presence of CO. Solvolysis reactions in the absence of additional gaseous CO were also attempted. The latter led to the discovery of several new compounds such as [Pd2(p>CO)2]-[Sb 2Fu] 2, cw-Pd(CO)2[Sb2Fn]2, and c«-Pt(CO) 2[Sb 2F n] 2 . Their characterization and spectroscopic features will also be discussed in this chapter. 6.2 Experimental Some of the experimental work (solvolysis reaction in SbF5 using metal chloride as a precursor) described here was done in collaboration with Sun Chau Siu.6 b Germaine Hwang carried out the solvolysis reaction of gold(III) fluorosulfate. The solvolysis reaction of mercury fluorosulfate was investigated by Professor Willner's group, and their results will be summarized in the discussion section. 6.2.1 Solvolysis Reactions in SbF5 in the Presence of CO 6.2.1.1 The Synthesis of [Pd(CO)4][Sb2Fn]2 from Pd[Pd(S03F)6] Palladium metal powder (103.8 mg, 0.9699 mmol) was used to prepare Pd(S03F)3 as previously reported by Lee and Aubke.7 After an almost quantitative yield of palladium fluorosulfate, Pd[Pd(S03F)6], was obtained, CO was introduced into the 100 mL one-part reactor, which was immersed in liquid nitrogen. About 500 mbar of CO pressure was applied to the reactor at liquid N 2 temperature. Then, the reactor was allowed to warm to room temperature, at which the pressure inside the reactor was approximately equal to 2 bar. A hot 170 water bath at 70-80 °C was used to reduce the viscosity of SbF5 and facilitate the magnetic stirring. White circles of solid products formed at the interface between CO gas and the red-brown suspension shortly after the bath temperature reached 80 °C. Within a few hours, the original red-brown suspension changed to an almost white suspension. The reaction was continued for another two days until a uniformly white suspension was obtained. The reactor was then cooled down, and gas phase products were collected at different temperatures and analyzed by IR spectroscopy. In addition to unreacted CO, the presence of C 0 2 and S 20 5F 2 was confirmed by an IR spectrum. The excess SbF5, together with the volatile by-products, was removed in vacuo first at room temperature overnight and then at 60 °C for another 4 hours. A white powdery product was obtained. Elemental analysis of the product gave the composition of [Pd(CO)4][Sb2Fn]2 (Analysis: Calculated for C 4F 2 20 4PdSb 4: C, 4.27%. Found: C, 4.22%). In a sealed capillary tube, the white compound turned orange at 140 °C and melted at 156 °C with gas evolution. 6.2.1.2 The Synthesis of [Pd(CO)4][Sb2Fn]2 from PdCl 2 In a glove box, 140.8 mg (0.79 mmol) of PdCl2 (a reddish-brown powder) was added to a one-part, 100-mL round bottom flask and dried in vacuo at 50 °C for 1 hour. After the addition of an excess of SbF5 by vacuum distillation, a CO pressure of approximately 500 mbar was admitted to the reactor with the flask cooled at liquid N 2 temperature. The reactor was sealed off and heated at 70-80 °C. After 1 week of reaction white solids appeared, but trace amounts of black particles remained. The mixture was allowed to react for another 2 weeks until a white suspension was obtained. Removal of all volatiles in vacuo at 60 °C left 171 behind a white solid product. The identification of the compound was confirmed by IR and Raman spectroscopy. 6.2.1.3 The Synthesis of [Pt(CO)4][Sb2FU]2 from Pt(S03F)4 Platinum tetrakis(fluorosulfate), Pt(S03F)4 (227.1 mg, 0.384 mmol), was prepared by the oxidation of platinum metal powder by S 2 0 6 F 2 in HS0 3F as described by Lee and Aubke.8 About 3 mL of SbF5 was added by vacuum distillation. Excess CO (6 mmol) was introduced to the reactor by cooling it to liquid N 2 temperature. A water bath at 60-80 °C was used to facilitate magnetic stirring. White products at the interface of gas and liquid were noticed within 2 hours after the reaction started. An almost white suspension was obtained in 12 hours. To ensure a complete reaction, the reactor was kept in the hot water bath under vigorous stirring for another 2 days when a homogeneous white suspension formed. The total CO uptake was 1.98 mmol, or 5.15 moles of CO per mole of Pt(S03F)4. The gas phase was found to contain CO, C0 2 , and S 2 0 5 F 2 by recording IR spectra at different temperatures. Removing all the volatiles at room temperature and at 60 °C produced a quantitative yield of [Pt(C0)4][Sb2Fn]2, confirmed by an IR spectrum. The white compound, [Pt(C0)4][Sb2Fu]2, melted to a brown liquid at 200 °C with gas evolution. 6.2.1.4 The Synthesis of [Au(CO)2][Sb2Fu] from Au(S03F)3 Au(S03F)3 was prepared by oxidation of the metal powder by S 2 0 6 F 2 in HS0 3F as previously described.9 About 3 mL of SbF5 was distilled onto 224 mg (0.494 mmol) of Au(S03F)3 in vacuo. The 50 mL round-bottom flask was warmed to room temperature. No reaction between the orange Au(S03F)3 and the clear, colorless SbF5 was observed. Carbon 172 monoxide (4.85 mmol) was admitted to the reaction vessel. The mixture was warmed to 65 °C to facilitate stirring. The solid present changed color from orange to brown after 10 minutes. White particles were visible in the mixture after 30 minutes. After 45 minutes, the solution turned dark yellow. The solid was gradually consumed at this stage. The solution was clear and colorless after 2 hours and no solid appeared to be present. Heating was stopped and the solution was left stirring for 20 hours. At ambient temperature the solution separated into a white, cloudy, viscous upper layer and a light brown, oily lower layer. Warming the mixture to 80 °C caused the two layers to combine to form a clear, colorless solution, but the brown oil returned upon cooling the reactor to room temperature. Infrared spectra were obtained of the volatile materials in the reactor at -196, -78, and 20.5 °C. The volatiles were found to be CO, C0 2 , S 20 5F 2 , and possibly Sb 2F 9S0 3F. The excess SbF5 was removed in vacuo, leaving a white solid. Approximately 3.18 mmol of CO was removed from the reactor. In total, 1.67 mmol CO appeared to have reacted with 0.0494 mmol of Au(S03F)3. The white solid was identified as [Au(CO)2][Sb2Fn]6a by its vibrational spectra. 6.2.1.5 The Synthesis of [Au(CO)2][Sb2Fn] from AuCl 3 To 100.3 mg (0.33 mmol) of finely ground, yellow-orange, crystalline AuCl 3, contained in a 100 mL round-bottom flask, 3 mL of SbF5 was added by vacuum distillation. A CO pressure of ca. 500 mbar was admitted to the reactor when it was still in liquid nitrogen bath. Upon warming the reaction mixture first to room temperature and then to 60-70 °C, the solids immediately turned black and then slowly changed to a white suspension. After heating for 5 days with magnetic stirring, no further color change or CO uptake was noted, and the 173 reaction was judged to be complete. The excess of SbF5 was removed in vacuo overnight, with the reactor at 55 °C. In the highly volatile gaseous fraction only CO and COF 2 could be detected by IR spectra. The white solid product was identified as [Au(CO)2][Sb2Fn]6a by its vibrational spectra. 6.2.2 Solvolysis Reactions in the Absence of Additional Gaseous CO 6.2.2.1 The Synthesis of [Pd2(n-CO)2][Sb2Fn]2 Orange crystalline [Pd2(p>CO)2](S03F)2 (27.2 mg) was finely ground and transferred into a 50 mL round-bottom reactor. Approximately 2 mL of SbF5 was added to the reactor by vacuum distillation. The reaction was allowed to proceed at room temperature without magnetic stirring because the extreme viscosity of SbF5 made effective magnetic stirring difficult at this temperature. Instead of magnetic stirring, mixing of the reagents was effected by swirling the mixture manually from time to time.. No noticeable physical change was observed during the first few weeks of the reaction. Over time, however, a slight color change was noticed, and the suspension became less viscous. After six months, a uniform bright yellow suspension was obtained while gently swirling the reactor. All the volatiles were then removed by vacuum distillation at room temperature, and a fine yellow powdery product was obtained. The weight of the product was found to be 67.3 mg by weight difference compared with 68.4 mg, expected of a quantitative yield of [Pd2(u.-CO)2j[Sb2Fn]2. Elemental analysis gave the composition of [Pd2(p>CO)2]rSb2Fn]2 (Calculated for CF„OPdSb2: C, 2.05%. Found: C, 2.05%). When heated in a capillary tube, [Pd2(p:-CO)2]-[Sb 2Fn] 2 did not show a distinct melting point. At 115 °C, the compound started to 174 decompose with noticeable gray spots, and the yellow compound became uniformly dark gray at 145 °C. 6.2.2.2 The Synthesis of as-Pt(C0)2(Sb2Fn)2 Solid Pt(S03F)48 (67 mg , 0.11 mmol) was dissolved in 3 mL of HS0 3F, and CO was introduced afterwards. The reductive carbonylation to prepare cw-Pt(CO)2(S03F)2 was carried out according to a previously reported procedure.10 When the reductive carbonylation was complete, most of the solvent, HS0 3F, was pumped off, and 3 mL of SbF5 was added by vacuum distillation. A warm water bath of 50 °C was used to facilitate magnetic stirring. A white cloudiness was noticed within 2 to 3 hours. The reaction was allowed to continue at 50 °C under vigorous stirring for 2 days. Some tiny crystals formed at the bottom of the reactor, after which the stirring speed was reduced. The crystals grew bigger after another 2 days. The volatiles were removed in vacuo at room temperature overnight, and then at 40 °C for another 1 hour. A few white crystals and some colorless waxy precipitate were obtained. An IR spectrum of the waxy precipitate, and a Raman spectrum of the crystals were recorded. In the CO stretching region, two strong peaks at 2230 and 2202 cm'1 were observed in the IR spectrum. Their counterparts (2236 and 2202 cm"1) were also found in the Raman spectrum of the crystalline product. 6.2.2.3 The Attempted Synthesis of <rw-Pd(C0)2(Sb2Fn)2 About 3 mL of SbF5 was distilled onto 0.70 mmol of cw-Pd(CO)2(S03F)2, which was prepared by a gas-solid reaction described in Chapter 3. A hot water bath at 70 °C was used to facilitate the magnetic stirring. Some bubbles were noticed shortly after the yellow 175 powdery cw-Pd(CO)2(S03F)2 was gradually mixed with SbF5. After 1 week, a homogeneous suspension was obtained. The color of the suspension did not change too much during the whole process of the reaction. The reaction was then stopped, and infrared spectra were recorded of the volatile materials in the reactor at -196, -78, and 20.5 °C. CO, C0 2 , and S 2 0 5 F 2 were found to be present among the volatiles. The excess SbF5 was removed in vacuo, leaving a yellow powdery solid. IR and Raman spectra were taken of the final product, and showed almost identical band positions and intensities with the newly obtained compound, [Pd2(p>CO)2][Sb2Fn]2. Elemental analysis also confirmed the composition (Calculated for CF„OPdSb2: C, 2.05%. Found: C, 2.12%). In order to prevent the apparent reduction of palladiumfll) to palladium(I), the solvolysis reaction was carried out at 50 °C. The physical appearance of the reaction did not show much difference from that observed when the reaction was carried out at 70 °C. The gas phase infrared spectrum at different temperatures also indicated the presence of CO, C0 2 , and S 20 5F 2 . However, the IR spectrum of the final product was more complicated than the spectrum of [Pd2(p>CO)2][Sb2Fn]2, obtained from the previous preparation at 70 °C. In addition to the bridging CO stretching band of [Pd2(p.-CO)2][Sb2Fn]2 at 2006 cm"1, two relatively strong bands at 2232 and 2216 cm"1 were also observed, which were considered to arise from cw-Pd(CO)2(Sb2F,,)2. In a subsequent reaction, the temperature of the water bath was further lowered. At 40 °C, the magnetic stirring was not very effective, and this was compensated by gently swirling of the reactor manually from time to time. It seemed that the initial yellow suspension 176 became very pale after 1 week of reaction. At -196 °C, the gas phase pressure was 0 mbar. IR spectra of the volatile materials were taken with the reactor being kept at -78 and 20.5 °C. C 0 2 and S 2 0 5 F 2 were found to be present among the volatiles. However, no CO was found either by the gas phase measurement or IR spectroscopy. The IR and Raman spectra indicated that the final product was a mixture, which consisted of the two known compounds, [Pd(CO)4][Sb2Fn]2 and [Pd2(u-CO)2][Sb2F„]2. 6.3 Results and Discussion 6.3.1 Solvolysis Reactions in SbFs in the Presence of CO - Improved Synthetic Routes to Metal Carbonyl Cations As can be seen from the preceding sections, all of the known metal carbonyl fluoroantimonate salts, [Au(CO)2][Sb2Fn]2,6a [Pd(CO)4][Sb2F„]2, and [Pt(CO)4][Sb2Fn]2,n can be produced by a one-step synthesis either from metal fluorosulfates or metal chlorides. The advantage of this direct synthesis over the previous two-step synthesis is apparent as can be seen from the following comparison. The synthesis of a metal carbonyl cation involving metal carbonyl fluorosulfates as precursors is often cumbersome and experimentally demanding. This is perhaps best illustrated by the original synthesis of [Au(CO)2][Sb2Fn]6a via Au(CO)S03F.1 2 Starting with the synthesis of gold(III) fluorosulfate9 according to 2Au + 3S 20 6F 2 H S O i F ' 6 0 ° c > 2Au(S03F)3 (6-1) 177 both the purification of the solvent, HS0 3 F 1 3 and the synthesis of S 2 0 6 F 2 (see Chapter 2) by the catalytic fluorination of sulfur trioxide with elemental fluorine are demanding procedures which require specialized equipment. The subsequent reductive carbonylation of Au(S03F)3 in HS0 3F is straightforward, facile, and completed within a few hours according to Au(S03F)3 + 3CO H S O i F >[Au(CO)2]+(golv) + C 0 2 + S 2 0 5 F 2 + S03F" (6-2) Nucleophilic displacement of one mole CO by S03F" on heating or on solvent removal according to [Au(CO) 2] + ( s o l v ) + S03F" -> Au(CO)S03F + C O ^ (6-3) will produce Au(CO)S03F in quantitative yield,12 but the pure compound is only obtained after sublimation. The final product, [Au(CO)2][Sb2Fu], is made by a solvolysis reaction in SbF5 in a CO atmosphere of 580 mbar according to Au(CO)S03F + CO + 4SbF5 H S O i F ' 8 0 ° c > [Au(CO)2j[Sb2Fu] + Sb 2F 9S0 3F (6-4) Two additional complications are encountered in the reductive carbonylations of Pt(S03F)48 and the mixed valency compound Pd[Pd(S03F)6].7 The rather slow reduction of Pt(S03F)4 in HS0 3F by CO produces a yellow solid intermediate of the composition of [Pt(CO)4][Pt(S03F)6], which has been isolated and characterized.14 This intermediate will convert completely under slightly more forcing conditions to white cw-Pt(CO)2(S03F)2 in HS0 3F in a CO atmosphere. The reductive carbonylation of Pd[Pd(S03F)6]7 in HS0 3F produces, in addition to cw-Pd(C0)2(S03F)2, a second isomer with bridging CO groups according to the IR spectrum (see Chapter 3) which has not been isolated as a pure compound. 178 However, both forms of Pd(CO)2(S03F)2 can be converted to [Pd(CO)4][Sb2Fn]2 in the presence of CO by solvolysis reactions in SbF5. These complications, the use of HS0 3F as a solvent and the necessity to isolate the carbonyl fluorosulfates as intermediates, are all avoided when metal fluorosulfates are used in a one-step solvolysis reaction in liquid SbF5 in a CO atmosphere. In addition, the direct solvolysis seems to proceed with comparative ease using a two-step synthesis via a metal carbonyl fluorosulfate. The whole reductive carbonylation usually takes 1 to 2 days under mild conditions, for example, Pd[Pd(S03F)6] + 9CO + 16SbF5 s 6 F s ' 7 0 - 8 0 ° c > 2 [Pd(CO)4][Sb2Fu]2 + C0 2 + S 2 O s F 2 + 4Sb2F9S03F (6-5) and Au(S03F)3 + 3CO + 4SbF5 S b F s ' 6 5 ° c ) [Au(CO)2][Sb2Fn] + C 0 2 + S 2 0 5 F 2 + Sb 2F 9S0 3F (6-6) The relatively low volatility of SbF5 and perhaps some trapped SbF5 in the solid product requires prolonged heating of the product mixture to ca. 65 °C in vacuo during the product isolation, but this causes no problems because [Au(CO)2][Sb2Fn]6a as well as [Pd(CO)4][Sb2Fn]2 and [Pt(CO)4][Sb2F11]211 show sufficient thermal stability at temperatures well above 100 °C. Although the simplified procedure was designed as an improved synthetic approach to prepare the known compounds such as [Au(CO)2][Sb2Fu],6a [Pd(CO)4][Sb2Fn]2, and [Pt(CO)4][Sb2Fn]2,n it has found two recent applications which have significantly expanded 179 the field of homoleptic metal carbonyl cations: (i) The solvolysis of Hg(S03F)2 in liquid SbF5 in a CO atmosphere has resulted in the isolation16 and subsequent structural characterization17 of [Hg(CO)2][Sb2Fu]2, the first thermally stable carbonyl derivative of a post-transition metal. The reaction proceeds according to Hg(S03F)2 + 2CO + 8SbF5 -> rHg(CO)2][Sb2Fn]2 + 2Sb2F9S03F (6-7) (ii) Reductive carbonylation of M(S03F)3, M = Ru 1 8 or Os, 1 9 in liquid SbF5 has allowed for the first time the isolation of a dipositive metal carbonyl cation [M(CO)6][Sb2Fn]2, M = Ru or Os (see Chapter 5). In the first case, carbonylation of Ffg(S03F)2 in HS0 3F has been unsuccessful, while in the case of ruthenium and osmium, metal carbonyl fluorosulfates have not yet been isolated and completely characterized. Therefore, in addition to being an improved synthetic route to the preparation of the known compounds, the one-step solvolysis reactions also resulted in the generation of three new metal carbonyl cations. These direct solvolysis reactions discussed above still require metal fluorosulfates as precursors. The preparation of binary metal fluorosulfates involves metal oxidation by S 20 6F 2 in HS0 3F as the reaction medium.7"9'15'18'19 Hence neither the use of HS0 3F nor the difficulties encountered in the synthesis of S 2 0 6 F 2 are completely avoided. Further investigation into other alternatives is desirable. Metal chlorides become new candidates for the development of a more simplified synthetic route because they are more readily available and their preparation does not involve S 2 0 6 F 2 or HS0 3F. A related precedent is seen in the reported solvolysis conversion of IC13 into ICl2[SbF6].20 The solvolysis reaction of AuCl3 was expected to proceed according to 180 AuCl 3 + 3C0 + 3SbF5 s b F s ' 6 5 ° c ) [Au(CO)2][Sb2Fn] + C0C12 + SbF4Cl (6-8) [Au(CO)2][Sb2Fn] is indeed formed and identified by comparison to an authentic sample.68 The mixed fluoride-chloride by-product formulated as SbF4Cl is, unlike Sb 2F 9S0 3F, 2 1 not fully characterized and corresponds in its composition to materials encountered during the original synthesis of SbF5 from SbCl5 and HF by Ruff and Plato22 and later described in more detail by Ruff23. The expected by-product COCl 2 , however, is not observed in a gas phase IR spectrum. Instead, carbonyl fluoride (COF2) is found to be a by-product by the IR spectrum. This is somewhat surprising since COCl 2 is usually obtained in the reductive carbonylation of AuCl3 to produce Au(CO)Cl. 2 4 , 2 5 ' 2 6 There are two pathways which can lead to the formation of carbonyl fluoride: (i) Since COF 2 can be prepared from COCl 2 (phosgene) and a mixture of Sb(III) and Sb(V) fluorides,27 it is possible that the initially formed phosgene is subsequently converted to carbonyl fluoride by reaction with SbF5, which is present in a vast excess, (ii) Alternatively, the initial conversion of the starting material AuCl 3 to AuF3 by SbF5 is possible. The synthesis of [Au(CO)2][Sb2Fn] from AuF3 and CO in the presence of liquid SbF5 is known to proceed via a black intermediate and to produce COF 2 as a by-product.68 The nature of this black intermediate is still unclear; nevertheless, the disappearance of the black color as the reaction progresses provides an indication of the completeness of the reaction - a uniformly white suspension is formed when the conversion has gone to completion. Black-brown intermediates are observed in the solvolysis reactions of PdCl2 and PtCl2, which proceed according to 181 MC12 + 4C0 + 6SbF5 s b F s ' 6 5 ° c ) [M(CO)4][Sb2Fn]2 + 2SbF4Cl (M = Pd or Pt) (6-9) In this case, the black-brown particles form as soon as the metal halide powders come in contact with SbF5 even before CO is added, and it takes several days before sizable quantities of white product form. Attempts to characterize the black intermediates by vibrational spectroscopy were unsuccessful. The opaque powders did not scatter Raman light very well and gave very poorly resolved IR spectra. As in the case of the solvolyses of AuCl 3 and AuF 3 , 6 a the gradual disappearance of the black intermediates allows monitoring of the course of the reaction. Since all carbonylation reactions of metal halides in liquid SbF5 appear to be heterogeneous processes, where starting materials, intermediates and final products are suspended rather than completely dissolved in liquid SbF5, it is necessary to react small amounts at a time, to use a large excess of SbF5, and to grind the solid metal chlorides very finely inside a glove box before use. As noted in the preceding discussion, the high viscosity of SbF5 causes problems, especially if a reaction has to be carried out at relatively low temperatures. This can be partly overcome in the case of solvolysis reactions that involve fluorosulfates since the by-product, Sb 2F 9S0 3F, 2 1 formed in solvolysis reaction, is found to reduce the viscosity of antimony(V) fluoride28 and hence facilitate the reaction. It is noticeable that the solvolysis of metal chlorides in SbF5 in a CO atmosphere proceeds very slowly for reactions where the oxidation state of the metal does not change as in the conversion of MC12, M = Pd or Pt, into [M(CO)4]2 + salts. Since in MC12 the divalent metal ion is in a square planar environment in both a- and P-forms,29 the conversion involves 182 square planar substitution, seemingly on solids. Thus the long reaction times are not surprising. The remaining four synthetic reactions discussed here involve reductive carbonylation and solvolysis. All are considerably faster and are suggested to proceed via initially formed "naked" metal ions as intermediates, which are strong Lewis acids that are stabilized by CO coordination to produce [Au(CO)2][Sb2Fn]6a from either AuCl 3 or Au(S03F)39 or [M(CO)4][Sb2Fn]2, M = Pd or Pt, from either Pd[Pd(S03F)6]7 or Pt(S03F)4.8 Therefore the advantage of using commercially available metal chlorides is in part offset by rather long reaction times and the necessity to react small quantities at a time as discussed above. 6.3.2 Solvolysis Reactions in SbF5 in the Absence of Gaseous CO While most solvolysis reactions in SbF5 in a CO atmosphere lead to previously known metal carbonyl compounds, the solvolysis reactions in SbF5 in the absence of gaseous CO have the potential of generating new metal carbonyl derivatives. By using c-[Pd2(u.-CO)2](S03F)2 as a precursor, the solvolysis reaction in the absence of additional CO produces the new palladium© carbonyl compound, c-[Pd2(p>CO)2]-[Sb 2Fn] 2. The objectives of pursuing this synthesis have been two-fold: (i) to identify the two missing vibrations of the cation [Pd2(u\-CO)2]2+ in c-[Pd2(p>CO)2](S03F)2; and (ii) to study the anion influence on the CO stretching vibrations. In order to avoid oxidation of palladium(I) by SbF5 and other possible side reactions, the solvolysis reaction is allowed to proceed at room temperature according to 183 c-[Pd2(^CO)2](S03F)2 + 8SbF5 s b F $ ' 2 5 ° c •> c-[Pd2(u-CO)2][Sb2Fn]2 + 2Sb2F9S03F (6-10) The molecular structure of c-[Pd2(p>CO)2](S03F)2 have been described in Chapter 3. However, in an attempt to provide a complete vibrational analysis for the cyclic [Pd2(|i-CO) 2 ] 2 + cation, which has D2h symmetry, two Raman active fundamentals, v5 and v6, have remained undetected. They are expected to be weak and to occur near 600 cm"1, and they may even be accidentally degenerate. Unfortunately, for [Pd2(p>CO)2][Sb2Fu]2 all bands (see Table 6-1) between 500 and 700 cm"1 are reasonably intense, both IR and Raman active, and are best attributed to Sb-F stretching vibrations according to precedents.6"'11,17 A clearer picture is found in the CO stretching region where the mutual exclusion principle is obeyed. In the IR spectrum the asymmetric CO stretching vibration v7(Biu) is shifted from 1977 cm"1 with a 1 3 C satellite band at 1940 cm"1 for the fluorosulfate to 2006 cm'1 with a 1 3 C satellite band at 1970 cm"1 for [Pd2(u-CO)2]rSb2Fn]2. The Raman active v^Ag) stretching vibration shifts from 2027 to 2048 cm"1 for the pair of compounds. Hence v(CO)a v has increased from 2002 cm"1 for [Pd2(u-CO)2](S03F)2 to 2027 cm"1 for [Pd2(p> C0>2][Sb2Fii]2. Both values are well outside the range of v(CO) between 1850 and 1700 cm"1 commonly observed for bidentate, symmetrically bridging CO groups in metal carbonyl , 30,31,32 compounds. The non-coincidence of IR and Raman active modes is consistent with the retention of D2h point symmetry for the [Pd2(p>CO)2]-metallo-cycle in both compounds. As in [Pd2(p> 184 Table 6-1 Vibrational Data for Solid c-fPd^-CCOJtSb^nk IR (cm1) Raman (cm1) Approximate Assignment 2048 vs v(12CO), A l g 2006 vs v(12CO), B l u 1970 vw v(13CO), B l u 734 vw 721 vs v(Sb-F) 708 vs v(Sb-F) 696 vs v(Sb-F) 682 vs 681 s v(Sb-F) 648 vs 643 vs v(Sb-F) 569 vs 571 m v(Sb-F) 492 m 409 m v(Pd2(CO)2), A l g 390 m 296 w, b 248 w 228 w 135 w 82 w Abbreviations: vs = vw = very weak, b very strong, s = strong, m = broad. = medium, w = weak, CO)2](S03F)2, where according to the molecular structure, bidentate, weakly bridging fluorosulfate groups are present, no vibrational coupling between individual cations is apparent in the v(CO) region. This is in contrast to the observations previously reported for 33 34 35 36 acetate derivatives ' and halide derivatives ' where vibrational coupling between [Pd2(p> CO) 2 ] 2 + units have been observed on account of relatively strong cation-anion interactions in these compounds, as discussed in detail in Chapter 3. 185 In contrast to the clean synthesis of [Pd2(p>CO)2][Sb2Fn]2 from the fluorosulfate precursor, [Pd2(u.-CO)J(S03F)2, the solvolysis reaction of cw-Pd(CO)2(S03F)2 in SbF5 in the absence of CO to produce c/5-Pd(CO)2(Sb2Fn)2 does not proceed readily. At 70 °C, the solvolysis reaction leads to a clean synthesis of [Pd2(n-CO)2][Sb2Fn]2 according to: 2cw-Pd(CO)2(S03F)2 + 8SbF5 s b F s t 7 0 ° c ) [Pd2(u.-CO)2][Sb2Fu]2 + CO + C 0 2 + S 2 0 5 F 2 + 2Sb2F9S03F (6-11) The gas phase infrared spectra acquired with the reactor at different temperatures show the presence of the gaseous by-products such as CO, C0 2 , and S 2 0 5 F 2 . The identity of the yellow solid compound has been confirmed by microanalysis, IR, and Raman spectra. Therefore, palladium(II) has apparently been reduced to palladium(I) during the solvolysis. Thus, this reaction corresponds to a decomposition of Pd(CO)2(S03F)2 in HS0 3F at room temperature, where single crystals of [Pd2(p.-CO)2](S03F)2 form according to 2Pd(CO)2(S03F)2 H S O l F ' 2 5 ° c > [Pd2(^-CO)2](S03F)2 + CO + C 0 2 + S 2 0 5 F 2 (6-12) At lower temperatures (50 °C) a mixture of cw-Pd(CO)2(Sb2Fu)2 and c-[Pd2(p> CO) 2][Sb 2F„] 2 forms after removal of all the volatile materials. Three bands at 2232, 2216, and 2007 cm"1 are found in the CO stretching region in the IR spectrum. The bridging CO stretching band of c-[Pd2(|i-CO)2][Sb2F,1]2 at 2007 cm"1 is easily recognizable, and the other two relatively strong bands at 2232 and 2216 cm"1 must have arisen from cis-Pd(CO)2(Sb2Fn)2 since the solvolysis reaction seems to have gone to completion as evidenced 186 by the IR spectrum. Therefore, two reactions appear to proceed separately according to Equations 6-11 and 6-13 cw-Pd(CO)2(S03F)2 + 8SbF5 sbFs' 5 0 °c > ri*-Pd(CO)2(Sb2Fn)2 + 2Sb2F9S03F (6-13) The target compound, cw-Pd(CO)2(Sb2Fn)2 does form as a major product via the solvolysis reaction in the absence of gaseous CO. However, a decomposition product, c-[Pd2(p>CO)2][Sb2Fn]2, is also present in the final product mixture. Separation of the mixture has been unsuccessful because a suitable solvent for recrystallization could not be found. Nevertheless, the IR spectrum of the mixture provides enough information to deduce the geometry of the metal carbonyl and to determine the average CO stretching frequency for cis-Pd(CO)2(Sb2Fn)2. Both the Aj mode (at 2232 cm"1) and the Bj mode (at 2216 cm"1) are shifted slightly to higher wavenumbers compared to the corresponding vibrations for cis-Pd(CO)2(S03F)2, which are found at 2228 and 2208 cm"1 respectively. The average CO stretching vibration is 2224 cm"1, 6 wavenumbers higher than that found for cis-Pd(CO)2(S03F)2. It was thought that at even lower temperatures the yield of cw-Pd(CO)2(Sb2Fu)2 might increase or that the pure compound could be obtained. Surprisingly, [Pd(CO)4][Sb2Fn]2 forms in addition to [Pd2(p>CO)2][Sb2Fu]2 when the temperature of the water bath is lowered to 40 °C, and no cw-Pd(CO)2(Sb2Fn)2 forms at all. It is obvious that the CO liberated from the decomposition reaction (Equation 6-11) combines with the remaining palladium(H) according to 187 a?-Pd(CO)2(S03F)2 + 2C0 + 8SbF5 S b F $ ! 4 0 ° c > [Pd(CO)4][Sb2Fn]2 + 2Sb2F9S03F (6-14) The gas phase IR spectrum and pressure measurements indicate that the overall solvolysis reaction proceeds according to 5c/>Pd(CO)2(S03F)2 + 24SbF5 S b F > ) 2[Pd2(n-CO)2][Sb2FM]2 + [Pd(CO)4][Sb2Fn]2 4- 2C0 2 + 2S 20 5F 2 + 6Sb2F9S03F (6-15) No attempts have been made to separate the apparent mixture since both compounds, [Pd20i-CO)2][Sb2Fn]2 and [Pd(CO)4][Sb2Fn]2 are known and have been characterized by vibrational spectroscopy. The solvolysis reaction of cw-Pt(CO)2(S03F)2 in SbF5 in the absence of CO was carried out under similar conditions to those of cw-Pd(CO)2(S03F)2, and the reaction produced the target compound according to cw-Pt(CO)2(S03F)2 + 8SbF5 s b F s ' 5 0 ° c > c«-Pt(CO) 2(Sb 2F n) 2 + 2Sb2F9S03F (6-16) Colorless crystals of cw-Pt(CO)2(Sb2Fn)2 were obtained as a final product of the solvolysis reaction. Attempts to determine its molecular structure have been unsuccessful so far because the crystals did not scatter X-rays well. The characterization of the compound is, therefore, based on vibrational spectroscopy and a comparison to its related precedent, cis-Pt(CO)2(S03F)2.10 The two CO stretching bands, 2230 and 2202 cm"1 in the IR and 2236 and 2202 cm"1 in the Raman spectrum, indicate that the C2v symmetry (cis configuration) is 188 retained in Pt(CO)2(Sb2Fn)2. As expected, the average CO stretching vibration, v(CO)a v is shifted about 20 wavenumbers higher compared to that of ds-Pt(CO)2(S03F)2.10 6.4 Summary and Conclusions As can be seen from the preceding discussion, antimony pentafluoride, SbF5, is a very important agent for the development of metal carbonyl cations and cationic metal carbonyl derivatives. Solvolysis reactions in SbF5 can be carried out under various conditions; in a CO atmosphere or in the absence of gaseous CO. In addition, both metal fluorosulfates and metal chlorides can be used as starting materials. Solvolysis reactions in SbF5 in the presence of CO using metal fluorosulfates or chlorides provide one-step syntheses to the previously known metal carbonyl salts [Au(CO)2][Sb2Fu]6a and [M(CO)4][Sb2Fn]2, M = Pd or Pt.11 This approach also permits the synthesis of the new compounds [Hg(CO)2][Sb2Fu]217 and [M(CO)6][Sb2Fn]2, M = Ru or Os, where metal carbonyl fluorosulfates are either unavailable or difficult to obtain as pure compounds. In the absence of gaseous CO, solvolysis reactions in SbF5 using metal carbonyl fluorosulfates as starting materials generated three new metal carbonyl derivatives. Two of them, [Pd2(p>CO)2][Sb2Fn]2 and cw-Pt(CO)2(Sb2Fn)2, were isolated and characterized by vibrational spectroscopy. A third compound was identified as cw-Pd(CO)2(Sb2F11)2 based on IR spectroscopy although its pure form was not obtained. The solvolysis of metal carbonyl fluorosulfates in SbF5 in the absence of additional CO replaces S03F" by Sb 2F n" in the metal carbonyl complexes. As a result of this anion 189 exchange, the CO stretching vibrations in these compounds are generally shifted to higher wavenumber. In addition, the energy difference between the symmetric and asymmetric CO stretches is reduced when the fluorosulfate groups are substituted by fluoroantimonate groups. These spectroscopic features are indicative of the further reduction of rc-back-donation in fluoroantimonate carbonyl derivatives compared to the corresponding fluorosulfate carbonyl derivatives. 190 References 1 Ruff, O.; Plato, W. Chem. Ber. 1904, 37, 673. 2 Fabre, P. L. ; Devynck, J.; Tremillon, B. Chem. Rev. 1982, 82, 591. 3 Olah, G. A. ; Prakash,; G. K. S.; Sommer, J. Superacids, John Wiley & Sons: New York, 1985, and references therein. 4 O'Donnell, T. A. Superacids and Acid Melts as Inorganic Chemical Reaction Media, VCH: Weinheim, Germany, 1993, and references therein. 5 Olah, G. A. ; Schleyer, P. v. R. (Eds.), Carbonium Ions, Wiley, New York, 1968, 1, 1970, 2, 1972, 3, 1973, 4, 1976, 5. 6 (a) Willner H , ; Schaebs, J.; Hwang, G.; Mistry, F.; Jones, R.; Trotter, J.; Aubke, F. J. Am. Chem. Soc. 1992, 114, 8972. (b) Siu, S. C. B.Sc. Thesis, the University of British Columbia, 1995. 7 (a) Lee, K. C ; Aubke, F. Can. J. Chem. 1979, 57, 2058. (b)Lee, K. C ; Aubke, F. Can. J. Chem. 1977, 55, 2473. 8 Lee, K. C ; Aubke, F. Inorg. Chem. 1984, 23, 2142. 9 Lee, K. C ; Aubke, F. Inorg. Chem. 1979, 18, 389. 1 0 Hwang, G.; Wang, C ; Bodenbinder, M.; Willner, H.; Aubke, F. J. Fluorine Chem. 1994, 66, 159. 11 Hwang, G.; Wang, C ; Aubke, F.; Bodenbinder, M.; Willner, H. Can. J. Chem. 1993, 71, 1532. 1 2 Willner, H . ; Aubke, F. Inorg. Chem. 1990, 29, 2195. 1 3 Barr, J.; Gillespie, R. J.; Thompson, R. C. Inorg. Chem. 1964, 3, 1149. 191 1 4 Hwang, G.; Bodenbinder, M . ; Willner, H . ; Aubke, F. Inorg. Chern. 1993, 32, 4667. 1 5 Mallela, S. P.; Aubke, F. Can. J. Chern. 1984, 62, 382. 1 6 Willner, H . ; Bodenbinder, M . ; Wang, C ; Aubke, F. 7. Chern. Soc, Chern. Commun. 1994, 1189. 1 7 Bodenbinder, M . ; Balzer-Jollenbeck, G.; Willner, H . ; Batchelor, R. J.; Einstein, F. B. W.; Wang, C ; Aubke, F. Inorg. Chern. 1966, 35, 82. 1 8 Leung, P. C ; Aubke, F. Can. J. Chern. 1984, 62, 2892. 1 9 Leung, P. C ; Wong, G. B.; Aubke, F. 7. Fluorine Chern. 1987, 35, 607. 2 0 Wilson, W. W.; Dalziel, J. R.; Aubke, F. J. Inorg. Nucl. Chern. 1975, 37, 665. 2 1 Wilson, W. W.; Aubke, F. 7. Fluorine Chern. 1979, 13, 431. 2 2 Ruff, O.; Plato, W. Ber. Dtsch. Chern. Ges. 1904, 37, 673. 2 3 Ruff, O. Ber. Dtsch. Chern. Ges. 1909, 42, 4021. 2 4 Manchot, W.; Gall, H. Ber. Dtsch. Chern. Ges. 1925, 58, 2175. 2 5 Kharash, M. S.; Isbell, H. S. 7. Am. Chern. Soc 1930, 52, 2919. 2 6 Belli Dell'Amico, D. ; Calderazzo, F.; Dell'Amico, G. Gazz. Chim. Ital. 1977, 107, 101. 2 7 Emeleus, H. J.; Wood, J. F. 7. Chern. Soc. 1948, 2183. 2 8 Cader, M . S. R.; Aubke, F. Can. J. Chern. 1989, 67, 1700. 2 9 Cotton, F. A. ; Wilkinson. G. Advanced Inorganic Chemistry; 5th ed.;Wiley: New York, 1988;p 877. 30 (a) Cotton, F. A. ; Wilkinson. G. Advanced Inorganic Chemistry; 5th ed.;Wiley: New York, 1988; p 58 and p 1035. (b) Pruchnick, F. P. Organometallic Chemistry of Transition Elements; Plenum: New York; 1990;p 23. 192 (c) Werner, H. Angew. Chem. Int. Ed. Engl. 1990, 29, 1077. 3 1 Kettle, S. F. A. Top. Curr. Chem. 1977, 71, 111. 3 2 Braterman, P. S. Metal Carbonyl Spectra; Academic: New York, 1975. 3 3 Stromnova, T. A. ; Kuz'mina, L .G. ; Vargaftik, M. N. ; Mazo, G. Ya.; Struchkov, Yu. T.; Moiseev, I. I. Izv. Akad. Nauk SSSR, Ser. Khim. 1978, 720. 3 4 Moiseev, I. I.; Stromnova, T. A. ; Vargaftik, M. N. ; Mazo, G. Ya.; Kuz'mina, L. G. ; Struchkov, Yu. T. J. Chem. Soc, Chem. Commun. 1978, 27. 3 5 Colton, R.; Farthing, R. H. McCormick, M. J. Aust. J. Chem. 1973, 26, 2607. 3 6 Goggin, P. L. ; Mink, J. J. Chem. Soc, Dalton Trans. 1974, 534. 193 CHAPTER 7 GENERAL CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK 7.1 Summary and General Conclusions The synthesis, structure, and spectroscopic properties of cationic metal carbonyl derivatives (including metal carbonyl cations) were investigated. These studies resulted in a better understanding of the M-CO bonding when the established synergetic bonding model failed to account for the high thermal stabilities of the new complexes. Conclusions relevant to each aspect of this research have already been given in the earlier chapters. An overview of the thesis and some general conclusions are presented below. A summary of the thermally stable, homoleptic metal carbonyl cations and metal carbonyl fluorosulfates synthesized over the past six years are listed in Table 7-1. As indicated by the footnote labelling in Table 7-1, the new compounds described here involve Pd(I), Pd(II), Ir(III), Ru(II), and Os(II). Alternative, simplified syntheses have been found for [Au(CO)2][Sb2F11]1 and [M(CO)4][Sb2Fn]2, M = Pd or Pt.2 Of the five structurally characterized metal carbonyl fluorosulfates and fluoroantimonates denoted by " * " in Table 7-1, four molecular structures, c-[Pd2(pL-CO)2](S03F)2, cw-Pd(CO)2(S03F)2, mer-Ir(CO)3(S03F)3, and [Ir(CO)5Cl][Sb2Fn]2 are discussed in this thesis. The structural characterization of cw-Pd(CO)2(S03F)2 revealed for the first time the important contribution 194 Table 7-1 Thermally Stable (Persistent) Cationic Carbonyl Derivatives Solid Compounds Solvated Cations (or Cationic Species) Group 12 [Hg(CO) 2][Sb 2F„] 2* Group 11 [Ag(CO)][Sb2Fu] Au(CO)S03F [Au(CO)2][Sb2Fn] (b) [Hg(CO)2]2+ in HS03F-3SbF5 [Ag(CO)]+, [Ag(CO)2]+ in HS03F-3SbF5 [Au(CO)]+, [Au(CO)2]+ in HS0 3F and HS03F-3SbF5 (a) Group 10 cw-Pd(CO)2(S03F)2*.(a) cw-Pd(CO)2(Sb2Fn)2 c-[Pd2(u-CO)2](S03F)2 c-[Pd2(p.-CO)2][Sb2Fn]2 [PdtCO^tSb.FuL 0 0'** m-Pt(CO)2(S03F)2 cw-PtCO) 2(Sb 2F„) 2 ( a ) [Pt(CO)4][Pt(S03F)6] [PKCO^ltSb.F,,]^ *> (a) (a) [Pd(CO)4]2+ in HS0 3F [Pt(CO)4]2+ in HS0 3F Group 9 mer-Ir(CO)3(S03F)3*'' [Ir(CO)5Cl][Sb2Fn]2*' [Ir(CO)6][Sb2Fn]3 (a) /ac-Ir(CO)3(S03F)3 in HS0 3F (a) (a) Group 8 [Os(CO)6][Sb2Fn]2 [Ru(CO)6][Sb2Fn]2(a) [Fe(CO)6][Sb2Fu]2 * Denotes structurally characterized compounds. ( a ) Synthesis and characterization are described in this thesis. ^ Simplified routes to the salt are described in this thesis. 195 that secondary inter- and intra-molecular interactions make towards the relatively high thermal stabilities of the compound. Similar secondary interactions were observed in the structures of mer-Ir(CO)3(S03F)3 and [Ir(CO)5Cl][Sb2Fn]2. These findings have contributed significantly to a better understanding of the bonding in the new cationic metal carbonyl derivatives. In addition, the synthesis and structural characterization of mer-Ir(CO)3(S03F)3 and [Ir(CO)5Cl][Sb2Fn]2 have expanded the range of thermally stable, predominantly o-bonded carbonyls to include iridium for the first time. The homoleptic carbonyl cations known so far are listed in Table 7-2. With one exception ([Pt(CO)4][Pt(S03F)6]),3 all the thermally stable carbonyl cations have [Sb2Fn]~ as a counter-anion, which points to the unique ability of this dinuclear anion to engage in secondary interactions as seen in the structures of [Hg(CO)2][Sb2Fn]24 and [Ir(CO)5Cl]-[Sb2Fn]2. Further contributions towards the knowledge of metal carbonyl chemistry from this thesis include square planar [Pd(CO)4]2+, cyclic [Pd2(li-CO)2]2+, the first regular octahedral dications [M(CO)6]2 + (M = Ru or Os), and [Ir(CO)5Cl]2+. The solvolysis of mer-Ir(CO)3(S03F)3 in SbF5 in the presence of CO produced [Ir(CO)6][Sb2Fn]3 as a major product. Pure [Ir(CO)6][Sb2Fu]3 was later obtained via IrF6 under similar conditions.5 As a result, it is now possible to correlate isoelectronic octahedral metal carbonyl species in the 5d series ranging from carbonyl anions [Hf(CO)6]2" and [Ta(CO)6]" via neutral carbonyl W(CO)6 to carbonyl cations [Os(CO)6]2+ and [Ir(CO)6]3+. As seen in Table 7-3, the CO bond strength, reflected by the stretching vibrational frequency and the corresponding force constant, increases consistently in the order [Hf(CO)6]2" < 196 Table 7-2 Homoleptic Metal Carbonyl Cations Group 12 Linear [Hg2(CO)2]2+ [Hg(CO)2]2+ Group 11 Linear [Cu(CO)]+ [Cu(CO)„]+* n = 2 to 4 [Ag(CO)]+* [Ag(CO)2]+* [Ag(CO)3]+* [Au(CO)]+ [Au(CO)2]+ Group 10 Square Planar [Pd2(CO)2]2+ [Pd(CO)4]2+ [Pt(CO)4]2+ Group 9 Octahedral [Ir(CO)6]3+ Group 8 Octahedral [Fe(CO)6]2+ [Ru(CO)6]2+ [Os(CO)6]2+ Group7 Octahedral [Mn(CO)6]+ [Tc(CO)6]+ [Re(CO)6]+ * Denote thermally unstable. [Ta(CO)6]" < W(CO)6 < [Re(CO)6]+ < [Os(CO)6]2+ < [Ir(CO)6]3+. The 1 3 C chemical shifts of the CO groups in these compounds move to higher field as the charge on the metal becomes more positive. The distribution of homoleptic metal carbonyls in the periodic table is illustrated in Figure 7-1. As seen in the Figure, highly reduced metal carbonyl anions6'7,8 are formed by early transition metals in Groups 4-6. Middle transition metals, Groups 6 to 8, form typical 197 O N C N <n O U + o u + o u ^—' rt v© o u o u o u « m N O O N r-C N C N C N C N S s wo C N O C N C N C N O N O O O OO wo O N . r-C N C N O N T t C N C N C N rti rti & 3!, & t— C N WO O N C N O O rti rti o C N C N C N O O C"~ >n -2-oo 4J-rti O N t -rti O N O N C N rti rti "•rt. rJ ° . ^CN t-~ O N o m oo C N wo e rt r t C N e o ^ — \ 60 ccT e 60 C N s N O C N *-< © ~ O N O N 3 6 rti _! u pd N O CD 0 1 rt o O N O N cm c •ci rt, OO CO ss <3 © e ' ~ C N f I 3 gp s I c _ •8 o C N P O N 3 UJ oo S O N r t cn C N oo" ©C N O O N 3 O 00 rt o Q o X <u c o CN CO r t r t O N I S oo O O N N O O N § a, t-T J3 13 ai 1 a, O0 xT 13 is >> H rt <u 3 PL, X) 00 0 U rt Vi a 1 3 Vi s s W0 <D O C S rt S . 9 198 THE TRANSITION METALS* (including some post-transition metals) Highly Reduced Typical Transition Metal Carbonyl Carbonyl Anions Metal Carbonyls Cations i i i I I I 3 4 5 6 7 8 9 10 11 12 21 22 23 24 25 26 27 28 29 30 Sc Ti V Cr Mn Fe Co Ni Cu Zn Scandium Titanium Vanadium Chromium Manganese iron Cobalt Nickel Copper Zinc 39 40 41 42 43 44 45 46 47 48 Y Zr Nb Mo Tc Ru Rh Pd Ag Cd Yttrium Zirconium Niobium Molybdenum Tecnetium Rutftenfum Rhodium Palladium Silver Cadmium 57 72 73 74 75 76 i l l 78 79 so La Hf Ta w Re Os Pt Au Hg Lanthanum Hafnium Tantalum Tunasten Rhenium Osmium Platinum Gold Mercurv t Overlap of all three M-CO bond types * Shaded area indicates thermally stable metal carbonyl cations were found. Figure 7-1 The distribution of metal carbonyl species in the periodic table. 199 metal carbonyl complexes where the metal oxidation state ranges from +1 to -1. Both types of these metal carbonyls obey the effective atomic number rule (EAN) with the exception of V(CO)6, which has 17 rather than 18 electrons associated with vanadium. Finally, predominantly a-bonded carbonyl cations and cationic carbonyl fluorosulfates are formed by late transition metals and post-transition metals (Groups 8-11). There are no clear-cut boundaries between the three types of metal carbonyls; however, iridium is so far unique in its ability to form all three types of carbonyls: (i) highly reduced carbonyl anion [Ir(CO)6]3",6 (ii) typical carbonyl Ir4(CO)12,9 and (iii) carbonyl cation [Ir(CO)6]3+. The synthesis and structural characterization of the first thermally stable, predominantly a-bonded Ir(III) carbonyl species is a highlight of this thesis. As illustrated in Figure 7-2, metal-CO bonds encountered in different situations can be categorized essentially into three types: (i) predominantly jc-bonded, (ii) synergically bonded, and (iii) predominantly a-bonded. The transition of the metal-CO bonding from mainly K-bonding to mainly a-bonding is gradual and continuous. Both the nature of the metal center and its oxidation state influence the M-CO bonding. As the metals (or metal cations) become more electron withholding or acquire higher positive charges (+2 and +3), the CO ligand becomes more " cyanide-like " (i.e. predominantly a-bonded), and behaves like a typical 2-electron donor ligand. The resulting geometries, square planar for [M(CO) 4] 2 +, M = Pd or Pt,2 ,3 and linear for [Au(CO)2]+ 1 and [Hg(CO)2]2 +,4 are 8 10 compatible with the d-electron configurations of d and d respectively. These geometries 200 7C (b) Typical transition metal carbonyls, v(CO) = 1850 - 2120 cm"1. (c) Metal carbonyl cations, v(CO) = 2150 - 2300 cm"1. Figure 7-2 Three types of metal-CO bonding and approximate ranges of v (CO), the oxidation state of metal center is indicated. 201 are expected in the coordination compounds of M(II), M = Pd or Pt (d ), and Au(I) and Hg(II) (d10). The effective atomic number rule is not obeyed by these carbonyl cations, for example, [Pd(CO)4]2+ and [Pt(CO)4]2+ have 16 electrons while [Au(CO)2]+ and [Hg(CO)2]2 + have 14 electrons associated with the respective metals. However, the departure from the EAN rule does not bring about special reactivities such as those observed in typical transition metal carbonyl chemistry.10 Hence homoleptic carbonyl cations display properties and bonding characteristics usually associated with coordination compounds while typical transition metal carbonyls and highly reduced carbonyl anions are better viewed as organometallic compounds. The extreme case of a-bonding-only in the M-CO moiety has never been attained in any of the known examples of metal carbonyl cations. Their CO stretching force constants are still lower than that exhibited by H C O + (ft — 21.3 X 102 N m"1), which is considered to be a benchmark for a measurement of 7c-back-bonding. The closest values were found for [Hg(CO)2][Sb2Fn]2 (21.0 x 102 N m"1) and [Ir(CO)6][Sb2F„]3 (20.8 x 102 N m"1).5 Based on the new findings reported in this research, as well as results from previous studies, answers for the questions posed at the beginning of this thesis are now attempted. (i) Why and under what conditions are cationic metal carbonyl derivatives thermally stablel Several factors influence the thermal stabilities of cationic metal carbonyls or metal carbonyl cations. The ideal conditions for the generation of cationic metal carbonyls include: (a) The metal center should have a strong tendency to form a covalent bond with the carbon 202 atom of the CO ligand, so that the transition metals are likely to form cationic carbonyl derivatives, (b) The counter-anion or anionic moiety should be weakly nucleophilic and less likely to compete with CO for a coordination site. The weak nucleophilicity will also result in a strong polar contribution to the metal-CO bond, (c) Conditions should exist for favorable secondary interactions between the carbon atoms of the CO ligand and electronegative atoms of an anion either through intra-molecular or inter-molecular contacts, where perhaps the latter is more important than the former. These secondary interactions have a dual effect on the thermal stabilities of metal carbonyl derivatives. They not only compensate the partial charge on the carbon atom of the CO ligand, but also increase the crystal lattice energy by engaging in three-dimensional network in the solid state. (ii) How is the metal-CO bonding best described? Will the synergic bonding model have to be replaced or modified) As discussed previously in this thesis, the contribution from rc-back-donation in cationic metal carbonyl derivatives prepared in this research, as well as in previous examples of gold(I)1 and platinum(II),2'3 is drastically reduced as reflected in their CO stretching force constants. At the other extreme, the c-dative contribution to the formation of metal-CO bond is insignificant in the highly reduced metal carbonyl anions. Nevertheless, the synergic bonding model can still be invoked to explain the formation of the majority of metal carbonyl complexes, especially those formed by middle-transition metals. Therefore, the bonding description is expanded, as shown in Figure 7-2, to allow a gradual transition from mainly 7t-back-bonding via synergic bonding to mainly a-bonding. From the large number and great variety of the metal carbonyl derivatives with the overall charge of the complexes ranging 203 from -4 to +3, it is easy to see that CO is an extremely versatile ligand. The realization of this versatility should overcome the psychological barrier induced by the conventional synergic bonding model and encourage more chemists to be involved in quest of new carbonyl compounds. (iii) How to decide upon the terminology of the cationic metal carbonyl derivatives with very high CO stretching frequencies'} The very high CO stretching frequencies and hence the anticipation of the absence of 7t-back-donation from the metal center in some cationic metal carbonyls prompted the terminology nonclassical metal carbonyls.4,11'12 The term implies that these compounds are different from previously known examples of metal carbonyls where the synergic bonding model applies. However, two ambiguities arise from this terminology: (a) The first examples of nonclassical metal carbonyls were discovered in 1868, predating the synthesis of the classical carbonyl Ni(CO)4 by 22 years, (b) Cationic metal carbonyl derivatives are not the only type of metal carbonyls that deviate from the synergic bonding model, as discussed in Chapter 1. In addition, a-bonding-only has not been achieved by any of the known examples of the cationic metal carbonyl compounds as aforementioned. Therefore, the term predominantly a-bonded metal carbonyl derivatives is proposed to denote metal carbonyl cations or cationic metal carbonyl derivatives obtained in this research and similar metal carbonyls. 204 7.2 Suggestions for Future Work Developments in this field have been rapid as can be seen from Tables 7-1 and 7-2. Among the noble metal carbonyl compounds, palladium and iridium carbonyls received the most extensive studies in this research. Altogether, five Pd carbonyls and four Ir carbonyls have been found, and new CO complexes of these two metals can still be expected. As for ruthenium and osmium, the exploratory work described here is just the start of a long journey. It is, therefore, evident that new compounds can be obtained and new synthetic approaches can be developed. Some specific suggestions are proffered for future work in this area. 7.2.1 Syntheses of New Cationic Metal Carbonyl Derivatives The reductive carbonylation of Ru(S03F)3,13 Os(S03F)3,1 4 and Rh(S03F)314 in fluorosulfuric acid should eventually lead to the isolation and characterization of the corresponding cationic metal carbonyl fluorosulfates. In addition, fluorosulfuric acid has been proven to be an excellent solvent for growing single crystals for X-ray diffraction studies. It is anticipated that molecular structures of these new metal carbonyl complexes will become available in the future. Using metal carbonyl fluorosulfates as precursors, the anion can be substituted by a fluoroantimonate anion via solvolysis reactions in SbF5. This synthetic approach will not only provide new compounds for structural and spectroscopic comparisons, but these complexes will also serve as precursors for new syntheses. One of the applications of these syntheses is the development of mixed ligand metal carbonyls in which other small ligands such as nitrosyl, phosphene, and olefins are also bonded to the metal center. 205 7.2.2 Syntheses of New Cationic Metal Carbonyl Salts Attempts to synthesize carbonyl cations of Ni in the 3d series and of Cd and Tl among the post-transition metals have been unsuccessful. Preliminary results of the reaction of Rh(S03F)3 with CO in HS0 3F indicates that the generation of the trication [Rh(CO)6]3+ is possible. This can be achieved by using either Rh(S03F)3 or RhF6 as a precursor. Since the metal fluorosulfate, like Ir(S03F)3, is very difficult to obtain by metal oxidation with S 20 6F 2 , the alternate route may be a better choice. Furthermore, ternary carbonyl cations of the type [M(CO) 5X]n +, M = Rh, Ir, Ru, or Os, X = Cl or F, and n = 1 or 2, should exist in addition to [Ir(CO)5Cl]2+, all with [Sb2F„]" as a counter-anion. 7.2.3 Substitutions of CO by Other Weak Nucleophiles On account of the dual role of CO as reducing agent and as a ligand, substitutions of CO by other weak nucleophiles may seem difficult. However, some exploratory work has already been started by using PF3 as a substitute ligand for CO. Other suitable, weakly basic ligands such as dinitrogen (N2), nitrosyl (NO), and various olefins may be considered as potential ligands for future investigations. 7.2.4 Theoretical Studies and Practical Applications Most efforts in this research have been concentrated on the synthesis and characterization of new compounds. It is, however, anticipated that these new metal carbonyls will also have practical applications, for example, in catalysis and gas purifications. Therefore, relevant thermodynamic and kinetic properties of these new compounds should be determined. 206 With the development of the simplified synthetic routes of cationic metal carbonyls described in this thesis, new types of metal carbonyl complexes should be available for both theoretical studies and practical application developments. It is hoped that more chemists, both academic and industrial, will be motivated to explore what promises to be an interesting and useful area of chemistry. 207 References 1 Willner H , ; Schaebs, J.; Hwang, G.; Mistry, F.; Jones, R.; Trotter, J.; Aubke, F. J. Am. Chern. Soc. 1992, 114, 8972. 2 Hwang, G. ; Wang, C.; Aubke, F.; Bodenbinder, M . ; Willner, H. Can. J. Chern. 1993, 71, 1532. 3 Hwang, G. ; Bodenbinder, M . ; Willner, H . ; Aubke, F. Inorg. Chern. 1993, 32, 4667. 4 Bodenbinder, M . ; Balzer-Jollenbeck, G.; Willner, H . ; Batchlor, R. J.; Einstein, F. W. B.; Wang, C.; Aubke, F. Inorg. Chern. 1996, 35, 82. 5 Bach, B.; Willner, H . ; Wang. C.; Rettig, S. J.; Trotter, J.; Aubke, F. to be published. 6 Ellis, J. E. Adv. Organomet. Chern. 1990, 31,1. 7 Warnock, G. F. P.; Sprague, J.; Fjare, K. L. ; Ellis, J. E. J. Am. Chern. Soc. 1983, 105, 672. 8 Chi, K. M . ; Frerichs, S. R.; Philson, S. B.; Ellis, J. E. J. Am. Chern. Soc. 1988, 110, 303. 9 Cotton, F. A. ; Wilkinson. G. Advanced Inorganic Chemistry; 5th ed.;Wiley: New York, 1988;p 1027. 1 0 Cotton, F. A. ; Wilkinson. G. Advanced Inorganic Chemistry; 5th ed.;Wiley: New York, 1988; p 1305, and references therein. 1 1 Hurlburt, P. K.; Rack, J. J.; Luck, J. S.; Dec, S. F.; Webb, J. D. ; Anderson, O. P.; Strauss, S. H. J. Am. Chern. Soc. 1994, 116, 10003. 1 2 Aubke, F.; Wang, C. Coord. Chern. Rev. 1994, 137, 483. 1 3 Leung, P.C.; Aubke, F. Can. J. Chern. 1984, 62, 2892. 1 4 Leung, P.C.; Wong, G. B.; Aubke, F. J. Fluorine Chern. 1987, 35, 607. 208 Appendix A Crystallographic Data and Atomic Coordinates for [Pd2(ji-CO)2](S03F)2 Table A - l Crystallographic Data for [Pd^-CCOJfSOaF^1 Compound [Pd2(ix-CO)2](S03F)2 Formula CF04PdS Formula weight 233.47 Crystal system Monoclinic Space group C2/c a, A 11.485 (1) b, A 8.255 (1) c, A 9.556 (1) Meg 91.94 (1) V, A3 906.3 (2) z 8 DcaZo g / c m 3 3.422 T, °C 21 \i, cnr1 44.91 Transmission factors 0.75-1.00 R(F) 0.061 RJF) 0.075 a Temperature 294 K, Rigaku AFC6S diffractometer, Mo K a (X = 0.71069 A) radiation, graphite monochromator, takeoff angle 6.0°, aperture 6.0 x 6.0 mm at a distance of 285 mm from the crystal, stationary background counts at each end of the scan (scan/background time ratio 2:1), c^F 2) = [S2(C+4fi) + (O.OIF2)2]/!^)2 (S = scan rate, C = scan count, B = normalized background count), function minimized 2w(|F0|-|Fc|)2 where w = 4F 0 2/a 2(F 0 2), R = Z||F0HFC||/Z|F01, Rw = [Z(w( | F01 -1 Fc | )2)/Z(wF0 2)]1 / 2, and gof = pw(F0|-|Fc|)2/(w-n)]1/2 Values given for R, Rw, and gof are based on those reflections with / > 3a(7). 209 Table A-2 Atomic Coordinates and B Values for [Pd2(H-CO)2](S03F)2 Atom X y z Beq Pd(l) 0.41737(3) 0.02552(4) 0.09417(3) 1.789(6) S(l) 0.28837(9) 0.3681(1) 0.2017(1) 1.93(2) F(l) 0.3654(4) 0.4972(4) 0.2763(4) 2.99(7) O(l) 0.3556(3) 0.2210(5) 0.2221(4) 2.60(7) 0(2) 0.1839(4) 0.3642(5) 0.2814(5) 3.35(8) 0(3) 0.2747(4) 0.4127(7) 0.0609(5) 3.19(8) 0(4) 0.4733(4) -0.3082(5) 0.0272(5) 3.35(8) C(l) 0.4857(4) -0.1730(5) 0.0136(5) 2.07(7) Beq = 1 % 1 ^Utja*aj*(araj) 210 Appendix B Crystallographic Data and Atomic Coordinates for cw-Pd(CO)2(S03F)2 Table B-l Crystallographic Data for the Structure Determination of d.y-Pd(CO)2(S03F)2 at 220K Formula PdS 2F 20 8C 2 Crystal system Monoclinic Formula weight 360.53 . Space group P2i/n a, (A) a 7.3697(11) p c (g cm"3) 2.642 MA) 14.7742(35) X(M6 Koq) (A) 0.70930 C (A) 8.3237(21) p:(Mo Ka) (cm"1) 25.2 P, (°) 90.52(2) Min-max 20 (°) 4-54 v, (A3) 906.3 Transmission * 0.620-0.719 z 4 Crystal dim. (mm) 0.15x0.17x0.22 RFC 0.027 J? d 0.034 a Cell dimensions were determined from 25 reflections (36° < 26 < 42°). b The data were corrected for the effects of absorption by the Gaussian integration method. c RF=L | (| F01 -1 Fc |) | /-1F 01, for 1464 data (I0 > 2.5o(/0)). dRWF=&(w(\Fo\-\Fc\)2)WwFo2)]m for 1464 data (I0 > 2.5o(/0)); w=[a(Fo)2+0.0002Fo2]-1. 211 Table B-2 Atomic Coordinates8 and Equivalent Isotropic Temperature Factors ( A 2 ) for m-Pd(CO)2(S03F)2 at 220K. Atom X y z Ueqh Pd 0.46944 (4) 0.140596(22) 0.19753( 4) 0.0229 S(l) 0.12649 (15) 0.18335 ( 8) -0.00469 (14) 0.0288 S(2) 0.36929 (18) -0.05813 ( 8) 0.27490 (14) 0.0319 F(l) 0.1073 (5) 0.08221 (19) -0.0545 ( 4) 0.0474 F(2) 0.5670 (5) -0.0855 (3) 0.3249 (4) 0.0661 0(11) 0.2084 (4) 0.17541 (24) 0.1576(4) 0.0328 0(12) 0.2456 (5) 0.22102 (22) -0.1204 (4) 0.0362 0(13) -0.0527 (5) 0.2170 (3) 0.0023 (5) 0.0448 0(21) 0.3703 (5) 0.03764 (22) 0.3269 (4) 0.0325 0(22) 0.3688 (6) -0.0692 (3) 0.1076(4) 0.0482 0(23) 0.2593 (8) -0.1108 (3) 0.3732 (6) 0.0597 O(l) 0.8518 (5) 0.0775 ( 3) 0.2755 (4) 0.0438 0(2) 0.6031 (4) 0.30614 (24) 0.0153 (4) 0.0387 C(l) 0.7126 (7) 0.1026(3) 0.2471 (5) 0.0312 C(2) 0.5559 (6) 0.2458 (3) 0.0803 (6) 0.0284 The general equivalent positions in the space group P2i/n are: x, y, z; -x, -y, -z; l/2-x, 1/2+y, l/2-z; 1/2+*, 1/2-y, 1/2 +z. b Ueq is the cube root of the product of the principal axes of the mean squared atomic displacement ellipsoid. 212 Appendix C Crystallographic Data and Atomic Coordinates for /ner-Ir(CO)3(S03F)3 Table C - l Crystallographic Data for the Structure Determination of mer-Ir(CO)3(S03F)3 at 200 K Formula IrS 3F 30 1 2C 3 Crystal system Monoclinic Formula weight 573.42 Space group P2}/c a, (k)a 8.476(1) Pc(gcm-3) 2.921 b, (A) 12.868(2) X(MoKctj)(A) 0.70930 c, (A) 12.588(1) u.(Mo Ka) (cm1) 107.6 p\ (°) 108.24 (1) Min-max 29 (°) 4 - 50 V, (A)3 1304.0 Transmission* 0.453 - 1.0 Z 4 Crystal dim. (mm) 0.20 x 0.34 x 0.36 R F C 0.022 R W F D 0.024 a Cell dimensions were determined from 25 reflections (38° < 28 < 50°). b The data were corrected empirically for the effects of absorption. CRF = X| (|FD|-1FC|) | / E | F 0 \ , for 2090 data (/„ >. 2.5o(/0)). D R W F = [ Z ( w ( \F O \ - \F C \ ) 2 ) / Z ( W F 2 ) ] 1 / 2 for 2090 data (IA >. 2.5a(/c)); w = 1. 213 Table C-2 Fractional Atomic Coordinates and Equivalent Isotropic Displacement Parameters (A)2 for mer-Ir(CO)3(S03F)3 at 200 K Atom X y z U a Ir 0.27018(3) 0.590208(17) 0.255878(18) 0.0144 S(l) 0.30798(20) 0.82664(13) 0.19034(14) 0.0242 S(2) -0.09454(18) 0.57702(13) 0.09385(13) 0.0225 SO) 0.62836(19) 0.62602(13) 0.42295(13) 0.0224 F(l) 0.1271(5) 0.8666(3) 0.1515(3) 0.0358 F(2) -0.1670(5) 0.6620(4) 0.1518(4) 0.0391 F(3) 0.5951(6) 0.7269(4) 0.4789(4) 0.0398 0(11) 0.3019(6) 0.7482(3) 0.2756(4) 0.0252 0(12) 0.3323(7) 0.7825(4) 0.0941(4) 0.0348 0(13) 0.4043(7) 0.9121(4) 0.2419(5) 0.0400 0(21) 0.0791(5) 0.6182(3) 0.1133(3) 0.0196 0(22) -0.1852(6) 0.5833(4) -0.0209(4) 0.0311 0(23) -0.0968(6) 0.4857(5) 0.1541(5) 0.0388 0(31) 0.4632(5) 0.5711(3) 0.3994(3) 0.0206 0(32) 0.6633(6) 0.6577(4) 0.3252(4) 0.0312 0(33) 0.7489(6) 0.5721(4) 0.5075(4) 0.0325 0(1) 0.2278(6) 0.3563(4) 0.2543(4) 0.0303 0(2) 0.4756(6) 0.5740(4) 0.0899(4) 0.0275 0(3) 0.0458(6) 0.6191(5) 0.4071(4) 0.0364 C(l) 0.2427(7) 0.4407(5) 0.2527(5) 0.0225 C(2) 0.4074(7) 0.5804(5) 0.1522(5) 0.0223 C(3) 0.1283(8) 0.6057(5) 0.3553(5) 0.0232 " Ueq is the cube root of the product of the principal axes of the mean squared atomic displacement ellipsoid. 214 Appendix D Crystallographic Data and Structural Parameters for [Ir(CO)5Cl][Sb2F11]2 Table D-l Crystallographic Data for pr(CO)5Cl][Sb2Fn]2a Empirical formula Formula weight Crystal color, habit Crystal dimensions Crystal system Lattice type C5ClF22Ir05Sb4 127.69 Colorless, irregular 0.20 x 0.40 x 0.45 mm Monoclinic Primitive No. of reflections used for unit cell determination (20 range) 25 (33.2 - 39.9°) Omega scan peak width at half-height Lattice parameters 0.32° a = 9.686(2) A b= 12.585(2) A c = 10.499(2) A j3 = 106.59(2)° V = 1226.5(4) A3 Space group Z value L\al c Fooo u\(MoKoc) R(f) Rw(f ) P2i (#4) 2 3.466 g/cm3 1132 100.55 cm"1 0.032 0.031 Temperature 294 K, Rigaku AFC6S diffractometer 215 Table D-2 Atomic Coordinates and Beq for [Ir(CO)5Cl][Sb2Fn]2 Atom X y z Ir(l) 0.74396(6) 0.5095 0.75201(5) 2.396(7) Sb(l) 0.7594(1) 0.13301(10) 0.50828(9) 3.55(2) Sb(2) 0.8311(1) 0.3168(1) 0.23180(9) 3.30(2) Sb(3) 0.6538(1) 0.7170(1) 0.25615(9) 3.50(2) Sb(4) 0.7387(1) 0.8764(10) -0.03153(9) 3.82(2) Cl(l) 0.9299(4) 0.4858(4) 0.6566(4) 4.43(9) F(l) 0.835(1) 0.2014(9) 0.3758(10) 5.7(3) F(2) 0.687(1) 0.074(2) 0.634(1) 11.8(5) F(3) 0.758(2) 0.018(1) 0.410(1) 10.2(4) F(4) 0.751(1) 0.2684(9) 0.5752(8) 6.2(3) F(5) 0.948(1) 0.122(1) 0.599(1) 9.3(4) F(6) 0.5724(9) 0.1550(10) 0.3998(8) 6.5(3) F(7) 0.818(1) 0.4118(8) 0.0914(8) 5.8(3) F(8) 0.955(1) 0.2250(9) 0.1820(9) 6.2(3) F(9) 0.710(1) 0.3930(9) 0.3015(9) 5.2(2) F(10) 0.678(1) 0.240(1) 0.131(1) 7.3(3) F(ll) 0.991(1) 0.375(1) 0.3481(10) 7.0(3) F(12) 0.675(2) 0.8218(9) 0.122(1) 7.8(4) F(13) 0.630(1) 0.6261(9) 0.3831(8) 5.7(2) F(14) 0.759(1) 0.6273(9) 0.1817(10) 6.5(3) F(15) 0.538(1) 0.8235(8) 0.2979(8) 5.2(2) F(16) 0.495(1) 0.673(1) 0.1295(8) 6.3(3) F(17) 0.810(1) 0.783(1) 0.364(1) 6.5(3) F(18) 0.802(2) 0.919(1) -0.172(1) 10.4(5) F(19) 0.700(1) 0.7386(8) -0.0958(8) 5.7(2) F(20) 0.773(1) 1.0116(9) 0.0525(9) 5.8(2) F(21) 0.918(1) 0.827(1) 0.063(1) 8.7(3) F(22) 0.554(1) 0.917(1) -0.109(1) 9.4(4) 0(1) 0.497(1) 0.5493(9) 0.883(10 5.2(3) 0(2) 0.543(1) 0.428(1) 0.4851(10) 4.6(3) 0(3) 0.970(1) 0.5890(9) 1.0088(9) 3.6(2) 0(4) 0.786(1) 0.2792(9) 0.8560(10) 4.2(3) 0(5) 0.725(1) 0.740(1) 0.6336(9) 4.5(3) C(l) 0.579(1) 0.534(1) 0.835(1) 3.3(3) C(2) 0.614(2) 0.456(1) 0.578(1) 3.1(3) C(3) 0.892(1) 0.562(1) 0.920(1) 2.9(3) C(4) 0.775(2) 0.360(1) 0.820(1) 3.0(3) C(5) 0.730(1) 0.658(1) 0.677(1) 2.8(3) Beq = -7i 2 (U„(aa*) 2 + U22(bb*)2 + U33(cc*)2 + 2U12aa*bb*cosy + 2U13aa*cc*cos|3 + 2U23bb*cc*cos a) 216 Table D-3 Bond Lengths and Bond Angles for [Ir(CO)5Cl][Sb2Fn]2 Bond Lengths (A) Atom Atom Distance Atom Atom Distance Ir(l) Cl(l) 2.321(3) Ir(l) e(i) 2.05(1) Ir(l) C(2) 2.01(1) Ir(l) C(3) 2.04(2) Ir(l) C(4) 2.00(2) Ir(l) C(5) 2.02(2) Sb(l) F(l) 1.946(9) Sb(l) F(2) 1.83(1) Sb(l) F(3) 1.78(1) Sb(l) F(4) 1.85(1) Sb(l) F(5) 1.81(1) Sb(l) F(6) 1.868(9) Sb(2) F(l) 2.089(10) Sb(2) F(7) 1.873(8) Sb(2) F(8) 1.846(9) Sb(2) F(9) 1.819(8) Sb(2) F(10) 1.83(1) Sb(2) F(H) 1.83(1) Sb(3) F(12) 1.980(10) Sb(3) F(13) 1.819(8) Sb(3) F(14) 1.837(9) Sb(3) F(15) 1.881(8) Sb(3) F(16) 1.81(1) Sb(3) F(17) 1.82(1) Sb(4) F(12) 2.007(9) Sb(4) F(18) 1.836(10) Sb(4) F(19) 1.86(1) Sb(4) F(20) 1.90(1) Sb(4) F(21) 1.84(1) Sb(4) F(22) 1.82(1) 0(1) C(l) 1.07(1) 0(2) C(2) 1.08(2) 0(3) C(3) 1.08(2) 0(4) C(4) 1.08(2) 0(5) C(5) 1.12(2) Bond Angles (°) Atom Atom Atom Angle Atom Atom Atom Angle Cl(l) Ir(l) C(l) 178.6(5) Cl(l) Ir(l) C(2) 85.8(4) Cl(l) Ir(l) C(3) 88.4(3) Cl(l) Ir(l) C(4) 88.6(4) Cl(l) Ir(l) C(5) 85.4(3) C(l) Ir(l) C(2) 94.2(5) C(l) Ir(l) C(3) 91.6(5) C(l) Ir(l) C(4) 92.8(5) C(l) Ir(l) C(5) 93.2(5) C(2) Ir(l) C(3) 174.2(6) C(2) Ir(l) C(4) 89.9(6) C(2) Ir(l) C(5) 90.1(5) C(3) Ir(l) C(4) 90.0(6) C(3) Ir(l) C(5) 89.3(5) C(4) Ir(l) C(5) 174.0(6) F(l) Sb(l) F(2) 177.9(8) 217 Table D-3 Bond Lengths and Bond Angles for nr(CO)5Cl][Sb2F11]2 (Continued) Atom Atom Atom Angle F(l) Sb(l) F(3) 83.7(5) F(l) Sb(l) F(5) 83.2(50 F(2) Sb(l) F(3) 98.(3) F(2) Sb(l) F(5) 97.4(6) F(3) Sb(l) F(4) 167.5(6) F(3) Sb(l) F(6) 85.4(6) F(4) Sb(l) F(6) 87.6(5) F(l) Sb(2) F(7) 174.8(4) F(l) Sb(2) F(9) 87.4(4) F(l) Sb(2) F(ll) 87.8(5) F(7) Sb(2) F(9) 94.1(4) F(7) Sb(2) F(ll) 97.1(5) F(8) Sb(2) F(10) 89.4(5) F(9) Sb(2) F(10) 90.3(5) F(10) Sb(2) F(ll) 171.3(7) F(12) Sb(3) F(14) 85.7(5) F(12) Sb(3) F(16) 85.6(5) F(13) Sb(3) F(14) 97.2(5) F(13) Sb(3) F(16) 95.0(5) F(14) Sb(3) F(15) 168.0(4) F(14) Sb(3) F(17) 94.8(5) F(15) Sb(3) F(17) 88.6(5) F(12) Sb(4) F(18) 176.8(6) F(12) Sb(4) F(20) 89.2(4) F(12) Sb(4) F(22) 88.2(6) F(18) Sb(4) F(20) 93.1(5) F(18) Sb(4) F(22) 94.2(7) F(19) Sb(4) F(21) 86.6(5) F(20) Sb(4) F(21) 92.2(5) F(21) Sb(4) F(22) 172.7(6) Sb(3) F(12) Sb(4) 156.5(6) Ir(l) C(2) 0(2) 179(1) Ir(l) C(4) 0(4) 177(1) Atom Atom Atom Angle F(l) Sb(l) F(4) 85.9(5) F(l) Sb(l) F(6) 89.4(4) F(2) Sb(l) F(4) 92.0(7) F(2) Sb(l) F(6) 90.0(5) F(3) Sb(l) F(5) 94.6(7) F(4) Sb(l) F(5) 91.0(6) F(5) Sb(l) F(6) 172.5(5) F(l) Sb(2) F(8) 83.8(4) F(l) Sb(2) F(10) 84.0(6) F(7) Sb(2) F(8) 94.6(4) F(7) Sb(2) F(10) 91.1(5) F(8) Sb(2) F(9) 171.2(4) F(8) Sb(2) F(ll) 86.9(5) F(9) Sb(2) F(ll) 92.2(5) F(12) Sb(3) F(13) 177.1(5) F(12) Sb(3) F(15) 83.2(4) F(12) Sb(3) F(17) 84.1(6) F(13) Sb(3) F(15) 94.0(4) F(13) Sb(3) F(17) 95.2(6) F(14) Sb(3) F(16) 87.3(5) F(15) Sb(3) F(16) 87.3(5) F(16) Sb(3) F(17) 169.3(6) F(12) Sb(4) F(19) 84.2(4) F(12) Sb(4) F(21) 84.6(6) F(18) Sb(4) F(19) 93.5(5) F(18) Sb(4) F(21) 93.1(6) F(19) Sb(4) F(20) 173.3(4) F(19) Sb(4) F(22) 91.8(6) F(20) Sb(4) F(22) 88.6(6) Sb(l) F(l) Sb(2) 153.2(6) Ir(l) C(l) 0(1) 176(1) Ir(l) C(3) 0(3) 179(1) Ir(l) C(5) 0(5) 178(1) 218 Table D-4 Non-bonded Contacts out to 3.60 A for [Ir(CO)5Cl][Sb2Fu]2 Atom Atom Distance ADC Ir(l) F(19) 3.379(10) 55601 Ir(l) F(6) 3.529(10) 65602 Cl(l) F(4) 3.22(1) 1 Cl(l) F(8) 3.48(1) 75602 Cl(l) F(20) 3.56(1) 74602 F(2) 0(2) 2.88(2) 64602 F(2) 0(4) 3.43(2) 1 F(3) F(17) 3.05(2) 54501 F(3) O(l) 3.38(2) 64602 F(3) C(l) 3.54(2) 64602 F(4) C(2) 2.72(2) 1 F(4) 0(2) 2.81(2) 1 F(4) F(14) 3.50(1) 64602 F(5) F(15) 3.10(2) 74602 F(6) C(5) 2.81(2) 64602 F(6) 0(5) 3.00(1) 64602 F(6) C(l) 3.14(1) 64602 F(6) F(13) 3.44(1) 64602 F(7) C(4) 2.84(2) 55401 F(7) 0(4) 2.93(1) 55401 F(7) F(14) 2.98(1) 1 F(7) F(18) 3.53(2) 74502 F(8) 0(3) 2.88(1) 74602 F(8 C(5) 3.10(2) 74602 F(8) F(20) 3.29(1) 54501 F(8) 0(4) 3.41(1) 55401 Atom Atom Distance ADC* Ir(l) F(15) 3.515(9) 64062 Ir(l) F(4) 3.570(10) 1 Cl(l) F(3) 3.32(1) 75602 Cl(l) F(21) 3.51(1) 74602 F(2) F(18) 2.80(2) 54601 F(2) F(13) 3.09(2) 64602 F(2) C(2) 3.45(2) 64602 F(3) F(15) 3.24(2) 54501 F(3) F(H) 3.47(2) 74602 F(3) 0(2) 3.58(2) 64602 F(4) C(4) 2.77(2) 1 F(4) 0(4) 2.87(1) 1 F(4) F(17) 3.04(2) 74602 F(5) F(H) 3.18(2) 74602 F(6) C(l) 2.90(2) 64602 F(6) C(2) 3.14(2) 64602 F(6) 0(2) 3.42(2) 64602 F(6) 0(2) 3.58(2) 1 F(7) C(3) 2.84(2) 55401 F(7) 0(3) 2.94(1) 55401 F(7) C(l) 3.38((2) 55401 F(7) F(21) 3.57(2) 74502 F(8) C(3) 2.91(2) 74602 F(8) 0(5) 3.15(2) 74602 F(8) F(21) 3.41(1) 74502 F(8) F(18) 3.42(2) 74502 Table D-4 Non-bonded Contacts out to 3.60 A for pr(CO)5Cl][Sb2Fn]2 (Continued) Atom Atom Distance ADC Atom Atom Distance ADC* F(9) F(22) 2.79(2) 64502 F(9) 0(2) 2.88(1) 1 F(9) F(13) 3.21(1) 1 F(9) F(14) 3.29(1) 1 F(9) C(2) 3.39(1) 1 F(10) F(16) 2.90(1) 64502 F(10) 0(1) 2.92(2) 64602 F(10) F(22) 3.12(2) 64502 F(10) F(20) 3.20(2) 54501 F(10) 0(4) 3.38(2) 55401 F(10) F(19) 3.58(2) 64502 F(ll) F(18) 3.14(2) 74502 F(ll) 0(5) 3.19(2) 74602 F(ll) F(17) 3.30(2) 74602 F(12) 0(1) 3.31(2) 65602 F(ll) 0(5) 2.90(1) 1 F(13) 0(2) 2.94(2) 1 F(13) C(5) 2.98(1) 1 F(13) C(2) 3.00(2) 1 F(14) 0(3) 3.13(1) 55401 F(14) C(3) 3.45(1) 55401 F(14) 0(1) 3.56(2) 55401 F(15) C(2) 2.78(2) 65602 F(15) 0(2) 2.92(1) 65602 F(15) C(4) 2.96(2) 65602 F(15) C(l) 3.06(2) 65602 F(15) 0(4) 3.14(2) 65602 F(15) O(l) 3.38(1) 65602 F(16) 0(1) 3.02(1) 55401 F(16) 0(4) 3.08(2) 65602 F(16) F(22) 3.25(2) 64502 F(16) F(20) 3.42(2) 64502 F(17) 0(5) 3.21(2) 1 F(18) 0(5) 2.99(2) 55401 F(18) 0(3) 3.19(2) 75602 F(19) C(5) 2.69(1) 55401 F(19) C(l) 2.84(2) 55401 F(19) C(3) 2.87(2) 55401 F(19) 0(5) 2.92(1) 55401 F(19) 0(1) 3.06(1) 55401 F(19) 0(3) 3.15(1) 55401 F(20) 0(3) 2.91(1) 75602 F(20) 0(1) 2.92(1) 65602 F(20) C(3) 3.24(2) 75602 F(21) 0(4) 2.81(1) 75602 F(21) C(4) 2.91(2) 75602 F(21) 0(3) 3.12(2) 55401 F(21) C(3) 3.46(2) 75602 F(22) 0(1) 3.05(2) 65602 0(2) 0(5) 3.47(2) 64602 0(3) 0(4) 3.38(2) 75702 220 * The ADC (atom designator code) specifies the position of an atom in a crystal. The 5-digit number shown in the table is a composite of three one-digit numbers and one two-digit number: TA (first digit) + TB (second digit) + TC (third digit) + SN (last two digits). TA, TB and TC are the crystal lattice translation digits along cell edges a, b, and c. A translation digit of 5 indicates the origin unit cell. If TA = 4, this indicates a translation of one unit cell length along the a-axis in the negative direction. Each translation digit can range in value from 1 to 9 and thus ±4 lattice translations from the origin (TA = 5, TB = 5, TC = 5) can be represented. The SN, or symmetry operator number, refers to the number of the symmetry operator used to generate the coordinates of the target atom. A list of symmetry operators relevant to this structure are given below. For a given intermolecular contact, the first atom (origin atom) is located in the origin unit cell and its position can be generated using the identity operator (SN = 1). Thus, the ADC for an origin atom is always 55501. The position of the second atom (target atom) can be generated using the ADC and the coordinates of the atom in the parameter table. For example, an ADC of 47502 refers to the target atom moved through symmetry operator two, then translated -1 cell translation along the a axis, +2 cell translations along the b axis, and 0 cell translations along the c axis. An ADC of 1 indicates an intermolecular contact between two fragments (e.g. cation and anion) that reside in the same asymmetric unit. Symmetry Operators: (1) x, y, z; (2) -x, 1/2 + y, -z. 221 Appendix E Selected Structural Parameters and the Molecular Structure of [Hg(CO)2][Sb2Fn]2* Table E - l Selected Distances (A) and Angles (°) for [Hg(CO)2][Sb2Fn]2 Hg-C 2.083(10) Hg-F(12) 2.595(5) O-C 1.104(12) 1.126" Hg-F(22)* 2.691(4) Sb(l)-F(l) 2.008(4) 2.013" Sb(2)-F(l) 2.035(4) 2.042" Sb(l)-F(12) 1.899(5) 1.911" Sb(2)-F(22) 1.890(5) 1.899" Sb(l)-F(13) 1.844(6) 1.874" Sb(2)-F(23) 1.834(5) 1.846" Sb(l)-F(14) 1.836(6) 1.858" Sb(2)-F(24) 1.869(4) 1.880" Sb(l)-F(15) 1.833(5) 1.850" Sb(2)-F(25) 1.847(5) 1.857" Sb(l)-F(16) 1.845(5) 1.865" Sb(2)-F(26) 1.847(5) 1.864" C-F(12)c 2.895(10) C-F(13) 3.002(10) C-F(16)rf 2.749(9) C-F(22)* 2.758(9) C-F(24) 2.653(9) 0--F(14)e 2.823(10) 0--F(15)* 2.922(9) 0--F(23/ 2.840(8) Hg-C-O 177.7(7) F(22)*-Hg-C 69.24(23) F(12)-Hg-F(22)* 98.88(16) Hg-F(12)-Sb(l) 129.2(3) F(12)-Hg-C 104.4(3) Hg--F(22)*-Sb(2)* 157.16(25) F(l)-Sb(l)-F(12) 85.88(21) F(l)-Sb(2)-F(22) 85.54(20) F(l)-Sb(l)-F(13) 85.9(3) F(l)-Sb(2)-F(23) 82.99(21) F(l)-Sb(l)-F(14) 86.65(23) F(l)-Sb(2)-F(24) 84.30(20) F(l)-Sb(l)-F(15) 84.83(23) F(l)-Sb(2)-F(25) 84.53(21) F(l)-Sb(l)-F(16) 178.5(3) F(l)-Sb(2)-F(26) 178.92(22) F(12)-Sb(l)-F(13) 83.5(3) F(22)-Sb(2)-F(23) 90.18(22) F(12)-Sb(l)-F(14) 171.7(3) F(22)-Sb(2)-F(24) 169.58(21) 222 Table E-l Selected Distances (A) and Angles (°) for [Hg(CO)2][Sb2F11]2 (Continued) F(12)-Sb(l)-F(15) 89.8(3) F(22)-Sb(2)-F(25) 87.74(21) F(12)-Sb(l)-F(16) 93.53(24) F(22)-Sb(2)-F(26) 93.65(23) F(13)-Sb(l)-F(14) 92.4(3) F(23)-Sb(2)-F(24) 90.89(23) F(13)-Sb(l)-F(15) 168.9(3) F(23)-Sb(2)-F(25) 167.47(23) F(13)-Sb(l)-F(16) 95.4(3) F(23)-Sb(2)-F(26) 96.30(24) F(14)-Sb(l)-F(15) 93.1(3) F(24)-Sb(2)-F(25) 88.97(22) F(14)-Sb(l)-F(16) 94.0(3) F(24)-Sb(2)-F(26) 96.53(23) F(15)-Sb(l)-F(16) 93.8(3) F(25)-Sb(2)-F(26) 96.16(24) Sb(l)-F(l)-Sb(2) 147.6(3) * See Bodenbinder, M . ; Balzer-Jollenbeck, G.; Willner, H . ; Batchelor, R. J.; Einstein, F. B. W.; Wang, C ; Aubke, F. Inorg. Chem. 1996, 35, 82. a Distances corrected for thermal motion using a riding model. b3l2-x, 112+y, V2-z; c l-x, -y, -z; d-l+x, y, z; e2-x, -y, l-z; fl-x, -y, l-z; F(15) Figure E-l The Molecular Structure of [Hg(CO)2][Sb2Fn]2. 223 

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:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0059520/manifest

Comment

Related Items