Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A study of group 6 and 8 metal sigma-carbonyl cations and related metal(II) hexafluoroantimonates(V) Sham, Iona Hiu Tung 2002

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

Item Metadata

Download

Media
831-ubc_2002-750787.pdf [ 15.45MB ]
Metadata
JSON: 831-1.0061269.json
JSON-LD: 831-1.0061269-ld.json
RDF/XML (Pretty): 831-1.0061269-rdf.xml
RDF/JSON: 831-1.0061269-rdf.json
Turtle: 831-1.0061269-turtle.txt
N-Triples: 831-1.0061269-rdf-ntriples.txt
Original Record: 831-1.0061269-source.json
Full Text
831-1.0061269-fulltext.txt
Citation
831-1.0061269.ris

Full Text

A STUDY OF GROUP 6 AND 8 M E T A L S I G M A - C A R B O N Y L CATIONS A N D R E L A T E D METAL(II) HEXAFLUOROANTIMONATES(V) by IONA HJTJ T U N G S H A M B. Sc. Honours (Chemistry), The University of British Columbia, 1997 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF - DOCTOR OF PFIJLOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES CHEMISTRY We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A July 2002 © Iona Hiu Tung Sham, 2002 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. The University of British Columbia Vancouver, Canada Department DE-6 (2/88) A b s t r a c t Predominantly a-bonded carbonyl ligands appear to cause large ligand field splittings and all Fe rj-carbonyl cations studied in this work are low spin. The Fe cation, [Fe(CO)e]2+ (structurally characterized [Sb2Fn]" or [SbFe]" salt), is diamagnetic. [Fe(CO)6][Sb2Fn]2 was synthesized by oxidative carbonylation of Fe(CO)s with X e F 2 in HF-SbFs and further purified by treatment of the crude product with F 2 in FfF. This procedure was compared to previous methods of obtaining [Fe(CO)6][Sb2Fn] 2 using AsFs and CI2 as oxidizing agents in SbF5 as solvent. Magnetic susceptibility measurements were used to determine the amount of paramagnetic impurity, identified as Fe[SbFe]2, in the samples. The thermolysis behaviours of [Fe(CO)6][Sb 2Fii] 2 and [Fe(CO)6][SbF6]2 were investigated using Differential Scanning Calorimetry. The formation of [Fe(CO)6][Sb2Fn]2 and the by-product Fe[SbF 6] 2 via an intermediate of the form Fe(CO)4X2 (X = CI or F) was postulated. It was found that M(CO)e (M = Mo and W) undergoes facile 2-electron oxidation when treated with SbFs at 40 - 60°C. Cr(CO)6 is likely to behave in the same way in SbFs at lower temperatures (20 - 25°C). For Mo and W, seven-coordinate carbonyl complexes, [{Mo(CO) 4}2(cis-u-F2SbF4)3]x[Sb2Fi,] x and [W(CO)6(FSbF5)][Sb 2Fii], were isolated and structurally characterized. For Cr, a transient carbonyl species was observed. The magnetic behaviours of metal(II) hexafluoroantimonates(V), M[SbFe]2 (M = Mn, Fe, Co, Ni , Cu and Pd), in general, are as predicted for metal ions in weak ligand fields of octahedral or pseudo-octahedral geometry. The experimental room temperature magnetic moments of these compounds are much closer to us, than U L + S or Uj. This is an indication of quenching of orbital 11 angular momentum by ligand fields, and shows that the metals in these systems cannot be modeled as "free" or "naked" metal ions. Addition reactions involving pyrazine (pyz) or 4,4'-bipyridine (4,4'-bipy) and several M[SbF6] (M = Cr, M n , Fe, Co, Ni and Cu) salts yielded mixtures of [M(pyz) n]X 2 or [M(4,4'-bipy)n]X2 complexes (X = F" and / or [SbF6]"); these complexes have different numbers, n, of pyz or 4,4' bipy ligands. The formation of pyz.2SbFs in the reactions involving pyz was observed. The crystal structure of pyz.2SbF5 shows the molecule to have D 2 i , symmetry. iii Table of Contents Abstract ii Table of Contents iv List of Tables x List of Figures xiv List of Abbreviations and Symbols xix Acknowledgements xxii Chapter 1 General Introduction 1 1.1 Homoleptic Carbonyl Cations and Their Derivatives 1 1.1.1 The Synergistic Bonding Mode of CO to a Transition Metal 1 1.1.2 Syntheses of Carbonyl Cations of Transition Metals 9 1.1.3 Aspects of Carbonyl Cations of Electron-Rich Transition Metals 15 1.2 Metal(II) Hexafluoroantimonates(V) 15 1.2.1 Syntheses of the Metal(II) Hexafluoroantimonates(V) 16 1.2.2 Addition Reactions of the Metal(II) Hexafluoroantimonates(V) with Potentially Bridging Ligands 19 1.3 Magnetochemistry 22 1.3.1 Introduction 22 1.3.2 Bulk Magnetic Properties of Substances 24 1.3.3 The Fundamental Equation of Magnetism 25 1.3.4 Magnetic Moments 26 1.4 Collaboration and Cooperation 27 1.4.1 Collaboration with Professor H. Willner's Group 27 1.4.2 Cooperation with Other Group Members of Dr. F. Aubke 27 iv 1.5 Research Directions 27 1.6 Outline of this Thesis 29 1.7 References 29 Chapter 2 General Experimental 33 2.1 Chemicals 33 2.2 Apparatus .- 36 2.2.1 Utility Vacuum Lines 36 2.2.2 Reaction and Storage Vessels 37 2.2.3 Other Special Apparatus and Equipment 41 2.3 Instrumentation 44 2.3.1 Vibrational Spectroscopy 44 2.3.2 Magnetic Susceptibility Measurements 45 2.3.3 Differential Scanning Calorimetry 46 2.3.4 Atomic Absorption Spectroscopy 49 2.3.5 Single Crystal X-Ray Diffraction 49 2.3.6 Microanalyses 49 2.4 References 49 Chapter 3 Homoleptic Carbonyl Cations of Iron 51 3.1 Hexakis(carbonyl)iron(II) Undecafluorodiantimonate(V), [Fe(CO)6][Sb2Fn]2, and Hexakis(carbonyl)iron(II) Hexafluoroantimonate(V), [Fe(CO)6][SbF6]2 51 3.1.1 Introduction 51 3.1.2 Syntheses and Characterization 58 3.1.3 Identification of Fe[SbF6]2 as the Paramagnetic By-Product in [Fe(CO) 6][Sb 2Fn] 2 Samples 61 3.1.4 Possible Formation of Fe[SbF6]2 70 3.1.5 Identification of Other By-Products 80 3.1.6 Improved Syntheses and Purification of [Fe(CO)6][Sb2Fn]2 and [Fe(CO)6][SbF6]2 87 3.1.7 Discussion 90 3.2 The Hypothetical Compound "[Fe(CO)6][Sb 2F n] 3" 95 3.3 Summary and Conclusions 96 3.4 References 98 Chapter 4 Carbonyl Cationic Complexes of Molydenum, Tungsten and Chromium 101 4.1 Introduction 101 4.2 Poly-Tetrakis(carbonyl)tri(cis-m-hexafluoroantimonato(V)molybdenum(II) Undecafluoroantimonate(V), [{Mo(CO) 4} 2(cis-u-F 2SbF 4)3]x[Sb 2Fii] x 104 4.2.1 Experimental 104 4.2.2 Characterization 105 4.2.3 Discussion of [{Mo(CO) 4} 2(cis-u-F2SbF 4)3]x[Sb 2Fn]x 110 4.3 Hexakis(carbonyl)hexafluoroantimonato(V)tungsten(II) Undecafluoroantimonate(V), [W(CO)6(FSbF5)][Sb2Fi,] 112 4.3.1 Experimental 112 4.3.2 Characterization 114 4.4 The Transient Carbonyl Species of Chromium 121 vi 4.4.1 Experimental 121 4.4.2 Characterization 122 4.4.3 Isolation and Molecular Structure Redetermination of SbF3 123 4.5 Discussion of the Reactions of M(CO)6 ( M = Mo, W or Cr) with SbFs 125 4.5.1 Comparison between [{Mo(CO)4}2(cis-u-F 2SbF 4)3]x[Sb 2Fii] x and [W(CO) 6(FSbF 5)][Sb 2F„] 125 4.5.2 Comparison of the Reaction of Cr(CO)6 in Liquid SbFs to those of M(CO) 6 (M = Mo and W) 128 4.5.3 Comparison of [{Mo(CO) 4} 2(cis-u -F 2SbF4)3]x[Sb 2F u]x and [W(CO)6(FSbF5)][Sb2Fn] with Other Homoleptic Carbonyl Cationic Complexes 129 4.6 Summary and Conclusions 132 4.7 References 133 Chapter 5 Metal(II) Hexafluoroantimonates(V) 136 5.1 Introduction 136 5.2 Syntheses 138 5.2.1 Synthesis via Reaction of Metal(II) Fluorides with SbF 5 in HF 138 5.2.2 Synthesis via Solvolysis of Metal(II) Fluorosulfates in excess SbF5 140 5.2.3 Synthesis via Oxidation of Metal by SbF5 in S0 2 . 141 5.2.4 Synthesis of Cr[SbF 6] 2 via Reaction of Cr(CO) 6 and SbF 5 143 5.3 Characterization 144 5.3.1 Structure Determination 144 vii 5.3.2 Vibrational Analysis 147 5.3.3 Elemental Analysis via Atomic Absorption Spectroscopy 152 5.3.4 Magnetic Susceptibility Measurements 154 5.3.5 UV-VIS Spectroscopy 176 5.4 Discussion 176 5.4.1 The Relationship between the Isolable Metal Cations as [SbF6]" Salts and the Transient Metal Cations that React with CO to form Carbonyl Cations 176 5.4.2 Parameters 178 5.4.3 Magnetic properties of the metal(II) hexafluoroantimonates(V) 182 5.5 Summary and Conclusions 185 5.6 References 186 Chapter 6 Addition Reactions of the Metal(II) Hexafluoroantimonates(V) 189 6.1 Introduction 189 6.2 Experimental 190 6.2.1 Reaction of the M[SbF 6 ] 2 complexes (M = Cr, Mn, Fe, Co, Ni , Zn, Pd and Ag) with pyz 190 6.2.2 Reaction of Cr[SbF 6] 2 or M[SbF 6 ] 2 .yS0 2 (M = Mn, Fe, Co or Ni) with 4,4'-bipy 194 6.2.3 Study of by-products formed in reactions of M[SbFe]2 complexes with donor ligands pyz or 4,4'-bipy 195 6.3 Characterization, Results and Discussion 198 6.3.1 Elemental Analysis 198 viii 6.3.2 Vibrational Analysis 201 6.3.3 Magnetic Susceptibility Measurements. 205 6.3.4 Other Characterization Attempts 207 6.4 Isolation and Characterization of pyz.2SbF5 (C4H4N2.2SDF5) 208 6.4.1 Synthesis and Characterization 208 6.4.2 Discussion 210 6.5 Summary and Conclusions 216 6.6 References 217 Chapter 7 General Summary and Suggestions for Future Work 220 7.1 General Summary 220 7.2 Suggestions for Future Work 222 7.3 References 224 Appendix 1 Crystallographic Data for the Structure Determination of Various Complexes 225 Appendix 2 Experimental Details for Differential Scanning Calorimetric (DSC) Studies on [ F e t C O M S ^ F , , ] ! and [Fe(CO)6][SbF6]2 239 Appendix 3 Magnetic Properties of Free Metal(II) Cations and Weak Field Metal(II) Coordination Complexes 250 Appendix 4 Magnetic Data for Various Complexes 257 ix List of Tables Table 2. 1 Sources, properties and preparations of superacids and other solvents 33 Table 2. 2 Sources and properties of other chemicals used in the addition reactions of naked metal(II) hexafluoroantimonates(V) 34 Table 2. 3 Sources, properties and preparations of other chemicals used in the synthesis of metal carbonyl cations - [Sb2Fn]" or -[SbFe]" salts 35 Table 2. 4 Sources, properties and preparations of other chemicals used in the synthesis and characterization of naked metal(II) hexafluoroantimonates(V) 36 Table 3.1 Vibrational assignments of the observed and calculated wavenumbers (cm"1), of the fundamentals for [Fe(CO)6] in [Fe(CO) 6][Sb 2Fii] 2 56 Table 3.2 Results of the Atomic Absorption Studies 68 Table 3.3 A comparison between thermal events and approximate onset temperatures of [Fe(CO) 6][Sb 2F u] 2 and [Fe(CO)6][SbF6]2 75 Table 3.4 Raman data: different samples of [Fe(CO)6][Sb2Fu]2 prepared with various oxidizing agents compared with published data 82 Table 3.5 Raman data: "crude" crystalline [Fe(CO)6][Sb2Fn]2 bulk sample obtained from experiment, of single crystals of [Fe(CO)e][Sb2Fn]2 and published data of [Fe(CO)6][SbF6]2 and of 6SbF3.5SbF5 85 Table 3.6 Structural and spectroscopic parameters for [M(CO)6][Sb 2Fn] 2 and [M(CO) 6][SbF 6] 2 (M = Fe, Ru and Os) 91 Table 4.1 Vibrational data of [{Mo(CO) 4} 2(cis-u -F 2SbF4)3]x[Sb 2F u]x in the C-0 stretching region 107 Table 4.2 C-0 stretching wavenumbers of [W(CO)6(FSbF5)][Sb2Fn] 118 Table 4.3 Vibrational data of the product from the reaction of Cr(CO)6 with liquid SbFs under a CO atmosphere 122 Table 4.4 Comparison of bond angles and distances of SbF3 from determination by Edwards and the present work 125 Table 4.5 Selected bond distances and angles for [{Mo(CO)4}2(cis-u -F 2SbF4)3] x[Sb 2Fn] x and [W(CO) 6(FSbF 5)][Sb 2F„] 127 Table 5.1 Characteristics of the starting reagents, the reaction mixtures and the products for the reactions of metal(II) fluorides with SbFs in FIF (metal = Fe, Co, Ni , Cu and Mn 139 Table 5.2 Temperatures, times, weights and colour changes for the reactions of metal(II) fluorosulfates in excess SbFs (metal = Pd and Ag) 141 Table 5.3 The colour of the reaction mixtures and the products obtained for each metal when oxidized by SbFs in S 0 2 (metal = Mn, Fe, Ni and Co) 142 Table 5.4 IR and Raman data for Cr[SbF6]2.6SbF3.5SbF5, CrF 2.2SbF 5 and 6SbF3.5SbFs 148 Table 5.5 IR and Raman data for Mn[SbF 6] 2 (250 - 800 cm"1) 149 Table 5.6 IR and Raman data for Fe[SbF 6] 2 and FeF2.2SbFs (290 - 800 cm"1) 149 Table 5.7 IR and Raman data for Co[SbF 6] 2 ,"Co[F(SbF6)]" and CoF 2.2SbF 5 (270 - 800 cm"1) 150 Table 5.8 IR and Raman data for Ni[SbF 6] 2 and NiF 2 .2SbF 5 (270 - 800 cm"1) 150 x i Table 5.9 IR and Raman data for Cu[SbF 6] 2 and CuF 2.2SbF 5 (270 - 800 cm"1) 151 Table 5.10 IR and Raman data for Pd[SbF 6] 2 151 Table 5.11 Results of the Atomic Absorption Studies 153 Table 5.12 Variables used in the calculations of the parameters in equations 5.8 and 5.9 and the calculated values for each metal ion 180 Table 5.13 Variables used in the calculations of parameters in equations 5.11 and 5.12,and the calculated values for each metal ion 181 Table 5.14 Magnetic behaviours of the M[SbFe]2 compounds (M = Cr, Mn, and Fe) 182 Table 5.15 Magnetic behaviours of the M[SbF6]2 compounds (M = Co, Ni, Cu and Pd) 183 Table 6.1 Experimental findings for reactions involving the M[SbF6]2 (M = Cr, Mn, Fe, Ni, Zn, Pd and Ag) compounds and pyz in S 0 2 191 Table 6.2 Observations for reactions of Cr[SbFe] 2, Co[SbF6]2 and Fe[SbF6]2-yS02 with pyz in S0 2 , where the precipitates were filtered off before volatiles were removed 192 Table 6.3 Experimental findings for the layering reactions of M[SbFe]2 (M = Cr, Mn, Fe, Co, Ni and Cu) with pyz 193 Table 6.4 Observations for reactions of Cr[SbF 6 ] 2 , and M[SbF 6 ] 2 y S 0 2 (M = Mn, Fe, Co and Ni) with 4,4'-bipy in S 0 2 194 Table 6.5 Results of elemental analysis for products, formulated as [M(pyz)n][SbF6]2, from the reactions involving the M[SbFe]2 (M = Cr, Mn, Fe, Co, Ni , Zn, Pd and Ag) compounds and pyz in S 0 2 199 xii Table 6.6 The obtained and calculated C, H and N contents in precipitates from the reaction of the M[SbFe]2 compounds with 4,4'-bipy in S0 2 , formulated as [M(4,4'-bipy)n][SbF6]2 200 Table 6.7 IR data: Comparison of the precipitates from reaction of K[SbFe]with pyz in SO2 with K[SbF 6], pyz and pyz.2SbF5 (500 - 1500 cm"1) 203 Table 6.8 Raman data: Comparison of the precipitates from reaction of K[SbFe]with 4,4'-bipy in SO2 with K[SbFe], 4,4'-bipy and the product from reaction of 4,4'-bipy and SbF5 in S 0 2 (200 - 1650 cm"1) 204 Table 6.9 Comparison of experimental P 3 0 0 obtained for [ML n][SbF6]2 compounds with the range of P 3 0 0 for typical coordination metal complexes and p s 207 Table 6. 10 Vibrational data of pyz.2SbF5 209 Table 6.11 Selected experimental and calculated structural parameters for pyz and some of its adducts with various Lewis acids 212 Table 6.12 Experimental and calculated bond parameters for the anion [Sb2Fn]" (point group D 4 h ) and the adduct pyz.2SbF5 (point groups D 2 d and D2i,) 214 X l l l List of Figures Figure 1. 1 Molecular Orbital Diagram of CO 2 Figure 1. 2 The distribution of the metal carbonyl anions, the neutral metal carbonyls and the metal carbonyl cations 3 Figure 1. 3 Schematic representation of the synergistic bonding model for (a) metal carbonyl anions (b) neutral metal carbonyls and (c) metal carbonyl cations 4 Figure 1. 4 Bonding interactions in metal carbonyl cations with fluoroantimonate anions in the solid state 8 Figure 1. 5 Structures of (a) pyrazine (C4H4N2, denoted by pyz) and (b) 4,4'-bipyridyl ( C 8 H i 0 N 2 , denoted by 4,4'-bipy) 20 Figure 1.6 A simple representation of (a)antiferromagnetic and (b) ferromagnetic behaviours in 1-dimension 23 Figure 2. 1 (a) Tubular one-part storage container (b) Round-bottomed one-part Pyrex reactor 37 Figure 2. 2 Storage vessel for S0F5 -.38 Figure 2. 3 One-part filtration vessel 38 Figure 2. 4 One-part V-shaped Pyrex reactor 39 Figure 2. 5 PFA reactor (a) side view (b) top view 40 Figure 2. 6 V-shape PFA reac 40 Figure 2. 7 Glass trap 41 Figure 2. 8 Sublimation apparatus ... .42 Figure 2. 9 Ampoule key 43 xiv Figure 2. 10 Gas cell for IR measurements on gaseous samples 45 Figure 2. 11 P V C sample holder for magnetic susceptibility measurements 46 Figure 2. 12 DSC sample holder: Al-pan with pierced lid 47 Figure 2. 13 DSC sample holder: stainless steel, gold-plated sealed crucible 47 Figure 2. 14 Apparatus used to open the stainless steel, gold-plated sealed crucible 48 Figure 2. 15 Apparatus used to open the crucible to allow the gas phase into a vacuum line 48 Figure 3. 1 Molecular Stmcture of [Fe(CO) 6][Sb 2Fn] 2 53 Figure 3. 2 Molecular Structure of [Fe(CO)6][SbF6]2 53 Figure 3. 3 Selected significant interionic contacts in a formula unit of [Fe(CO>,][Sb2Fn]2 54 Figure 3.4 Selected significant interionic contacts in a formula unit of [Fe(CO)6][SbF6]2 55 Figure 3.5 Magnetic moments of various samples of [Fe(CO)6][Sb2Fn]2 and Fe[SbF 6] 2 62 Figure 3.6 Magnetic behaviour of purified [Fe(CO)6][Sb2Fn]2 compared to different models 65 Figure 3.7 Magnetic behaviour of [Fe(CO)6][SbF6]2, prepared from purified [Fe(CO)6][Sb2Fn]2, compared with different models 66 Figure 3.8 DSC Plot of [Fe(CO)6][Sb2Fn]2: Events from 25 - 250 °G- 71 Figure 3.9 The molecular structure of [Ir(CO)6][SbF6]3.4HF 72 Figure 3.10 DSC Plot of [Fe(CO)6][SbF6]2: Events from 25 - 300 °C 74 xv Figure 3.11 Proposed reaction pathway for the formation of Fe[SbF6]2 in preparations of [Fe(CO)6][Sb2Fn]2 78 Figure 3.12 Vibrational spectra of [Fe(CO)6][Sb2Fn]2 81 Figure 3.13 Vibrational spectra of [Fe(CO)6][SbF6]2 81 Figure 3.14 Raman Spectrum of crude [Fe(CO)6][Sb2Fn] 2 crystalline solids in the bulk sample 84 Figure 3.15 Raman spectrum of purified [Fe(CO)6][Sb2Fu]2 (after treatment with F 2 in HF) 88 Figure 3.16 Raman spectrum of [Fe(CO)6][SbF6]2 obtained from purified [Fe(CO) 6][Sb2F„]2 89 Figure 4.1 The repeat unit of the polymeric cation [{Mo(CO)4}2(cis-|a -F 2 SbF 4 ) 3 ] + 105 Figure 4.2 The stereoview of the unit cell of [{Mo(CO)4}2(cis-u-F 2 S b F 4 ) 3 ] x [ S b 2 F n ] x 106 Figure 4.3 Magnetic Behaviour of [{Mo(CO) 4} 2(cis-u - F 2 S b F 4 ) 3 ] x [ S b 2 F i i ] x from 2 - 300 K 109 Figure 4.4 A formula unit of [W(CO) f,(FSbF5)][Sb2Fii] 114 Figure 4.5 An idealized structure of [W(CO) 6(FSbF 5)]+ 115 Figure 4.6 Magnetic Behaviour of [W(CO)6(FSbF 5 )][Sb2F n ] from 2 - 300 K 119 Figure 4.7 The molecular structure of SbF3 124 Figure 5.1 Relationship between the structures of Ag[SbF6J2 and Pd[SbF6]2 146 x v i Figure 5.2 The experimental magnetic data of Cr[SbFe]2 compared to various models for octahedral Cr(II) complexes 156 Figure 5.3 Experimental magnetic moments of Mn[SbFe]2 compared to the Zero Field Splitting Models 158 Figure 5.4 The experimental magnetic moments of Fe[SbFe]2 compared to calculated moments from the Figgis Model of Single Ion Effects 159 Figure 5.5 Tetragonal elongation of octahedrally coordinated Fe in Fe[SbF6]2 161 Figure 5.6 The experimental magnetic moments of different samples of Co[SbFe]2 compared to calculated moments from the Figgis Model of Single Ion Effects 163 Figure 5.7 Experimental magnetic moments of Ni[SbF6]2 compared to the Zero Field Splitting Models with and without Molecular Field Correction 165 Figure 5.8 A comparison of experimental magnetic moments obtained here for Ni[SbFe]2 with values from the literature 167 Figure 5.9 Experimental magnetic data of Cu[SbF6]2 compared to the Curie Model with Molecular Field Correction and the Linear Chain Model 168 Figure 5.10 A comparison of experimental magnetic moments of Cu[SbFe]2 obtained here with values in the literature 170 Figure 5.11 Experimental magnetic moments of Pd[SbFe]2 compared to the Zero Field Splitting models 172 Figure 5.12 A comparison of experimental magnetic moments of Pd[SbF6]2 obtained here with literature values 173 Figure 5.13 The magnetic susceptibility and moment of Pd[SbFe]2, Pd(S03F)2 from present work and a previous study 175 xvu Figure 5. 14 The distribution of metals that form isolable M[SbF6]2 compounds, those that form carbonyl cations as -[Sb2Fn]' salts and the transient metal cations stabilized by CO 178 Figure 6.1 The experimental magnetic moment for products obtained from the reactions involving M[SbFe]2 complexes and pyz in SO2 206 Figure 6.2 The molecular structure of pyz.2SbFs 208 Figure 6.3 The [Sb 2F„]" anion in [Au(CO)2][Sb2Fn] 213 xviii List of Abbreviations and Symbols dco C-0 bond length v c o C-0 stretching wavenumber M W molecular weight b.p. boiling point m.p. melting point p density cm centimeters mm millimeters PFA polyfluoroacetamide PVC polyvinylchloride mL millilitre L litre mmol millimole mol mole mg milligram g gram mbar millibar TIP temperature independent paramagnetism Temp. temperature °C degree Celsius K degree Kelvin IR Infrared Spectroscopy f.Ra Raman Spectroscopy xix A A Atomic Absorption Spectroscopy ESR Electron Spin Resonance Spectroscopy S spin (j.s spin-only magnetic moment p. 3 0 0 magnetic moment at 300K j i B Bohr Magneton v wavenumbers fco C-0 stretching force constant Nm"1 Newton per meter H homogeneous magnetic field B magnetic induction M magnetic dipole moment per unit volume K volume magnetic susceptibility X g gram magnetic susceptibility X M or x molar magnetic susceptibility J exchange parameter D Zero Field Splitting parameter zJ Molecular Field Exchange parameter F value goodness of fit (of calculated values to experimental values) F = V[(l/N)xE{[(Xcalc-Xexp,)/Xexpt]2}] where N = number of data, Xcaic = calculated value of magnetic susceptibility and Xexpt= experimental value of magnetic susceptibility 5 splitting of the T orbital wavefunction xx C,„s free ion spin-orbit coupling constant X spin-orbit coupling constant of metal ion v distortion parameter, defined as 8 / X k derealization parameter or orbital reduction factor A parameter from the Figgis Model for Single Ion Effects that indicates the strength of crystal field splitting (1.5 is the weak field limit and 1.0 is the strong field limit) ref. reference eq. equatorial ax. axial br. bridging A angstrom HOMO highest occupied molecular orbital L U M O lowest unoccupied molecular orbital DFT Density Functional Theory xxi Acknowledgements I would like to express my deep gratitude to Dr. F. Aubke and Dr. R. C. Thompson, my research supervisors, for their inspiration and guidance throughout the years of my graduate study. Sincere thanks are due to our collaborator, Dr. H. Willner, who allowed me to work and learn at his laboratory in Duisburg, Germany during the summer of 2 0 0 0 . Here is a special thank you for Dr. M . Bodenbinder and Dr. R. Brochler, who passed on their expertise and unique experimental techniques to me. Thanks are also extended to all the members, past and present, of our research groups, for their enlightening discussions and many valuable suggestions. Dr. M . Fryzuk, Dr. M . Wolf and Dr. K. A. R. Mitchell are thanked for reading this manuscript. Gratitude are expressed to other members of the faculty and staff of this department for their assistance in various aspects of this project, in particular the staff at the X-ray lab, the microanalysis lab and glassblowing shop. I am indebted to Miss Lara Morello, Mrs. Erin MacLachlan, Mrs. Tracey Kennedy, Miss Carolyn Moorlag and Miss Cecilia Stevens for reading parts of this thesis, for their support, encouragement and pleasant friendship, as well as for all the wonderful discussions - scientific and non-scientific. Finally 1 would like to dedicate this thesis to my parents and my husband, Garrick, and thank them for all their love and support through all the years. xxii Chapter 1 General Introduction 1.1 Homoleptic Carbonyl Cations and Their Derivatives 1.1.1 The Synergistic Bonding Mode of CO to a Transition Metal Carbon monoxide is viewed as the most important and versatile ligand in organometallic chemistry of transition metals.1 CO can act as a monodentate or multidentate ligand to metals in various oxidation states. According to a model originally proposed by Dewar, Chatt and Duncanson,1 the metal-CO bonding consists of two components which interact synergistically: 1. a-donation from the HOMO, the 5a M O of CO to a metal orbital of a symmetry 2. 7t-backbonding from the metal atom into the L U M O , the 2K orbital, of CO. The Molecular Orbital diagram of free CO is shown in Figure 1.1. In general, homoleptic metal carbonyl complexes may be divided into three groups: 1. Neutral metal carbonyl complexes, M x (CO) n (x = 1,2,3...), where the oxidation state of the metal is 0. These metals are found primarily in groups 6 to 10. 2. Metal carbonyl anions, [M(CO)„]'"', where the oxidation state of the metal is -1 to -3. The metals are usually found in groups 4 - 9 , and may be mono- or polynuclear. 3. Metal carbonyl cations, [M(CO)„]m +, where the oxidation state of the metal range from +1 to +3. The metals are typically in groups 6 -12 . The distribution of these three types of metal carbonyl complexes is shown in Figure 1.2. 1 Groups in which metals form all three types of carbonyl complexes 4 5 6 7 8 " f ' 9 10 11 12 ~ i Ti v 1 ' J Cr Mn Fe Co Ni ; j Cu Zn Zr Nb ! Mo ! I Tc 1 i Ru Rh Pd i i Ag Cd Hf Ta w I " Re i L Os Ir Pt Au i j Hg Metal Neutral M e t a l Carbonyl Metal Carbonyl Anions Carbonyls Cations (groups 4 - 9) (groups 6-10) (groups 6-12) Figure 1.2 The distribution of the metal carbonyl anions, the neutral metal carbonyls and the metal carbonyl cations. Figure 1.3 3 provides a schematic representation of orbital overlap of the metal and a terminally bonded CO. The synergistic mechanism can be explained in this linear M-C-0 arrangement as follows:1 The dative o-donation from the 5a HOMO of CO to an empty o type orbital on the metal leads to a high electron density on the metal, unless the metal is a cation with a charge of +2 or higher. Thus, there is a tendency for the metal to reduce its electron density by donating electrons back towards the ligand. This is possible through interactions between the it-symmetry metal orbital and the antibonding 2lt LUMO's of CO. 3 (a) Metal Carbonyl Anions v(CO) < 2080 cm" 1 (predominantly 71-bonded) o 5-(b) Neutral Metal Carbonyls 2080 cm"1< v(CO) < 2180 cm •1 o (c) Metal Carbonyl Cations v(CO) > 2180 cm" 1 (predominantly a-bonded) o Figure 1.3 Schematic representation of the synergistic bonding model for (a) metal carbonyl anions (b) neutral metal carbonyls and (c) metal carbonyl cations.3 The relative thickness of the arrows indicates the relative strength of the bond. In Figure 1.3, the partial charges on the CO ligands are shown. Due to these charges, interionic contacts are directed towards the O atom in carbonyl anions and the C atom in carbonyl cations.3 The limiting cases are highly reduced metal carbonyl anions with dominating n-backbonding and the metal carbonyl cations where the CO ligands are predominately a-bonded. The synergic bonding in zero valent metal carbonyls may tend either way depending on the oxidation state of the metal and its position in the d-block.3 One important point to note is that in this bonding representation, neither a "71-only" nor a "a-only" bonding should exist. The A l -and a- components are regarded as complementary and reinforce each other. Observable changes in C-0 bond length, dco, and high CO bond orders are very small, while the C-0 stretching wavenumber, v c o , is very sensitive to these changes. Therefore, vibrational spectroscopy, in addition to X-ray crystallography, is an important structural tool in the characterization of metal carbonyls.4 Moreover, the CO stretching vibrations of most CO derivatives are intense bands that fall into an uncluttered spectral region, whereas the M-C bands are occasionally of low intensity and fall into a more cluttered region of the vibrational spectrum.1'5 1. Neutral Metal Carbonyl Complexes Neutral metal carbonyls were the first known type of carbonyl complexes. They are formed by metals in the middle of the d series (Groups 6 to 10),3 as the low formal charge on the metal (normally 0) is needed to optimize the synergic bonding. The molecular geometries of neutral mononuclear complexes are readily predictable from VSEPR Theory. They are usually octahedral, trigonal-bipyramidal or tetrahedral, M(CO) n . 3 There is generally strict observation of the 18-electron or the Effective Atomic Number (EAN) rule, with V(CO)6 a rare exception with its electron count of 17.6 V(CO)6 is also the only paramagnetic metal carbonyl complex. 5 Examples in this group are M(CO) 6 (M - Cr, Mo, W), M(CO) 5 (M = Fe, Ru, Os) and M(CO) 4 (M = Ni , Pd and Pt).8*9 Many neutral, polynuclear homoleptic metal carbonyls are also known, such as Mn 2(CO)io, Co2(CO)g and Ir4(CO)i2, which all have M - M single bonds and mostly with terminal CO ligands.1 As M - C 7t-backbonding weakens the C - 0 bond, the C - 0 stretch usually shifts from 2143 cm"1 for free C O 7 down to 2125 - 1850 cm"1 for terminal monodentate C O . 5 The thermal stability of these compounds also varies. 2. Metal Carbonyl Anions The preparation of highly reduced metal carbonyl anions by reduction in basic solvents is one of the more recent developments beyond the typical mononuclear homoleptic carbonyl complexes.10 Their general formula is [M(CO)„]m" (M = metal from groups 4 to 9, m = 2 to 4, n = 3 to 6). Their geometries are usually similar to those of neutral carbonyls (octahedral, trigonal-bipyramidal and tetrahedral),3 but trigonal-planar anions of Co, Rh and Ir (n = 3, m = 3) have also been described.10 In the low-valent cases, these highly reduced metal carbonyl anions obey the 18-electron rule. With the high electron density at the metal and strong 7i-back-bonding, the C - 0 stretch usually shifts down to 1750 - 1400 cm"1 for the 2- to 4- anions.10 These anions have mainly been studied by vibrational spectroscopy, but some structures are known as wel l . 1 0 3. Metal Carbonyl Cations Another recent synthetic development in this area is the preparation of homoleptic carbonyl cations of electron-rich metals in superacid media. 3 ' 1 1" 1 3 The general formula of these cations is [M(CO)„]m + (M = metal from groups 6 to 12, m = 1, 2, 3; n - 2, 4, 6, corresponding to linear, square planar and octahedral coordination).3 The carbonyl cations form stable salts with 6 [SD2F11]" or [SbFe]" counteranions; these anions are extremely weak nucleophiles which are still capable of stabilizing electrophilic cations. The carbonyl cations in groups 10 - 12 do not obey the 18-electron rule, e.g. [Hg(CO) 2] 2 + has 14 electrons,14 and [M(CO) 4 ] 2 + , where M = Pd or Pt has 16 electrons.15 The linear and square planar geometries of the above examples are unprecedented in metal carbonyls, but the corresponding cyanide complexes are known. For this type of CO complex, the C-0 stretches are usually higher than that of free CO (2143cm"1).7 It is generally believed that this increase in the C-0 bond force constant is due, in large part, to polar bond contribution, a bond polarization of the CO bond induced by M n + , and to a minor degree the removal of electron density from the slightly antibonding 5a HOMO of C O . 5 ' 1 6 Without the 7t-back-bonding compensation of electron derealization from the metal, the M-C-O arrangement approaches 8+5+8-. This polar interaction makes the M-C bond relatively weak and the C-0 bond very strong. In addition, for the predominately a-bonded carbonyl cations, interionic interaction competes with 71-backbonding. Significant C--F interactions between C of the CO ligand and F of the counteranion [Sb2Fn]" or [SbFe]" involve external electron donation into the n* molecular orbital of C O . 1 3 A schematic representation of the solid-state bonding interactions in metal carbonyl cations with fluoroantimonate anions is shown in Figure 1.4.13 Since the metal carbonyl cations are complexes of electropositive carbon, • They are formed and exist only in superacid solutions, or, stabilized by superacid anions in the solid state. • They form extended structures in the solid state by C--F interionic contacts. • The M-CO bond approaches a pure a-bond, because 71-back-donation becomes insignificant when the oxidation state of M is +2 or +3. 7 • They form derivatives, when one or more CO ligands are replaced by a uni-negative anion such as F", CI" SO3F" and SbFe" etc. anion n = 1 to 5 cation m = 1 to 3 11= Hg, Au, Pel, Pt, Rh, Ir, Fe, Ru, Os, Mo or W Figure 1.4 Bonding interactions in metal carbonyl cations with fluoroantimonate anions in the solid state.13 At low oxidation states of the metal, e.g. [Au(CO)2] +, only very weak interionic interactions are expected due to the presence of appreciable 7t-backbonding. In this case, the counteranion [SbiFn]" should have D 4 h symmetry and behave like a spectator ion. As the oxidation state of the metal increases to +2 or +3, 71-back-donation becomes negligible and a-donation becomes dominant, e.g. for [Hg(CO)2] and [Ir(CO)6J . Simultaneously stronger and more interionic C--F contacts are found. 1 3 ' 1 4 The dioctaheral [SbaFn]" anion distorts, usually towards Ci symmetry by bending about the Sb-F-Sb bridge and rotation of the equatorial SbF 4 groups, to 8 bond polarization facilitate the formation of strong C--F contacts. Some limitations are expected for octahedral complexes, however, due to steric reasons.13 The ligand field strength of a-bonded CO, with reduced 7t-backbonding, is unknown in classical coordination chemistry, because the high ligand field strength of 7t-acceptor ligands is caused by the Tt-accepting ability of the ligand. Since a-bonded metal carbonylcations exist only in superacids like HF-SbFs, it has not been possible to record ligand field spectra and to arrive at an unambiguous assignment for any of the cations. [Fe(CO)6] is the only thermally stable cation in the first row (3d) transition series where lODq is considerably smaller than for 4d and 5d metals (see Chapter 3). Therefore it is reasonable to assume the existence of metal carbonyl cations with paramagnetic ground states. 1.1.2 Syntheses of Carbonyl Cations of Transition Metals One of the subjects of investigation in this work is the metal carbonyl cations of Fe, Cr, Mo and W. The details of synthetic procedures and characterization for these specific compounds are discussed in Chapters 3 and 4. In this chapter, the general aspects of syntheses of homoleptic metal carbonyl cations of electron-rich metals in superacid media are presented. 1. General Syntheses The generation of these homoleptic metal carbonyl cations in superacid media is accomplished by: 3 i . Solvolysis and addition of CO to highly reactive metal cations i i . Reduction and carbonylation of metal salts in a higher oxidation state, using CO as a reducing agent and ligand i i i . Oxidation of neutral mono- or multinuclear metal carbonyls or derivatives in the presence of CO 9 Since the homoleptic carbonyl cations of electron-rich metals of the general type [M(CO) n ] m (n = 2, 4, 6; m = 1, 2, 3) exist almost exclusively with [Sb2Fn]~ or [SbF6]~ as counteranion, liquid SbFs is the reaction medium or one of the media of choice.3 /. Solvolysis and Carbonylation In this case, the oxidation state of the metal remains the same during the reaction. An example is the synthesis of [Hg(CO)2][Sb 2Fii] 2: 1 5 Hg(S0 3 F) 2 + 2CO + 8 S b F 5 -> [Hg(CO) 2 ][Sb2Fii] 2 + 2Sb 2F 9(S0 3F) (eq. 1.1) The counteranion [Sb2Fn]" is produced during solvolysis. The by-product Sb 2F9(S0 3F) is sufficiently volatile to be easily removed from the product under dynamic vacuum.3 In general, solvolysis and carbonylation allows simple and facile product isolation and gives quantitative yields, but the most suitable precursors are not commercially available.1 3 ii. Reductive Carbonylation In this case reduction of the metal salt happens during solvolysis. CO acts as both the reducing agent and the ligand whereas SbFs becomes the reaction medium as well as the source of counteranion [Sb2Fn]".3 In the synthesis of [Pt(CO)4][Sb2Fn]2, 1 5 all the by-products are volatile and can be separated readily: Pt(S0 3F) 4 + 5CO + 8 S b F 5 -> [Pt(CO) 4][Sb 2Fn] 2 + C 0 2 + S 2 0 5 F 2 + 2Sb 2F 9(S0 3F) (eq. 1.2) In this reaction, C 0 2 and S2O5F2 are formed from the decomposition of CO(S0 3F)2. Reductive carbonylation of metal fluorosulfates and fluorides is both elegant and versatile, but the most suitable precursors are, again, not commercially available.1 3 10 /'//. Oxidative Carbonylation Oxidative carbonylation is a more recent development in the preparation of metal carbonyl cations. This method is used to synthesize [M(CO) n ] m + salts for metals of groups 6, 8 and 9, such as M = Fe, Mo, W, Rh, Ir etc. Addition of CO is accompanied by the oxidation of neutral metal carbonyl or derivatives by an external oxidizing agent, or by SbFs, which also serves as the source of counteranions [Sb2Fn]" or [SbF6]". One external oxidizing agent that can be used for the synthesis of [Fe(CO)6][Sb2Fn]2 is XeF 2 , and the reaction medium in this case is HF-SbF5. The details of synthesis and characterization of [Fe(CO)6][Sb2Fn]2 1 7 and related compounds are reported in Chapter 3. The choice of reaction medium of HF-SbFs over liquid SbFs is also discussed in the next section. An advantage of oxidative carbonylation over the two methods previously mentioned is the use of commercially available precursors. Nevertheless, this advantage is partly offset by the requirement of unusual reagents such as XeF 2 or AsF 5 , the necessity of product isolation by crystallization and the experimental difficulties associated with working in SbFs or FLF-SbFs.1 3 Other examples of homoleptic metal carbonyl cations that are prepared from the oxidation by SbF5 are [{Mo(CO)4h(cis-u.-F2SbF4)3]x[Sb2Fu]x 1 8 and [W(CO)6(FSbF 5)][Sb 2F„] 1 9 (see Chapter 4). 2. Superacids as Reaction Media The most widely accepted definition of a superacid is any Bransted acid system that is stronger than 100% sulfuric acid or any Lewis acid system that is stronger than aluminum chloride.1 2 SbF5, HF-SbFs, HSO3F and HF-HSO3F are all superacids that are used as reaction media for homoleptic metal carbonyl cations of electron-rich metals. 11 /. Antimony Pentafluoride, SbFs SbFs, a good oxidizing and fluorinating agent, is a liquid between 7 and 149.5°C. 1 2 Therefore, it is particularly favoured as a reaction medium among other Lewis acidic fluorides. 3 ' 2 0 It has a polymeric structure with cis-F-bridges as confirmed by 1 9 F N M R , 1 2 and is very viscous. SbFs is also a component of two important conjugate superacid systems, magic acid, HSC^F-SbFs and fluoroantimonic acid, H F - S b F s . 1 1 ' 1 2 ' 2 0 These conjugate superacid systems, as well as SbFs, are used widely in the generation of many carbocations and inorganic cations. They are also sources of the counteranions [SbFe]" or [Sb2Fn]" in these cases.20 As SbFs is widely regarded as the strongest Lewis acid, 1 2 [SbF6]" and [Sb2Fn]" are conversely extremely weak nucleophiles capable of stabilizing very electrophilic cations. There are a number of advantages to using SbF 5 as a reaction medium: • Carbonylation reactions in liquid SbF5 require only mild reaction conditions. Only moderate temperatures (20 - 80 °C) and CO pressures (0.5 to 2 atm) are required.20 Handling is easy as glass reactors and vacuum lines can be used. • Carbonylation reactions are favoured in liquid SbFs as CO can act as both the ligand and a 2-electron reducing agent in SbFs. Moreover, there is no evidence of free or coordinated CO forming any stable complex with SbFs. 2 0 • SbFs can act as the oxidizing agent in addition to being the reaction medium at the same time. • Stabilization by secondary interionic contacts is very important for cations with electrophilic sites, such as the homoleptic metal carbonyl cations of electron rich metals, [Hg(CO)2][Sb2F1 1]2 1 4 and [M(CO)4][Sb2Fii]2 (M = Pt and Pd) . 1 5 The counteranions [Sb2Fn]" and [SbFe]" generated from SbFs are very efficient in facilitating such secondary contacts. In the metal carbonyl [Sb2Fn]" salts, secondary contacts to the 12 electropositive C-atom of the carbonyl group appear to produce extended, 3-dimensional networks. This raises the lattice energy of the solid to such an extent that CO release would become an unfavourable process.20 There are, however, some disadvantages in using SbFs as the reaction medium. Since it is a poor ionizing solvent and a viscous liquid, the starting materials and products of many carbonylation reactions are merely suspended rather than dissolved.3 Also, the oxidizing ability of SbFs can sometimes be an unfavourable aspect. Some disadvantages to its uses as sole reaction medium are: • These heterogeneous reactions are very time consuming, with reaction times varying from days to weeks. • Single crystal growth is extremely difficult. • By-products can form from partial reduction of SbFs in the oxidative carbonylation reactions even when a stronger external oxidizing agent is used. Whereas pure SbF5 offers these problems in the carbonylation reactions, many of these disadvantages disappear when the corresponding conjugate acids are used as the reaction media. ii. Fluoroantimonic Acid (Hydrogen Fluoride - Antimony Penlqflnoride), HF-SbFs HF-SbFs is the strongest liquid Bransted-Lewis superacid and has the widest acidity range. The acidity of HF is increased by an addition of even a small amount of SbF 5 . 1 2 The predominant species in a dilute solution of HF-SbFs at higher concentration of SbFs are: SbF5(soiv) + 2HF U SbF 6- ( s o l v ) + H 2 F + (eq. 1.3) At lower SbFs concentration, the major anionic species is [Sb2Fn]":12 2SbF5(Soiv) + 2HF Sb 2 F,,- ( g o l v ) + H 2 F + (eq. 1.4) 13 Since HF is a good ionizing solvent, the solubility of starting materials and products of a number of carbonylation reactions is much higher in HF-SbFs than in pure SbF 5. Therefore, the reaction phase is homogeneous; the reaction time is thus shortened, and single crystal growth becomes possible. Moreover, reactions can often be carried out at even lower temperatures than in SbFs. In a few cases, the yield is also improved when a reaction is carried out in HF-SbFs rather than in pure SbFs. 1 7 The formation of by-products due to partial oxidation of SbFs can, in some cases, be reduced. An obvious drawback of using HF-SbFs is that metal and/or fluoro-plastic reactors as well as vacuum lines made of metal must be used since HF attacks glass. The syntheses of [Fe(CO)6][Sb2Fii]2, [{Mo(CO)4}2(cis-u-F2SbF4)3]x[Sb2Fii]x 1 8 and [W(CO) 6(FSbF 5)][Sb 2Fi,] 1 9 illustrate these points (see Chapters 3 and 4). iii. Alternate superacid media and solvents Fluorosulfuric acid, HSO3F and magic acid, HS03F-SbF 5, are other superacid media commonly used in the synthesis of homoleptic metal carbonyl cations [Sb2Fn]~ or [SbFs]" salts. As with HF-SbFs, the solubilities of the reaction mixtures are higher in HSO3F and HSOsF-SbFs than in pure SbFs. Magic acid also allowed the growth of single crystals and thus the determination of structures for metal-carbonyl cations such as [Hg(CO)2][Sb2Fn]2.14 Apart from superacid media, only a few other solvents find use in the synthesis and characterization of some transition metal carbonyl cations. However, these solvents are usually only suitable for one or a few particular compounds. Liquid sulfur dioxide, SO2, is one such solvent. [Au(CO) 2][Sb 2F,i] 2 1 is very soluble in liquid S0 2 , whereas [M(CO)4][Sb2F,,]2 (M = Pt and Pd) 1 5 are very insoluble. [Hg(CO)2][Sb2Fn] 1 4 decomposes in SO2 with evolution of CO. No substitution of CO by weakly coordinating S 0 2 has been reported for 14 [Au(CO)2][Sb2Fn], while the carbonyl compound was transformed into [Au(NCCH3)2][SbFe] in the donor solvent acetonitrile.21 SO2 also forms a 1:1 adduct with SbF 5 , 2 2 which is sometimes present as a by-product in the product mixtures. 1.1.3 Aspects of Carbonyl Cations of Electron-Rich Transition Metals Homoleptic metal carbonyl cations with paramagnetic ground states are unknown in metal carbonyl chemistry.1-3'1 0 At the beginning of this project, cations of the type [Fe(CO)e] n + (n = 2 or 3) seemed to be the best candidates. Thus one of the objectives is to improve the synthesis of the Fe carbonyl cations in terms of yield and purity. Another is to use magnetic susceptibility measurements along with other characterization techniques such as vibrational studies, x-ray crystal structure determination and DSC studies to fully characterize these compounds. This work completes the full characterization of [Fe(CO)6][Sb2Fn]2 and [Fe(CO)6][SbF6]2,17 to date the two most extensively studied homoleptic carbonyl cations of a transition metal. Studies of magnetic behaviours are also used to probe the electronic structures of two other carbonyl cations, [{Mo(CO)4}2(cis-u-F2SbF4)3]x[Sb2F„]x 1 8 and [W(CO) 6(FSbF 5)][Sb 2F 1i]. 1 9 These results are presented in Chapters 3 and 4. 1.2 Metal(II) Hexafluoroantimonates(V) In this work, "naked" metal cations are defined as those cations in binary combination with very weakly nucleophilic anions in the solid state or in weakly coordinating solvents. There are different formulations of metal complexes in the form of metal(II) hexafluoroantimonates(V), as there is frequently a lack of definitive structural information about these compounds.23 The two most generally accepted formulations are M[SbF6J2, with a divalent metal cation in association 15 with the hexafluoroantimonate anions [SbFe]";2 4"2 7 and MF 2.2SbFs, a SbFs adduct of the divalent metal fluoride.2 6 The initial incentive to study such systems stemmed from earlier work which showed that the M[SbF6]2 compounds are possible transients and by-products in the preparation of the homoleptic metal carbonyl salts, [M(CO)n][Sb 2Fu]m or [M(CO) n][SbF 6]m (e.g. M = Fe in the 2+ oxidation state, n = 6, m = 2 , 1 7 details in Chapter 3). Thus it became necessary to magnetically characterize these impurities in order to interpret the magnetism and explain the formation of the carbonyl cations. In this work the formulation M[SbF 6 ] 2 is used, as this typeof compound is compared to the [M(CO)n][Sb2Fii]m or [M(CO)„][SbF6]m complexes. 1.2.1 Syntheses of the Metal(II) Hexafluoroantimonates(V) Divalent metal hexafluoroantimonates are synthesized by four general methods:23 1. Reaction of a metal(Il) fluoride with SbFs in HF or SO2 2 6 2. Reaction of a metal(II) fluorosulfate in excess SbFs 2 4 3. Oxidation of a metal by SbFs in S 0 2 2 5 4. Direct fluorination of a metal with F 2 in the presence of SbFs /. Reaction of Metal(II) Fluoride with SbFs in HF or SO2 When anhydrous M F 2 (M = Mg, Cr, Fe, Co, Ni, Cu, Ag, Zn, Cd or Pb) is mixed with SbFs in anhydrous HF or SO2 at room temperature, a crystalline binary compound formulated as MF 2-2SbF 5 by Gantar et. al is obtained. 2 3 ' 2 6 An obvious advantage to this procedure is that the products are obtained as crystalline solids, and the growth of single crystals suitable for X-ray structural determination is promising, as for the case of A g : 2 6 16 HF or S0 2 M F 2 + 2SbF5 > MF 2 .2SbF 5 (eq. 1.5) (M = Mg, Cr, Fe, Co, Ni , Cu, Ag, Zn, Cd or Pb) It was indicated that these compounds exhibit more bands in the IR and Raman spectra than those expected for an isolated [SbFe]" in the Oi, point group. 2 6 One possible problem is that the M F 2 lattice may be incompletely broken up under the reaction conditions and thus forms M[F(Sb 2Fn)]. This complex is a structural isomer of M[SbF 6] 2 , and hence not distinguishable by elemental analysis.24 Moreover, moving from ionic [SbFe]" to coordinated [SbF6]* in these compounds, the vibrational spectra get more complex. Although this procedure may or may not lead to a mixture of products such as M[Fn(SbF6)2.n] and others discussed above, the intended product is M[SbF 6] 2 . The alternative formulation M[SbF6]2 will be used hereafter to denote all metal(II) hexafluoroantimonates(V) including compounds obtained from this preparation. 2. Reaction of Metal(II) Fluorosulfate in excess SbFs The solvolysis of M(S03F) 2 (M = Sn, Ni, Pd, Cu, Ag or Zn) in excess SbF5 yields the corresponding M[SbF 6 ] 2 according t o : 2 4 ' 2 8 M(S0 3 F) 2 + 6SbF5 2 5 ' 6 0 ° C > M[SbF 6 ] 2 + 2Sb 2 F 9 S0 3 F (eq 1.6) As mentioned earlier, Sb 2F9S0 3F is very volatile and can be readily removed in dynamic vacuum. The downside of this clean route to M[SbF 6 ] 2 is the long reaction time required (10 -14 days) due to the limited solubility of M(S0 3 F) 2 in SbF5. 3. Oxidation of Metal by SbFs in SO2 This method involves the oxidation of a metal (Mn, Fe, Co or Ni) by SbFs in S 0 2 . 2 5 For M = Mn, Fe or Ni: 17 s o 2 6M + 23SbF5 > 6M[SbF 6] 2 + 6SbF3.5SbF5 (eq. 1.7) It has also been reported that Co[F(SbF6)] was obtained from the following reaction:25 S 0 2 6Co+17SbF 5 > 6Co[F(SbF6)] + 6SbF3.5SbF5 (eq. 1.8) Results from vibrational analysis and magnetic susceptibility measurements in this work, however, suggest that Co[SbFe]2 was actually obtained from this preparation (see Chapter 5 for details). An advantage of using SbFs as the oxidizing agent is that the Sb(V)/Sb(III) couple is not strong enough to oxidize metals to beyond the +2 state.24 However, quantitative separation of the by-product 6SbF3.5SbFs and the excess reagent metals from the intended products M[SbF6J 2 can be difficult. Furthermore, metal fluoride, M F 2 , or mixed fluoride-hexafluoroantimonates of the type M[Fn(SbF6)2.n] may also be produced.2 4 4. Direct Fluorination of Metal with F2 in the presence of SbFs The direct fluorination of a metal with F 2 in SbFs to obtain the metal(II) hexafluoroantimonate(V) is restricted only to N i . 2 3 ' 2 9 Ni[SbFe]2 can be obtained as: ^T. „ ^ 270°C. 250 atm ^ T . r „ , „ , Ni + 2SbF5 + F 2 > Ni[SbF 6] 2 (eq. 1.9) An obvious advantage to this route is that there are no by-products. Nevertheless, as higher oxidation states are accessible, direct fluorination may result in oxidation beyond the +2 state.23 The metal cations in these metal(II) hexafluoroantimonates(V) should be in solid state environments that approximate the free, or naked state. Therefore, the study of their magnetic properties is not only essential to the understanding of the corresponding metal carbonyl cations, but also interesting in the comparison of the magnetism to what is expected 18 theoretically for "free" metal cations. These studies are shown and discussed in detail in Chapter 5. 1.2.2 Addition Reactions of the Metal(II) Hexafluoroantimonates(V) with Potentially Bridging Ligands Another aspect of interest in this work is the addition of potentially bridging ligands, pyrazine (1,4-diazine) and 4,4'-bipyridyl, to the metal(II) hexafluoroantimonates(V). Variable temperature magnetic susceptibility measurements show that for many first row transition metals, the M[SbF6]2 compounds are paramagnetic. Some of these compounds even show very weak magnetic interactions between the metal centres. The potential for magnetic exchange interactions between metal cations being propagated via the bridging pyrazine or 4,4-bipyridyl systems constructed from the metal hexafluoroantimonates is of great interest, as is the engineering of one- (1-D), two- (2-D) and three-dimensional (3-D) extended structures.3 0'3 1 /. Pyrazine and 4,4 '-bipyridyl as Potentially Bridging Ligands Pyrazine (C4H4N2, denoted as pyz) and 4,4'-bipyridyl (CgHioN2, denoted as 4,4'-bipy) are both linear, bifunctional ligands. Their structures are shown in Figure 1.5. Both of these ligands are good candidates for molecular building blocks due to their rod-like rigidity and length. 3 2 For polymeric structures reported that contain pyz and 4,4'-bipy, it was found that factors like the choice of solvents, the specific counteranion, the oxidation state and the coordination preference of the metal cation can influence the specific framework structure of the product system (open structures vs. interpenetrated lattices).32 For example, there is a dramatic difference between the double-layer structure of the solvent-inclusion compound [Ag(pyz)2][Ag2(pyz)5](PF6)3-2CH2Cl2 and the layered structure of the solvent-free compound 19 [Ag(pyz)2][PFe]. 3 2 - 3 4 Changing the counteranion from [PFe]" to [SbF6]" also change the structure from the double layer Ag-pyz-PF 6 to the 3-D non-interpenetrating cubic framework of [Ag(pyz)3][SbFg]. 3 2 ' 3 4 In terms of oxidation states and coordination preference of the metal, it can be seen that C u 2 + normally coordinates as a distorted octahedron, and forms square 2-D networks with pyz or interpenetrated 2-D networks with 4,4'-bipy, while C u + can take on various coordination modes.3 2 /= \ / = \ / = \ N N N />—(\ N (a) pyz (b) 4,4'-bipy Figure 1.5 Structures of (a) pyrazine (C4H4N2, denoted by pyz) and (b) 4,4-bipyridyl (C8H10N2, denoted by 4,4-bipy) 2. Metal(II) Hexafluoroantimonates(V) as Building Blocks The advantage to using metal(II) hexafluoroantimonates(V) as building blocks for the construction of extended systems is the exclusion of incomplete substitution leading to mixed ligand complexes, since there are effectively no ligands being replaced. The blocks of metal cations can simply be linked by the bridging ligands in chains or even more extended structures. In the hope of generating extended cationic lattices, a comparison of the magnetic behaviour of the product to that of the starting materials, the M[SbF6J2 compounds, should be interesting. The study of magnetic exchange interactions in such systems is a topic of considerable current interest and thus, the structures and magnetic properties of new solid state materials produced through this chemistry are thoroughly investigated. The studies and the results are discussed in detail in Chapter 6. 20 3. Addition Reactions of Metal(II) Hexafluoroantimonates(V) with Pyrazine or 4,4 -Bipyridyl Addition reactions of M[SbF 6]2 (M = Cr, Mn, Fe, Co, N i and Cu) with pyz or 4,4'-bipy were studied using two preparative methods, in a weakly coordinating solvent or via layering of one solution over another. i. Addition Reactions in a Weakly Coordinating Solvent In early attempts to prepare extended lattices from the M[SbFs]2 compounds and pyz or 4,4'-bipy, the reagents were simply mixed in SO2. In a weakly coordinating solvent such as SO2, the metal cations should be only very weakly coordinated, and the addition of conventional coordinating ligands such as pyz and 4,4'-bipy should be facile and clean. However, experiments show that these reactions are not so straightforward. The situation is complicated by the reactions between the ligands and the anion [SbF6]~. Elemental analysis shows that mixtures of products with different number of ligands for each metal are obtained. A by-product, C 4 H 4 N 2 . 2 S D F 5 , has also been identified for the reactions involving pyz. The synthesis and characterization of the products, along with the identification of the by-product are reported in Chapter 6. ii. Layering of one Solution over another It was found in this work that crystalline solids of the type [M(C4H4N2)n][SbF6]2 could be obtained by layering an anhydrous ethanol solution of M[SbF6]2 over a dichloromethane solution of pyz. 21 For both of these preparative methods, attempts were made to grow crystals of the products. Crystal structures of the products are determined whenever possible. Variable temperature magnetic susceptibility measurements on the products were made and the results were analyzed. Vibrational analysis was also performed. These results and discussion are presented in Chapter 6. 1.3 Magnetochemistry Magnetic susceptibility measurements are used in this work as tools for studying the magnetic behaviour of the homoleptic carbonyl cationic complexes and metal(II) hexafluoroantimonates(V). This section serves as a brief introduction to magnetochemistry as it relates to the characterization of complexes prepared in the present work. 1.3.1 Introduction Magnetochemistry is the study of the effects of applied magnetic fields on samples. It probes the ground electronic state of the complex being studied and is complementary to electronic spectroscopy, which probes the excited electronic states. The primary interest of magnetochemistry is the study of paramagnetic materials. 1. A Qualitative Description of Magnetism Diamagnetism is the response of moving paired electrons in an atom to the application of an external magnetic field; whereas paramagnetism arises from the angular momentum of unpaired electrons. The angular momentum of an electron comes from two sources, the orbital angular momentum, L, and the spin of the electron, S. At the microscopic level, the spin and orbital 22 motions of the electron generate microscopic magnetic fields. Therefore, the electron behaves as an "atomic bar magnet" and interacts as such with an external applied field. 2. Types of Samples There are two types of paramagnetic samples in magnetochemistry: magnetically dilute and magnetically concentrated samples. The paramagnetic centres, which can be atoms or ions, act independently in magnetically dilute samples. In magnetically concentrated systems, the magnetic field produced by one paramagnetic centre interacts with those of its neighbours. According to the type of interactions, magnetically concentrated systems can be sub-classified into various groups. The two most common types of interactions are antiferromagnetic and ferromagnetic. In an antiferromagnetic system, the dipoles of neighbouring atoms tend to align antiparallel to each other; whereas in a ferromagnetic system, the dipoles tend to align parallel. Figure 1.6 is a simple representation of antiferromagnetic and ferromagnetic behaviour in 1-dimension. These interactions can be through 1, 2 or 3-dimensional extended lattices, and the extent of interactions can also vary. The ordering can be short range, restricted within a domain of the sample, or long range, with alignment across the domains. u t m u antiferromagnetic behaviour ferromagnetic behaviour Figure 1.6 A simple representation of (a) antiferromagnetic and (b) ferromagnetic behaviours in 1-dimension. The electron spins are represented by arrows. 23 1.3.2 Bulk Magnetic Properties of Substances If a substance is placed in a homogeneous magnetic field, H , the field within the substance is usually different from the free space value. 3 5" 3 7 The magnetic induction B, of the field within the substance, is B = H + A H (eq. 1.10) where A H is the field produced by magnetic polarization of the substance. At this point, the substance has been magnetized. A H is related to the intensity of magnetization, M , which is the magnetic dipole moment per unit volume. In c.g.s. units, A H = 4 T I M (eq. 1.11) The volume magnetic susceptibility, K , is given by, K = M / H (dimensionless) (eq. 1.12) Hence all experimental methods for determining magnetic susceptibility depend upon the measurement of M / H . The gram magnetic susceptibility, Xg, is given by, Xg = K / p (cmV) (eq. 1.13) where p is the density in the unit of g cm"3. The molar magnetic susceptibility, XM, is then X M = Xg x M W (cm3 mol"1) (eq. 1.14) In eq. 1.14, M W is the molecular weight of the substance in g mol"1. The molar magnetic susceptibility is the most commonly used susceptibility index, and is often abbreviated as just magnetic susceptibility or %. In principle, % is the algebraic sum of two contributions,38 X = X D + X P (eq. 1.15) where %D and %? represent the diamagnetic and paramagnetic susceptibilities. x D ' s negative while x P is positive. When x ° dominates, the sample is said to be diamagnetic and is repelled 24 by the external magnetic field. In contrast when %p dominates, the sample is said to be paramagnetic and is attracted by the applied field. Diamagnetism is an underlying property of matter and is always present even when it is masked by the paramagnetism.38 Therefore, measured magnetic susceptibilities must be corrected for diamagnetism. 1.3.3 The Fundamental Equation of Magnetism In order to treat the susceptibility data obtained from magnetic studies quantitatively, some definitions are necessary. Since magnetism is quantum mechanical in origin, such definitions are derived from quantum mechanics. Van Vleck developed an equation that can give a closed form expression of the magnetic susceptibility: 3 5 ' 3 7 ' 3 8 NS n (E n ( 1 ) 2 / kT - 2E n ( 2 ))exp(-E n ( 0 )/kT) X = (eq. 1.16) 2 nexp(-E n ( 0 )/kT) where E n ( 0 ) are the energies of the zero-field eigenvalues of the spin Hamiltonian, E n ( 1 ) are the first-order Zeeman energies which are equal to <i|/n | H Z E I v|/n>, and En(2) are the second-order Zeeman energies which are E J | <i|/n I H Z E I H>n> 12 / (E n ( 0 ) - E m ( 0 ) ) ] . In these expressions, v|/n are the eigenvectors corresponding to the energy eigenvalues, E n ( 0 ) , and H Z E is the Zeeman operator which accounts for the interactions between the magnetic field and the electronic angular momenta. In the Van Vleck equation (eq. 1.16), N is the Avogadro's constant and k is the Boltzmann constant. This equation is only valid in the magnetic field range where the M vs. H plot is linear. From eq. 1.16, it is obvious that % is temperature dependent. However, i f the ground state of a complex couples with thermally non-populated, low-lying excited paramagnetic states, a temperature independent component can arise. This contribution is often called Temperature 25 Independent Paramagnetism (TEP). Since these excited states are not populated, there is no temperature dependence in their contribution to the magnetic susceptibility.36 TIP is often of the same order of magnitude as the diamagnetism but of opposite sign. 3 8 1.3.4 Magnetic Moments It is the convention to define the effective magnetic moment, uefror simply p, as 3 5 u = (3k /Nu B 2 ) 1 / 2 x (xT) 1 / 2 (eq. 1.17) = 2.828 (xT)1/2 (in u B , Bohr Magnetons) where k is the Boltzmann constant, 1.38044><10"16 erg deg"1 mol"1; N is the Avogadro's constant, 6.0249xl0 2 3 atoms mol"1; % is the molar magnetic susceptibility in cm 3 mol"1 and T is the temperature in K. The unit ps is the Bohr Magneton:3 5 u B = (eh / 47tmc) = 0.927xl0"2 0 erg gauss"1 (eq. 1.18) where e is the electronic charge; h is the Planck's constant, 6.6256x 10"27 erg s; m is the mass of the electron and c is the speed of light, 2.9979xl0 1 0 cm s"1. As stated above, paramagnetism arises from the angular momentum of unpaired electrons and the angular momentum comes from the orbital angular momentum, L, and the spin of the electron, S. Paramagnetic transition metal complexes often possess electronic configurations or low-symmetry geometries such that the orbital angular momentum in these systems is partially or completely quenched; thus the paramagnetism observed is primarily due to the spin angular momentum.39 Therefore, it is sometimes very useful to compare the measured effective moments to the spin-only moments, u s = 2[S(S+l)] 1 / 2 (eq. 1.19) 26 For some metal(II) hexafluoroantimonates(V), M[SbF6]2, such as M = Fe or Co, the orbital angular momentum is only partially "quenched". The orbital contribution for each individual metal cation in the M[SbF6]2 compounds is discussed in Chapter 5. 1.4 Collaboration and Cooperation 1.4.1 Collaboration with Professor H. Willner's Group Previous research work on homoleptic carbonyl cations in Dr. F. Aubke's group here at UBC was based on close cooperation with Professor H. Willner's group at Gerhard Mercator Universitat in Duisburg, Germany. This cooperation has continued into the present and extends to the work on [Fe(CO)6][Sb2Fn]2 and [Fe(CO)6][SbF6]2. The syntheses of these carbonyls were first conducted by Dr. B. Bley and Dr. E. Bernhardt of Prof. Willner's group. The initial syntheses were repeated, and the purification procedure was developed as well as further characterization such as vibrational studies, magnetic susceptibility measurements and DSC studies were carried out by myself. These studies were carried out here at U B C and during a month of work in Prof. Willner's group in Duisburg in 2000. 1.4.2 Cooperation with Other Group Members of Dr. F. Aubke The synthesis of [{Mo(CO) 4 }2(cis-u-F 2 SbF 4 ) 3 ] x [Sb2Fii] x was first conducted by Dr. R. Brochler. The preparation of [W(CO)6(FSbF5)][Sb2Fii] was a joint effort by Dr. R. Brochler and myself. The methods were repeated and the magnetic susceptibility measurements in this work were performed by myself. 27 1.5 Research Directions To date, the only known paramagnetic carbonyl species is V(CO)6. The main goal of this project is to synthesize and study paramagnetic cationic carbonyl species. One starting point is to use a metal with an odd number of d-electrons. The most promising candidate at the beginning of this project was Fe 3 + (d5); it was believed that an extra IR band in the C-0 stretching region of [Fe(CO)6]2 + could be due to [Fe(CO)6] 3 + . 4 0 ' 4 1 Therefore, attempts were made to synthesize and isolate homoleptic carbonyl cations of Fe . Octahedral divalent 3d transition metal(II) carbonyl cations with high spin are also promising paramagnetic candidates. d 6 octahedral metal cations may have a paramagnetic high spin configuration (with a 5T2 g ground state), or a diamagnetic low spin configuration ( ' A ] g ground state). Low lODq is required to achieve a 5 T 2 g ground state. It is plausible for predominantly o-bonded CO to act as a weak field ligand, due to the significant reduction of 71-backbonding which is a key factor for CO to be a strong field ligand. Of the known homoleptic carbonyl cations with a d 6 configuration, the only realistic target is the carbonyl cation of the 3d metal Fe, [Fe(CO)6J . The other two known octahedral species, [M(CO)6] , where M = Ru and Os, have metals in the second and third row, and hence lODq's are much higher . A slightly less likely target involves a possible decomposition product of [Fe(CO)6] . Consider the hypothetical decomposition mode: [Fe(CO) 6] 2 + — • [Fe(CO) 4] 2 + + 2CO (eq. 1.20) The hypothetical decomposition species [Fe(CO) 4] 2 + should be tetrahedral, and therefore high spin. To explore the possibility of paramagnetic decomposition products of [Fe(CO)e] , the thermal decompositions of [Fe(CO)6][Sb2Fn]2 and [Fe(CO)6][SbF6J2 were studied by Differential Scanning Calorimetry (DSC). 28 Whereas [M(CO) 6 ] n + (n = 1, 2, 3), [M(CO) 4] n (n = 1, 2) and [M(CO) 2] n (n =1,2) have all very regular geometries Oh, D4h and D«>h, this is not the case for seven-coordinate d 4 metal carbonyl species. These seven-coordinate complexes can have complicated, low symmetry structures with possible low lying, thermally accessible excited states or even ground states that may be paramagnetic. Therefore, the carbonyl species of Cr, Mo and W were included in this study. During the course of research, it was found that metal(II) hexafluoroantimonates(V) were closely related to the apparent paramagnetic results obtained for the metal carbonyl cationic species and their formation. Therefore, the magnetism and chemistry, in particular addition reactions, of the M[SbFe]2 compounds were also investigated. 1.6 Ou t l i ne o f this Thesis Following the general introduction in this chapter, the general experimental aspects are presented in Chapter 2. Chapters 3 and 4 deal with an examination of the metal carbonyl cations of Fe and the triad W, Mo and Cr. Chapter 5 describes an investigation of metal(II) hexafluoroantimonates(V) and their relationships to the corresponding metal carbonyl cations. The addition reactions of these M[SbF6]2 compounds are presented in Chapter 6. Finally, a general summary and suggestions for future work are presented in Chapter 7. 1.7 References 1) Cotton, F. A. ; Wilkinson, G. Advanced Inorganic Chemistry; 5th ed.; Wiley: New York, 1988. 2) Shriver, D. F.; Atkins, P. W.; Langford, C. H. Inorganic Chemistry; 6th ed.; Oxford University Press, 1990. 3) Willner, H. ; Aubke, F. Angew. Chem. Int. Ed. Engl. 1997, 36, 2402. 29 4) Wang, C. Ph. D Thesis; The University of British Columbia: Vancouver, 1996, pp 223. 5) Aubke, F.; Wang, C. Coord. Chem. Rev. 1994, 137, 483. 6) Holland, G. F.; Manning, M . C ; Ellis, D. E.; Trogler, W. C. J. Am. Chem. Soc. 1983, 105, 2308. 7) Nakamato, K. Infrared Spectra of Inorganic Compounds and Coordination Compounds, 1 st ed.; Wiley: New York, 1963. 8) Moskovits, M . ; Ozin, G. A. Cryochemistry; Wiley: New York, 1976. 9) Mond, L. ; Langer, C ; Quincke, F. J. Chem. Soc. 1890, 749. 10) Ellis, J. E. Adv. Organomet. Chem. 1990, 31, 1. 11) O'Donnell, T. A. Superacids and Acid Melts as Inorganic Chemical Reaction Media; V C H Publishers: Weinheim, 1993. 12) Olah, G. A.; Prakash, G. K. S.; Sommer, J. Superacids; John Wiley & Sons: New York, 1985. 13) Willner, H. ; Aubke, F. Angew. Chem. Int. Ed. Engl, submitted. 14) Bodenbinder, M . ; Balzer-Jollenbeck, G.; Willner, H. ; Batchelor, R. J.; Einstein, F. W. B.; Wang, C ; Aubke, F. Inorg. Chem. 1996, 35, 82. 15) Willner, H. ; Bodenbinder, M . ; Brochler, R.; Hwang, G ; Rettig, S. J.; Trotter, J.; von Ahsen, B.; Westphal, U . ; Jonas, V.; Thiel, W.; Aubke, F. J. Am. Chem. Soc. 2001, 123, 588. 16) Goldman, A. S.; Krogh-Jespersen, K. J. Am. Chem. Soc. 1996, 118, 12159. 17) Bernhardt, E.; Bley, B.; Wartchow, R.; Willner, H. ; Bill, E.; Kuhn, P.; Sham, I. H . T.; Bodenbinder, M . ; Brochler, R.; Aubke, F. J. Am. Chem. Soc. 1999,121, 7188. 18) Brochler, R.; Bodenbinder, M . ; Sham, I. H . T.; Rettig, S. J.; Aubke, F. Inorg. Chem. 1999, 38, 3684. 19) Brochler, R.; Sham, I. H. T.; Bodenbinder, M . ; Schmitz, V.; Rettig, S. J.; Trotter, J.; Willner, H. ; Aubke, F. Inorg. Chem. 2000, 39, 2172. 30 20) Wang, C ; Hwang, G.; Siu, S. C ; Aubke, F.; Bley, B.; Bodenbinder, M . ; Bach, C ; Willner, H. Eur. J. Solid State Inorg. Chem. 1996, 33, 917. 21) Willner, H. ; Schaebs, J.; Hwang, G.; Mistry, F.; Jones, R.; Trotter, J.; Aubke, F. J. Am. Chem. Soc. 1992, 114, 8972. 22) Ainsley, E. E.; Peacok, R. D.; Robinson, P. C. Chem. Ind. 1951, 1117. 23) Aubke, F.; Cader, M . S. R.; Mistry, F. Transition Metal Derivatives of Strong Protonic Acids and Superacids; Olah, G. A., Chambers, R. D. and Prakash, G. K. S., Ed.; John Wiley & Sons, Inc., 1992, pp 43. 24) Cader, M . S. R.; Aubke, F. Can. J. Chem. 1989, 67, 1700. 25) Dean, P. A. W. J. Fluorine Chem. 1975, 5, 499. 26) Gantar, D.; Leban, I.; Frlec, B.; Holloway, J. H. J. Chem. Soc. Da/ton Trans. 1987, 2379. 27) Miiller, M . ; Muller, B. G.; Abstract of ll"1 European Symposium on Fluorine Chemistry. Bled, Slovenia, September 17 -22, 1995; pp 127. 28) Brochler, R., Personal Communication. 29) Christe, K. O.; Wilson, W. W.; Bougon, R. A.; Charpin, P . F l u o r i n e Chem. 1987, 34, 287. 30) Carlucci, L. ; Ciani, G.; Proserpio, D. M . ; Sironi, A. Angew. Chem. Int. Ed. Engl. 1995, 34, 1895. 31) Robson, R.; Abrahams, B. F.; Batten, S. R.; Gable, R. W.; Hoskins, B. F.; Liu, J. Supermolecidar Architecture, 1992; Vol. 449, Chapter 19. 32) Lu, J.; Paliwala, T.; Lim, S. C ; Yu, C ; Niu, T.; Jacobson, A. J. Inorg. Chem. 1997, 36, 923. 33) Carlucci, L. ; Ciani, G.; Proserpio, D. M . ; Sironi, A. Inorg. Chem. 1995, 34, 5698. 34) Carlucci, L. ; Ciani, G.; Proserpio, D. M . ; Sironi, A. Angew. Chem. Int. Ed. Engl. 1995, 34, 1895. 31 35) Mabbs, F. E.; Machin, D. J. Magnetism and Transition Metal Complexes; Chapman and Hall: London, 1973. 36) O'Connor, C. J. Magetochemistry - Advances in Theory and Experimentation; Lippard, S. J., Ed., 1982; Vol. 29, pp 203. 37) Figgis, B. N . ; Hitchman, M . A. Ligand Field Theory and Its Applications; Wiley-VCH, 2000. 38) Kahn, O. Molecular Magnetism; V C H Publishers, Inc., 1993. 39) Ehlert, M . K. Ph. D Thesis; The University of British Columbia: Vancouver, 1992. 40) Bley, B. ; Willner, H. ; Aubke, F. Inorg. Chem. 1997, 36, 158. 41) Bley, B. Ph. D. Thesis; Universitat Hannover: Hannover, 1997. 32 Chapter 2 General Experimental General experimental techniques as well as sources (and purification, where necessary) of the chemicals, apparatus and instrumentation used in this work are described in this chapter. The specific syntheses and procedures will be presented in later chapters. 2.1 Chemicals Tables 2.1 - 2.4 list the chemicals used in this work: Table 2.1 Sources, properties and preparations of superacids and other solvents Chemicals, Sources and Properties1 Fluorosulfuric acid, HSO3F Orange County Chemicals or General Chemical Corp. (technical grade), colourless liquid, M W = 10.067 g/mol, b.p. = 162.7 °C, m.p. = 88.98 °C, p = 1.726 g/mL (25 °C) Bis(fluorosulfuryl)peroxide, S2O6F2 Prepared by previous group members of Dr. F. Aubke, pale yellow liquid, M W = 198.118 g/mol, b.p. = 67.1 °C, m.p. = -55.4 °C, p = 1.2 g/mL (25 °C) Sulfur Dioxide, S 0 2 Matheson Gas Products (anhydrous), colourless gas, colourless liquid under its own vapour pressure, M W = 34.06 g/mol, b.p. = -10 °C, m.p. = -72.7 °C, p = 2.927 g/L (gas), 1.434 g/mL (liquid) Preparation or Purification • purified by double distillation under dry nitrogen2 • the constant boiling fraction at 164 °C was collected into a glass, one-part storage vessel (Figure 2.1 a) • prepared from F2 and SO3 using AgF 2 as a catalyst and purified as described earlier3'4 • small amounts were stored in a glass, one-part vessels (Figure 2.1a) • sealed ampoules were used for long-term storage • transferred in vacuo into a glass, one-part storage vessel (Figure 2.1a) • dried over P4O10 33 Antimony pentafluoride, SbF5 Ozark-Mahoning or Scientific Industrial Association (P & M Ltd.) (technical grade), viscous, colourless liquid, M W = 216.74 g/mol, b.p. = 149.5 °C, m.p. = 7 °C, p = 2.99 g/mL (25 °C) • purified by distillation under dry nitrogen into a round bottom flask • air and trace amounts of HF were removed by transferring the distilled SbFs to a cooled storage vessel (see Figure 2.2) under dynamic vacuum Hydrogen fluoride, HF Matheson Gas Products (anhydrous), colourless gas, colourless liquid under its own vapour pressure, M W = 20.006 g/mol, b.p. = 20 °C, m.p. = -83.35 °C, p = 0.000818 g/mL Stored over SbFs to remove H2O as [H 30][Sb2F„]2. Ethanol, CH 3 CH 2 OH anhydrous (absolute, 99%), colourless liquid, M W = 46.068 g/mol, b.p. = 78.29 °C, m.p. =-114.14 °C, p = 0.7893 g/mL (20 °C) dried by refluxing over CaCl 2 Dichloromethane, CH2CI2 B D H , colourless liquid, sensitive to light, M W = 84.932 g/mol, b.p. = 40 °C, m.p. = -97.2 °C, p = 1.3266 g/mL (20 °C) dried by refluxing over CaCb or CaO Table 2.2 Sources and properties of other chemicals used in the addition reactions of naked metal(H) hexafluoroantimonates(V). Chemicals Sources and Properties1 Pyrazine (1,4-diazine), C4H4N2 Aldrich Chemical Company Inc. (99+%), volatile, white crystalline solids M W = 80.09 g/mol, b.p. = 115 °C (768 mmHg), m.p. = 54 °C 4,4'-bipyridyl (4,4'-dipyridine), CioHgN2 Aldrich Chemical Company Inc. (98%), off-white powder, M W = 156.19 g/mol, b.p. = 305 °C, m.p. = 111 - 114 °C 34 Table 2.3 Sources, properties and preparations of other chemicals used in the synthesis of metal carbonyl cations - [SbiFuf or -[SbF^f salts Chemicals, Sources and Properties1 Preparation or Purification Gold powder, Au ABCR, Karlsruhe, Germany or Alfa (99.99%) flamed-dried under dynamic vacuum Iron pentacarbonyl, Fe(CO)s BASF, dark brown liquid, sensitive to light • purified by trap-to-trap distillation (Figure 2.7 shows the trap used) • stored in a glass, one-part vessel (Figure 2.1a) in the absence of light Arsenic pentafluoride, AsF 5 Ozark-Mahoning, gas at ambient temperature N / A (Not Applicable) Chlorine, Ch Matheson Gas Products, gas at ambient temperature N / A Xenon(Ii) fluoride, XeF2 volatile, white solid synthesized from photolysis of equimolar mixtures of Xe and F 2 5 Xenon, Xe Messer Griesheim, gas at ambient temperature N / A Fluorine, F2 Air Products Inc. or Solvay A G (technical grade), gas at ambient temperature N / A Carbon monoxide, CO Linde Gases or Praxair (C. P. grade, 99.5%), gas at ambient temperature passed through a trap (Figure 2.7) at -196 °C to remove trace amounts of CO2, H2O and other impurities Tungsten hexacarbonyl, W(CO)6 Pressure Chemical Co., volatile, white solid moisture on surface removed by sublimation Molybdenum hexacarbonyl, Mo(CO)6 Pressure Chemical Co., volatile, white solid moisture on surface removed by sublimation Chromium hexacarbonyl, Cr(CO)6 Pressure Chemical Co., volatile, white solid moisture on surface removed by sublimation 35 Table 2.4 Sources, properties and preparations of other chemicals used in the synthesis and characterization of naked metal(II) hexafluoroantimonates(V). Chemicals Sources and Properties Silver shots, Ag * Alfa or Ventron (99.9999%), shiny silver solids Palladium powder, Pd * John Matthey Catalog Company (99.95%), black powder Zinc powder, Zn * Fisher Scientific (99.2 %), grey powder Manganese powder, Mn * Matheson Coleman & Bell, black powder Iron powder, Fe * B & A, black powder Cobalt powder, Co * Fischer Scientific (99.28 %), black powder Nickel powder, Ni * PCR Inc., black powder Manganese(II) fluoride, MnF 2 Alfa (anhydrous, 99%), white powder Iron(II) fluoride, FeF2 Alfa (anhydrous, 99%), yellow powder Cobalt(II) fluoride, CoF 2 Alfa (anhydrous, 99.99%), pink powder Nickel(II) fluoride, NiF 2 Alfa (anhydrous, 99%), greenish-yellow powder Copper(II) fluoride, CuF 2 Alfa (anhydrous, 99.5%), white powder Potassium Hexafluoroantimonate(V), KSbF 6 Alfa (anhydrous, 98% min), white powder * The metals were flame-dried under dynamic vacuum before use. 2.2 Apparatus 2.2.1 Utility Vacuum Lines Volatile materials were manipulated in a glass or metal vacuum line of known volume. The glass vacuum line was used unless HF was involved in the reaction, in which case the metal line was required. The glass line was equipped with a pressure gauge (Setra 280E, Setra Systems Inc.), valves with Teflon stopcocks (Kontes or Young, used without grease), and BIO ground glass joints. The ground glass connections were lubricated with Fluorolube grease series 25-1OM, CF 2Cl(CF 2CFCl) nCF2Cl (Halocarbon Products Corporation). The metal line was 36 equipped with a pressure transducer attached to a micron gauge, and stainless steel needle valves. Teflon coated stir bars and magnetic stirrers were used to enable mixing of the reagents. 2.2.2 R e a c t i o n a n d S t o r a g e V e s s e l s 1. Glassware Purified or pre-treated reagents, such as Fe(CO)s and SO2, were stored in tubular one-part containers (Figure 2.1a). Reactions were carried out in round-bottomed one-part Pyrex reactors of 50, 100 or 250 mL capacities (Figure 2.1b). A specially designed vessel with two outlets (Figure 2.2) was used for the storage of SbFs. Reactive reagents, e.g. S2O6F2, or products for long-term storage were flame-sealed in glass ampoules and stored at liquid nitrogen temperature. When filtration was required during a synthesis, a one-part filtration vessel equipped with a frit (class D) between two round bottom flasks (Figure 2.3) was used. A one-part V-shaped Pyrex reactor (Figure 2.4) was used when decanting or washing of a solid was required during a preparation. Figure 2.1 (a) Tubular one-part storage container (b) Round-bottomed one-part Pyrex Teflon Stopcocks (a) reactor. 37 Tef lon S t o p c o c k s B10 g r o u n d g l a s s c o n e B10 g r o u n d g l a s s s o c k e t Figure 2.2 Storage vessel for SbFs. Teflon Stopcocks Figure 2.4 One-part V-shaped Pyrex reactor. 2. PFA Reactors PFA (polyfluoroacetamide) reactors consisting of 120 mL or 250 mL PFA bulbs with a NS 29 socket standard taper (Bohlender) in connection with a stainless steel NS 29 cone standard taper and a needle valve (Figure 2.5) were employed for reactions involving HF. Both parts of the reactor were pressed together with a metal flange such that the reactor was leak tight without using grease. A V-shaped reactor made entirely out of PFA (Figure 2.6) was used when decanting or washing of a solid was required during a preparation. Anhydrous HF (which contained a small amount of SbFs as the drying agent) was stored in a calibrated PFA tube with one end sealed and the other end connected to a stainless steel needle valve. 39 Figure 2.5 PFA reactor (a) side view (b) top view. 3 P F A tubes Figure 2.6 V-shape PFA reactor. 40 2.2.3 Other Special Apparatus and Equipment 1. Trap A glass trap (Figure 2.7) was used for the purification of CO(g). Usually two or more traps were used in series for trap-to-trap purification of other chemicals such as Fe(CO)s. B14 ground glass cone Teflon Stopcocks B14 ground glass socket ai 41 2. Sublimation Apparatus A sublimation apparatus (Figure 2.8) was used for vacuum sublimation as means of purification or crystal growing of moisture-sensitive products. It could be opened inside the dry box for recovery of the material. Teflon Stopcock Figure 2.8 Sublimation apparatus. 42 3. Ampoule Key Reactive reagents or products for long-term storage were flame-sealed in glass ampoules of 6.0 mm outer diameter. An ampoule key (Figure 2.9) was used for opening and sealing of these ampoules under vacuum.6 vacuum B14 ground . glass socket stressor B14 ground glass socket B14 metal taper rubber o-ring metal ring metal ring ampoule Figure 2.9 Ampoule key. 43 4. Inert Atmosphere Box A l l non-volatile, moisture-sensitive materials were manipulated inside an inert atmosphere box (Vacuum Atmospheres Corp. Dry box Model DL-001-S-G Dri-Lab equipped with a Model HE-493 Dri-Train filled with molecular sieves.) The dry box was filled with " L Grade" nitrogen and was regularly regenerated by heating the sieves and copper catalyst in the Dri-Train. 5. Balances Reactions were monitored by weight difference in the reaction vessels whenever possible. A Mettler Gramatic analytical balance #1-910 (max. 200 g; ± 0.2 mg) was used for weighing of reagents and products in glass reactors. PFA reactors were weighed on a top-loading Sartorius Type 1104 balance (max. 100 g; ± 0.5 g). A Mettler PC440 balance was used for weighing inside the inert atmosphere box. 2.3 Instrumentation 2.3.1 Vibrational Spectroscopy 1. Infrared Spectroscopy (IR) Infrared spectra were recorded at room temperature on an Bruker IFS66v FT or Bomem Michelson M B 102 FT-IR spectrometer. Solid samples were crushed between AgBr plates inside the dry box. Gaseous samples were introduced via the vacuum manifold into a gas cell equipped with silicon windows (Figure 2.10). A cold finger on the gas cell permitted the condensation of volatile materials into the cell. Spectra were recorded from 5000 - 80 cm"' with a spectral resolution of 2 cm"1 (for Bruker) or 4 cm'1 (for Bomem). 44 B10 ground glass cone g lass wa = 2mm silicon IR windows Teflon Stopcock Figure 2.10 Gas cell for IR measurements on gaseous samples. 2. Raman Spectroscopy (Ra) Raman spectra were recorded at room temperature with a Bruker FT Raman accessory mounted on an optical bench of the RFS 100/S instrument. Ground solid samples were sealed in melting point capillaries under nitrogen or argon atmosphere. Spectra were recorded from 5000 - 80 cm"1 with a spectral resolution of 2 cm"1. 2.3.2 Magnetic Susceptibility Measurements Magnetic susceptibility measurements were performed in the temperature range of 2 - 300 K at a field of 10,000 G using a Quantum Design (MPMS) SQUID magnetometer. Solid samples weighing about 40 mg were placed in a PVC sample holder of constant cross-sectional density (Figure 2.11). Details of the PVC holder are described elsewhere.7 Susceptibilities were corrected for the background signal from the sample holder at all temperatures. The susceptibilities were also corrected for the diamagnetism of all atoms using Pascal's constants.8 45 Extension rods (connected by friction fit) 5.5 cm 3.8 cm sample chamber (closed by friction fit) Height = 3 mm inner diameter = 4 mm 3.8 cm 5.0 cm Figure 2.11 PVC sample holder for magnetic susceptibility measurements. 2.3.3 Differential Scanning Calorimetry (DSC) DSC studies were performed using a Netzsch DSC 204 Calorimeter. 10 - 50 mg of solid samples were loaded into an Al-pan with pierced lid (Figure 2.12) or a 100 uL stainless steel, gold-plated sealed crucible (Figure 2.13) in the dry box. The samples were heated from 25 °C to 400 °C. The apparatus used to open and close the crucible is shown in Figure 2.14, and that used to open the crucible to allow the gas phase into a vacuum line is shown in Figure 2.15. 46 Pin Hole Pressed Down to Seal Top A l Pan Pressed Down to Sea l Bottom A l Pan Figure 2.12 DSC sample holder: Al-pan with pierced lid. View from the side: Pin Hole Closed by screwing the Cap onto the Crucible Stainless Steel Cap Gold Thin Plate (for sealing the crucible under the pin hole) Gold Plated Stainless Steel Crucible View from the top: Stainless Steel Cap Pin Hole View from the bottom: Gold Plated Stainless Steel Crucible Marks for securing cap and crucible in tools used during opening and closing Figure 2.13 DSC sample holder: stainless steel, gold-plated sealed crucible. 47 ] L Stainless l ^ ^ l / Steel Crucible Figure 2.14 Apparatus used to open the stainless steel, gold-plated sealed crucible. O - R i n g s -P in Metal joint (B14, male) Sta in less Steel Crucib le Figure 2.15 Apparatus used to open the crucible to allow the gas phase into a vacuum line. 48 2.3.4 Atomic Absorption Spectroscopy (AA) Atomic Absorption studies were performed using a Varian AA-875 instrument with suitable lamps for different samples. Samples were fed to the spectrometer in the form of aqueous solutions. Use of the spectrometer in the advanced analytical chemistry teaching laboratory of UBC was courtesy of B. Clifford. 2.3.5 Single Crystal X-Ray Diffraction Single crystal X-ray diffraction studies were carried out on a Rigaku AFC6S diffractometer using Mo Koc radiation of 0.71069 A wavelength. Crystal structures were determined by S. J. Rettig, B. O. Patrick and J. Trotter of the X-Ray Crystallography Laboratory in the Department ofChemistryatUBC. 2.3.6 Microanalyses Carbon, hydrogen and nitrogen microanalyses were carried out by P. Borda of the Microanalytical Facility in the Department of Chemistry at UBC. The instrument employed was a Carlo Erba Elemental Analyzer Model 1106. 2.4 References 1) CRC Handbook of Chemistry and Physics; 81st ed.; CRC Press LLC, 2000. 2) Barr, J.; Gillespie, R. J.; Thompson, R. C. Inorg. Chem. 1964, 3, 1149. 3) Wang, C. Ph. D. Thesis; The University of British Columbia: Vancouver, 1996. 4) Mistry, F. Ph. D. Thesis; The University of British Columbia: Vancouver, 1993. 5) Holloway, J. H. Chem. Commun. 1966, 22. 6) Gombler, W.; Willner, H. J. Phys. E. Sci. Instrum. 1987, 20, 1286. 49 7) Ehlert, M . K. Ph. D Thesis; The University of British Columbia: Vancouver, 1992. 8) Konig, E. Landolt-Bdrnstein, New Series; Hellwege, K. H. and Hellwege, A. M , Ed.; Springer-Verlag: Berlin, 1966; Vol. II. 50 Chapter 3 Homoleptic Carbonyl Cations of Iron 3.1 Hexakis(carbonyl)iron(II) Undecafluorodiantimonate(V), [Fe(CO)6][Sb2Fn]2, and Hexakis(carbonyl)iron(II) Hexafluoroantimonate(V), [Fe(CO)6] [SbF6]2 3.1.1 Introduction 1. General Comments on Collaboration The study of [M(CO)e] 2 + salts of group 6 metals (M = Fe, Ru and Os) has been a part of our collaboration with Prof. Willner's research group for some time. Research on the carbonyl cations of Ru and Os was started by C. Wang (Ph. D. thesis, UBC, 1996), who synthesized the [M(CO)6][Sb2Fn]2 compounds (M = Ru and Os), and continued by C. Bach (Ph. D. thesis, Hannover, 1998), who solved the molecular structures of these compounds. B. Bley (Ph. D. thesis, Hannover, 1997) extended the study to Fe and prepared [Fe(CO)6][Sb2Fn]2. The improved synthesis of [Fe(CO)6][Sb2Fu]2 and conversion to [Fe(CO)6][SbF6]2 was then achieved by E. Bernhardt (Postdoc, Duisburg) and subsequently by myself in the present work. My initial objective was the synthesis and characterization of paramagnetic hexacarbonyl cations of Fe(II) and Fe(III). Three synthetic routes to [Fe(CO)6][Sb2Fn]2 have been found, of which two give paramagnetic products. All of these syntheses were reinvestigated in the current work, and a purification procedure was developed. Further characterization such as vibrational studies, magnetic susceptibility measurements and DSC studies were also carried out. 2. Account of Research Accomplishments with respect to the [Fe(CO)rJ2+ salts i. Synthetic Methods All three synthetic routes to [Fe(CO)6][Sb2Fn]2 involve the oxidation of Fe(CO)s, either by AsFs or CI2 in SbF5, or by XeF 2 in HF-SbF 5 in the presence of CO. The two former 51 preparations give paramagnetic products while the latter yields diamagnetic [Fe(CO)e] salts. The oxidative carbonylation of Fe(CO)s by AsF 5 or Cb as external oxidizing agents in liquid SbFs to produce [Fe(CO)6][Sb2Fn]2 was first formulated as:1 1 atm CO, SbF 5 Fe(CO) 5 + AsF 5 + CO + 5SbF5 * [Fe(CO) 6][Sb2F„] 2 + AsF 3 .SbF 5 (eq. 3.1) 60 -80 °C 1 atm CO, SbF 5 Fe(CO) 5 + C l 2 + CO + 6SbF5 4 0 _ 8 0 oC* [Fe(CO)6][Sb2Fn]2 + 2SbF 4Cl (eq. 3.2) An improved synthesis of [Fe(CO)6][Sb2Fn]2 using XeF 2 as the oxidizing agent was studied recently:2 Fe(CO) 5 + XeF 2 + CO + 4SbF, l a t m C O , 5 0 ° C ^ r F e ( C o) 6 ][Sb 2 F,i]2 + Xe (eq. 3.3) HF-SbF 5, 2 days This preparation produced crystalline [Fe(CO)6][Sb2Fn]2. [Fe(CO)6][SbF6]2 crystals could be obtained subsequently by washing with H F : 2 anhydrous HF [Fe(CO)6][Sb2Fii]2 + 2HF * [Fe(CO)6][SbF6]2 + 2(HF.SbF5) (eq. 3.4) With this method, single crystals of [Fe(CO)6][Sb2Fn]2 and [Fe(CO)6][SbF6]2 were isolated and the structures of both carbonyl compounds were solved.2 //. Molecular Structures and Extended Structures The molecular structures of [Fe(CO)e][Sb2Fn]2 and [Fe(CO)6][SbF6]2 are shown in Figures 3.1 and 3.2.2 The crystal data can be found in Appendix 1. 52 F4 Figure 3. J Molecular Structure of [FefCOJej'fSb^FjjJ2 (ellipsoids are drawn at 50% probability).2 Figure 3.2 Molecular Structure of [Fe(CO)6][SbFe]2 (ellipsoids are drawn at 50% probability).2 [Fe(CO)6j[SbF6]2 crystallizes in a tetragonal space group PA/mnc (no. 128) whereas [Fe(CO)6][Sb2Fn] 2 crystallizes in the monoclinic space group P2\/n (no. 14), the same space group as all other structurally characterized [Sb2Fn]" salts with dipositive homoleptic metal carbonyl cations.2'3 [Fe(CO)6][SbFe]2 is the first fully characterized example of a homoleptic metal carbonyl cation stabilized by the [SbFe]" counteranion rather than by [ S b 2 F n ] " . 2 The 53 [SbFe]" anions in [Fe(CO)6][SbF6]2 show a very slight tetragonal distortion from regular Oh geometry, while the [SD2F11]" anions in [Fe(CO)6][Sb2Fn]2 are tetragonally distorted from ideal D4h symmetry. The two SbFs moieties are bent about the bridging F atom with an Sb-F-Sb angle of 148.5°; and the two SbF4 planar groups are rotated relative to each other with a dihedral angle of 36.8°. The [Fe(CO) 6] 2 + cations in both [Fe(CO) 6][Sb 2Fii] 2 and [Fe(CO)6][SbF6]2 are essentially octahedral, and their relevant bond lengths and angles are virtually identical in both salts. This is also evident from the identical vibrational spectra observed for the cations in the two salts.2 The long Fe-C bonds (an average of 1.911(5) A) 2 and short C-0 bonds (an average of 1.104(5) A 2 , see Appendix 1), which are shorter than that in free CO (1.1281 A), 4 suggest that CO is primarily a-bonded to Fe 2 + with minimal 7t-backbonding.2 There are a number of secondary interionic contacts in [Fe(CO)6][Sb2Fn]2 and [Fe(CO)6][SbF6]2 (see Figures 3.3 and 3.4).2 Significant secondary interionic contacts are defined as those 10% shorter than the sum of the van der Waals radii. 5 These contacts are Figure 3.3 Selected significant interionic contacts in a formula unit of [Fe(CO)g]'[Sb2F] 1J2. slightly more significant in [Fe(CO) 6 ][Sb2Fii] 2 than in [Fe(CO)6][SbF6]2. 54 2.853(5) ,T^1 / \ 3.061(5) 2.842(5) Figure 3.4 Selected significant interionic contacts in a formula unit of [Fe(CO)6][SbFe]2.' It can be seen in Figure 3.3 that there is no significant contact between the cation and the SbFs moiety of the [Sb2Fn]" anion in the formula unit of [Fe(CO)6][Sb2Fn]2- This is consistent with the fact that [Fe(CO)6][Sb2Fn]2 can easily be converted into [Fe(CO)e][SbF6]2 by repeated washing with HF. It can be seen in Figures 3.3 and 3.4 that the contacts are primarily due to C— F and to a lesser extent O—F interactions. The interactions involve all six C atoms of the 2+ [Fe(CO)e] cation in [Fe(CO)6][Sb2Fn]2 but only four of the six CO groups for [Fe(CO)6][SbFe]2. The contact distances are also longer and thus the interactions are presumably weaker than in other homoleptic carbonyl cations. Within the formula unit of [Fe(CO)6][Sb2Fii]2 and [Fe(CO)6][SbF6]2, there is about one C--F contact per CO ligand in contrast to approximately five such contacts in [Hg(CO)2][Sb2Fn]2, 6 and four to five in [M(CO)4][Sb2Fn]2 where M = Pd, Pt. 7 Multiple F—CO contacts involving up 55 to four contacts per F atom in the [Sb2Fn]" salt are observed, whereas only single or double contacts have been observed in the other homoleptic carbonyl cations.6'7 Slight departures from linearity observed for the Fe-C-0 group in both [Fe(CO)6] 2 + cations are caused by significant interionic F--C contacts.2 Similar observations are reported for all salts of metal carbonyl cations.3 The lesser extent of multiple F--CO contacts and the only slight non-linearity for the Fe-C-0 group show that there is a great intrinsic stability of the octahedral cation [Fe(CO)6] 2 + relative to [M(CO) 4 ] 2 + (M = Pd or Pt) and [Hg(CO) 2] 2 +. iii. Vibrational Analysis The fundamentals of [Fe(CO)e]2+ have all been experimentally determined with position confirmed by calculations (see Table 3.1).2 Table 3.1 Vibrational assignments of the observed (obs.) and calculated (calc.) wavenumbers, v(cni'), of the fundamentals for [FefCOJe]2* in [Fe(CO)6][Sb2Fu]2.2 V Assignment Observed ( c m 1 ) a Calculated ( c m 1 ) b V i , Vco A i g 2241.2 2221 V3, VCO Eg 2219.5 2188 v 6 , Vco T, u 2204 2173 V7, SFeCO T i u 586 608 Vio, SFeCO T 2 g 500.9 508 Vl2, 5FeCO T 2 u 468 478 V8, VFeC T,„ 380 380 V4, VFeC Eg 361 355 V 2 , V F e C Aig 347.1 361 V5, 5FeCO Tig [336] 338 V l l , ScFeC T, u 170 117 V9, ScFeC T 2 g 138 98 Vl3, ScFeC T 2 u 114 77 a : Values in square brackets are derived from overtones. b : BP86/ECP2 calculations.2 56 The band positions for the fundamentals of [Fe(CO)6] are independent of the anion.^ These cation bands can be clearly differentiated from the anion bands. This is also true for the isostructural complexes [M(CO)6] 2 + (M = Ru or Os), as will be discussed in Section 1 of 3.1.7. Furthermore, the CO stretching bands are not subject to vibrational mixing. 2 iv. The [Fe(CO)6f+ Problem The possibility of generating "[Fe(CO)6][Sb2Fn]3" when C l 2 was used as the oxidizing agent has been reported.1 In the report, a single C-0 stretching band at 2256 cm"1 in the vibrational spectrum was erroneously assigned to [Fe(CO)6]3 + by analogy to V6 (Ti u) for [Ir(CO)e]3+ at 2254 cm"1.1 It was later found that the band belonged to the vco of [C1C0] + . 8 3. Contributions of the Current Work The impetus for this work was to produce and study paramagnetic carbonyl cationic species. Therefore, the nature of paramagnetism in [Fe(CO)6][Sb2Fn]2 was investigated. The initial syntheses of [Fe(CO)6][Sb2Fn]2 and [Fe(CO)6][SbF6]2 were repeated, and the preparation of [Fe(CO)6][Sb2Fii]3 was attempted. Paramagnetism in these [Fe(CO) 6] n + (n = 2, 3) samples was studied using magnetic susceptibility measurements; and thermal decompositions were studied using differential scanning calorimetry (DSC). Since these measurements required bulk samples of known composition and high purity, improved syntheses and purification of the Fe carbonyl compounds were also investigated. In this chapter, the by-products in the [Fe(CO)6][Sb2Fn]2 and [Fe(CO)6][SbF6]2 samples are identified, and the compositions of the bulk materials are determined. Pathways by which the by-products may form are also proposed. In addition, a brief summary of evidence concerning the presence of [Fe(CO)6][Sb2Fi i ] 3 in the samples studied is included. 57 4. Paramagnetism in Metal Carbonyl Cations Paramagnetic metal carbonyl cations are thus far unknown. This is not overly surprising, because almost all carbonyl cations are formed by 4d and 5d metals, where spin paired complexes are common.9 In addition, all these cations have even numbered d electron configurations (d 4, d 6, d 8 and d 1 0). Octahedral [Fe(CO) 6] 2 + ( ' A i g ground state) is the only example of a carbonyl cation formed, as -[Sb2Fn]~ or [SbF6J~ salts, by a 3d metal. The Fe 2 + centre is d 6, and would be paramagnetic in high spin complexes but diamagnetic in low spin systems. In contrast, the Fe 3 + centre, such as in [Fe(CO)6]3 +, is d 5 and paramagnetism is expected whether the configuration is high or low spin.3 With significantly reduced 7i-back bonding, the ligand field strength of predominately o-bonded carbonyl ligands is expected to be considerably weaker. Therefore, the existence of octahedral Fe(II) carbonyl cations with high spin configurations cannot be ruled out. Another type of carbonyl species likely to be paramagnetic is possible decomposition products of [Fe(CO)6]2 +, such as [Fe(CO)4]2 +, provided that the thermal decomposition of the cation occurs with stepwise loss of CO. This hypothetical decomposition product should be tetrahedral and, therefore, is high spin. Hence the initial goal of this project was to synthesize complexes containing the cations [Fe(CO)6J2 + and [Fe(CO)6]3 + in very pure form, and to characterize the components using magnetochemical techniques and DSC in the hope of finding evidence for [Fe(CO)6-n]2 +, which could conceivably be paramagnetic. 3.1.2 Syntheses and Characterization 1. Synthetic Reactions i. Preparation of [Fe(CO)6j'/S^F;iJ2 with AsF5 as the Oxidizing Agent Experimental In a typical experiment, a 100 mL one-part glass reactor was charged with 0.1834 g (0.9381 58 mmol) Fe(CO)5. Then 480 mbar (in a 215 mL vacuum line, 4.3 mmol) AsFs, 1 mL (14 mmol) SbF 5 and finally 450 mbar (in a 215 mL vacuum line, 4.0 mmol) CO was added in vacuo at -196 °C. The reaction mixture was then stirred at 60 - 80 °C for two hours. When the yellow colour had vanished, all volatiles were removed in vacuo from the white mixture and an off-white solid was obtained. The product [Fe(CO)6][Sb 2Fn] 2 was collected as an off-white powder (0.87 g, expected 1.059 g, 82% yield) after the by-product AsF 3 .SbF 5 had been removed by repeated washing with S 0 2 at room temperature. Comments Repeated synthetic experiments gave the same product, identical to the published data of [Fe(CO)6][Sb 2Fn] 2, as identified by IR and Raman spectroscopy.1 The samples produced in this manner were strongly paramagnetic. The cause of the paramagnetism will be discussed later in this chapter. The residue of the isolated by-product, obtained after removal of all volatiles, was identified by IR spectroscopy as S0 2 .SbF 5 . 1 0 This suggested a displacement reaction of AsF 3 by S 0 2 during washing. ii. Preparation of [Fe(CO)6j'[Sb2Fj iJ2 with CI2 as the Oxidizing Agent Experimental In a typical experiment, a 100 mL one-part glass reactor was charged with 0.2093 g (1.068 mmol) Fe(CO)5. Then 650 mbar (in a 215 mL vacuum line, 5.8 mmol) chlorine gas, 1 mL (14 mmol) SbFs and finally 300 mbar (in a 215 mL vacuum line, 2.7 mmol) CO were added in vacuo at -196 °C. The reaction mixture was stirred at 40 - 80 °C for eight hours. During this time the colour changed from yellow to white. All volatiles were then removed and an off-white powder was obtained (0.9028 g, expected 1.206 g, 74.9% yield). 59 Comments It has been previously reported that in the IR spectra of samples prepared using this method, a band at 2256 cm"1, assigned to [Fe(CO)6][Sb 2Fii] 3, appears in addition to those of [Fe(CO)6][Sb2Fn] 2 . 1 Repeated experiments in the current study gave products with the same C-O stretching frequencies as those published for [Fe(CO)6][Sb2Fn]2, with the IR band at 2256 cm"1 absent. The samples produced in this manner were also paramagnetic as will be discussed. ///. Preparation of [Fe(CO)6][Sh2Fi jJ2 with XeF2 as the Oxidizing Agent Experimental In a typical experiment, 15mLHF, 1.52 g (9.1 mmol) X e F 2 and 12mL(160 mmol) SbFs were condensed into a 250 mL PFA reactor. When a homogeneous solution was obtained, 1.00 g (5.1 mmol) Fe(CO)5 was added in vacuo at -196 °C. Then 530 mbar (12 mmol) CO was introduced into the system. The mixture was stirred at -50 °C in an ethanol bath for 10 minutes, and the colour changed from white to yellowish. Stirring was then continued at 50 °C for two days. Upon removal of volatiles in vacuo, a mixture of white crystalline solids and powder were obtained and characterized by Raman spectroscopy. The crystalline products (5.62 g, expected 5.6 g) were isolated from the powder in the dry box. A portion of these crystalline solids was then washed with a (v/v 1:1) mixture of HF-SbFs in a V-shaped PFA reactor to remove the powder on the surface of the crystals. Comments While it was possible to isolate single crystals of [Fe(CO)6][Sb2Fn] 2 , the Raman spectrum in the Sb-F stretching region (500-700 cm'1) for the bulk sample of crystals differed from the published spectra obtained from single crystal (see Section 2 of 3.1.5).2 This indicated the 60 presence of impurities in the bulk sample. The identification of these by-products and the purification of [Fe(CO)6][Sb2Fn]2 are discussed in Section 3.1.6. iv. Preparation of [Fe(CO)6JfSbFsj'2 obtained from the Reaction with XeF2 Experimental Crystalline [Fe(CO)6][Sb2Fn] 2 (4.625 g, 4.097 mmol, prepared as described above) was loaded into the main arm of a V-shaped PFA reactor. 5 mL HF was then added in vacuo. The resulting suspension was stirred briefly, and the crystals were allowed to settle. The supernatant solution was then decanted into the sidearm. Subsequently, HF was condensed back into the main arm. This washing process was repeated four times. After washing, 3.905 g (5.616 mmol) of solids consisting of white powder and a few small white crystals were obtained in the main arm. Comments As for [Fe(CO)6][Sb 2Fii] 2, the Raman spectrum of [Fe(CO)6][SbF6]2 differed in the Sb-F region from previously published spectra,2 indicating the presence of impurities in the sample (see Section 2 of 3.1.5). The purification of [Fe(CO)6][SbFe]2 is discussed in Section 3.1.6. 3.1.3 Identification of Fe[SbF6]2 as the Paramagnetic By-Product in [Fe(CO)6][Sb2Fn]2 Samples 1. Magnetic Susceptibility Measurements i. Magnetic susceptibility measurements on [Fe(CO)6][Sb2F]j]2 Molecular structure determinations and vibrational analysis indicate that the [Fe(CO)e]2+ cation in both [Fe(CO)6][Sb2Fn] 2 and [Fe(CO)6][SbF6J2 is octahedral. Fe(II) has a d 6 configuration, with an octahedral geometry. There are no unpaired electrons in the low spin case (ground state 61 = 1 A i g ) but 4 unpaired electrons in the high spin case (spin, S = 2; spin-only magnetic moment, U s = 4.9 P B ; ground state = 5 T 2 g ) . The experimental magnetic moment found at 300 K for [Fe(CO)6][Sb 2Fii] 2 from the AsF 5 reaction (see eq. 3.1) is 3.1 p B and that from the C l 2 reaction (see eq. 3.2) is 3.2 P B , which correspond to neither of the two cases (see Figure 3.5). 01 E o 2 c O) B 2 0 o O ° oooooooooo ooooo oooooo 0 50 100 150 200 Temperature (K) 250 300 350 O Experimental moment for a powdered sample obtained from single crystals [FeCCOyfSbjF^k prepared using XeF 2 • Experimental moment for a powder sample of [Fe(CO)g][Sb2F^ ^2 prepared using C l 2 A Experimental moment for a powder sample of [Fe(CO)g][Sb2F-| -| ]2 prepared using AsF 5 Q Experimental moment for a powder sample of Fe[SbF 6] 2 (Solid) Calculated moment of a [ F e ^ O ^ n S b ^ . , ] . , sample containing 38% Fe[SbF 6] 2 (fixed: TIP = 4X10 4 cm 3/mol. fitted: F = 0.0106) (Dashed) Calculated moment of a [Fe^OJJ tSb jF^k sample containing 34% Fe[SbF ] (fixed: TIP = 4x10"4 cm3/mol, fitted: F = 0.00903) Figure 3.5 Magnetic moments of various samples of [Fe(CO)6][Sb2Fj 1]? and FefSbFt]? (2 300K 62 On the other hand, a single crystal sample prepared from XeF2 is found to be essentially diamagnetic (see Figures 3.6 and 3.7). These findings suggested the presence of a paramagnetic by-product in [Fe(CO)6][Sb2Fn]2 for samples prepared from AsF 5 and CI2. Fe[SbF6]2 is suspected to be the paramagnetic by-product; and the reasoning behind this assumption is presented in Section 3.1.4. From the magnetic behaviour of the [Fe(CO)6][Sb2Fn]2 samples prepared using CI2 or AsFs as the oxidizing agent, it is estimated that 38% and 34% of Fe[SbFe]2 is present in the respective samples, assuming that the TIP is 0.0004 cm 3 / mol (see Figure 3.5). The goodness of fit (F-value) was measured by the least-squares fitting function, F = A / [ ( 1 / N )XE { [(Xcaic-XexptVXexpt]2}], where N is the number of data, Xcaic is the calculated value and Xexpt is the experimental value of the magnetic susceptibility: Satisfactory fits (F = 0.0106 and 0.00903) were obtained for both cases. In Figure 3.5, the magnetic moments of various samples of [Fe(CO)6][Sb2Fn]2 are also compared to those of Fe[SbF 6] 2 from 2 - 300 K. It should be noted that the temperature dependence of magnetic moments of the [Fe(CO)6][Sb2Fn]2 samples from CI2 and AsFs are typical for high spin Fe 2 + with a 5 T 2 g ground state in an octahedral environment. A detailed analysis of the magnetic behaviour of Fe[SbF6]2 and other M[SbFe]2 complexes is presented in Chapter 5. At this point, the comparison suggests that all preparations of [Fe(CO)6][Sb2Fn]2 produce variable amounts of Fe[SbFe]2 in the product mixtures, since a similar temperature dependence is noted in all samples. Also, the high spin Fe 2 + found in Fe[SbF6]2 seems to be the only paramagnetic species present in various [Fe(CO)6][Sb2Fn]2 samples. An unambiguous identification of Fe[SbFe]2 in the product mixture by vibrational spectroscopies would be difficult. Although the IR and Raman spectra of Fe[SbF6]2 are known, 1 1 most of the bands fall between 540 - 704 cm"1 and 273 - 285 cm"1 (the Sb-F stretching and deformation regions). 63 These regions o f the [Fe(CO)6][Sb2Fn] 2 spectra are extremely cluttered. Therefore, the unambiguous identification of Fe[SbFe]2 in these samples is only possible using magnetic susceptibility measurements. The magnetic susceptibility measurements on purified bulk samples of [Fe(CO)6][Sb2Fii]2 and [Fe(CO)6][SbF6]2 show that the compounds are essentially diamagnetic (XM = 2 X 10"4 cm 3 / mol, see Figures 3.6 and 3.7). (For purification of bulk samples, see Section 3.1.6). This corresponds to low-spin Fe(II) with a ' A i g electronic ground state. There is, however, residual weak paramagnetism present. Again, it is assumed that the residual weak paramagnetism in the purified [Fe(CO)6][Sb 2Fn] 2 and [Fe(CO)6][SbF6]2 samples may be due to a small amount of Fe[SbFe]2 present, which adds to the small amount of Temperature Independent Paramagnetism (TIP) of the 'Aig state. 9 ' 1 2 Since the paramagnetic contribution is low, a clear identification of the impurity is not possible. The magnetic behaviour of [Fe(CO)6][Sb 2Fn] 2 (Figure 3.6) is compared to several models, and the two models that the data best fit are: (1) Model 1: A diamagnetic Fe(II) ground state with 0.48% spin-only Fe 2 + (u s = 4.9 OB), a theoretical species as impurity. A TIP of 3.63x10"4 cm3/mol is fitted. (2) Model 2: A diamagnetic Fe(II) ground state with 0.5% Fe[SbFe]2 as impurity. A TIP of 3.6xl0"4 c m V m o l is fixed. Based on the F values (0.112 vs. 0.0639), it is found that the experimental data fit Model 1 (with 0.48% spin-only Fe 2 +) better than Model 2 (with 0.5% Fe[SbF6]2), especially from 2 - 200 K. For Model 2, the data fit well only from 70 - 200 K. There is no significant difference in terms of values of calculated susceptibilities of the two fits mentioned from 70 - 200 K. This suggests 64 that Fe[SbF6]2 may be the bulk paramagnetic impurity in the sample. It should be noted that, however, the curves generated from both models are very similar (see Figure 3.6). The value of TIP fitted (0.00036 cm 3 / mol) is comparable to that of K4[Fe(CN)6] (6xl0" 4cm 3/mol). 9 1.1 0.2 { 0.1 J . 1 . . . . . 1 0 50 100 150 200 250 300 350 Temperature (K) O Experimental moment of a [Fe(CO) 6][Sb 2F 1 1] 2 sample (prepared with XeF 2 as the oxidaizing agent and purified by treatment with F 2 in HF) (Dashed) Calculated moment of a [ F e ^ O y f S b j F ^ k sample containing 0.5% Fe[SbF 6] 2 with TIP correction (fixed: TIP = 3.6 x 1CT4 cm 3/mol, fitted: F = 0.112) (Solid) Calculated moment of a [Fe(CO) 6][Sb 2F 1 1] 2 sample containing 0.48(1)% spin-only Fe 2 * (fitted: TIP = 3.63(8) x 10"4 cm 3/mol, F = 0.0639) Figure 3.6 Magnetic behaviour of purified [Fe(CO)6][Sb2Fi j]i compared to different models (2 - 300K) 6 5 1.4 0.2 J 1 1 1 , , , 1 1 0 50 100 150 200 250 300 350 Temperature (K) O Experimental moment of [Fe(CO)6][SbF6]2 (Dashed) Calculated moment of a [Fe(CO)6][SbF6]2 sample containing 0.80(4)% spin-only F e 2 + (fitted: TIP = 6.2(3) x 10"4 cm 3/mol, F = 0.162) (Solid) Calculated moment of a [Fe(CO)6][SbF6]2 sample containing 1.2% Fe[SbF 6] 2 with TIP correction (fixed: TIP = 5 x 1 0 ^ cm 3/mol, fitted: F = 0.0755) Figure 3.7 Magnetic behaviour of [Fe(CO)6][SbF6]2, prepared from purified [Fe(CO)6][Sb7Fu]2, compared with different models (2 - 300K). 66 /'/. Magnetic Susceptibility Measurements on [Fe(CO)6][SbF6]2 The magnetic behaviour of [Fe(CO)6][SbF6]2 (Figure 3.7) determined using bulk powder samples, is compared to several models, and the two models that the data best fit are: 1. Model 1: A diamagnetic Fe(II) ground state with 0.80% spin-only Fe 2 + (p s = 4.9 pe), a theoretical species as impurity. A TIP of 6.2xl0"4 cm3/mol is fitted. 2. Model 2: A diamagnetic Fe(II) ground state with 1.2% Fe[SbF6]2 as impurity. A TIP of 5xl0" 4 cm3/mol is fixed. The experimental data fits Model 2 (with 1.2% Fe[SbF6]2) well, especially from 20 - 300 K. Again, this suggests that Fe[SbF6J2 may be the bulk paramagnetic impurity. The fitted value of TIP (5xl0" 4 cm 3 / mol) is also comparable to that of K 4[Fe(CN) 6] (6x10 - 4 c n r V m o l ) . 9 2. Iron Content Analysis i. A tomic A bsorption Spectroscopy (AA) Samples of [Fe(CO)6][Sb2Fn] 2 and Fe[SbFe]2 were analyzed in aqueous solution of known concentration. For all samples containing Fe 2 + in this work, a clear yellow solution was obtained when the samples were dissolved in deionized water. Two methods of analysis were used: calibration curve and standard addition. a. Calibration curve Calibration curves were constructed using five standard aqueous solutions of FeS0 4 .7H 2 0 (3.77xl0"5 - 1.89xl0"4 M). Absorbance was plotted against the concentration of Fe for each standard solution. The absorbance of the analyte samples was recorded and the concentration of Fe(II) in each sample was interpolated from the calibration curve. 67 b. Standard Addition For standard addition, a fixed volume of a particular sample solution was added to the five standard aqueous solutions of FeS0 4 .7H 2 0 (3.77xl0"5 - 1.89xl0"4 M). The absorbance of each solution was plotted against concentration of FeS04.7H 20 in each standard. The plot was then extrapolated to find out the concentration of Fe(II) in the original sample at the y-intercept. Table 3.2 Results of the Atomic Absorption Studies (mole % of Fe)a Sample b Calibration Curve Standard Addition Theoretical % of Fe in sample [Fe(CO)6][Sb2Fn]2 from AsFs 2.0% 1.9% 2.5% n. a.d 5.0% [Fe(CO) 6][Sb 2Fu]2 from C l 2 2.7% 2.4% 3.3% 2.6% 5.0% Fe[SbF 6] 2 c 9.9% 10.1% 12.0% n. a.d 10.6% a: Absolute uncertainty in results is + 0.5%. b : All samples were from different preparations, except that the same samples were used for the second calibration curve study and the first standard addition study. c : Samples were made from reacting Fe metal powder with liquid SbFs in S 0 2 according to eq. 1.7 in Chapter 1. d : n. a. = not applicable The calibration curve is the most common and direct method for determining the Fe(II) content in the sample, whereas the standard addition method may be used to correct the matrix 68 difference between the standard and the sample solutions.13 The different values from the second calibration curve study and the first standard addition study suggest that the matrix difference between the standards and the samples can significantly affect the results. The results are unexpected, as the percentage of Fe found experimentally is generally lower than the theoretical values. Due to the presence of Fe[SbF6]2, the Fe content in the samples should be higher than calculated. This may be an indication of non-Fe containing by-products, in addition to Fe[SbF6J2, in the samples. Details of the by-products associated with each preparation of [Fe(CO)6][Sb2Fn]2 are presented later in this chapter. It can also be seen that samples prepared by the same method at different times may contain different amounts of each by-product. //. Titrations Complexometric titrations with EDTA (using indicator variamine blue) and potentiometric titrations14 carried out in an attempt to determine the Fe content in samples of [Fe(CO)6][Sb2Fn]2 were unsuccessful. Detection of the endpoint was virtually impossible, perhaps due to interference of some F and/or Sb containing species. 3. Microanalysis The microanalysis for carbon in [Fe(CO)6][Sb2Fi i ] 2 was unsatisfactory. The observed value of 0.48% is much lower than the calculated value of 6.38%. This is probably because the high fluorine content of the samples, which may cause CF4 to be produced during combustion. Therefore, it can be concluded that elemental analyses for [Fe(CO)6][Sb2Fn]2 are difficult and not overly precise, and thus should not be considered reliable. 69 3.1.4 Possible Formation of Fe[SbF6]2 1. Thermal Decomposition of[Fe(CO)6][Sb2Fn/2 and [Fe(CO)6][SbF6]2 Both [Fe(CO) 6][Sb 2Fi,] 2 and [Fe(CO)6][SbF6]2 are stable up to approximately 150 °C. When a bulk sample of [Fe(CO)e][Sb2Fii]2 is heated to 150 °C in a melting point capillary tube, a clear colorless liquid, identified as SbFs by its vibrational spectrum and physical appearance, is initially formed: [Fe(CO) 6][Sb 2F n] 2 — [Fe(CO)6][SbF6]2 + SbF5 (eq. 3.5) On further heating, the evolution of CO is noted. The decomposition may be formulated as the following: [Fe(CO)6][SbF6]2 — Fe[SbF 6] 2 + 6CO (eq. 3.6) Details of thermal events accompanying the heating of [Fe(CO)6][Sb2Fn]2 and [Fe(CO)6][SbFe]2 were studied using Differential Scanning Calorimetry (DSC). 2. Differential Scanning Calorimetric (DSC) Studies on [Fe(CO)6][Sb2Fii]2 and [Fe(CO)6][SbF6]2 i. DSC Studies on [Fe(CO)6][Sb2Fii]2 (see Appendix 2 for experimental details of the DSC Studies) The DSC plot for [Fe(CO)6][Sb 2F n] 2 , from 25 - 250 °C, is shown in Figure 3.8. The assignments of the thermal events of [Fe(CO)6][Sb 2Fn] 2 are as follows: (1) The small peak with onset at about 100°C (Event #1) is tentatively assigned to liberation of FfF from the lattice. The isolation of [Ir(CO)6][SbF6]3.4HF and its structural characterization1 5'1 6 provide conclusive evidence for the electrophilic character of the carbonyl C-atom. With HF coordinated to C over F, the structure gives a clear indication of how superelectrophilic metal carbonyl cations are solvated in HF-SbFs (see Figure 3.9). 70 Since all [M(CO)„] cations (n = 2, 4, 6; m = 2,3), among them [Fe(CO)6] , are stable in HF-SbFs, 3 ' 1 7 the formation of HF solvates is very probable. As discussed in the previous sections, HF and HF-SbFs are used extensively in the preparation of [Fe(CO)6]2+ salts with both [SbFe]" and [Sb2Fn]~ as anions. Although no solvated HF was found in the crystal structure of [Fe(CO)6][SbF6]2 or [Fe(CO)6J[Sb2Fi i]2, the samples used for these DSC studies are bulk samples prepared in HF-SbFs. Therefore, the retention of solvated HF in these samples is plausible. Furthermore, this event is found to be irreversible, consistent with the liberation of a volatile product. It should be noted that, however, identification of HF in the gas phase by IR in glass cells is very difficult. Event #3: o ' — ' — 1 — 1 — 1 — 1 — ' ' 1 ' ' ' ' ' 1 ' 1 ' ' ' ' ' 50 100 150 200 250 Temperature (°C) Figure 3.8 DSC Plot of[Fe(CO)6J[Sb2F11J2: Events from 25 - 250 °C. 71 <gR>0(1) #C(1) Figure 3.9 The molecular structure of [Ir(C0)6][SbF$]3AHF. 1 6 The [SbF6J- anions and the other three F atoms of HF are omitted for clarity. (2) The shoulder with onset at about 170 °C (Event #2) is assigned to the decomposition of the anion: [Sb2Fn]- - [SbF6]" + SbF 5 ( 1 ) (eq. 3.7) This assignment is supported by the fact that a viscous liquid, presumably SbFs, is observed as bulk samples are heated to 150 °C. After the sample has been heated to 190 °C, the Raman spectrum shows strong fluorescence with the observation of only one weak, broad peak of 653 cm"1. This peak is assigned to the A i g Sb-F stretch of [SbF6]".2 The presence of [SbFe]" is consistent with the proposed decomposition of [SbaFn]" (see eq. 3.7). Moreover, this event is absent in the DSC curve for [Fe(CO)6][SbF6]2. This event is also irreversible. (3) Event #3, identified by the large peak with onset at approximately 185 °C, is assigned to the decarbonylation (complete loss of CO) of [Fe(CO)6]2 + and a disproportionation reaction of CO: 72 2 C O ( g ) ^ C 0 2 ( g ) + C ( s ) (eq.3.8) This disproportionation reaction is the reverse of the Boudouard Equilibrium, and is exothermic by 38.8 kcal on heating of C O . 1 8 Evidence for its involvement here was obtained from analysis of the gaseous products on heating the sample above 250 °C and of the black solid residue, presumably carbon (see Appendix 2 for details). ii. DSC Studies on [Fe(CO)6][SbF6]2 The DSC plot for [Fe(CO)6][SbF6]2, from 25 - 250 °C, is shown in Figure 3.10. The assignments of the thermal events of [Fe(CO)6][SbFe]2 are as follows: (1) Event #1, with onset at 95 °C, is irreversible and is assigned to liberation of HF from the lattice as discussed for [Fe(CO)6][Sb2Fn]2. (2) The reversibility of Event # 2, with onset at 150 °C, suggested that it may be due to a structural phase transition as follows: <150°C: a phase 150 - 170°C: a phase ±5 p phase >170°C: P phase (3) Event # 3, with onset at 180 °C, is irreversible and involves the overlap of at least two peaks. This event can be assigned to the loss of CO, since the onset temperatures and irreversibility, along with the comparison with the thermal events of [Fe(CO)6][Sb2Fn]2 suggest that decarbonylation of [Fe(CO) 6] 2 + in [Fe(CO)6][SbF6]2 begins at about 180 °C. CO evolution was observed in the gas phase from 220 °C and up, and no C-0 stretch for a metal coordinated CO ligand was observed at this point. This indicates that a significant amount of CO has been lost before 220 °C. From 240 °C up, the gas phase was found to 73 contain both CO and C 0 2 and the solid residue was black, suggesting disproportionation of CO (see eq. 3.8). Event # 3 : mW/mg 0 50 100 150 200 250 300 Temperature (°C) Figure 3.10 DSC Plot of [Fe(CO)6] [SbF6]2: Events from 25 - 300 °C A comparison between thermal events and approximate onset temperatures of [Fe(CO)6][Sb2Fn]2 and [Fe(CO)6][SbF6]2 is shown in Table 3.3. For both carbonyl compounds, the thermal events of liberation of HF from the lattice, decarbonylation of the cation and disproportionation of CO occur at about the same temperatures. The thermal decompositions are not reversible. The conversion of [Fe(CO)6][Sb2Fn]2 into [Fe(CO)6][SbF6]2 (see eq. 3.7) at about 170 °C is not cleanly separated from the decarbonylation of [Fe(CO)6]2 +, hence this conversion by controlled pyrolysis is not a practical route to [Fe(CO)6][SbFe]2- The washing of [Fe(CO)6][Sb2Fii]2 with HF to give [Fe(CO)6][SbF6]2 remains the preferred method. 74 Table 3.3 A comparison between thermal events and approximate onset temperatures of [Fe(CO)6][Sb2Fn]2 and [Fe(CO)6][SbF6]2 Events / Approximate Onset Temperature (°C) [Fe(CO)6][Sb2F„]2 [Fe(CO)6][SbF6]2 Liberation of HF from lattice is suggested 100 95 Phase Transition N . 0 * 150 Decomposition: [ S b 2 F n ] ~ —> [SbFe]" + SbFs(i) 170 N . O * Decarbonylation of [Fe(CO)6] Disproportion: 2 C O ( g ) ^+ C 0 2 ( g ) + C(S) 185 180 *: N.O. = not observed The decarbonylation of [Fe(CO)6][SbFe]2 occurs at about 180 °C; therefore, it can be concluded that thermal decomposition of the carbonyl complex alone is not responsible for the large amounts of Fe[SbF6]2 found in the samples, which were prepared at 40-80 °C. For both compounds, the main event, which consists of featureless, partly overlapping peaks, suggests complete loss of CO according to: [Fe(CO)6][SbF6]2 -» Fe[SbF 6] 2 + 6CO (eq. 3.9) There are no distinct steps observed, and no well-defined intermediate cations of the type [Fe(CO)6- n] 2 + are detected or judged to be isolable as: [Fe(CO)6][SbF6]2 -» [Fe(CO)6.n][SbF6]2 + nCO (eq. 3.10) While CO is formed during thermal decomposition, CO2 clearly results from a secondary process (see eq. 3.8). No COF 2 is observed as a decomposition product for either of the two Fe compounds, while a significant amount of this material is formed in the decomposition of [M(CO)4][Sb2Fn]2 (M = Pt or Pd). 7 This is due to the interionic interactions being relatively 75 less important, with less C--F contacts established in the solid-state structure, for [Fe(CO)6][Sb2Fii]2 and [Fe(CO)6][SbF6]2 than for [M(CO) 4][Sb 2F n]2 (M = Pt or Pd). 7 Also, the DSC results indicate that Fe(II) is a stable oxidation state. Fe[SbFe]2 and, at higher temperatures, eventually FeF 2 forms as decomposition products. [Fe(CO)6][SbF6]2 appears to undergo a structural phase transition while no such event is observed for [Fe(CO)6][Sb2Fn]2. This suggests that the phase transition in [Fe(CO)6][SbFe]2 may be related to a different packing of the symmetrical, essentially octahedral cations and anions. At a high temperature of 150 °C, the effect of secondary interionic contacts may be weakened. The slightly elongated cationic octahedron and the slightly compressed anionic octahedron may be further distorted axially, or become more regularly octahedral during the phase transition. 2. Pathways for the Formation of FefSbFrfz At this point, the possibility of detecting a high spin configuration of [Fe(CO)e]2+ or [Fe(CO)e]3+ can be ruled out (see Section 1 of 3.1.3). Fe[SbF 6] 2 is identified as the sole paramagnetic by-product, through the similarity of its magnetic moments temperature dependence compared to those of the [Fe(CO)6][Sb 2Fn] 2 samples prepared with CI2 and AsFs. Furthermore, Fe[SbFe]2 was obtained in early attempts to prepare [Fe(CO)e]2+ from high spin FeX 3 (X = CI or SO3F) by reductive carbonylation.1 The low spin complexes M(SC«3F)3 (M = Ru or Os) can be reduced to give the corresponding hexacarbonyl complexes [M(CO)6][Sb2Fn]2 • A reason for this different behaviour is found in the magnetic ground states of the precursors. Both M(S03F)3 (M = Ru or Os) are low spin complexes and have magnetic properties consistent with a 2T2 g octahedral ground state between 300 and 80 K , 1 9 > 2 0 whereas FeX 3 (X = CI or S0 3 F) 2 1 are high spin complexes with magnetic properties consistent with a 6 A i g ground state. In case of M(S0 3 F) 3 (M = Ru or Os), the eg set of orbitals (d z 2 and d x 2- y 2) would be available as acceptor orbitals to 76 interact with CO. Hence, formation of [Fe(CO)6] with a low spin ground state would require a large amount of additional (pairing) energy for a change from high to low spin during reduction, a step that is not necessary for Ru(III) or Os(III).17 Instead of [Fe(CO)6][Sb2Fi i ] 2 , the reduction of Fe(S03F)3 gives quantitatively Fe[SbF6]2, which is also a high spin complex. The thermal decomposition of [Fe(CO)6][Sb2Fn]2 yields first [Fe(C0) 6][SbF 6] 2 and then Fe[SbF 6] 2. This decomposition at 180 °C, as the DSC curve shows, is strongly endothermic and provides 2"t" sufficient energy for a low spin ([Fe(CO)e] ) to high spin conversion. However, this thermal decomposition does not occur at temperatures at which [Fe(CO)6][Sb2Fn]2 is synthesized (40 -80 °C). Therefore, [Fe(CO) 6] 2 + cannot be the source of large amounts of Fe[SbFe]2 found in the samples. By comparing the different sets of conditions, it was found that the formation of Fe[SbFe]2 can best be avoided by the use of XeF2 in HF-SbFs for the oxidative carbonylation of Fe(CO)s into [Fe(CO)6][Sb2F] i ] 2 . It is notable that the reaction or oxidation that yields cleaner products takes place in a different solvent system, strongly ionizing HF-SbFs. From this, a possible reaction pathway can be postulated. The formation of [Fe(CO)6][Sb2Fn]2 is thought to proceed via the possible intermediates Fe(CO) 4Cl 2 or "Fe(CO) 4 F 2 " : 2 2 ' 2 3 SbF5 Fe(CO) 5 + C l 2 • Fe(CO) 4Cl 2 + CO (eq. 3.11) SbF5 Fe(CO) 5 + AsF 5 : • "Fe(CO) 4F 2" + AsF 3 + CO (eq. 3.12) HF SbF Fe(CO) 5 + XeF 2 : — ^ "Fe(CO) 4F 2" + Xe + CO (eq. 3.13) The known complexes, Fe(CO) 4X 2 (X = C I , Br and I), are unstable, with stabilities increasing in the order of CI < Br < I . 2 2 - 2 3 Fe(CO) 4Cl 2 can be prepared from Fe(CO) 5 and C l 2 2 2 > 2 3 or C O C l 2 2 4 while "Fe(CO) 4F 2" has not been observed 2 5 The compound [Fe(CO) 5Cl]Cl may be a possible intermediate; however, this compound is unlikely to be obtained under these reaction 77 conditions since it can only be synthesized at low temperature (about -100 °C) in liquid H Q . 2 6 Also, [Fe(CO)sF]F has not been observed. There may be two competing reactions for the Fe(CO)4X2 intermediates in excess SbFs (see Figure 3.11): 1. Decomposition leading to Fe[SbFe]2 via: Fe(CO) 4 X 2 -> FeX 2 + 4CO (X = CI or F) (eq. 3.14) FeCl 2 + 2SbF5 -> FeF 2 + 2SbF 4Cl (eq. 3.15) FeF 2 + 2SbF5 -> Fe[SbF 6] 2 (eq. 3.16) 2. Addition of CO to form [Fe(CO)6][Sb2Fi,]2: Fe(CO) 4Cl 2 + 6SbF5 + 2CO -» [Fe(CO)6][Sb2Fii]2 + 2SbF 4Cl "Fe(CO) 4F 2" + 4SbF5 + 2CO -> [Fe(CO)6][Sb2F,i]2 Support for eq. 3.17 comes from the observation that Fe(CO)4Cl2 can be converted to [Fe(CO)6][Sb2Fn]2 in rather low yie ld . 2 2 - 2 3 Fe(CO) 5 CI, (in S b F 5 ) or A s F s (in S b F 5 ) or X e F , (in H F - S b F 5 ) (eq.3.17) (eq. 3.18) Fe(CO) 4 X 2 (X = CI or F) +2SbF 5 F e d , 2 C O S b F 5 o r H F - S b F 5 +2SbF, competing reactions [FefCOJISb.FJ, Fe[SbF 6] 2 Figure 3.11 Proposed reaction pathway for the formation of FefShFffi in preparations of [Fe(CO)6][Sb2Fn]2. 78 When Fe(CO) 5 (Fe(0)) is oxidized to the intermediates Fe(CO) 4Cl 2 or "Fe(CO) 4F 2" (Fe(II)), n-backdonation becomes unfavourable due to the increased charge on iron. This labilizes the CO ligands and makes the intermediates unstable and prone to decomposition. For the reactions with AsFs and C l 2 , SbFs is used as the reaction medium. In excess SbFs, the formation of Fe[SbFe]2 from FeX 2 (X = CI or F) is favoured. Furthermore, addition of CO in conversion into [Fe(CO)6] 2 + appears to be slow under the heterogeneous reaction condition in SbFs, due to the poor ionizing ability and viscosity of SbF5. This allows sufficient time for the intermediates to convert to the paramagnetic by-product Fe[SbFe]2. As a result, a significant amount of Fe[SbF 6] 2 is produced in these reactions employing SbF5 as the reaction medium. On the other hand, HF-SbFs was used as the solvent for the oxidation of Fe(CO)s with XeF 2 . HF-SbFs is strongly ionizing and provides homogeneous reaction conditions. Therefore the reaction is expected to be much faster than in SbFs alone. Fe(CO)s and XeF 2 reacts to form the intermediate thought to be "Fe(CO) 4F 2" (see eq. 3.13). Since the reaction is homogeneous, the intermediate "Fe(CO) 4F 2" is quickly converted into [Fe(CO)e]2+ in HF-SbFs (according to eq. 3.18) before decomposition into FeF 2 occurs (see eq. 3.14). Also, since the amount of SbFs is decreased, the decomposition into Fe[SbF6]2 becomes less favourable (see eq. 3.16). Consequently, only small amounts of the paramagnetic Fe[SbFe]2 are found as impurity in the product. These postulates are supported by the results of magnetic susceptibility measurements, which are consistent with the presence of 34% and 38% of Fe[SbF 6] 2 respectively in the [Fe(CO)6][Sb 2Fn] 2 samples prepared with AsF 5 and C l 2 but only 0.4% in that with XeF 2 . Support also comes from DSC studies (see Section 2 of 3.1.1), where [Fe(CO)e]2+ salts with 79 either [Sb2Fn]" or [SbFe]" as the anion are thermally stable up to almost 185 °C. As previously discussed, decomposition of [Fe(CO)6][Sb2Fn]2 or [Fe(CO)6][SbFe]2 during synthesis (at 40 -80 °C) to give Fe[SbF 6] 2 is highly improbable. Furthermore, once Fe[SbF6]2 is formed, it cannot be converted back into the carbonyl product. This can be seen from the unsuccessful attempts to synthesize [Fe(CO)6][Sb2Fn]2 from Fe[SbFe]2.1 It is proposed that oxidation of Fe(CO)s takes place before carbonylation. Support comes from the conversion of W(CO)6 into [W(CO)6(FSbF5)][Sb2Fn], where a transient metal carbonyl cation, [W(CO)e]2 +, appears to form initially. The intended reaction, addition of CO to give [W(CO)7] 2 + , does not occur; the addition of [SbFe]" to the W centre is observed to give [W(CO) 6(FSbF 5)] + instead. The complex is isolated as [W(CO)6(FSbF5)][Sb2Fn].17'27 3.1.5 Identification of Other By-Products 1. The AsFs / ShFs and Cl2 / ShF5 Systems For the study of [Fe(CO)6][Sb2Fu]2 and [Fe(CO)6][SbFe]2, Raman spectra are more useful than IR, because Raman bands in the Sb-F region are better resolved than the IR bands, as shown in the reference spectra (Figures 3.12 and 3.13).2 The asymmetric C-0 stretching frequency observed for the samples synthesized with AsFs, CI2 and XeF 2 (powder and single crystals) are all at 2204 cm"1 (Ti„) and the symmetric stretches are at 2220 cm"1 (E g) and 2241 cm"1 (Ai g ) . These values are identical to published data.1'2 However, the IR band at 2256 cm"1 (with a shoulder at 2250 cm"1) reported previously for the sample prepared from CI2 is absent. This band, erroneously assigned to [Fe(CO)6] 3 + before,1 is likely the C-0 stretching band of [ClCO] + . 8 In this case, the only carbonyl complex present in the samples is of Fe(II). The Sb-F stretches for all samples are very similar. However, there are some differences in the spectra, 80 especially in the lower wavenumber region of the vibrational spectra, for [Fe(CO)6][Sb2Fn]2 when synthesized from the different oxidizing agents (see Table 3.4). Figure 3.13 Vibrational spectra of [Fe(CO)(,][SbF6]2.2 81 Table 3.4 Raman data:" different samples of [Fe(CO)6][Sb2Fi j]2 prepared with various oxidizing agents compared with published data2 (200 - 700 cm'1, only the Sb-F deformation and stretching regions are shown). powder / powder / single crystals / published data b 2 Cl 2 AsF s XeF 2 207.4 (vw) 226.6 (w) 234 (vw) 235.6 (w) 282 (m) 275.5 (w) 293 (m) 298 (w) 299.3 (m) 332.6 (vw) 391 (s) 441 (m) 591.3 (w, sh) 601 (m) 601 (m) 598 (m) 598.1 (m) 601 (vw, sh) 647 (m) 651 (s) 647 (m) 647.9 (s) 668 (m) 668 (m) 670 (m) 670.5 (s) 687 (m) 687 (w) 687 (w) 686.8 (m) 694.5 (m) a : Intensities are given in parentheses: v = very, w = weak, m = medium, s = strong, sh = shoulder, b = broad b,: The cation bands (347, m; 361, w, sh; 500.9 ww) are not listed. The Raman spectra for the [Fe(CO)6][Sb2Fn]2 samples made from AsF 5 and single crystals of [Fe(CO)6][Sb2Fn]2 prepared with XeF 2 are consistent with the published data,1'2 whereas two additional bands at 391 and 441 cm"1 are obvious in the Raman spectra for the samples synthesized using C l 2 (see Table 3.4). These bands may be the Sb-Cl bands of the substituted anion, [Sb2Fn_nCln]" (in comparison to the 399 and 444 cm"1 bands of SbF 3 Cl 2 2 8 ) . For the reaction using AsFs (see eq. 3.1), removal of the reduced by-product, AsF3.SbF 5, 2 9 presents another problem. This by-product can be removed from the product mixture by solvent 82 extraction with SO2 to produce the thermally less stable SCh-SbFs 1 0 and ASF3 in a displacement reaction; however, the washing process results in partial removal of SbFs from the anion [SbiFn]" to give some [SbFe]". When CI2 is used, a solid by-product of the approximate composition SbF4Cl, which can only be removed by sublimation, was formed (see eq. 3.2). 2. The XeF2/HF-SbFs System It has been found that the synthetic route using XeF 2 (see eq. 3.3) is preferred. Crystalline products, which allow for the determination of the molecular structure by X-ray diffraction, have been produced by this route. The reduced by-product, Xe, is volatile and can easily be removed in vacuo. There is now only a trace amount of paramagnetism due to Fe[SbF6J,2 observed (see Section 1 of 3.1.3). However, the vibrational spectra of bulk materials obtained in this manner raise doubts regarding the bulk composition of the reaction product as [Fe(CO)6][Sb2Fn]2, as there is a difference in the vibrational spectrum in the Sb-F stretching region (550 - 700 cm"1) in the reported data (see Figures 3.12 2 and 3.14). Instead of having six bands in the Sb-F stretching range (590-710 cm"1) of medium to low intensity, there are now strong bands at 634 and 655 cm"1 for the bulk sample with the latter also found in [SbF6]" (vi, A i g ) 2 and in the mixed valency compound 6SbF 3.5SbF 5. 3 ( ) In addition, v 2 of [SbF6]" at 570.7 cm"1 2 is also observed. Moreover, the intensity of the Sb-F bands is exceptionally strong compared to the C-0 bands. These suggest that there may be other Sb and F containing species, in addition to the counteranion, [SD2F11]". 6SbF3.5SbF5 and [SbF6]" are the likely impurities present in the crude crystalline sample. Unfortunately, all these species have bands in the Sb-F stretching region that are very close to each other. Identification of bands from individual species in the mixture is thus difficult. 83 I 1 1 1 1 — i 1 1 1 1 1 r 2500 2300 700 500 300 100 Wavenumber (cm-1) Figure 3.14 Raman Spectrum of crude [Fe(CO)6][Sb2Fu]2 crystalline solids in the bulk sample. In Table 3.5, some of the Raman bands for the crude product in the bulk sample are compared to published data for [Fe(CO) 6][Sb 2Fii]2, 2 [Fe(CO) 6][SbF 6] 2 2 and 6SbF3.5SbF5. 3 0 The bands that are likely to be of the cation [Fe(CO)6] 2 + are omitted in Table 3.5; and the bands listed should all be due to Sb-F stretching and deformation modes. 84 Table 3.5 Raman data:" "crude " crystalline [Fe(CO)6] [Sb2Fi /]2 bulk sample obtained from experiment, of single crystals of [Fe(CO)ej'[Sb2Fu]'2 and published data of [Fe(CO)(,]'[SbF^J'2 2 and of6SbF3.5SbF5 3 0 (200 - 700 cm') "crude" crystalline solids single crystals of single crystals of 6SbF3.5SbF5 3 0 of[Fe(CO)6][Sb2F„]2 rFe(CO) 6][Sb 2F„l 2 b 2 [Fe(CO)6][SbF6]22 209 (vw) 207.4 (vw) 226.6 (w) 235.6 (w) 275.5 (w) 289 (m) 299.3 (m) 332.6 (vw) 281.4 (s) 282 (mw) 294 (w) 528 (w) 566 (m) 570 (w) 591.3 (w, sh) 598.1 (m) 601 (vw, sh) 570.7 (m) 577 (m) 622 (w) 634 (vs) 647.9 (s) 652 (vs) 655 (vs) 652.6 (vs) 659 (s) 668 (s) 670.5 (s) 686.8 (m) 694.5 (m) 681 (w) : Intensities are given in parentheses: v = very, w = weak, m = medium, s = strong, sh -shoulder, b = broad b : The cation bands (347, m; 361, w, sh; 500.9 vvw) are not listed.2 i. The Presence of [Fe(CO)6][SbFe]2 in the [Fe(CO)(,][Sb2F]/J2 Sample As mentioned in the previous section, one of the possible by-products in the bulk sample of [Fe(CO)6][Sb2Fn]2 is [Fe(CO)6][SbF6]2. [Fe(CO) 6][Sb 2Fii]2can readily be converted into [Fe(CO)6J[SbF6]2, because the SbFs moiety of [Sb2Fi 1]" is relatively unimportant in the involvement of interionic contacts of the complex (see Figure 3.3). The anion [SbF6]" is clearly identified by the E g mode at 570 cm"1, and less unambiguously by the 655 cm"1 band, most likely due to the A j g mode of [SbFe]".2 Very strong bands at 634 and 655 cm"1 may also be attributed to 6SbF3.5SbF5 (see Figures 3.12 to 3.14 and Table 3.5). It should be noted that in HF-SbFs the formation of [SbFe]" is likely, especially when the concentration of SbFs is low. This is shown by the following equilibrium: HF [Sb2Fn]"(soiv) + HF < > HF-SbF 5 + [SbF 6]" ( s o I v ) (eq.3.19) SbF5 ii. The Presence of 6SbF3.5SbF$ in the [Fe(CO)e][Sb2Fi/]2 Sample Another by-product in the bulk sample of [Fe(CO)6][Sb2Fn]2 appears to be 6SbF3.5SbFs. This by-product results from oxidation of Fe(CO)s by SbFs to give SbF 3 , which subsequently forms an adduct with SbFs. Although oxidation of Fe(CO)s by SbFs had not been previously observed and is not expected in the presence of a much stronger oxidizing agent such as XeF2, the oxidation by SbFs might occur when the amount of XeF 2 in the system is decreased. A well-known side reaction of XeF2 with SbF 5 would give a number of adducts, of which yellow solid XeF2.2SbF5 is structurally characterized.3 1'3 2 The molecular structure of XeF2.2SbFs is interpreted as [XeF][Sb2Fn], which has a Xe-F stretch at 621 cm"1 in the Raman spectrum.31 Unfortunately, this band falls into a region in the Raman spectrum of crude [Fe(CO)6][Sb2Fn]2 where all the other Sb-F stretches are observed, making its unambiguous identification difficult. Therefore, xenon difluoride can exist in the system in several forms: XeF 2( S Oiv.), [XeF]+(SOiv.), solid [XeF][Sb2Fn] and possibly other adducts or salts formed with SbFs. XeF 2( S 0iv.) and [XeF]+(S Oiv.) are both good oxidizing agents, and are able to oxidize Fe(0) into Fe(II). The formation of the [XeF][Sb2Fn] salt, on the other hand, decreases the amount of XeF2(soiv.) or [XeF]+(S0|V.). Therefore, it is plausible that oxidation of Fe(CO)s by SbF 5 occurs because some XeF 2 is likely to undergo side reactions with SbFs, which is seen to decrease the amount of 86 XeF2 available for oxidation. In summary, the bulk composition of [Fe(CO)6][Sb2Fn]2 product may include Fe[SbF 6] 2, [Fe(CO)6][SbF6]2, 6SbF 3 .5SbF 5 and XeF 2.2SbF 5. 3.1.6 Improved Syntheses and Purification of [Fe(CO)6] [Sb2Fn]2 and [Fe(CO)6] [SbF6]2 It was previously found that no reaction is observed when single crystals of [Fe(CO)6J[Sb 2Fi \] 2 were treated with fluorine in an attempt to prepare the Fe(III) salt, [Fe(CO)6][Sb2Fi i ] 3 . 2 The oxidative stability of [Fe(CO)6][Sb2Fn]2 towards fluorine forms the basis for the following purification procedure: The "crude" crystals [Fe(CO)6][Sb2Fn]2, prepared as described in Section 1 of 3.1.2, were transferred to a 100 mL round bottom PFA vessel and subsequently 8 mL HF and 1 bar F 2 were introduced into the reactor. The mixture was stirred at room temperature for two days. Upon removal of volatiles, a white powder remained. This was identified as highly pure [Fe(CO)6][Sb2Fu]2. A Raman spectrum (Figure 3.15) that was identical to the reported one (Figure 3.12)2 was obtained for this product upon purification by fluorination. An attempt to purify the crude [Fe(CO)6][SbF6]2 product, prepared as described in Section 1 of 3.1.2, was carried out by introducing 6 mL HF and 1 bar F2 onto the solids in a round bottom PFA flask. Highly pure [Fe(CO) 6][Sb 2Fn] 2 (2.672 g) in the form of a white powder was obtained upon removal of all volatiles. Some of the purified [Fe(CO)6][Sb2Fj\]2 (0.864 g, 0.765 mmol) was then converted back to [Fe(CO)6][SbF6]2 (0.515 g, 0.741 mmol, 96.8% yield) by washing with HF as described above. Raman spectra (Figure 3.16) of [Fe(CO)6][SbFe]2 synthesized from purified [Fe(CO)6][Sb2Fn]2 are also identical to those of its single crystals (Figure 3.13) 2 87 CM I 1 1 1 - I 1 1 1 1 1 1 1 2500 2300 800 600 400 200 Wavenumber (cm- 1) Figure 3.15 Raman spectrum of purified [Fe(CO)6][Sb2Fii]i (after treatment with Fi in HF) During the work-up step in purification of [Fe(CO)6][Sb2Fii]2, F2 was introduced onto the product dissolved in HF. With [Fe(CO)6][SbF6]2 and 6SbF3.5SbF5 present in the sample, SbFs could be produced from 6SbF3.5SbF5 under F2 atmosphere in HF. Subsequently, there is enough SbF 5 for the conversion of [Fe(CO)6][SbF6]2 into [Fe(CO)6][Sb2F,i]2: 8 8 2(6SbF3.5SbF5) + 12F2 -> 22SbF5 (eq. 3.20) 2SbF5 + [Fe(CO)6][SbF6]2 - * [Fe(CO)6][Sb 2F„] 2 (eq. 3.21) As a result, purified powder of [Fe(CO)6][Sb2Fn]2 is obtained, regardless of whether [Fe(CO)6][SbF6]2 or [Fe(CO)6][Sb2Fn] 2 was the major component in the initial product mixture. From this preparation of [Fe(CO)6][Sb2Fn]2, [Fe(CO)6][SbF6]2 was obtained in high yield (96.8%) by the method of repeated washing with HF. 1.5H 1.0 0.5 c 0 +=* c c co £ CM CM rr CM CM CD i— CM IT) C O IS r r 2500 2300 700 500 Wavenumber (cm"1) 300 100 Figure 3.16 Raman spectrum of [Fe(CO)6][SbF(,]2 obtained from purified [Fe(CO)(,] [Sb2Fy//? 89 3.1.7 Discussion /. Comparison of the Carbonyl Cations of Fe with those of Os, Ru The synthesis of hexacarbonylaithenium(II) undecafluorodiantimonate(V), [Ru(CO)6][Sb2Fn] 2 and hexacarbonylosmium(II) undecafluorodiantimonate(V), [Os(CO)6][Sb2Fn] 2 via reductive carbonylation of the respective M(S03F) 3 (M = Ru or Os) in liquid SbF5 under a CO atmosphere was first described by Wang et a l . : 3 3 2M(S0 3 F) 3 + 16SbF5 + 13 CO SbFsrn, 1 atm CO 60 - 90 °C, 2 - 4 days 2[M(CO) 6][Sb 2F,,] 2 + C 0 2 + S2O5F2 + 4Sb 2F 9(S0 3F) (eq. 3.22) (M = Ru or Os) These carbonyl salts are collected as white, moisture-sensitive solids, with a very similar appearance to [Fe(CO)e][Sb2Fn]2. Reductive carbonylation of metal trisfluorosulfates in a strongly acidic medium, which was first employed in the generation of [Au(CO)2]+ 3 4 in H S 0 3 F and then extended to [M(CO) 4][Sb 2Fn]2 (M = Pd or Pt) 7 in SbF5, is also found to be very effective for the synthesis of [M(CO)6 ] [Sb2Fn] 2 (M = Ru or Os). 3 3 However, attempts to extend this approach to the synthesis of [Fe(CO)6][Sb2Fn]2 have not been successful, and Fe[SbFe]2 was formed in all attempts. One possible explanation for this failure (concerning the spin states of the starting complexes, metal trisfluorosulfates, and the product carbonyls) has been presented in the Section 3.1.4. An improved synthesis of [Os(CO)6][Sb2Fn] 2 has also been described recently, 3 5- 3 6 where reductive carbonylation of OsF6 by CO is carried out in liquid SbFs. This is a unique route for Os due to the non-existence of MF6 (M = Fe or Ru) . 3 3 The corresponding -[SbFe]" salts, hexacarbonylruthenium(II) hexafluoroantimonate(V), [Ru(CO)6][SbFe]2 and hexacarbonylosmium(II) hexafluoroantimonate(V), [Os(CO)6][SbFe]2, have also been prepared by repeatedly washing the corresponding carbonyl [ S b 2 F n ] ' salts with 90 H F , 3 5 employing the procedure described earlier for the Fe salts. As in the case of the [Fe(CO)6][Sb2Fi i ] 2 synthesis, the use of HF-SbFs instead of SbFs produces a mixture of [M(CO) 6 ] 2 + (M = Ru or Os) with both [Sb2Fn]" and [SbF6]" as anions 3 5 (see eq. 3.19). Single crystals of [M(CO) 6][Sb 2Fn] 2 and [M(CO) 6][SbF 6] 2 (M = Ru or Os) have also been obtained from recrystallization of the respective carbonyl salts in HF-SbFs solutions.35 A l l of the cations in these salts are octahedral. The structural and spectroscopic parameters of the carbonyl -[Sb2Fn]~ and -[SbFe]" salts for the triad (M = Fe, Ru and Os) are summarized in Table 3.6. Table 3.6 Structural and spectroscopic parameters for [M(CO) 6][Sb2Fi / J2 and [M(CO)6][SbF6]2 (M = Fe, Ru and Os)2'35 Compound Structural parameters Spectroscopic Space Unit Cell Average Average parameter Group Volume M-C C O Average vC-O (A3) Distance (A) Distance (A) (cm1) [Fe(CO) 6][Sb 2F u] 2 Fl\ln 1198.4(2) 1.911(5) 1.104(6) 2215 [Fe(CO)6][SbF6]2 PAImnc 850.5(2) 1.908(7) 1.108(9) 2215 [Ru(CO) 6][Sb 2F n] 2 Plxln 1215.1(2) 2.039(5) 1.094(10) 2214 [Ru(CO) 6][SbF 6] 2 PAImnc 853.1(2) 2.024(5) 1.101(7) 2214 [Os(CO) 6][Sb 2F n]2 P2xln 1210.8(2) 2.027(5) 1.102(7) 2209 [Os(CO)6][SbF6]2 PAImnc 850.3(2) 2.022(10) 1.104(7) 2209 From this table, one can see that the structural and spectroscopic parameters for all six salts are very similar. Structural and spectroscopic properties of [M(CO)6J (M = Fe, Ru and Os) are 91 independent of the anions. Both triads are isostructural. The vibrational spectra of the cations are practically identical with respect to the band positions and intensities in the C-0 stretching region.1 Also, the CO stretching bands are not subject to vibrational mixing. 3 3 The average M - C and C-0 distances are virtually identical as well. Furthermore, the three [Sb2Fn]" or [SbFs]" salts have the same space groups with cell dimensions that are almost independent of the metal. Preliminary studies show that the DSC curves for [M(CO)6][Sb2Fn] 2 (M = Fe, Ru and Os) are very similar, with thermal stabilities in the order of Fe < Ru < Os . 3 3 Both [M(CO) 6][Sb 2Fii]2 (M = Ru or Os) solids shrank at about 170 °C and decomposed at 300 °C to black residues,33 whereas [Fe(CO)6][Sb2Fn] 2 solid gives off SbF5(i) at about 150 °C to become [Fe(CO)6][SbFe]2 and starts turning black at a lower temperature of approximately 185 °C. It has been suggested that "[Os(CO)6]3+" can be synthesized from the reductive carbonylation of OSO4 by CO in SbFs, due to an IR band observed at 2253 cm"1 in addition to the vco of [Os(CO) 6] 2 + at 2209 cm" 1 . 2 2 However, as in the case for "[Fe(CO)6]3+", this suggestion is found to be in error. The 2253 cm"1 band was later reassigned to the fully characterized compound tetrakis(carbonyl)dioxoosmium(VI) cation, trans-[Os0 2(CO)4] 2 +. 3 7 The discussion on "[Fe(CO) 6] 3 +" is presented in Section 3.2. 2. Conversion of [Sb2Fu[ into [SbFtf As has been mentioned, the -[SbFe]" salts, [M(CO)6][SbF6]2 (M = Fe, Ru or Os) have been prepared successfully by repeatedly washing the corresponding carbonyl [Sb2Fn]" salts with H F . 3 5 This conversion is also successful for bis(carbonyl)gold(I) undecafluorodiantimonate(V), [Au(CO)2][Sb2Fn]. The synthesis of [Au(CO)2][Sb2Fn] occurs in three steps: 3 4 92 Step 1: Reduction of Au(S0 3 F) 3 by CO in H S 0 3 F Au(S0 3 F) 3 + 3CO ( g ) [Au(CO)2]+ (soiv.) + C 0 2 ( g ) + S 2 0 5 F 2 + S 0 3 F (eq. 3.23) 25 °C Step 2: Nucleophilic substitution of CO by S0 3F" HS03F [Au(CO) 2] + ( s o l V) + S0 3F" • Au(CO)S0 3 F ( s ) + C O ( g ) (eq. 3.24) 80 C Step 3: Conversion of Au(CO)S0 3F in the presence of CO into [Au(CO) 2][Sb 2Fn] Au(CO)S0 3 F + CO + 4SbF5 S b F s » [Au(CO) 2][Sb 2F n] + Sb 2 F 9 S0 3 F (eq. 3.25) In an attempt to recrystallize [Au(CO)2][Sb2Fn] in acetonitrile, substitution of CO was observed:34 - 2 C O [Au(CO) 2][Sb 2F n] + 3CH 3 CN 2 5 oC * [Au(NCCH 3) 2][SbF 6] + SbF 5 .CH 3 CN (eq. 3.26) Notice that in this case the anion is converted from [Sb2Fn]" into [SbFe]". This suggests that conversion of [Au(CO)2][Sb2Fn] into [Au(CO)2][SbF6] in a non-coordinating solvent is feasible. Thus, in the current work the following reaction was attempted: HF [Au(CO) 2][Sb 2F u] + HF ^[Au(CO) 2][SbF 6] + HF-SbF 5 (eq. 3.27) Unlike [Fe(CO)6][Sb2Fn]2 which is insoluble in HF, [Au(CO) 2][Sb 2Fn] is very soluble. Therefore instead of washing the [Sb 2Fn]' salt with HF as in the Fe(II) case, white solids of [Au(CO) 2][Sb 2Fn] were dissolved in HF and stirred at room temperature for about 15 minutes. A purple solid was collected after all volatiles were removed in vacuo. The IR spectra of products collected from this procedure all show a single C-0 stretch at 2195 cm"1 (vs. 2217 cm"1 for [Au(CO) 2][Sb 2F u] 3 4 and 2220 cm"1 for [Au(CO) 2][UF 6]. 3 8) The single band is consistent with the Dooh geometry of [Au(CO) 2] +. Instead of having a number of bands at 500 - 700 cm"1, characteristic of [Sb2Fn]", a single Sb-F stretch characteristic of octahedral [SbFg]" was found at 654 cm"1. The purple colour of the product was attributed to formation of very fine colloidal gold(0) from the partial decomposition of [Au(CO) 2][Sb 2Fn], and the 93 product could possibly be a mixture of [Au(CO)2J [SbFe] with colloidal gold. Unfortunately, the colloidal gold was too fine to be physically separated from the presumably white powder of [Au(CO)2][SbF6]. Even stirring the reactants only briefly at a temperature as low as -60 °C cannot prevent the partial decomposition. No bands in the C-0 stretching region could be observed in the Raman spectrum of the product because of the tremendous fluorescence in the C-0 stretching region possibly due to the colloidal gold. The only band observable in the Raman spectrum is an intense single Sb-F stretch at 640 cm"1, which is consistent with [SbFe]". In comparing the products arising from the treatment of [Fe(CO)6][Sb2Fi i]2 and [Au(CO)2J[Sb2Fn] in HF, one can see that one of the major differences is the partial decomposition of [Au(CO)2][Sb2Fn]. The resistance of [Fe(CO)6]2 + to decomposition can be attributed to its great intrinsic stability and also the additional stabilization of the cation by secondary interionic contacts, as described in Section 2 of 3.1.1. These contacts in [Fe(CO)6][Sb2Fn]2 are mainly due to C--F interactions, which involve all six C atoms of the [Fe(CO) 6] 2 + cation in [Fe(CO)6][Sb2F ] 1]2.2 In contrast, facile substitution of CO by C H 3 C N in [Au(CO) 2][Sb 2Fn] was observed.34 It was also found that [Au(CO) 2][Sb 2F n] can only be heated to about 130 °C without detectable loss of mass or evolution of CO, and formation of COF2 gas, in addition to metallic gold. 3 4 These observations indicate the relative instability of the Au carbonyl cation compared to the Fe carbonyl cation. In contrast, attempts to convert [M(CO)4][Sb2Fii]2 (M = Pd, Pt) 7 into the corresponding [SbF6]" salts were unsuccessful.39 This is probably due to the secondary interionic contacts being too strong for [M(CO)4][Sb2Fii]2 (M = Pd, Pt). As mentioned in Section 2 of 3.1.1, the contact distances for [Fe(CO)6][Sb2Fn]2 are longer and thus the interactions presumably weaker than in 94 [M(CO) 4][Sb 2Fn]2 (M = Pd, Pt). Within the formula unit, there is about one C--F contact per CO group rather than approximately four in [M(CO)4][Sb2Fn]2 where M = Pd, Pt. 7 3.2 The Hypothetical Compound "[Fe(CO)6][Sb2Fi,]3" When the synthesis of [Fe(CO)6][Sb2Fi i ] 2 was first investigated by Bley et al. 1 using CI2 as the oxidizing agent, it was suspected that an intense IR band at 2256 cm"1 suggested the formation of [Fe(CO) 6] 3 +. If "[Fe(CO) 6] 3 + " does exist, the Fe centre is in the +3 oxidation state. It would be reasonable to believe that, having six carbonyl ligands, the d 5 Fe(III) would be in its low spin state with an octahedral geometry. Therefore, there should be one unpaired electron in the t 2 g set which causes Jahn-Teller distortion. The split IR bands of C-0 stretch (with a shoulder) observed were regarded as one of the observations providing evidence for this distortion. This is a rare case as a Jahn-Teller distortion due to degeneracy in the t 2 g set is rarely detected spectroscopically. Other evidences of Jahn-Teller distortion include a large splitting in the 5 7 Fe Mossbauer spectrum and an anisotropic g-value obtained from the EPR spectrum.1 However, magnetic measurements done in this work on samples provided, courtesy of Dr. B. Bley, indicated a magnetic behaviour closer to the high spin case of Fe(III) than to the low spin case. Since Jahn-Teller distortion is impossible for high spin Fe(III) which has no degeneracy in the ground state, the composition of the sample "[Fe(C0) 6][Sb 2Fi i ] 3 " was in doubt. On the basis of the subsequent magnetic studies on [Fe(CO)6][Sb2Fn]2 performed in the current work, it is concluded that the paramagnetism of the "[Fe(CO)6][Sb2Fn]3" sample is largely due to Fe[SbF6]2- The similarity of the temperature dependence magnetic behaviours of the "[Fe(CO)6][Sb2Fn]3" sample and Fe[SbF6]2, which are typical for high spin Fe 2 + in a 5 T 2 g 95 ground state and in an octahedral environment, supports the conclusion. The results of the Mossbauer spectroscopic studies and the EPR studies can also be explained by the presence of Fe(II) in form of high spin, paramagnetic Fe[SbF6J2-It should be noted that, "[Fe(CO)6][Sb2Fn]3" was claimed to be produced only by oxidation with CI2, but not with any other oxidizing agents. Furthermore, no Raman bands corresponding to "[Fe(CO)6][Sb2Fi i]3" were ever observed. Afterwards, investigations by Bernhardt et al. identified the complex related to the particular C-0 stretching frequencies in the IR spectrum to be [C1C0] + . 8 The sample of "[Fe(CO)6][Sb2Fn]3" is determined, in fact, to be a mixture of [Fe(CO)6][Sb2Fii]2, [ClCO][Sb 2Fn] and Fe[SbF 6] 2; the latter being responsible for the large amount of paramagnetism observed in magnetic susceptibility measurements. Attempts to produce [Fe(CO)6] have so far been unsuccessful. 3.3 Summary and Conclusions The initial incentive to study [Fe(CO)6]n + (n = 2 or 3) species was to characterize new paramagnetic carbonyl species. As the carbonyl ligands are predominantly a-bonded, a question was raised as to whether or not they could be regarded as weak field ligands. It was found that predominantly a-bonded carbonyl ligands cause large ligand field splittings and the Fe species formed are low spin. The n = 2 species, [Fe(CO)6]2 +, as the [Sb2Fn]" or the [SbF6]" salt, was found to be diamagnetic. The n = 3 salt, which in the low spin configuration would be paramagnetic, was never isolated. [Fe(CO)6][Sb2Fi t ] 2 of improved purity was synthesized by the improved oxidative carbonylation of Fe(CO)s with XeF 2 in HF-SbF 5 and further purified by treatment with F2 in HF. This synthetic procedure was compared to previous methods of obtaining [Fe(CO)6][Sb2Fn]2 using AsF 5 and C l 2 as oxidizing agents in SbF 5 as solvent. The 96 transformation of [Fe(CO)6][Sb2Fn] 2 into [Fe(CO)6][SbF6]2 by washing with HF provides the first thermally stable [SbFg]" salt of a homoleptic metal carbonyl cation. Molecular structure determination, vibrational analysis and magnetic susceptibility measurements confirm the A j g ground state of the essentially octahedral [Fe(CO)g] cations in both salts. There are fewer and weaker C—F interionic contacts in [Fe(CO)e]2+ and in isostructural [M(CO )6 ] 2 + (M = Ru or Os) salts, thus only slight departures from linearity for the corresponding M - C - 0 (M = Fe, Ru or Os) group, than in [M(CO)4][Sb 2Fii] 2 (M = Pd or Pt) or [Hg(CO) 2][Sb 2Fn] 2. Magnetic susceptibility measurements have subsequently been used to monitor the amount of paramagnetic impurity, identified as Fe[SbFe]2, in the samples. Also, only octahedral homoleptic [M(CO )6 ] 2 + (Fe, Ru or Os) exist with either [Sb2Fn]" or [SbF6]" as counteranions; there is an equilibrium with both [Sb2Fn]" and [SbF(;]~ present in HF-SbF 5. By-products such as Fe[SbF 6] 2, [Fe(CO)6][SbF6]2, 6SbF3.5SbF5 in samples of [Fe(CO)6][Sb 2Fn] 2 were also studied. Thermolysis behaviours studied using Differential Scanning Calorimetry confirmed that, unlike most other homoleptic metal carbonyl cation [Sb2Fn]" salts, [Fe(CO)6][Sb 2Fn] 2 releases SbF5 before CO is given off upon heating at about 185 °C. Thermal decomposition of [Fe(CO)e][Sb2Fu]2 with stepwise loss of CO did not occur and no CO containing decomposition products can be isolated. It was also concluded, from the DSC results, that [Fe(CO)e]2+ is not the source of Fe[SbF6]2. The reaction pathway for the formation of [Fe(CO)e]2+ and the by-product Fe[SbFe]2 via an intermediate of the form Fe(CO)4X2 (X = CI or F) was therefore postulated. A phase transition for [Fe(CO)6][SbFe]2 at about 150 - 170 °C was also detected with DSC. 97 The suggestion1 about the formation of [Fe(CO)6]3+ during the oxidation of Fe(CO)s with Cb is found to be in error. While results from IR, Mossbauer and EPR spectroscopies are found to be suggestive of [Fe(CO)6]3+, magnetic susceptibility measurements indicate otherwise. The composition of the samples initially believed to correspond to [Fe(CO)6][Sb2Fn]3 is found to consist mainly of [Fe(CO)6][Sb2Fn]2, [ClCO][Sb2Fn] and Fe[SbF6J2- There is no evidence to date for the existence of the [Fe(CO)6]3+ cation, and hence [Ir(CO)6]3+ 1 5 remains as the sole homoleptic carbonyl cation with a metal in the 3+ oxidation state. 3.4 References 1) Bley, B.; Willner, H. ; Aubke, F. Inorg. Chem. 1997, 36, 158. 2) Bernhardt, E.; Bley, B.; Wartchow, R.; Willner, H. ; B i l l , E.; Kuhn, P.; Sham, I. H . T.; Bodenbinder, M . ; Brochler, R ; Aubke, F. J. Am. Chem. Soc. 1999,121, 7188. 3) Willner, H. ; Aubke, F. Angew. Chem. Int. Ed. Engl. 1997, 36, 2402. 4) Herzberg, G. Spectra of Diatomic Molecules; 2nd ed.; Van Nostrand: Toronto, Canada, 1966. 5) Bondi, A . J. Phys. Chem. 1964, 68, 441. 6) Bodenbinder, M . ; Balzer-Joellenbeck, G.; Willner, H. ; Batchelor, R. J.; Einstein, F. W. B.; Wang, C ; Aubke, F. Inorg. Chem. 1996, 35, 82. 7) Willner, H. ; Bodenbinder, M . ; Brochler, R.; Hwang, G.; Rettig, S. J.; Trotter, J.; von Ahsen, B.; Westphal, U. ; Jonas, V. ; Thiel, W.; Aubke, F. J. Am. Chem. Soc. 2001,123, 588. 8) Bernhardt, E.; Willner, H. ; Aubke, F. Angew. Chem. Int. Ed. Engl. 1999, 38, 823. 9) Figgis, B. N . ; Hitchman, M . A . Ligand Field Theory and its Applications; Wiley-VCH, 2000. 10) Aynsley, E. E.; Peacock, R. D.; Robinson, P. L. Chem. Ind. 1951, 1117. 98 11) Gantar, D.; Leban, I.; Frlec, B.; Holloway, J. H. J. Chem. Soc. Dalton Trans. 1987, 1987, 2379. 12) Mabbs, F. E., Machin, D. J. Magnetism and Transition Metal Complexes; Chapman and Hall: London, 1973. 13) Skoog, D. A.; Leary, J. J. Principles of Instrumental Analysis; 4th ed.; Saunders College Publishing, 1992. 14) Schwarzenbach, G.; Flaschka, H. Complexometric Titrations; Methuen & Co. Ltd.: London, 1969. 15) Bach, C ; Willner, H. ; Wang, C ; Rettig, S. J.; Trotter, J.; Aubke, F. Angew. Chem., Int. Ed. Engl. 1996, 35, 1974. 16) von Ahsen, B.; Berkei, M . ; Henkel, G.; Willner, H. ; Aubke, F. J. Am. Chem. Soc. 2 0 0 2 , accepted. 17) Willner, H . ; Aubke, F. Angew. Chem. Int. Ed. Engl. 2 0 0 2 , 124, 0000, in press. 18) Sidgwick, N . V. Chemical Elements and Their Compounds; Oxford University Press: London, 1962; Vol. 1. 19) Leung, P. C ; Wong, G. B.; Aubke, F. J. Fluorine Chem. 1987, 35, 607. 20) Leung, P. C ; Aubke, F. Can. J. Chem. 1984, 62, 2892. 21) Goubeau, J. M . , Milne, J. B. Can J. Chem. 1967, 45, 2321. 22) Bley, B. Ph. D. Thesis; Universitat Hannover: Hannover, 1997. 23) Hieber, W. Advances in Organometallic Chemistry; Stone, F. G. A. and West, R., Ed.; Academic Press: New York, 1970, Vol 8, pp 1. 24) Jaluvka, I ; Zima, J. Chem. Zvesti. 1968, 22, 281. 25) Doherty, N . M . ; Hoffman, N . W. Chem. Rev. 1978, 91, 553. 26) Iqbal, Z.; Waddington, T. C. J. Chem. Soc. 1968, 16, 709. 27) Brochler, R.; Sham, I. H. T.; Bodenbinder, M . ; Schmitz, V.; Rettig, S. J.; Trotter, J.; 99 Willner, H . ; Aubke, F. Inorg. Chem. 2000, 39, 2172 -2177. 28) Siebert, H . Anwendwigen der Schwingungsspektroskopie in der Anorganischen Chemie; Springer-Verlag: Berlin / Heidelberg, Germany, 1966. 29) Birchall, T.; Dean, P. A. W.; Valle, B. D.; Gillespie, R. J. Can. J. Chem. 1973, 51, 667. 30) Nandana, W. A. S.; Passmore, J.; White, P. S. J. J. Chem. Soc. Dalton. Trans. 1985, 1623. 31) McRae, V. M . ; Peacock, R. D.; Russell, D. R. Chem. Comm. 1969, 62. 32) Stadky, F. Noble Gases; Emeleus, H. J . , FRS and Gutmann, V. , Ed.; Butterworths: London, 1972; Vol. 3. 33) Wang, C ; Bley, B.; Balzer-Joellenbeck, G.; Lewis, A. R.; Siu, S. C ; Willner, H . ; Aubke, F. J. Chem. Soc, Chem. Commun. 1995, 2071. 34) Willner, H. ; Schaebs, J.; Hwang, G.; Mistry, F.; Jones, R.; Trotter, J.; Aubke, F. J. Am. Chem. Soc. 1992, 114, 8972. 35) von Ahsen, B.; Bernhardt, E.; Bach, C ; Wartchow, R.; Willner, H. ; Aubke, F., Personal Communications. 36) von Ahsen, B.; Bach, C ; Pernice, H.; Willner, H. ; Aubke, F. J. Fluorine Chem. 2000, 102, 243. 37) Bernhardt, E.; Willner, H. ; Jonas, V.; Thiel, W.; Aubke, F. Angew. Chem. Int. Ed. Engl. 2000, 39, 168. 38) Adelhelm, M . ; Bacher, W.; Hohn, E. G.; Jacob, E. Chem. Ber. 1991, 124, 1559. 39) Aubke, F.; Willner, H. , Personal Communications. 100 Chapter 4 Carbonyl Cationic Complexes of Molybdenum, Tungsten and Chromium 4.1 Introduction The group 6 metals Cr, Mo and W form neutral homoleptic metal carbonyls of the type M(CO)6 (M = Cr, Mo or W). 1 They are strictly octahedral and obey the effective atomic rule (EAN) of 18 valence electrons. More recently a limited number of carbonyl metalates, of the type [M(CO) 4] 4" and [M(CO) 5] 2" (M = Cr, Mo or W), 2 have been added to this list. However, the molecular structures of these anions are not as well known as those of the neutral M(CO)6 (M = Cr, Mo or W) species.1-3"5 At the outset of this study, carbonyl cations of group 6 metals were unknown, but a number of metal carbonyl halides,6 such as M(CO) n F m (M = Mo: n = 2, m = 2; n = 4, m = 2; n = 3, m = 3; n - 3, m = 4; M = W: n = 4, m = 2),7>8 [{M(p-X)X(CO) 4} 2] (M = Mo, W; X = CI, Br, 1)9.10 a n c [ [MXY(CO) 3 (NCMe) 2 ] (M = Mo, W; X = Y = Br, I ; X = CI, Br, Y = I or X = CI, Y = GeCl 3 , SnCb), 6 ' 1 had been reported. In this chapter, the synthesis and characterization of [{Mo(CO) 4 }2(cis-p-F2SbF 4 ) 3 ] x [Sb 2 Fii] x , 1 3 [W(CO) 6(FSbF 5)][Sb2F,i] 1 4 and a transient carbonyl complex of Cr are presented. These compounds were prepared by the oxidation of M(CO)6 (M = Mo, W, Cr) with liquid SbF 5, where SbFs is the oxidizing agent, the solvent, and also the source of the SbF6_ ligands and the [Sb2Fn]" counteranions. [{Mo(CO)4}2(cis-p-F 2 SbF 4 ) 3 ] x [Sb 2 Fii] x was the first compound studied. Its synthetic route evolved from the project of a colleague, Mr. D. Freidank. The project was the attempted synthesis, from Mo(CO)6, of [(CO) 4Mo=Mo(CO) 4][Sb 2Fi i ] 4 , which would have been the first binuclear metal carbonyl cation with a metal-metal quadruple bond. Freidank's proposed, ultimately unsuccessful synthetic 101 approach is shown in Scheme l . 1 5 Scheme 1: Mo(CO) / CH 3 CH 2 OH ^ Mo 2 (CH 3 C0 2 ) 4 * H S ° 3 F CO, SbF5 mixture of many species, attempt abandoned SbF, "shortcut" Mo 2 (S0 3 F) 4 * CO, SbF5, heat - S b 2 F 9 ( S 0 3 F ) mixture of CO-containing species separation by sublimation MoOF 4 * = known compounds Another colleague, Dr. R. Brochler, took up the project and successfully prepared the complex [{Mo(CO) 4} 2(cis-u-F 2SbF 4) 3] x[Sb 2F u] x by a "shortcut" from Mo(CO) 6 with the overall equation 14 SbF, 5(1) 12xMo(CO) 6 + 41xSbF 5 ^ c > 2 4 h o u r s eflMoCCO^JzCcis-u^SbF^JxtSbzFujx + 4xCO + x(6SbF3.5SbF5) (eq. 4.1) The synthesis of [W(CO)6(FSbF 5)][Sb 2Fn]2 from W(CO)6 was then investigated, in the current work, in cooperation with Dr. R. Brochler: 1 4 HF-SbFs, 2 atm CO 6W(CO) 6 + 29SbF5 • 6[W(CO) 6(FSbF 5)][Sb 2F 1 1] + 6SbF3.5SbF5 (eq. 4.2) 40°C, 1 day Single crystals of [{Mo(CO) 4} 2(cis-u-F 2SbF 4) 3] x[Sb 2F n] x and [W(CO) 6(FSbF 5)][Sb 2F 1 1] were isolated and their molecular structures were solved. 1 3 ' 1 4 In addition, vibrational spectra, 102 magnetic susceptibility measurements and thermal decomposition results were obtained and are presented here. It appears that Mo(CO)6 and W(CO)6 react in liquid SbF5 in a very similar manner shown in Scheme 2 below. Scheme 2: Step 1: Oxidation Step 2: Polycondensation +2e\ -SbF3* -CO, -SbF5 M(CO) 6 - [M(CO) 6(FSbF 5)][Sb 2F„] - [{M(CO) 4}2(p-F 2SbF4)3]x[Sb 2F 1 1] x excess SbF5 (M = Mo or W) (isolable for M = W) (isolable for M = Mo) *: SbF3, the reduced by-product, forms adducts such as 5SbF3.5SbF5 when SbF5 is present in excess. [W(CO)6(FSbF5)][Sb2Fn] is isolable but the Mo analogue is not, and is observed as a by-product. Conversely, polymeric [{W(CO)4} 2(p-F 2SbF 4) 3] x[Sb 2Fn] x is found only as a by-product during the thermal decomposition of [W(CO)6(FSbF 5)][Sb 2Fn]. Attempts to prepare a cationic Cr carbonyl complex from Cr(CO)6 have not been successful. A transient Cr carbonyl species has been observed as an off-white solid. Results of vibrational analysis suggest that the Cr carbonyl species is centrosymmetric, and the oxidation state of Cr appears to be close to +2. This carbonyl species was formed along with SbF3. The latter was converted to 6SbF3.5SbF5 in excess SbFs, as for the Mo and W analogs. Due to its thermal instability and similar solubility in HF or HF-SbFs as 6SbF3.5SbFs, a clean separation of the two entities and crystallization of this Cr carbonyl species was not successful. 103 4.2 Poly-Tetrakis(carbonyI)tri(cis-|i-hexafluoroantimonato(V)molybdenum(II) Undecafluoroantimonate(V), [{Mo(CO)4}2(cis-^F2SbF4)3]x[Sb2F11]x 4.2.1 Experimental [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fn]x is first prepared by the oxidation of Mo(CO)6 in liquid SbF5:13,14,16 In a typical experiment, 15 mL of SbFs were condensed onto 264 mg of freshly sublimed Mo(CO)6 (1 mmol) in a 50 or 100 mL one-part glass reactor. The reaction mixture, which turned yellow immediately, was heated at 60 °C for 24 hours. Upon removal of all volatiles in vacuo, a yellow powder was obtained. This was found to be a mixture of [{Mo(CO)4}2(cis-p-F2SbF 4 )3]x[Sb 2 F n ] x and 6SbF3.5SbF5. Single crystals of [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fi i ] x were obtained by dissolving the product mixture in HF-SbF 5 (20 volume % of SbFs) in a PFA reactor. Lowering the temperature from 40 °C at a rate of 10 °C per hour produced yellow-orange crystals of [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fii]x and colourless crystals or powder of 6SbF3.5SbFs. The moisture-sensitive crystals were isolated by removing all volatiles in vacuo. Alternatively, when HF-SbFs was used as the solvent for the reaction instead of SbFs, the reaction proceeded at 40 °C with the formation of crystalline [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fi i ] x and 6SbF3.5SbFs upon lowering the temperature at a rate of 10 °C per hour. 104 4.2.2 Characterizations 1. Molecular Structure Determination Description of the Molecular Structure A repeat unit of the polymeric cation [{Mo(CO)4}2(cis-|4.-F2SbF4)3]+ is shown in Figure 4.1. The crystal data can be found in Appendix 1. Both molybdenum ions are seven-coordinate, which is common for Mo(II) complexes. 1 3 > 1 7 The coordination polyhedron is best described as a square antiprism or 4:3 geometry, where a trigonal MoF 3 and approximately square pyramidal Mo(CO)4 moiety share a common apex.1 3 The two molybdenum centres and the two u-F2SbF2 moieties form an eight-membered ring. The cyclic [{Mo(CO)4}2(cis-u-F2SbF4)3] units are further linked by cis-u.-F2SbF 4 bridges into chains. In addition, there are weak C--F contacts within the cation as well as interionic C--F contacts to the bent [Sb2Fn]" anion, as will be described in the next section. Figure 4.1 The repeat unit of the polymeric cation [{Mo(CO)4}2(cis-ju-F2SbF4)i]+ (ellipsoids are drawn at 50% probability). The shaded area is the eight-membered ring formed by the two Mo centres and the two ju-F2SbF4 moieties.13 105 Discussion of the Molecular Structure [{Mo(CO)4}2(cis-u.-F2SbF4)3] x[Sb2Fii] x is the first and so far only polymeric derivative of a metal carbonyl cation where eight-membered metallocycles are linked by p-F2SbF4 into chain like cations.13 The C-0 bond lengths fall into the range of 1.089(11) - 1.136(11) A. 1 3 The Mo-C distances, 2.140(5) - 2.196(5) A, are correspondingly long. 1 3 While the C-Mo-C angles alternate between acute (70.2(4)° and 75.3(4)°) and relatively wide (111.6(4)° and 113.9(4)°), the Mo-C-0 angles depart from linearity by 1.3 - 5.2° as found in [Hg(CO ) 2 ] 2 + . 1 3 ' 1 8 This is probably due to the observed interionic and intramolecular contacts that involve both O and C atoms of the carbonyl groups.1 3 The secondary contacts are about 0.20 - 0.40 A shorter than the sum of the van der Waals radii 1 9 between the F atoms of the [Sb2Fn]" or the cis-SbFe' ligands and O or C atoms of the carbonyl groups.1 3 The stereoview of the unit cell of [{Mo(CO)4}2(cis-p-F2SbF 4 ) 3 ] x [Sb2Fn] x is shown in Figure 4.2. Figure 4.2 The stereoview of the unit cell of [{Mo(CO)4}2(cis-p,-F2ShF4)3]x[Sb2Fn]x. 106 2. Vibrational Analysis There are four C-0 stretching bands in each of the IR and Raman spectra of [{Mo(CO)4}2(cis-u-F2SbF4)3] x[Sb2Fn] x. This indicates that none of the C-0 groups is equivalent. The coincidence of IR and Raman bands also shows that the cation lacks a centre of symmetry. The observed angle alternation, along with the coincidence of vco in the IR and Raman spectra suggest that the local symmetry of the pyramidal Mo(CO)4 moiety is approximately C i . 1 4 The vibrational data are shown in Table 4.1. Table 4.1 Vibrational data* of [{Mo(CO)4}2(cis-ju-F2SbF4)3] x[Sb2Fn] x in the C-O stretching region.13 IR bands (cm"1) Raman bands (cm1) 2156 (ms) 2156 (vs) 2105 (m,sh) 2105 (vs) 2092 (s,sh) 2088 (m,sh) 2086 (vs) 2085 (m) *: Intensities are given in parentheses: v= very, s = strong, m = medium and sh = shoulder. The average vco of 2110 cm"1 and the calculated C-0 stretching force constant,/^,, of 18.0 x 102 Nm"1 are slightly lower that those of free CO (2143 cm"1 and 18.6 x 102 Nm" 1 ) . 1 3 They are lower than those observed for most of the predominantly a-bonded homoleptic carbonyl cations.2 0 There are two reasons for these low values.1 3 First, the net charge on each Mo is only +1/2 compared to, e.g., +2 for [Hg(CO)2J where the vco is 2279.5 cm" . In comparison, v<x> for neutral Mo(CO)6, ranges from 1990 - 2112 cm" 1 , 2 1 and the fco is 17.1 x 102 Nm"1 2 2 . Secondly, some 7t-backbonding, and secondary interionic and intramolecular contacts 107 combine to weaken the C-0 bond. Despite these factors, the C-0 stretching wavenumbers of [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fn]x are still among the highest for Mo(II) carbonyl species.13 3. Microanalysis The microanalysis for carbon in [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fu]x was unsatisfactory. The observed value of 3.88% is lower than the calculated value of 6.09%. 1 3 This is probably because the high fluorine content of the sample causes C F 4 to be produced during combustion14 even when CuO is added as a combustion aid. In addition, the separation of the product from the by-product 6SbF3.5SbF5 may not have been complete in the bulk sample used for analysis. Therefore, the microanalysis data for [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fn]x cannot be considered reliable. 4. Magnetic Susceptibility Measurements For magnetic susceptibility measurements, single crystals of [{Mo(CO)4} 2(cis-u-F 2 SbF 4)3] x [Sb 2 Fn] x , physically separated from 6SbF3.5SbF5, were used. Figure 4.3 shows plots of the magnetic susceptibilities and magnetic moments of this molybdenum compound. Possible ground states for this d 4 system are S = 0, 1 or 2. The moment approaches zero at 0 K, implying a spin-paired S = 0 ground state for the compound. The increase in moment with increasing temperature suggested a thermally populated S = 1 and / or S = 2 excited state; however, attempts to model the magnetic data accordingly failed. A model which considers Temperature Independent Paramagnetism (TIP) as the major contribution of the paramagnetism was used successfully in fitting the data. 2 3 ' 2 4 The model used assumes that a fraction, P, of the sample is present as S = 2 paramagnetic impurity with (1-P) constituting the bulk sample, the 108 latter having an S = 0 ground state with a TIP paramagnetic component. The calculated susceptibility utilizing this model is given by the equation % M = (P x 0.75 g / T) + [(1-P) x TIP]. 0.0016 0.0014 4 0.0012 T_ 0.0010 o E co E gf 0.0008 Jo. —» a. o CO 13 CO o 0.0006 i co c 0.0004 0.0002 0.0000 Experimental Magnetic Susceptibility (Solid) Fitted Magnetic Susceptibility Experimental Magnetic Moment (with 1 % error bars) (Dashed) Fitted Magnetic Moment (fixed: g = 2.0; fitted: TIP = 0.00030(0) c m 3 mol" 1, P = 0.080(0)%, F = 0.078) P o r w ^ >oo 0.9 0.8 h0.7 0.6 1 o.5 r" c CD E o 0-4 c CD CO 2 0.3 0.2 0.1 0.0 50 100 150 200 Temperature (K) 250 300 350 Figure 4.3 Magnetic Behaviour of [{Mo(CO)4}2(cis-VL-F2SbF4)i]x[Sb2Fn]xfrom 2 - 300 109 Fitted values of TIP and P are 0.00030(0) cm 3 / mol and 0.080(0)% respectively. The value of g is fixed at 2.0. The agreement between calculated and measured moments and susceptibilities is shown visually in Figure 4.3, and is measured by the goodness of fit parameter, F. F = V[(l/N)xE{[(xcaic-Xexpt)/Xexpt]2}], where N is the number of data, Xcaic is the calculated value and Xexpt is the experimental value of the magnetic susceptibility. The F-value is 0.78. In summary, the analysis indicates that [{Mo(CO)4}2(cis-p-F2SbF 4)3] x[Sb2Fii] x is spin-paired with an S = 0 ground state and a paramagnetic component arises from Temperature Independent Paramagnetism. The sample studied also had a small amount (<0.1%) of a, presumably, S = 2 paramagnetic impurity in it. This S = 2 impurity is possibly an unidentified Mo(II) species. 5. Thermal Decomposition Studies For the thermal decomposition studies, crystals of [{Mo(CO)4}2(cis-p-F2SbF4)3] x[Sb 2Fn] x were heated in a one-part glass reactor. The crystals were stable up to 140 °C. At approximately 150 °C, melting and decomposition took place. The gas-phase IR spectra of volatiles given off at 150 °C indicated the presence of CO, and B F 3 as well as SIF4 from the reaction of the glass.1 3 Liquid SbFs formed and some colourless crystals sublimed. These crystals were tentatively identified by vibrational spectroscopy as M 0 O F 4 . 1 3 The residue was olive in colour. 4.2.3 Discussion of [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fn]x /. Structure The predominantly a-bonded metal carbonyl derivative, polymeric [{Mo(CO)4h(cis-p-F2SbF4)3] x[Sb2Fii] x, has no precedent among known CO-containing'coordination compounds of molybdenum 6 ' 1 1 ' 1 2 ' 1 7 on four accounts.13 Firstly, before [{Mo(CO)4}2(cis-u-110 F2SbF4)3]x[Sb2Fn]x was obtained, all known homoleptic metal carbonyl cations and their derivatives20 were mononuclear or binuclear, e.g. [Pd2(p-CO)2]2+.20 Secondly, previously reported molybdenum(II) carbonyl halide derivatives were either monomers or ligand-bridged di-, tri-, or tetra-mers 6 ' 1 1- 1 2 ' 1 7 Thirdly, [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fn]x is the first cationic seven-coordinate group 6 metal carbonyl derivative prepared. All previously studied predominantly a-bonded cationic metal carbonyl derivatives are linear (d 1 0), square planar (d8) or octahedral (d 6 ) . 2 0 In the case of molybdenum, it is d 4 with the first seven-coordinate 4:3 geometry encountered for this type of compound. Finally, this is the first time that predominantly a-bonded metal carbonyl complexes have had two extremely weakly nucleophilic fluoroantimonate(V) anions, SbF6_ as cis-p-F 2SbF 4 ligand and [Sb2Fn]" as counteranion simultaneously present. The bidentate cis-F-bridging SbF6_ ligand is responsible for two unusual structural features in this compound: the eight-member trans-Mo2F2Sb2F2 rings with fluoro bridges between alternating Mo and Sb atoms, and the further linking of the metallocycles into chains. The significance of secondary interionic contacts for [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fii]x has been discussed in Section 1 of 4.1.1. The observed formation of a metallocycle and polymers via F2-SbF4 bridges, as well as the interionic and intramolecular C--F contacts, reflect the high electrophilicity and Lewis acidity of the pyramidal Mo(CO)4 moiety. 2. Synthesis In an attempt to prevent the formation of 6SbF3.5SbF5 in the synthesis of [{Mo(CO)4h(cis-p-F2SbF4)3]x[Sb2Fn]x, CI2 was used as an oxidant instead of SbFs. In this case, a yellow solid was obtained. The bands in the C-0 stretching region of the IR spectrum were identical to those of 111 [{Mo(CO)4}2(cis-|o,-F2SbF4)3]x[Sb2Fn]x synthesized using SbFs as the oxidant. Therefore, the two products should have the same Mo(II) carbonyl cation. However, the region below 750 cm"1 was different for the two products. The silver nitrate test indicated that chloride was present in the product obtained via CI2 oxidation. Chlorides may have displaced some of the fluorides in the [Sb2Fn]" counteranion. There were small differences in the vibrational spectra of products obtained in repeated experiments, suggesting that the number of fluorine atoms displaced was irreproducible. Therefore, oxidation by chlorine is not considered a good route to bulk, homogeneous samples of [{Mo(CO)4}2(cis-p-F2SbF4)3]+. In some cases, small amounts of the monomer [Mo(CO)6(FSbFs)][Sb2Fn] were obtained as a by-product in the synthesis of [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fii]x,14 as indicated by the vibrational spectra. The positions of vibrational bands of the monomer [Mo(CO)6(FSbFs)][Sb2Fi 1] in the CO stretching region are very similar to those of [W(CO)6(FSbF5)][Sb2Fi 1], as are discussed in Section 4.3. Unfortunately, it has not been possible to obtain the monomer, [Mo(CO)6(FSbF5)][Sb2Fii], in pure fo rm. 1 3 ' 1 4 On the other hand, the progress of the reaction and the composition of the Mo product can be monitored by CO evolution, since the reaction proceeds without an external CO pressure (see Scheme 2). 4.3 Hexakis(carbonyl)hexafluoroantimonato(V)tungsten(II) Undecafluoroantimonate(V), [W(CO)6(FSbF5)] [Sb 2F„] 4.3.1 Experimental Synthesis and Comments In a typical experiment,14 10 mL of SbFs were condensed onto 179.5 mg of freshly sublimed W(CO)6 (0.51 mmol) in a 50 or 100 mL one-part glass reactor. The reaction mixture, which 112 turned bright yellow immediately, was heated at 40 °C for a week. Upon removal of all volatiles in vacuo, a mixture of bright yellow [W(CO)6(FSbF 5 )][Sb2Fn] and 6SbF 3 .5SbF5 powder (11A A mg) was obtained. The expected weight was 730.4 mg. The difference between the actual and calculated weight was attributed to residual SbFs in the product mixture. 1 4 No release of CO was observed, in the gas phase IR spectrum, during the reaction. In order to prevent any loss of CO from W(CO)6, the reaction may be performed in a CO atmosphere. Single crystals of [W(CO)6(FSbF5)][Sb2Fn] were obtained by adding 5 mL of SbFs, followed by 10 mL of HF onto 352 mg of freshly sublimed W(CO) 6 (1 mmol) in a 100 mL PFA reactor.14 The CO pressure was adjusted to 2 atm by addition of CO gas. On warming to 40 °C, the reaction mixture turned bright yellow immediately. The temperature was maintained for 24 hours and then gradually lowered at a rate of about 1 °C per hour. During the cooling, yellow crystals were produced on the reactor wall. When the crystals grew to about 0.3 - 5.0 mm in length, all volatiles were removed in vacuo. Extremely moisture-sensitive crystals of [W(CO)6(FSbF5)][Sb2Fn] (320 mg) were obtained along with crystalline and powdered 5SbF3 .5SbF5. The isolated yield was approximately 30%. It should be noted that [W(CO)6(FSbF5)][Sb2Fn] is unique among metal carbonyl cations because there is no gain or loss of CO during the synthesis. In the presence of excess CO, the intermediate cation [W(CO)6] is still stabilized by the weakly nucleophilic SbF6~, which functions as a monodentate FSbF 5 ligand, rather than by CO. Furthermore, all attempts to substitute the FSbFs ligand by CO have so far failed, and the homoleptic cation [W(CO)7J remains unknown. 113 4.3.2 Characterization 1. Molecular Structure Determination A formula unit of [W(CO)6(FSbF5)][Sb2Fn] is shown in Figure 4.4. 1 4 The crystal data can be found in Appendix 1. The simultaneous presence of SbFe" as ligands and [SbaFn]" as counteranion in the same compound is uncommon. The only precedent is [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fn]x 1 3 discussed in Section 4.2, where SbFg" functions as a cis-F-bridging bidentate ligand. Figure 4.4 A formula unit of [W(CO) 6(FSbFs)J[Sb?F} j] (ellipsoids are drawn at 50% probability).14 The bridging groups of the cation and anion are highlighted. 114 The seven-coordinate cation, [W(CO)6(FSbFs)]+, is best described as a distorted, C 2 V capped trigonal prism formed by six CO ligands, with a quadrilateral face capped by bridging fluorine of the monodentate FSbF5" ligand (see Figure 4.5). 1 4 Structural analogues are found in isocyanide complexes of Mo(II) and W(II) of the type [M(CNR) 7 ] 2 + or [M(CNR) 6 X] + (M = Mo or W, R = tert-butyl and X = CI, Br, or I) . 2 5 Figure 4. 5 An idealized structure of [W(CO)6(FSbF5)] . The seven-coordinate cation is described as a distorted, C2v capped trigonal prism formed by six CO ligands, with a quadrilateral face capped by bridging fluorine of the monodentate FSbF5~ ligand. 115 In [W(CO)6(FSbF5)][Sb2Fii], there are two short W-C distances of 1.996(9) and 2.021(7) A to C(l)-0(1) and C(2)-0(2), which are located on the unique edge of the prism (see Figure 4.4). The two corresponding C - 0 distances are quite long (1.160(10) and 1.155(8) A). The remaining four W-C distances to CO ligands in the quadrilateral plane are longer. They range from 2.130(10) to 2.150(9) A. Conversely these four C - 0 distances are short and vary from 1.109(10) to 1.128(10) A. Therefore, there seems to be reduced TC-backbonding within the distorted quadrilateral plane.1 4 There are also significant bond angle distortions in the [W(CO)e(FSbF5)]+ cation. 1 4 The C-W-C angles for the CO groups cis to each other fall into three groups of about 75°, 102° and 115°. Also, the W-C -0 angles depart from linearity by up to 5°. The strictly octahedral molecule W(CO) 6 studied by electron diffraction26, has W-C and C - 0 distances of 2.058(3) and 1.148(30) A respectively.26 This provides a strong contrast to the highly distorted W(CO) 6 moiety of the [W(CO) 6(FSbF 5)] + cation. There is a striking similarity between the Sb-F-Sb bridge in the anion and the W-F-Sb bridge in the cation of [W(CO)6(FSbF 5)][Sb 2Fn]. 1 4 Both of the bridging groups are highlighted in Figure 4.4. The bridge angle of the cation, 149.7(3)°, is identical to that of the anion, 149.4(3)°, within experimental error. If allowance is made for the slightly larger effective radius of W(II), the distances within the two bridging units are nearly identical; the Sb-Fbridging distances of 2.034(5) and 2.022(5) A are only marginally longer than that in the cation at 2.040(4) A . 1 4 The implication is that the transient cation [W(CO)e]2+ is a powerful Lewis acid, comparable in strength to the Lewis superacid, SbFs. 116 The W-Fbridging bond length of 2.109(5) A is unusually short and comparable to the W-Fterminai distance of 2.081(1) A reported for W m (PMe 3 ) 4 H 2 (OH 2 )F . 1 4 > 2 7 The remaining Sb-Fte™™! distances in the cation [W(CO) 6(FSbF 5)] + are between 1.801(6) - 1.868(5) A. For the anion [Sb2Fn]", a similar range of 1.811(6) - 1.864(6) A is found. In both cases, the range of bond lengths is small. Furthermore, lengthening of the terminal Sb-F bonds seems to involve F atoms engaging in intraionic (F(2)) or interionic (F(12), F(14) and F(16)) interactions with C atoms of the CO ligands. 2. Vibrational Analysis All C-0 stretching frequencies of [W(CO)6(FSbF5)][Sb2Fii] are IR and Raman active. All degeneracies are also removed, which indicates that the local symmetry of the W(CO)6 group is C i 1 4 The six fundamentals observed for the six inequivalent CO's in [W(CO) 6(FSbF 5)][Sb 2Fii], along with two bands attributed to [{W(CO)4}2(cis-u-F 2 SbF 4 )3] x [Sb 2 Fii] x are shown in Table 4.2. (The formation of [{W(CO)4}2(cis-u.-F2SbF 4 )3 ] x [Sb 2 Fi i ] x is discussed in Section 5 of 4.1.1). The observed average C-0 stretching frequency of about 2125 cm"1 for [W(CO)6(FSbF5)][Sb2Fn] is 115 cm"1 higher than the value of 2010 cm"1 for W(CO) 6 , 2 1 but is slightly lower than the value 2143 cm"1 for free C O . 2 8 This indicates that [W(CO)6(FSbF5)][Sb2Fn] has significantly reduced 7t-backbonding14 compared to the starting material, W(CO)6. It should be noted that even though the oxidation state of W is +2, the formal charge in the [W(CO)6(FSbF5)] + cation is reduced to +1 by the anionic ligand FSbF5". In the region below 750 cm"1, the vibrational spectrum of [W(CO)6(FSbF5)][Sb2Fn] is uninterpretable.14 There is a near coincidence of bands due to the [Sb2Fn]" anion and the 117 FSbFs" ligand. The W-C stretching and W-C-0 deformation bands occur in this region as well. The low symmetry of the W(CO)6 moiety is expected to result in extensive band proliferation.1 Table 4.2 C-0 stretching wavemimbers * of [W(CO)6(FSbF5)][Sb2Fj J.14 Complex IR bands (cm 1) Raman bands (cm"1) [W(CO) 6(FSbF 5)][Sb 2F„] 2205 (vw) 2206 (s) [W(CO) 6(FSbF 5)][Sb 2F„] 2170 (w,sh) 2170 (s) [{W(CO) 4} 2(cis-p-F 2SbF 4)3]x[Sb 2F 1 1]x 2157 (w,sh) 2157 (vw) [W(CO)6(FSbF5)][Sb2F„] 2147 (s) 2146 (s) [{W(CO) 4} 2(cis-p-F 2SbF 4) 3]x[Sb 2F„] x 2102 (vw,sh) 2098 (vw) [W(CO) 6(FSbF 5)][Sb 2F„] 2088 (s) 2084 (m,sh) [W(CO) 6(FSbF 5)][Sb 2F u] 2075 (s,sh) 2078 (s) [W(CO) 6(FSbF 5)][Sb 2F n] 2064 (s) 2060 (s) *: Intensities are given in parentheses: s = strong, m = medium, w = weak, v = very and sh = shoulder. In summary, in contrast to other known metal carbonyl cations,20 the molecular structure of [W(CO)6(FSbF5)]+ cannot be deduced from vibrational spectra on account of the low symmetry (Ci). 3. Magnetic Susceptibility Measurements The powdered samples of [W(CO)6(FSbF5)][Sb2Fn] used for magnetic susceptibility measurements contain 6SbF3.5SbF5. Thus, diamagnetic corrections have been made for the W complex as well as for 6SbF3.5SbFs according to the ratio of one W to one 6SbF3.5SbF5 118 obtained from the balanced chemical equation. Figure 4.6 shows plots of the magnetic susceptibilities and magnetic moments of [W(CO)6(FSbF5)][Sb2Fn] vs. temperature. At 300 K, the magnetic moment is 1.5 pe-0.10 0.08 ^ o £ 0.06 E o Q_ CD « 0.04 CO o "-»—' 0) c D5 03 ° 0.02 0.00 o • B • o 0 Do o o 8 o o • • o • • • o ° ° ° o o o o o o o o o o o o o o o o o o o o o o Experimental Magnetic Susceptibility • Experimental Magnetic Moment 0 50 100 150 200 250 Temperature (K) 300 350 Figure 4.6 Magnetic Behaviour of [W(CO)6(FSbF5)][Sb2Fu] from 2 - 300 K. 119 The W(II) centre in [W(CO)6(FSbF5)][Sb2Fu] is d 4. As with Mo(II) in [{Mo(CO) 4} 2(cis-p-F2SbF4)3]x[Sb2Fii]x, possible ground states for this system are S = 0, 1 or 2. This experimental magnetic moment at 300 K is slightly lower than the theoretical spin-only moment of 2.8 pe for the S = 1 state (see Figure 4.6). It should be noted that there is less certainty in the composition of the W sample than the Mo sample, due to the presence of 6SbF3.5SbF5 and possibly SbF3 or its other adducts with SbFs. There is also an unknown, although small, amount of the condensation product, [{W(CO)4} 2(cis-p-F 2SbF 4) 3] x[Sb 2Fii] x. Therefore, results of magnetic studies on [W(CO)6(FSbF5)][Sb2Fn] are less reliable than those of [{Mo(CO) 4} 2(cis-p-F 2SbF 4) 3] x[Sb 2Fn] x . For this reason, no attempt to further analyze the magnetic data for this compound was made. 4. Microanalysis As with the [{Mo(CO) 4} 2(cis-p-F 2SbF 4) 3] x[Sb 2Fn] x sample, microanalysis for carbon in [W(CO)6(FSbF5)][Sb2Fn] was unsatisfactory. The experimental values of 5.5 - 6.0% were lower than the theoretical value of 6.92%.1 4 This is probably due to the high fluorine content of the sample, causing production of CF 4 as a by-product during combustion,14 and the presence of by-product 6SbF3.5SbFs as in the case of [{Mo(CO) 4} 2(cis-p-F 2SbF 4) 3] x[Sb 2Fn] x. 5. Thermal Decomposition Studies The controlled thermal decomposition of [W(CO)6(FSbF5)][Sb2Fn] was studied by heating the yellow crystals in a pre-weighed one-part glass reactor.14 The crystals are stable until about 100 °C. Upon further heating, partial elimination of CO was detected in the gas phase and formation of SbFs as a colourless liquid was noted. Changes observed in the vibrational spectra suggest a condensation reaction according to: 1 4 120 2x[W(CO)6(FSbF5)][Sb2Fn] • temperature > 100°C [{W(CO)4}2(cis-p-F2SbF4)3]x[Sb2F1i]x + 4xC0 + xSbF 5 (eq. 4.3) During this condensation, monodentate FSbFs" is converted to bidentate u-F2SbF4, and one of the [Sb2Fn]" anions decomposes to SbFs and p-F2SbF4. The decomposition was followed by the weighing of the solid residue and monitoring the amount of CO evolution into a vacuum line of known volume. The decomposition product has a very similar band distribution to [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fi,]x in the C-0 stretching range of the IR spectrum. 1 3- 1 4 Unfortunately, it has not been possible to isolate this thermal decomposition product in pure form, but all indications are that the structure is identical to that of the molybdenum species discussed above. 4.4 The Transient Carbonyl Species of Chromium 4.4.1 Experimental In a typical experiment, 10 mL of SbFs was condensed onto 182.6 mg of freshly sublimed Cr(CO)6 (0.8298 mmol). CO gas was then added to a pressure of 2 atm. The white reaction mixture was stirred at room temperature for 24 hours. Upon removal of all volatiles in vacuo, 1082 mg of white powder was obtained. This white powder contained a carbonyl species as indicated by the CO stretching bands in the vibrational spectra (see Section 4.4.2). Attempts to produce single crystals of the carbonyl species in FfF-SbFs have been unsuccessful. Colourless crystals were only observed in solution and could not be isolated. When the reaction mixture of Cr(CO)6, SbF5 and CO was held slightly above room temperature (at 30 °C), Cr[SbF6]2 was obtained as the main product. 121 4.4.2 Characterizations The microanalysis for carbon in the product formed from the reaction of Cr(CO)6 with liquid SbF 5 under CO atmosphere was 2.93%. The vibrational data of the product are shown in Table 4.3. There is a single C-0 stretching band at 2088 cm"1 in the IR spectrum and two at 2169 and 2124 cm"1 in the Raman spectrum. The number of bands and the non-coincidence of the IR and Raman spectra suggest that a carbonyl species with either D4h or Oh symmetry is present in the product. Table 4.3 Vibrational data* of the product from the reaction of Cr(CO)e with liquid SbFs under a CO atmosphere. IR bands (cm 1) Raman bands (cm"1) 487 (w) 550 (m) 561 (w) 590 (w,sh) 622 (s,sh) 636 (m) 649 (s) 667 (ms) 665 (vs) 692 (m) 686 (m) 706 (vs) 697 (w) 719 (s) 727 (s) 2088 (s) 2124 (s) 2169 (vs) *: Intensities are given in parentheses: v = very, s = strong, m = medium, w = weak and sh = shoulder. This product, formed from the reaction of Cr(CO)6 with liquid SbF 5 under CO atmosphere, was unstable at room temperature under static vacuum or dry nitrogen atmosphere, and eventually decomposed into Cr[SbFe]2 and SbFs with evolution of CO gas. 2 9 122 In summary, the results for the reaction of Cr(CO)6 in SbFs are preliminary, and no meaningful conclusions can be drawn. Although results of vibrational analysis are consistent with production of a thermally unstable carbonyl species of Cr, the actual composition of the product cannot be determined. As for [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fii]x and [W(CO)6(FSbFs)][Sb2Fn], molecular structure determination is required to identify this carbonyl species. 4.4.3 Isolation and Molecular Structure Redetermination of SbF3 1. Isolation ofSbF3 Colourless single crystals of SbF3 were isolated in the process of washing the product described above, obtained from the reaction of Cr(CO)6, SbF5 and CO, with HF. The single crystal was identified by x-ray diffraction as SbF3 in terms of its cell parameters. The structure of SbF3 had been reported by A. J. Edwards 3 0 in 1970. The redetermination here reveals little discrepancy with Edwards' report, and is an improvement only in terms of the smaller uncertainty range obtained. The isolation of SbF3 is evidence that an oxidation of Cr(CO)6 has taken place while Sb(V) was reduced to Sb(III). The initial Sb(III) reduced product was SbF3, which formed the adduct 6SbF3.5SbF5 in excess SbF5. In 6SbF3.5SbFs, the SbF3 moiety is not strongly bound to the SbF5 moiety. The fact that SbF3 can be isolated by washing with HF, is analogous to the conversion of [M(CO)6][Sb 2Fii] 2 (M = Fe, Ru or O s ) 3 1 ' 3 2 into the [SbF6]' salt by washing with HF, where the secondary contacts are weak in these complexes. HF 6SbF3.5SbF5 + 5HF • 6SbF3 + 5HF-SbF5 (eq. 4.4) HF rM(CO) 6][Sb 2F,,] 2 + 2HF ^ [M(CO) 6][SbF 6] 2 + 2HF-SbF 5 (eq. 4.5) 123 Both of the reactions involve an extrusion of SbFs, either from an adduct 6SbF3.SbF5 or an anion [Sb2Fi i]". The formation of the superacid HF-SbFs appears to be the driving force. 2. Molecular Structure of SbF$ Figure 4.7 shows the molecular structure of SbF3 with selected bond angles and distances as well as secondary contacts determined in the present work. 1(1) Figure 4. 7 The molecular structure of SbF$ (ellipsoids are drawn at 50% probability). Selected bond angles, distances and secondary contacts are shown. 124 All three Sb-F distances are about 1.93 A (see Figure 4.7), thus SbF3 has the expected C3V symmetry within the uncertainty range. There are three neighbouring fluorine atoms 2.58 A from Sb to complete the distorted octahedral arrangement. Table 4.4 shows a comparison between the bond parameters determined by Edwards 3 0 and in the present work. As mentioned above, there are few discrepancies between the two sets of data. The uncertainty range is smaller in this work. A detailed list of parameters from both determinations can be found in Appendix 1. Table 4.4 Comparison of bond angles and distances of SbFs from determination by Edwards30 and the present work. Edwards (1970) This work (2001) Bond distances (A) Sb-F(l) 1.90(2) 1.934(3) Sb-F(2) 1.94(2) 1.939(4) Bond angles (°) F(l)-Sb-F(2) 88.9(1.5) 87.21(9) F(l)-Sb-F(2*) 72.9(1.5) 84.2(2) 4.5 Discussion of the Reactions of M(CO)6 (M = Mo, W or Cr) with SbF5 4.5.1 Comparison between [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fii]x and [W(CO) 6(FSbF 5)][Sb 2F„] /. Synthesis and Thermal Stability It was found that W(CO)6 and Mo(CO)6 react with SbFs in a similar manner: initial oxidation to the salt [M(CO)6(FSbF 5)][Sb2Fn] (M = Mo or W), followed by a condensation reaction with partial elimination of CO and SbFs to form cis-p-F2SbF4 bridges and produce polymeric [{M(CO)4}2(cis-u-F2SbF4)3]x[Sb2Fii]x (M = Mo or W). It appears that the two reactions in SbF5 125 differ only in the thermal stabilities of the products.14 In one instance, the product isolated is polymeric [{Mo(CO)4}2(cis-u-F2SbF4)3]x[Sb2Fii]x, while in the other case monomeric [W(CO)6(FSbF5)][Sb2Fn] was isolated. 1 3 ' 1 4 The polymerization and elimination of CO appears to be more facile for Mo, while the W(CO)6 moiety in [W(CO)6(FSbF5)][Sb2Fn] has a greater thermal stability because heating to 100 °C is required to effect condensation.14 To prevent partial decomposition of [W(CO)6(FSbF5)][Sb2Fn], the reaction of W(CO)6 with SbFs is best performed in a CO atmosphere.14 On the other hand, monitoring the CO evolution is one way to monitor the reaction compositions. For both Mo and W, the reaction in HF-SbFs can afford crystalline products, and in some cases single crystals suitable for X-ray diffraction structure determination were produced. The isolated reduced by-product in both cases is 6SbF3.5SbF5. This is evidence that SbF5 is the oxidant,1 4 the source of SbF6_ ligand and [Sb2Fn]" counteranion, and the reaction medium. 2. Structural Features and Internal Bond Parameters Although both [{Mo(CO)4}2(cis-n-F2SbF4)3]x[Sb2Fn]x and [W(CO)6(FSbF 5)][Sb 2Fii] were the main products from the reaction of M(CO)6 (M = Mo or W) with SbF5, their coordination geometries differ. As has been discussed before, the W(II) complex is best described as a Ozv capped trigonal prism, 1 4 while the Mo(CO)4F3" segment of the Mo(II) complex may be described as a square antiprism or 4:3 geometry.13 There are precedents for both geometries among structurally characterized halocarbonyl complexes of Mo(II) and W(II) . 6 ' 1 1 Nevertheless, the 4:3 geometry is less common for Mo(II) coordination compounds.12 126 Some selected bond distances and angles are listed in Table 4.5. The wide spread in M - C and, to a lesser extent C-0 bond distances, and the wide variations in C-M-C angles show that the M(CO) n moieties (M = Mo or W; n = 4 or 6) are highly distorted. Distortions are expected for the C 2 v capped trigonal prismatic structure and have been observed previously.12 However, in this case the local symmetry is reduced to Ci in both metal carbonyl moieties. This is in agreement with the observation of four and six C-0 stretching vibrations that are IR and Raman active. 1 4 The observed spread in M-C distances and CO stretching vibrations is more pronounced for the W(CO)6 group than for the Mo(CO)4 moiety in the complexes. Table 4.5 Selected bond distances and angles for [{Mo(CO)4}2(cis-fx-F2SbF4)3]x[Sb2F\ i]x and [W(CO)6(FSbF5)][Sb2F,,].14 [{Mo(CO)4}2(cis-p-F 2SbF 4)3]x[Sb 2F„] x [W(CO)f , (FSbF 5 ) ] [Sb 2 F 1 1 ] (a) Bond Distances (A) W - C : 1.996(9)-2.150(9) M o - C : 2.021(1)-2.052(10) W-Caven.se: 2.095(6) M0-Caverage: 2.036(10) C - O : 1.109(10) - 1.160(1) C - O : 1.089(11) - 1.136(11) C-Oaven.se: 1.130(10) C-O average: 1.113(10) W - p - F : 2.109(5) Mo-p-F: 2.140(5) - 2.196(5) Sb-p-F( -W): 2.010(4) Sb-p-F(-Mo): 1.933(6)- 1.988(5) (anion) Sb-p-F-Sb: 2.034(5) and 2.022(5) Sb-p-F-Sb: 2.051(6) and 2.014(6) (b) Bond Angles (degrees/ C - W - C : 69.3(4) to 77.2(3), 101.5(3), 117.1(3) C - M o - C : 70.2(4) and 75.3(4), 111.6(4) and 113.9(4) S b - F - W : 149.7(3) Sb-F-Mo: 142.1(3)- 162.8(13) W - C - O : 174.8(8)- 179.3(8) M o - C - O : 174.8(9) - 178.7(9) (anion) Sb-F-Sb: 149.4(3) Sb-F-Sb: 150.0(3) : for adjacent CO groups. 127 For both of the Mo and the W carbonyl cations, the average M - C (M = Mo or W) distances are long while the average C-0 distances are short compared to that of a free CO molecule. This is consistent with the average C-0 stretches of 2125 and 2110 cm'1 for the W and Mo cations respectively, which are both slightly below that of gaseous CO (2143 cm" 1). 2 8 There appears to be a modest reduction in 7t-backdonation in both cations when compared to the synergistic bonding situation found for M(CO)6, (M = Mo or W ) . 2 1 ' 2 6 The pyramidal Mo-(u.-F)3 group is significantly distorted from C 3 v local symmetry. 1 3 ' 1 6 The Mo-u.-F distances range from 2.140(5) to 2.196(5) A while the F-Mo-F bond angles are rather acute and vary from 73.2(2)° to 77.6(2)°. 1 3 The monodentate FSbFs" ligand coordinates more strongly to W in [W(CO)6(FSbF5)][Sb2Fn] than does the bidentate F 2SbF 4" ligand to Mo in [{Mo(CO) 4} 2(cis-u-F 2SbF 4) 3][Sb 2F 1i]. 1 3 This is reflected in the observed Sb-u-F(-W) distance of 2.010(4) A in [W(CO)6(FSbF5)][Sb2Fn] whereas the Sb-u-F(-Mo) distances are between 1.933(6) and 1.969(5) A. 4.5.2 Comparison of the Reaction of Cr(CO)6 in Liquid SbF5 to those of M(CO) 6 (M = Mo and W) 1. Synthesis and Thermal Stability All three of the carbonyl complexes, [{Mo(CO)4}2(cis-u.-F2SbF4)3]x[Sb2Fii]x, [W(CO)6(FSbF5)][Sb2Fn] and a transient Cr carbonyl species are formed by oxidation of the corresponding M(CO)6 (M = Mo, W or Cr) in liquid SbFs. The Mo and W complexes were obtained at slightly elevated temperature (60 and 40 °C respectively). As the temperature increased, the W compound underwent condensation to form [{W(CO)4}2(cis-u-F 2SbF4) 3] x[Sb 2Fu]x 1 4 On the other hand, the Cr species can only be synthesized at room 128 temperature. Slightly higher temperatures, such as 30 °C, yield Cr[SbFe]2 instead. Therefore, the thermal stability of these carbonyl species increases from Cr to W down the triad. As a result, the reactions of W(CO)6 and Mo(CO)6 yield isolable products but the reaction of Cr(CO) 6 does not. When HF-SbFs was used as the solvent, single crystals of [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb 2Fii] x and [W(CO)6(FSbF5)][Sb2Fn] and the by-product 6SbF3.5SbF5 were isolated. 1 3 ' 1 4 When HF was used in the reaction of the Cr species, the CO ligands in the complex were gradually lost and powdered Cr[SbFe]2 and 6SbF3.5SbFs were obtained as a mixture upon removal of volatiles. 2. Vibrational Spectra The cations in [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2F,,]x and [W(CO) 6(FSbF 5)][Sb 2Fii] lack a centre of symmetry. There are 4 and 6 coincident C-0 stretching bands in the IR and Raman spectra, respectively. For the Cr species, the one IR band and two Raman bands are mutually exclusive. These bands suggest a centrosymmetric geometry of D4h or Oh symmetry, which is extremely unusual and puzzling. Although the structure of the Cr complex is unknown, it can be speculated from the vibrational data that there are four or six CO ligands. 4.5.3 Comparison of [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fii]x and [W(CO)6(FSbF5)][Sb2Fij] with Other Homoleptic Carbonyl Cationic Complexes 1. Synthesis Unlike the situation involved in the formation of other of homoleptic metal carbonyl cations and their derivatives in liquid SbFs, 2 0 a solid product mixture, instead of a single solid product, 129 results for Mo and W. 1 3 > 1 4 As discussed above, the other products also form as carbonyl fluororantimonato cations. In the syntheses of [{Mo(CO)4}2(cis-u-F2SbF4)3]x[Sb2Fii]x and [W(CO)6(FSbF5)][Sb2Fn], 6SbF3.5SbFs crystals were isolated and identified by cell parameter analysis. 1 3 ' 1 4 Due to the similar solubilities of the Mo and W carbonyl cationic complexes and 6SbF3.5SbF5 in HF-SbFs, a clean, quantitative separation was not achieved.1 4 The main products are identified by colour and physically separated with forceps. However, their crystal structures are solved. 1 3 ' 1 4 2. Structural Features and Internal Bond Parameters As with most other thermally stable homoleptic carbonyl cationic complexes,2 0 [{Mo(CO)4}2(cis-p.-F2SbF4)3]x[Sb2Fii]xand [W(CO)6(FSbFs)][Sb2Fii] contain the superacid anion [Sb2Fn]" as counteranion. However, the formation of polymers and the simultaneous presence of SbFe" ligand(s) and [Sb2Fn]" counteranion as observed in the present work are unprecedented features for cationic carbonyl complexes 2 0 The vibrational spectra of [{Mo(CO)4}2(cis-u-F2SbF4)3]x[Sb2Fn]x and [W(CO) 6(FSbF 5)][Sb 2F 1i], on account of the low symmetry of the W(CO) 6 and the Mo(CO) 4 moieties, are difficult to interpret due to extensive band proliferation. In sharp contrast, for the highly symmetric [Fe(CO) 6] 2 + cation in [Fe(CO)6][Sb2F,,]2 and [Fe(CO) 6][SbF 6] 2, 3 2 all 13 fundamentals have been detected experimentally (see Chapter 3). Considerably longer M-C distances, higher V(co)average values, and higher C-0 stretching force constants are found for the structurally characterized [Sb2Fn]" salts with the metal carbonyl cations [Hg(CO) 2 ] 2 + , 1 8 [Ir(CO) 5 Cl] 2 + , 3 3 and [Fe(CO) 6 ] 2 + . 3 2 The v(CO)average of [Hg(CO) 2] 2 + is 130 as high as 2280 cm"1 and the fco value is 21.0 x 102 N m" 1 . 1 8 Here the extent of 7c-backbonding is affected by the formal charge of the central metal, which is +2 for Hg and F e , 1 8 ' 3 2 - 3 3 and also +3 for Ir as opposed to only +1 in [W(CO)6(FSbF5)][Sb2Fii] and +1/2 in [{Mo(CO) 4} 2(cis-p-F 2SbF4)3][Sb 2Fn]. Although the observed C - 0 stretching frequencies of [{Mo(CO) 4} 2(cis-p-F 2SbF 4) 3] x[Sb 2Fn] x and [W(CO)6(FSbF 5)][Sb 2F u] are lower than those of most homoleptic metal carbonyl cations, they are the highest reported for Mo(II) and W(II) carbonyl derivatives. 6 ' 1 1 ' 2 1- 3 4 There are a number of similarities between the structurally characterized metal carbonyl-[Sb2Fn]" sa l t s 1 8 ' 3 2 ' 3 3 and the two complexes, [{Mo(CO) 4} 2(cis-u-F 2SbF 4) 3] x[Sb 2Fii] x and [W(CO) 6 (FSbF 5 )][Sb 2 Fii]. 1 3 ' 1 6 Firstly, the M - C - 0 bond angles depart from linearity and fall between 174.8(8)° and 179.3(8)°. Secondly, the di-octahedral anion [Sb2Fn]" in [{Mo(CO) 4 } 2 (cis-p-F 2 SbF 4) 3 ]x[Sb 2 F u ] x and [W(CO)6(FSbF5)][Sb2Fn] is distorted with Sb-F-Sb bridge angles of 150.0(3)° and 149.4(3)°, and dihedral angles of 19° and 30° respectively. Also, the SbF4" groups of the anions of both complexes are staggered. 1 3' 1 6 Very similar bridge angles and rotational distortions are found for the anions in [Hg(CO) 2][Sb 2Fn] 2 , 1 8 [Ir(CO) 5Cl][Sb 2Fn], 3 3 [Fe(CO)6][Sb2Fn]2 3 2 and [H 3 0 ] [Sb 2 F n ] . 3 5 These distortions are attributed to significant interionic contacts20 or to asymmetric O - H - F hydrogen bonds.3 5 The significant secondary interionic contacts involve primarily the C atoms of the carbonyl groups and equatorial F atoms of the [Sb2Fn]" anion in metal carbonyl -[Sb 2Fn]" sal ts . 1 8 ' 3 2 ' 3 3 In the case of the Mo(II) and W(II) complexes, there are fewer C—F contacts, although there are also intra-cationic C—F contacts that involve F atoms of the F 2SbF 4" and FSbFs" ligands. There is about one significant C—F contact per carbonyl group in each case. 1 3 ' 1 6 For 131 [W(CO)6(FSbF5)][Sb2Fn], the secondary contacts are 0.45 to 0.20 A shorter than the sum of the van der Waals rad i i . 1 4 ' 1 9 In [{Mo(CO)4}2(cis-u.-F2SbF4)3]x[Sb2Fii]x, weaker contacts are noted which range from 0.2 to 0.1 A shorter than the sum of the van der Waals rad i i 1 3 ' 1 9 . The reduced number and strength of secondary C--F contacts for [{Mo(CO)4}2(cis-u.-F 2SbF 4)3] x[Sb 2Fi 1] x ,[W(CO)6(FSbF 5)][Sb 2F u] compared to other metal carbonyl-[Sb2Fn]" sa l t s 1 8 ' 3 2 ' 3 3 reflect differences in the formal charge of the central metal. Moreover, in open and less crowded cations like linear [Hg(CO) 2] 2 + 1 8 or square planar [M(CO )4 ] 2 + (M = Pd or Pt) , 3 6 secondary interionic contacts with the [Sb2Fn]" anions are stronger and more numerous 1 8 ' 3 6 than in the seven-coordinated Mo(II) and W(II) salts. 1 3 ' 1 6 4.6 S u m m a r y and Conclus ions It was found that M(CO)6 (M = Mo and W) undergoes facile 2-electron oxidation when treated with SbFs under mild conditions (40 - 60 ° C ) . 1 3 ' 1 4 The oxidation may be performed in an excess of SbFs or in HF-SbFs, where crystalline products are obtained. Cr(CO)6 is likely to behave in the same way in SbFs, at lower temperatures (20 - 25 °C). For Mo and W, seven-coordinate carbonyl complexes [{Mo(CO)4} 2(cis-|i-F 2SbF4) 3] x[Sb 2Fn] x and [W(CO)6(FSbFs)][Sb2Fn] were isolated as the main products. For Cr, a transient cationic carbonyl complex was observed. This species decomposed upon removal of volatiles in vacuo. When FfF was involved in the reaction, Cr[SbFe]2 was formed as a solid after the decomposition. However, neither the transient carbonyl species nor Cr[SbF 6] 2 could be isolated in pure form. 132 4.7 References 1) Cotton, F. A.; Wilkinson, G. Advanced Inorganic Chemistry; 5th ed.; Wiley: New York, 1988. 2) Ellis, J. E. Adv. Organomet. Chem. 1990, 31, 1. 3) Heinemann, F.; Schmidt, H. ; Peters, K.; Thiery, D. Z Kristallogr. 1992, 198, 123. 4) Mak, T. C. W. Z Kristallogr. 1984, 166, 277. 5) Jones, L. H. ; McDowell, R. S.; Goldblatt, M . Inorg. Chem. 1969, 8, 2349. 6) Baker, P. K. Adv. Organomet. Chem. 1996, 40, 45. 7) O'Donnell, T. A.; Phillips, K. A. Inorg. Chem. 1973, 12, 1437. 8) O'Donnell, T. A. Superacids and Acid Melts as Inorganic Chemical Reaction Media; V C H Publishers: Weinheim, 1993. 9) Anker, M . W.; Colton, R.; Tomkins, 1. B. Pure Appl. Chem. 1968, 18, 23. 10) Colton, R. Coord. Chem. Rev. 1971, 6, 269. 11) Baker, P. K. Chem. Soc. Rev. 1998, 27, 125. 12) Melnik, M . ; Sharrock, P. Coord. Chem. Rev. 1985, 65, 49. 13) Brochler, R.; Freidank, D.; Bodenbinder, M . ; Sham, I. H. T.; Willner, H. ; Rettig, S. J.; Trotter, J.; Aubke, F. Inorg. Chem. 1999, 38, 3684. 14) Brochler, R.; Sham, I. H . T.; Bodenbinder, M . ; Schmitz, V. ; Rettig, S. J.; Trotter, J.; Willner, H. ; Aubke, F. Inorg. Chem. 2000, 39, 2172. 15) Freidank, D.; Aubke, F., Personal Communication. 16) Brochler, R., Personal Communication. 17) Leigh, G. J.; Richards, R. L. Comprehensive Coordination Chemistry; Wilkinson, G., Ed.; Pergamon Press: London, U . K., 1987; Vol. 3, pp 1265. 18) Bodenbinder, M . ; Balzer-Joellenbeck, G ; Willner, H. ; Batchelor, R. J.; Einstein, F. W. B.; 133 Wang, C ; Aubke, F. Inorg. Chem. 1996, 35, 82. 19) Bondi, A. J. Phys. Chem. 1964, 68, 441. 20) Willner, H. ; Aubke, F. Angew. Chem. Int. Ed. Engl. 1997, 36, 2402. 21) Adams, D. M . Metal-Ligand and Related Vibrations; Edward Arnold (Publishers) Ltd.: London, 1967. 22) Jones, L. H. Inorganic Vibrational Spectroscopy; Marcel Dekker: New York, 1971; Vol. 1. 23) Mabbs, F. E., Machin, D. J. Magnetism and Transition Metal Complexes; Chapman and Hall: London, 1973. 24) Figgis, B. N . ; Hitchman, M . A. Ligand Field Theory and its Applications; Wiley-VCH, 2000. 25) Lam, C. T.; Novotsy, M . ; Lewis, D. L.; Lippard, S. J. Inorg. Chem. 1978, 17, 2127. 26) Arnesen, S. P.; Seip, H. M . Acta Chem. Scand. 1966, 20, 2711. 27) Green, M . L. H. ; Parkin, G.; Mingquin, C ; Prout, K. J. Chem. Soc, Chem. Commun. 1984, 1400. 28) Nakamoto, K. Infrared and Raman Spectra of Inorganic and Coordination Compounds; 4th ed.; Wiley: New York, 1986. 29) Siebert, H. Anwendungen der Schwingungsspektroskopie in der Anorganischen Chemie; Springer-Verlag: Berlin / Heidelberg, Germany, 1966. 30) Edwards, A. J. J. Chem. Soc. (A) 1970, 2751. 31) von Ahsen, B.; Bach, C ; Pernice, H.; Willner, H. ; Aubke, F. J. Fluorine Chem. 2000, 102, 243. 32) Bernhardt, E.; Bley, B.; Wartchow, R.; Willner, H. ; Bill, E.; Kuhn, P.; Sham, I. H . T.; Bodenbinder, M . ; Brochler, R.; Aubke, F. J. Am. Chem. Soc. 1999, 121, 7188. 33) Bach, C ; Willner, H. ; Wang, C ; Rettig, S. J.; Trotter, J.; Aubke, F. Angew. Chem., Int. Ed. Engl. 1996, 35, 1974. 134 34) Braterman, P. S. Metal Carbonyl Spectra, Academic Press: New York, 1977. 35) Zhang, D.; Rettig, S. J.; Trotter, J.; Aubke, F. Inorg. Chem. 1996, 35, 6113. 36) Willner, H . ; Bodenbinder, M . ; Brochler, R.; Hwang, G.; Rettig, S. J.; Trotter, J.; von Ahsen, B.; Westphal, U. ; Jonas, V.; Thiel, W.; Aubke, F. J. Am. Chem. Soc. 2001, 123, 588. 135 Chapter 5 Metal(II) Hexafluoroantimonates(V) 5.1 Introduction As described in Section 1.2 of Chapter 1, this work is concerned with so-called isolable naked metal cations, defined as cations in binary combination with very weakly coordinating anions such as [SbF6]" in the solid state or in weakly coordinating solvents. The metals which form isolable metal(II) hexafluoroantimonate(V) salts are mainly from the 3d series (Cr, Mn, Fe, Co, N i and Cu), and infrequently from the late 4d series (Pd and Ag). These M[SbFe]2 compounds are occasionally found as impurities in or decomposition products of complex homoleptic carbonyl cation salts with [Sb2Fn]". For example, Fe[SbFe]2 is a paramagnetic impurity in some preparations of [Fe(CO)6][Sb2Fi i ] 2 ; it is also present as the decomposition product of the carbonyl compounds (see Chapter 3). Cr[SbF6]2 is sometimes obtained together with a Cr carbonyl cationic complex in the reaction involving Cr(CO)6 and SbF 5 (see Chapter 4). Several synthetic routes to the metal(II) hexafluoroantimonates(V) have been reported,1"3 and the molecular structures of a few M[SbF6J2 complexes (where M = Ag, M n and Pd) have been determined previously.3"5 However, variable temperature magnetic studies were only reported for several M[SbF6]2 complexes (M = Ni , Cu and Pd).1 Therefore, the magnetic properties of these compounds were studied in this work. Magnetic susceptibility measurements on the metal(II) hexafluoroantimonates(V), prepared here, indicated that products from different synthetic routes have different magnetic properties in the case of nickel and palladium. The results also support the conclusion that the Fe[SbFe]2 is present as by-products in iron carbonyl product mixtures (see Chapter 3). These magnetic characterization have permitted interpretation of the magnetic properties of the carbonyl complexes. 136 Another interesting aspect of these metal cations is that, in their solid-state environments, they might be anticipated to approximate the free ion state. The ground electronic states of the M[SbFe]2 complexes studied here, using magnetic susceptibility measurements, are compared to the theoretical free ions ground states in this chapter. Aside from the isolable metal cations, there exists also another type of metal cation that is only transient. The transient cations are from the 4d, and more so from the 5d series. They include Os 2 + , Ir 3 +, Pt 2 + , Au + , Hg 2 + , R u 2 + and Pd 2 + . These cations exist in superacids and will be carbonylated to give linear d 1 0 (M = Au or Hg), square planar d 8 (M = Pd or Pt) and octahedral d 6 (M = Ir, Ru or Os) homoleptic metal carbonyl cations. The transient metal cations can best be formed by reductive carbonylation of metal fluorides or fluorosulfates in SbFs or HF-SbFs, and are stabilized by CO. For example, 2IrF6 + 12SbF5 + 15CO — 2[Ir(CO)6][Sb2Fn]3 + 3COF 2 In a similar manner, O s F 6 and PtFc are converted to [Os(CO)6][Sb2Fn] 2 and [Pt(CO)4][Sb2Fn]2. In the synthesis of homoleptic carbonyl compounds, such as [Fe(CO)6][Sb2Fn]2, 6 oxidation of the metal takes place before carbonylation7 (see Chapter 3). Support for this comes from the conversion of W(CO)6 into [W(CO)6(FSbFs)][Sb2Fn], where a transient metal carbonyl cation, [W(CO)6] 2 + , appears to form initially.8 The intended reaction, addition of CO to give [W(CO) 7] 2 + , does not occur; the addition of [SbF6]" to the W centre is observed to give [W(CO) 6(FSbF 5)] + instead. The complex was isolated as [W(CO )6 (FSbF5)][Sb2F u ] 7 (see Chapter 4). No M[SbFe]2 salts, except for Pd[SbF6]2, of these metals are known; and the transient cations can only be isolated in the form of a carbonyl salt under these conditions. In this chapter, we will focus on the characterization and chemistry of the isolable M[SbF6]2 salts of Cr, Mn, Fe, Co, Ni, Cu, Pd and Ag. 137 5.2 Syntheses The divalent metal hexafluoroantimonates, M[SbFe]2 (M = Mn, Fe, Co, Ni, Cu, Zn, Pd and Ag), synthesized in this work were prepared according to published methods or employing the same reagents in modified published procedures.1"3 The methods of preparation were: 1. Reactions of metal(II) fluorides with SbFs in HF 2. Solvolysis of metal(II) fluorosulfates in excess SbF5 3. Metal oxidation by SbF5 in S 0 2 4. The reaction of Cr(CO)e with SbF5 These preparations are described in the following sections. 5.2.1 Synthesis via Reaction of Metal(II) Fluorides with SbFs in H F In this method, metal(II) hexafluoroantimonates(V) are prepared through an acid-base reaction in HF-SbF 5 : 3 HF M F 2 + 2SbF5 • M[SbF 6 ] 2 (eq. 5.1) In a typical experiment, 2 mL (28 mmol) of SbFs was added to about 0.3 g (about 3 mmol) M F 2 (M = Fe, Co, Ni, Cu 3 and Mn) in a PFA reactor at -196 °C. HF (5 mL) was then added to the mixture. The reaction mixture was stirred briefly at -60 °C. Afterwards, it was slowly warmed to -15 °C and held at this temperature for a week. In the reactions of Mn, Co, Ni and Cu, crystalline solids formed during this time. Upon removal of the volatiles under dynamic vacuum at -15 °C, the products were collected in the form of crystalline solids or powders; they were identified, by vibrational spectroscopy, as the corresponding M[SbFe]2 compounds (M = Mn, Co, Ni and Cu). 3 In an earlier report, the reactions were carried out in fluoroplastic reaction tubes with valves, where the reaction mixtures were warmed to room temperature upon addition of H F . 3 The volatiles were immediately removed at room temperature. Crystals were 138 prepared by slow concentration of the HF solutions at elevated temperatures.3 It was found in the current work that even slowly concentrating the HF solution at or above room temperature usually produced crystals that were too small for X-ray diffraction studies. Table 5.1 summarizes the visual characteristics of the reagent fluorides, the reaction mixture and the product for each metal. Table 5.1 Characteristics of the starting reagents, the reaction mixtures and the products for the reactions of metal(II) fluorides with SbFs in HF (metal = Fe, Co, Ni, Cu andMn). Starting Reagent Reaction Mixture Product FeF 2, pale pink powder light pink colour of the solution fades to give a white precipitate in a colourless solution white powder CoF 2 , bright pink powder bright pink solution, no precipitate pink crystals N1F2, bright yellow powder pale yellow solution, no precipitate light yellow crystals CuF 2 , white powder colourless solution, no precipitate colourless crystals MnF 2 , white powder colourless solution, no precipitate colourless crystals Recrystallization was carried out if only small crystals were obtained from the initial preparation of the M[SbF6]2 (M = Mn, Co, Ni and Cu) compounds. A sample (0.33 - 0.35 g) was dissolved in 2 mL HF in a PFA reactor. Volatiles were removed from the solution in vacuo until the first solids appeared. The solution was then slowly cooled to -15 °C at about 1 °C per hour, and held at -15 °C for five days. Upon removal of all volatiles in vacuo at -15 °C, the crystals were collected. These crystals were stable at room temperature in a moisture-free environment. In the preparation of Ni[SbFe]2 and Cu[SbFe]2, some well-formed crystals were 139 isolated and subjected to X-ray diffraction analysis, and identified by their unit cell dimensions as [H 30][SbF6]. 9 (Details are presented in Section 5.3.1). To ensure that samples of Cu[SbF6J2 contained no by-products of the type Cu[Fn(SbF6)2-n], samples of Cu[SbF6J2 were stirred in liquid SbFs at 50 °C for a week. HF or SO2 was then added and the mixture was stirred for another week. Upon removal of all volatiles, the IR spectra of the product after this treatment were found to be the same as those of the starting material, Cu[SbF6]2- The weights of the products were found to slightly exceed that of the starting materials (by approximately 3.0% in weight). This could be due to residual SbFs or a small amount of S0 2 .SbF 5 . 1 0 5.2.2 Synthesis via Solvolysis of Metal(II) Fluorosulfates in excess SbFs The M[SbF6]2 compounds (M = Pd and Ag) were synthesized according to:1 M(S0 3 F) 2 + 6SbF5 2 5 ' 6 ° ° C > M[SbF 6 ] 2 + 2Sb 2 F 9 S0 3 F (eq. 5.2) from M(S0 3 F) 2 (M = Pd and Ag) in SbF 5. The fluorosulfate, Pd(S03F)2, was prepared from the reaction between Br 2 and Pd2(S03F)6; the latter fluorosulfate was obtained from mixing Pd and S 2 0 6 F 2 in"HS0 3 F. 1 1. 1 2 Ag(S0 3 F) 2 was prepared from Ag and S 2 0 6 F 2 in H S 0 3 F . 1 3 Excess SbF 5 (approximately 10 mL, 140 mmol) was added to M(S0 3 F) 2 (M = Pd or Ag; 400 -900 mg, 1 - 3 mmol) in a one-part glass reactor at -196 °C. The reaction mixture was then stirred at slightly elevated temperatures (25 - 50 °C). Detailed temperatures, times, weights and colour changes of the reactions are shown in Table 5.2. In both cases, the measured weight of the product slightly exceeded the theoretical yield. This was ascribed to residual SbFs in the product. 140 Table 5.2 Temperatures, times, weights and colour changes for the reactions of metal(Il) fluorosulfates in excess SbFs (metal = Pd and Ag). M Reaction temperature Reaction time Weight and colour of reagent, M(S0 3F) 2 Weight and colour of product, M[SbF6]2 Theoretical yield Pd 50 °C 14 days 443.7 mg 876.5 mg 841.9 mg (1.457 mmol), (1.517 mmol), purple powder light grey powder Ag 25 °C 10 days 829.9 mg (2.712 mmol), brown powder 1593.0 mg (2.750 mmol), white powder 1571.4 mg 5.2.3 Synthesis via Oxidation of Metal by SbF5 in S0 2 This method involves the oxidation of a metal (Mn, Fe, Co or Ni) by SbFs in SO2.2 6M + 23SbF5 — — > 6M[SbF 6] 2 + 6SbF3.5SbF5 (eq. 5.3) In an earlier report, it was stated that this reaction for Co yields Co[F(SbF6)] according to: 2 SQ 2 6Co+17SbF 5 > 6Co[F(SbF6)] + 6SbF3.5SbF5 (eq. 5.4) It was found in the present work that only Co[SbF6J2, not Co[F(SbF6)], is produced using this route. The product Co[SbFe]2 was identified by its appearance, vibrational spectra and magnetic properties (see Section 5.3). In a typical experiment,2 1 mL of SbFs was added to 1 g of metal powder (metal = Mn, Fe, N i and Co) in a one-part glass filtration vessel, with two round bottom flasks connected by a frit 141 bridge, at -196 °C. Excess liquid SO2 (approximately 10 mL) was then added. The reaction was exothermic and the vessel became warm. Vigorous bubbling was observed as the reaction mixture was warmed from -196 °C to room temperature. The solution was stirred at room temperature for 24 hours. In each case, the product mixture consisted of excess metal powder and a precipitate of initially white 6SbF3.5SbFs in addition to a solution of the intended product. Table 5.3 gives the colour of the reaction mixture for each metal. The solution was separated from the solid 6SbF3.5SbFs and excess metal powder by filtration through the frit into the other side of the vessel. Upon removal of all volatiles, the product was collected in powder form as an adduct of SO2. To remove all remaining SO2, the powder was heated at 120 °C under dynamic vacuum to constant weight. Table 5.3 The colour of the reaction mixtures and the products obtained for each metal when oxidized by SbFs in SO2 (metal = Mn, Fe, Ni and Co) Metal Reaction mixture Product Mn dark blue mixture —» grey mixture —> black powder (Mn) and white solids in colourless solution Mn[SbF 6] 2, white powder Fe yellow mixture —» green mixture —> black powder (Fe) and green solids in dark green solution Fe[SbF 6] 2, off-white powder Ni green mixture —> black powder (Ni) and green solids in yellow-green solution Ni[SbF 6] 2, light yellow powder Co violet mixture —> dark violet mixture —» grey powder (Co) and violet solids in violet solution Co[SbF 6] 2, violet powder 142 This procedure is slightly modified from Dean's procedure.2 The one-part filtration vessel used in this work is similar to the reaction vessel used in Dean's experiment, where the vessel consisted of two 100 mL round-bottomed flasks, open at the top via % inch outer diameter glass tubing and joined at the neck by an arm having in it a fine frit and a constriction.2-1 4 SbFs was added to the metal in the dry box instead of addition via condensation as in this work. The open ends were temporarily closed with Teflon fittings at this point. Upon condensation of SO2, the two lA inch glass tubes were sealed off under vacuum. The products, which were soluble in SO2, were separated from any insoluble materials by filtration through the fine glass frit as in this work. To remove the volatiles from the product, SO2 was frozen in the flask which is closest to the constriction and the constriction was sealed off . 2 ' 1 4 The products, collected as solvated materials, were then transferred to glass vessels in the dry box. The unsolvated products were obtained by pumping the solvated materials in vacuo at only 50 °C instead of at 120 °C as in the present work.2 5.2.4 Synthesis of Cr[SbF 6 ] 2 via Reaction of Cr(CO) 6 and SbF 5 In early attempts to synthesize carbonyl cations of Cr from the reaction of Cr(CO)6 in SbFs (see Chapter 4), Cr[SbFe]2 was obtained instead. In a typical reaction, 2 mL (28 mmol) of SbFs was condensed onto 0.22 g (1 mmol) of Cr(CO)6 in a one-part glass reactor at -196 °C. No immediate reaction was observed at room temperature. On stirring, the mixture became slightly yellowish and more viscous. The mixture was stirred at 50 °C for two days, and during this time became a creamy light yellow solution. Upon removal of the volatiles at 50 °C, a pale yellow powder was obtained. This was the 1:1 stoichiometric mixture of Cr[SbF6]2 and 6SbF3.5SbF5, resulting from the redox reaction: 143 Cr(CO) 6 + 3SbF5 » Cr[SbF 6] 2 + SbF3 + 6 C 0 ( g ) (eq. 5.5) 6SbF3 + 5SbF5 — • 6SbF3.5SbF5 (eq. 5.6) It was later determined that the target Cr carbonyl species can be obtained under different reaction conditions (see Chapter 4). In an attempt to separate Cr[SbF 6] 2 and 6SbF3.5SbF5, the mixture was transferred into a one-part filtration vessel. Excess liquid S 0 2 was condensed onto the powder at -196 °C. As the mixture was warmed to room temperature, the Cr[SbFe]2 dissolved in S 0 2 to produce a clear, light orange solution mixed with 6SbF3.5SbFs as a white solid. The solution was filtered through the frit. Upon removal of the volatiles in vacuo, the filtrate yielded an off-white powder. Unfortunately, there was a lot of noise in the vibrational spectrum of this off-white powder, and the complete removal of 6SbF3.5SbFs from Cr[SbF 6] 2 could not be confirmed by vibrational spectroscopy. 5.3 Characterization 5.3.1 Structure Determination As mentioned previously, single crystals of Ni[SbFe]2 and Cu[SbFe]2 were thought to have been isolated in this work. However, data from the X-ray diffraction studies of both sets of crystals were not consistent with these compositions. The results, instead, gave the unit cell dimensions of [H30][SbFe].9 To confirm that the metals nickel or copper were present in the bulk crystalline samples, qualitative analyses were performed. As a brief test, 5-10 mg of the crystals were hydrolyzed with 20 mL of deionized water in each case. The light yellow crystals of Ni[SbF6]2 dissolved in water to give a yellowish-green solution, the characteristic colour of 144 [Ni(H20 )6] ; and the colourless crystals of Cu[SbF6]2 dissolved in water to give a light blue solution, the colour of [Cu(H20 )6 ] 2 + . To further confirm the presence of the metals, additional tests were performed.15 To isolate the nickel in Ni[SbFe]2 by precipitation, N H 4 O H was added to the yellowish-green aqueous solution dropwise until the solution was alkaline, converting [ N i ( H 2 0 ) 6 ] 2 + to [Ni(NH3)6] 2 +. Dimethylglyoxime (3 drops) was then added. Nickel was collected as bis(dimethylglyoximato)nickel(II) ( N I C S H H ^ C M ) , a salmon coloured precipitate. To isolate the copper in Cu[SbF6]2 by precipitation, the light blue aqueous solution was made alkaline by adding N H 4 O H . The solution turned from light blue into deep blue as [Cu(H20 ) 2 (NH3) 4 ] 2 + formed. Acetic acid ( H C 2 H 3 0 2 ) , 5 M , was added dropwise until the deep-blue colour disappeared, then 2 drops of 0.2 M K4Fe(CN)6 were added. A red precipitate of C u 2 F e ( C N ) 6 was then collected from the solution. Although the bulk metal tests are positive, data from x-ray crystal studies do not agree with the presence of the divalent metal cations in the structures. Therefore, it was concluded that for each metal, the bulk sample but not the isolated single crystal contains the metal. There are presently only three crystal structures solved to date for M[SbF6]2 compounds, namely those of silver(II) hexafluoroantimonate(V),3 palladium(II) hexafluoroantimonate(V) 4 and manganese(II) hexafluoroantimonate(V).5 In the report where the crystal structure of the silver salt is published, silver(II) hexafluoroantimonates(V) is formulated as AgF2 .2SbFs. 3 In this complex, distorted [SbFg]~ octahedrons form a cubic close-packed array and Ag ions occupy one half of the octahedral holes. The Ag cation is surrounded by six F atoms in a distorted octahedral arrangement.3 The structure of Pd[SbF6]2 is reported to be similar to 145 Ag[SbF6]2-4 This structure can also be described in terms of a distorted hexagonal close packing of F atoms, with Sb occupying the octahedral sites; in contrast to Ag[SbFe]2, a disordered distribution of the metal, Pd, is on the octahedral sites. A figure depicting the layered structure and the relationship between Pd[SbFe]2 and Ag[SbFe]2 is shown (see Figure 5.1).4 The crystal structure of Mn[SbFe]2 also has the F atoms forming a hexagonal closed-packed array, but only one fourth of octahedral holes are filled with Sb(V) and Mn(II). 5 The result is a layered structure. Figure 5.1 Relationship between the structures of Ag[SbF6]2 and PdfSbF^Ji-4 The layered structures can also be seen here. Another metal(II) hexafluoroantimonate(V) compound studied was Ni[SbF6]2, where the structure was deduced from the x-ray powder data.1 6 The structure of Ni[SbF6]2, having a hexagonal unit cell, was found to be similar to Mn[SbF6]2- It can be related to the prototype 146 structure of Li[SbF6j by occupation of every other octahedral hole with N i . 1 6 Since the use of X-ray crystal structure determination to study the M[SbF6]2 systems was unpromising in the current work, indirect methods such as vibrational spectroscopy and magnetic susceptibility measurements were employed. 5.3.2 Vibrational Analysis In this section, the vibrational spectra of the metal(II) hexafluoroantimonates(V) prepared in the current work are compared to those reported in the literature.1'3 In Table 5.4, the vibrational spectra of the 1:1 stoichiometric mixture Cr[SbFe]2-6SbF3.5SbF5 prepared in this work are compared to those of 6SbF3.5SbF5 1 7 and QF2. 2SbFs as reported earlier.3 The product mixture obtained here has Sb-F bands that correspond to literature values for both CrF2. 2SbFs and 6SbF3.5SbF5; this shows that both entities are in the sample. The vibrational spectra of Mn[SbF6]2 have not been reported previously. However, the Sb-F bands observed for Mn[SbFe]2 prepared in this work (Table 5.5) are close to those of the other M[SbF 6 ] 2 compounds (Tables 5.4 to 5.10). In Tables 5.6 to 5.10, the vibrational spectra of the M[SbF6]2 compounds (M = Fe, Co, N i , Cu and Pd) prepared in this work are compared to the published spectra.1'3 It can be seen that the intense Sb-F bands obtained here correspond to the literature values. The IR and Raman frequencies of the "Co[F(SbF6)]" and Co[SbF6]2 samples prepared in the present work are shown in Table 5.7. The vibrational spectra of "Co[F(SbFe)]" have not been previously reported in the literature. It was, however, found in this work that "Co[F(SbF6)]", synthesized from reacting Co with SbFs in S 0 2 as described in an earlier report,2 has vibrational spectra 147 extremely similar to those of CofSbTVh- Results of the magnetic susceptibility measurements also indicate similar magnetic behaviour for both of the samples (see Section 5.3.4). The colour of the two samples is also very alike. From these results, we have concluded that both samples were Co[SbFe]2 (see Section 5.3.4 for details), and that Co[F(SbFe)] was not formed in the reactions carried out and described in the current work. Table 5.4 IR and Raman data for Cr[SbF6]2.6SbF3.5SbF5, CrF2.2SbF5 3 and 6SbF3.5SbF5.17 1:1 stoichiometric mixture of Cr[SbF6]2.6SbF 3.5SbF 5 (this work) CrF 2.2SbF 5 3 6SbF3.5SbF5 1 7 IR* Raman* IR* Raman* Raman (350 - 800 cm"1) (280 - 800 cm"1) 291 (w) 298 (11) 294(w) 502 (w) 527 (w) 522 (w) 520 (m) 524 (8) 549 (sh,m) 556 (m) 555 (m) 567 (m) 560 (w) 564 (7) 566(m) 577(m) 589 (mw) 601 (mw) 605 (m) 600 (w) 635 (m) 630 (m) 622(w) 647 (s) 648 (100) 652(vs) 659(s) 669 (m) 666 (sh,m) 681(w) 698 (w) 692 (20) 706 (vs) 694 (vs) 704 (22) 710 (s) 711 (w) 729 (s) 723 (m) 726 (36) Notes for Tables 5.4 - 5.10: The bands are in wavenumbers (cm") and the estimated intensities are given in parentheses. # : v = very, w = weak, m = medium, s = strong, sh = shoulder and br = broad; *: relative intensity. 148 Table 5.5 IR and Raman data for Mn[SbF6]2 (250 - 800 cm'). Mn[SbF 6] 2 (this work) IR* Raman11 268 (vw) 279 (m) 288 (m) 300 (mw) 308 (mw) 528 (m) 539 (m) 622 (mw) 631 (mw) 643 (s) 690 (m) 698 (s) 698 (w) 722 (sh,m) 723 (m) Table 5.6 IR and Raman data for Fe[SbF6]2 and FeF2.2SbF5 3 (290 - 800 cm'1). Fe[SbF 6] 2 (this work) FeF 2.2SbF 5 3 IR" Raman IR" Raman* (350-800 cm"1) (280 - 800 cm"1) 264 (w) 268 (28) 282 (s) 292 (80) 543 (w) 540 (m) 562 (19) 581 (vw) 611 (m) 618 (m) 647 (s) 669 (w) 671 (vs) 670 (m) 667 (100) 694 (m) 685 (m, sh) 700 (s) 703 (m) 702 (vs) 704 (33) 713 (s) 724 (m) 149 Table 5.7 IR and Raman data for Co[SbF6]2, "Co[F(SbF6)] " and CoF2.2SbF5 3 (270- 800 cm'1). Co[SbF 6] 2 (this work) "Co[F(SbF6)j" (this work) CoF 2.2SbF 5 3 IR* Raman* IR* Raman* IR* Raman* (280 - 800 (280 - 800 cm'1) cm'1) 283 (vw,sh) 296 (m) 296 (m) 460 (w, sh) 21A (22) 286(22) 480 (8.2) 519 (vw) 557 (w, sh) 558 (9.2) 569 (s) 590 (s) 589 (m) 592 (s) 604 (s) 601 (m) 621 (m) 650 (w) 671 (m) 670 (vs) 672 (m) 670 (s) 670 (vw, sh) 664 (100) 711 (s) 707 (s) 711 (s) 707 (m) 692 (vs) 738 (s) 737 (s) Table 5.8 IR and Raman data for Ni[ShF6]2 and NiF2.2SbFs 3 (270 - 800 cm'1). Ni[SbF6] 2 (this work) NiF 2.2SbF 5 3 IR* (350- 800 cm"1) Raman* (280 - 800 cm"1) 297 (w) 304 (w) IR* Raman* 300 (25) 524 (m) 527 (ms) 580 (s) 576 (25) 612 (20) 673 (s) 673 (vs) 675 (ms) 674 (100) 704 (vs) 709 (m) 723 (w) 710 (vs) 712(37) 739 (s) 740 (sh) 150 Table 5.9 IR and Raman data for Cu[SbF6]2 and CuF2.2SbF5 3 (270- 800 cm'1). Cu[SbF 6] 2 (this work) CuF 2.2SbF 5 3 IR* (350-800 cm"1) RamanU (280 - 800 cm"1) 284 (m) 291 (m) IR* 450 (sh) Raman* 274 (20) 287 (30) 525 (m) 544 (10) 550 (m) 560 (m,sh) 612 (m,sh) 564 (15) 632 (mw) 636 (m,sh) 656 (m) 669 (m) 668 (s) 665 (100) 681 (w) 676 (s) 706 (vs) 703 (vs) Table 5.10 IR and Raman data for Pd[SbF6]2. 1 Pd[SbF 6 ] : j (this work) Pd[SbF 6] 2 1 IR* (350- 1300 cm"1) Raman (280 - 800 cm"1) IR* 290.7 (m) 285 (w) -300 (w) 335 (m,sh) 345 (m) 523 (m) 522 (ms) 542 (m) 550 (w,sh) 580 (m) 570 (m) 591 (m) 612 (s) 615 (ms) 633 (sh,m) 630 (m) 642 (m) 669 (w) 667 (ms) 700 (sh,s) 698 (s) 706 (vs) 716 (vs) 719 (m) 724 (sh,s) 730 (s,sh) 884 (w) 1110(w) 1236 (sh,w) 1249 (mw) 151 There are some IR bands (884, 1110, 1236, 1249 cm"1) in the region that correspond to typical S-F and SO3 symmetric and asymmetric stretches of metal fluorosulfates11>12,18 f o r m e Pd[SbFe]2 sample prepared in this work (see Table 5.10). These bands are different from the bands reported for the starting reagent, Pd(S03F)2 (VS-F at 860 cm"1, v s y m s o , (Ai) at 1090 cm"1 and VasymSOj (E) at 1240 cm" 1). 1 1 In the product, SO3F" appears to stay bridging 1 1 but v a S y m s o 3 is split by about 13 cm"1; the other two bands are shifted to higher wavenumbers by about 20 cm"1 due to the introduction of [SbFe]". Therefore, there was an incomplete conversion of Pd(S03F)2 into Pd[SbF6]2- However, the bands of (S0 3F)" are generally of low intensity and the bulk composition of the sample appears to be mostly Pd[SbF6]2-A l l of the Sb-F stretching frequencies, which are very intense bands in the vibrational spectra, are observed at similar wavenumbers for the different M[SbF6]2 compounds (M = Cr, Fe, Co, N i , Cu, Ag and Pd) prepared in this work. They correspond to values reported in the literature for these compounds.1'3 There are differences observed in the lower intensity bands between different M[SbFe]2 compounds; nonetheless, compounds obtained here are reasonably consistent with those reported in the literature.1'3 This indicates that the metal(II) hexafluoroantimonates(V) compounds prepared in this work have vibrational spectra which are consistent with their formulation as M[SbF6]2 compounds. 5.3.3 Elemental Analysis via Atomic Absorption Spectroscopy (AA) An attempt was made to determine the purities of Fe[SbFe]2 and Cr[SbF6]2, prepared in this work, by metal analysis. Samples were dissolved in deionized water to give pale yellow solutions. Two methods of analysis were used: calibration curve and standard addition (see Chapter 3 for detailed descriptions of the two methods). Five standard aqueous solutions of 152 FeS0 4 .7H 2 0 (3.77xl0"5 - 1.89xl0"4 M) for the study of Fe[SbF 6] 2 and CrCl 3 .6H 2 0 (5.98xl0"6 -1.80xl0"5 M) for Cr[SbF6]2 were used in each analysis by both methods. Table 5.11 Results of the Atomic Absorption Studies (mole % of metal)" Sample b Calibration Curve Standard Theoretical % Addition of metal in sample Fe[SbF 6] 2 c 9.9% 10.1% 12.0% 10.6% 6Cr[SbF6]2.6SbF3.5SbF5 d 4.5% - 5.8% 5.7% a : The expected absolute uncertainty is ±0.5%. b : Fe[SbFg]2 samples were prepared by reacting Fe metal powder with liquid SbFs in S 0 2 as described in Section 5.2.3. Cr[SbF6J2 samples were prepared by reacting Cr(CO)6 with liquid SbFs at 50 °C as described in Section 5.2.4. c: These data were given in Chapter 1 in discussing the purity of the carbonyl compounds [Fe(CO) 6][Sb 2Fn] 2 and [Fe(CO)6][SbF6]2. d : The reduced by-product, SbF3, obtained in the preparation of Cr[SbF6]2 is assumed to be in the form of 5SbF 3.5SbFs. 1 7 The sample labeled Cr[SbFe]2 is therefore a mixture of Cr[SbFe]2 and 6SbF3.5SbF5 in the ratio of six to one; the ratio is determined from the balanced equation (see eq. 5.5 and 5.6): 6Cr(CO) 6 + 23SbF5 • 6Cr[SbF 6] 2 + 6SbF3.5SbF5 + 36CO ( g ) (eq. 5.7) The variation between values from the calibration curve and the standard addition studies (Table 5.11) suggest that the matrix difference between the standards and the samples can substantially affect the results.19 Differing results obtained for the same sample also suggest that there may 153 be interference due to the by-products of the hydrolysis. Clearly, these attempts to determine sample purity were unsuccessful. 5.3.4 Magnetic Susceptibility Measurements The results of magnetic susceptibility measurements on the metal(II) hexafluoroantimonates(V) are presented in this section. The observed behaviours are compared to theory for ions in weak ligand fields. Weak field terms will be used rather than free ion terms in the discussion of energy levels of the metal ions. The weak field approximation is considered appropriate because: 1. The F atom of [SbF6]" is considered to be weakly coordinated to the metal centre. This is supported by the long Ag-F distances (2.095 - 2.431 A) compared to the Sb-F distances (from 1.830 - 1.965 A) found for Ag[SbF 6 ] 2 . 3 2. F" is a weak field ligand. 2 0 Given these circumstances, the ligand field splitting of the d orbitals on the metal ion should be relatively small and the metal(II) hexafluoroantimonates(V) should be spin free, or high spin. It should be noted that the metal ions in these M[SbF6]2 complexes are weakly coordinating but not truly "naked". The "no ligand environment" description is the extreme case of a very weak ligand environment. 1. Cr[SbF6]2 The samples of Cr[SbF6]2 used for magnetic susceptibility measurements contain the reduced by-product, SbF3. It was assumed that all SbF3 in the sample was in the form of 6SbF3.5SbF5. Therefore, diamagnetic corrections have been made for Cr[SbF6]2 as well as for 6SbF3.5SbF5 according to the ratio of six Cr[SbF6]2 to one 6SbF3.5SbF5 obtained from eq. 5.7. The 154 experimental magnetic moment of Cr[SbFg]2 at 300 K of 4.39 U.B is lower than the spin-only value of 4.9 U.R. The experimental magnetic susceptibilities increase as the temperature decreases from 300 K to 2.8 K, then decrease as the temperature decreases from 2.8 K to 2.0 K (Figure 5.2). This is indicative of antiferromagnetic behaviour. The Cr(II) centre in Cr[SbF6]2 has a d 4 configuration and is subject to Jahn-Teller distortion.2 0 Hence a perfectly octahedral geometry is not expected for this species. Nonetheless, in the discussion that follows, an octahedral geometry is assumed as a first approximation. For an octahedral geometry with a high spin configuration, there are four unpaired electrons (S = 2) and a 5 E g ground state. With an E ground state, the magnetic moment of Cr[SbF6]2 is expected to be less than the spin-only value, as the metal ion has a less than half-filled shel l 2 0 ' 2 1 (see Appendix 3). Zero Field Splitting and Temperature Independent Paramagnetism (TIP) are also expected to affect the moment values.2 0 In a first attempt to fit the magnetic data, the Zero Field Splitting Model for S = 2 2 2 was employed. In this attempt, g was fixed at a calculated value of 1.98 (see Section 5.4.2) and TIP was fixed at a calculated value of 8.2xl0"5 mol/cm3 (see Section 5.4.2). Using this model, the calculated magnetic moments are much higher than the experimental values (see Figure 5.2). In a second attempt to fit the data, a Molecular Field Correction 2 2 was applied to the Zero Field Splitting Model for modeling the antiferromagnetic exchange which appears to be present. Although this gives lower calculated moments, they are still quite high compared to the experimental values (see Figure 5.2). Also, the fitted value of D is found to be zero, indicating no zero field splitting. This, in fact is impossible as this is a Jahn-Teller ion . 2 0 Furthermore, no satisfactory fit can be obtained for Cr[SbFg]2 assuming a tetrahedral geometry either. 155 5.0 0.30 4 0.25 -o E " 0 . 2 0 .a -4—» Q_ Q) § 0.15 co o 'A-' CD C D o.io H O 0.05 H o.oo H <5 CD O o o o o o O Experimental Magnetic Susceptibility • Experimental Magnetic Moment (Dashed) Calculated Moment for Zero Field Splitting with T I P (fixed: g = 1.98 and T I P = 8.2 x 10"5 cm 3 /mol, fitted: D = 17(1)cm"1, F = 0.213) (Sol id) Calculated Moment for Zero Field Splitting with T I P and Molecular Field Correction (fixed: g = 1.98 and T I P = 8.2 x 10"5 cm 3 /mol, fitted: D = 0(0)cm"1, zJ = -1.4(8), F = 0.150) n 1 1 1 1— 0 50 100 150 200 Temperature (K) T" T 250 300 350 Figure 5.2 The experimental magnetic data of CrfSbFJ? compared to various models for octahedral Cr(II) complexes. The magnetic susceptibilities peak at 2.8 K, indicative of antiferromagnetic behaviour. 156 It should be noted that the composition of the sample, formulated as 6Cr[SbF6]2-6SbF3.5SbF5, is determined based on the assumption that all of the reduced by-product was present in the form of 6SbF3.5SbF5. In fact, SbF3 or adducts of SbF 3 with SbFs in other ratios could also be present. Unfortunately, a satisfactory analysis of the magnetic data for Cr[SbFe]2 cannot be obtained here, on account of the questionable composition of the sample. Other structural characterization methods, such as X-ray crystal diffraction study, would be needed for a better understanding of the structure and magnetic properties of this material. 2. Mn[SbF6]2 The experimental magnetic moment of Mn[SbF6]2 at 300 K of 5.93 U B agrees with the spin-only value of 5.9 U B , as is expected for a A ) g term (Figure 5.3).20>21 The Mn(II) centre in Mn[SbF6]2 has a d 5 configuration and an octahedral geometry.5 For a high spin configuration, this should 6 2"F give five unpaired electrons (S = 5/2) and a A i g ground state. No TIP is expected for Mn and the g value is anticipated to be 2.0 (see Appendix 3). 2 0> 2 1 Therefore, g was fixed at 2.0 in attempts of theoretical model fittings to the experimental data. The data show magnetic moments closer to those given by the S = 5/2 Zero Field Splitting Mode l 2 2 with Molecular Field Correction 2 2 than those without the latter correction. Without the Molecular Field Correction, the calculated moments are higher than the experimental values at low temperatures. With the correction, the model fits the experimental moments within an 1.5% error limit (Figure 5.3). The goodness of fit (F-value) was measured by the least-squares fitting function, F = V[(l/N)xl{[(%caic-Xexpt)/Xexpt]2}], where N is the number of data, Xcaic is the calculated value and X e x p t is the experimental value of the magnetic susceptibility. The F-value decreases from 0.332 to 0.0274 when the correction is applied, indicating a better fit. D and zJ were set as variables and fitted to be about 1 cm"1 and -0.038(7) respectively. This indicates the presence of weak 157 (as shown by the small magnitude of zJ) antiferromagnetic behaviour. Hence, Mn[SbF6]2 is found to have characteristics of both zero field splitting and weak antiferromagnetism. Furthermore, the observed pefrof at 5.93 ps at 300 K is very close to the value 5.91 ± 0.09 pe (at 298 K) reported previously.2 6.0 H 5.5 H 5.0 H § 4 . 5 -o CD 14.0 -\ Experimental Magnet ic Moment (with 1.5 % error bars) (Dashed) Calcu la ted Moment for Zero Field Splitting (fixed: g = 2.0, fitted: D = 6cm" 1 and F = 0.332) 3.5 H (Solid) Calcu la ted Moment for Zero Field Splitting with Molecular Field Correct ion (fixed: g = 2.0, fitted: D = 1cm" 1 , z J = -0.0038(7) and F = 0.0274) 3.0 H 1 2.5 "~l 1 1 1— 50 100 150 200 Temperature (K) 250 300 350 Figure 5.3 Experimental magnetic moments of Mn[SbF(,]2 compared to the Zero Field Splitting Models. The 1.5% error bars were determined from uncertainty in experimental data and from results of measurements on different samples of Mn[SbF6]2-158 3. Fe[SbF6/2 The experimental magnetic moment of 5.38 U.B for Fe[SbF6]2 at 300 K (Figure 5.4) is quite high compared to the theoretical spin-only value of 4.9 | i B - This is expected for this case, due to the effects of first order orbital contribution. 2 0 ' 2 1 5 . 4 5.2 5.0 4 . 8 4 . 6 4 . 4 r 4 . 2 c CD o 4 . 0 o ! 3.8 c CO 5 3.6 3 . 4 3.2 3.0 2.8 2.6 O Experimental Magnetic Moment (with 1% error bars) (Dashed) Calculated Moment from Figgis Model (v = -2) (Solid) Calculated Moment from Figgis Model (v = -3) (Dotted) Calculated Moment from Figgis Model (v = -5) 50 100 150 200 Temperature (K) 250 300 350 Figure 5.4 The experimental magnetic moments ofFe[SbF(,]2 compared to calculated moments from the Figgis Model of Single Ion Effects (k — -100 cm'1, A = 1.5, k = 0.7). The 1% error bars were determined from uncertainty in experimental data and from results of measurements on different samples ofFe[SbF(,]2. 159 The Fe(II) centre in Fe[SbFe]2 has a d 6 configuration. For a high spin octahedral geometry there should be four unpaired electrons (S = 2) and a 5 T 2 g ground state. Zero Field Splitting and TIP are also expected 2 0 ' 2 1 (see Appendix 3). Attempts to fit the magnetic data from 2 to 300 K to the Zero Field Splitting,2 2 Linear Chain 2 3 and Sheet2 4 Models failed, indicating that none of these models alone can be used to explain the magnetic behaviour of Fe[SbF(;]2. On the other hand, the Figgis Model of Single Ion Effects for the 5 T 2 g term 2 1, which regards the metal centre as an isolated complex ion and takes the orbital contribution of the metal ion as well as deviation from regular octahedral geometry into account, fits the data within an error limit of 1% from 15 to 290 K (see Figure 5.4). (This model does not allow calculation of moments outside the 15 - 290 K range). The fit gives a distortion parameter, v of-3 and a derealization, or orbital reduction, factor, k of 0.7. The parameter v is defined as 8 / X where 8 is the splitting of the T term due to distortion from regular Oh symmetry and X is the spin-orbit coupling constant. 2 0 ' 2 1 The X value of-100 cm"1 was used, as calculated from the free ion spin-orbit coupling constant, ^ n s , of Fe 2 + (see Section 5.4.2). The sign of v is indicative of a particular form of distortion. Since X is negative as defined for metal ions having a more than half-filled shell, 8 has to be positive to give a negative value of v. During distortion of Oh geometry towards D 4 |,, the T 2 g term splits into B 2 g + E g . 2 5 When 8 is positive, the ground term is the doubly degenerate E g term. In effect, the d x z and d y z orbitals become lower in energy than the dxy orbital. Therefore the distortion is a tetragonal elongation along the z-axis 2 0 (Figure 5.5). The magnitude of v, 3, indicates a small to moderate splitting of the T term. The parameter k is a measure of the effective reduction of the orbital angular momentum of a central metal ion caused by the derealization of electrons from the t 2 g orbitals 160 of the ion onto the donor atoms of the "ligands". 2 0 ' 2 1 The presence of orbital contributions raises the magnetic moment above the spin-only value. 2 0 ' 2 1 Therefore, having a u.300 of 5.38 u.B which is greater than u.s = 4.9 \XB is consistent with this orbital contribution explanation. In fact, this experimental u.300 falls into the range of experimental u.3oo values for typical high spin Fe(II) complexes extending from about 5.0 to 5.6 ps- 2 1 It is also quite close to that of the high spin octahedral complex with weak field ligands, (NH4)2Fe(S04)2-6H 20 at 5.47 U . B 2 1 Furthermore, u.3oo value of 5.38 \XB found here is in good agreement with that of Fe[SbF6]2 reported earlier (5.42 ±0.08 u B at295 K ) . 2 B T 2g 2g 8 (positive) Clx2.y2, dz2 ^ x z ' ^ y z ' ^ x y - t - J / - t -— ^ x z ' ^ y z Ground term = T. tetragonal elongation along z axis D 4h 2g Ground term = E„ Figure 5.5 Tetragonal elongation of octahedrally coordinated Fe in Fe[SbFe]2 161 4. Co[SbF6]2and "Co[F(SbF6)J" The experimental magnetic moment for Co[SbF6]2 of 5.28 p B at 300 K (see Figure 5.6) is higher than the theoretical spin-only value of 3.9 pB- The Co(II) centre in Co[SbF6]2 has a d 7 configuration. For an octahedral geometry with high spin configuration this should give three unpaired electrons (S = 3/2) with a 4 T i g ground state. For a T i g ground state, the magnetic moment at 300 K is expected to be higher than the spin-only value for a more than half-filled shell, such as Co 2 + , due to first order orbital contributions.2 0'2 1 The Figgis Model of Single Ion Effects for weak field complexes of 4 T i g terms without electron derealization (A = 1.5 and k = 1.0),21 using X of-140 cm"1, fits the experimental magnetic moment data from 20 to 300 K within 1% error (see Figure 5.6, note that the model does not allow calculation of moments below 20 K). The maximum possible value of X calculated from <^ ns for the free ion Co 2 + is-172 cm" (see Section 5.4.2). Modeling with a A. of-172 cm"1 gives moments higher than the experimental values. Since metal centres of complexes usually have X values that are slightly lower than A, of the free i o n , 2 0 ' 2 1 lower values of X were used in modeling, and the value of-140 cm"1 in fact gave a good fit (Figure 5.6). In the final fit, v is found to be 0, indicating that there is no significant distortion from Oh symmetry. The parameter A is a measure of the strength of the crystal field splitting which ranges from 1.0 to 1.5. The A value of 1.5 is the weak field l imi t , 2 0 ' 2 1 which is the suitable for the metal(II) hexafluoroantimonates(V). The experimental P300 at 5.28 PB falls into the range of those for typical high spin C o 2 + complexes (4.3-5.3 p B ) , 2 1 and is also very close to that of (NH 4)2Co(S04)2 .6H 20 (p3oo = 5.10 p B), another high spin octahedral complex with weak field ligands.21 162 5.4 -5.2 -5.0 4.8 -4.6 -4.4 -f 4.2 -CD | 4.0 H o "a> 3.8 H c CD TO 5 3.6 -3.4 -3.2 -3.0 2.8 2.6 i u i O Experimental Magnetic Moment for sample prepared from CoF 2 and SbF 5 in HF (with 1% error bars) (Dashed) Calculated Moment from Figgis Model (v = 2) (Solid) Calculated Moment from Figgis Model (v = 0) (Dotted) Calculated Moment from Figgis Model (v = -2) Experimental Magnetic Moment for sample prepared from Co and SbF 5 in S 0 2 (with 1 % error bars) 100 200 300 Temperature (K) 400 500 Figure 5.6 The experimental magnetic moments of different samples of Co [SbFs] 2 compared to calculated moments from the Figgis Model of Single Ion Effects (X- -140 cm'1, A = 1.5, k -1.0). The samples prepared from Co and SbF5 in S02 was initially thought to be Co[F(SbFt;)]. The 1 % error bars were determined from uncertainty in experimental data andfrom results of measurements on different samples of Co[SbF(,]2. 163 As mentioned previously, t h e vibrational spectra of the "Co[F(SbFe)]" sample (prepared from Co and SbFs in SO2) and Co[SbFe]2 (prepared from CoF 2 and SbF 5 in HF) are extremely similar. The two samples also have very similar appearance and identical vibrational spectrum (see Section 5.3.2). Hence it is concluded both samples are in fact Co[SbFe]2. To investigate this, the magnetic susceptibility data of "Co[F(SbF6)]" were treated as Co[SbFe]2 in calculations. From Figure 5.6, it can be seen that the two samples have magnetic moments that are almost the same (within 1% error). Therefore, there is strong evidence that the reaction of Co and SbF5 in SO2 yields Co[SbF6]2 in the present work, rather than Co[F(SbFe)] as was reported in the literature.2 5. Ni[SbF6]2 The experimental magnetic moment for Ni[SbFe]2 of 3.40 pe at 300K (Figure 5.7) is higher than the spin-only value of 2.8 p B , as predicted for N i 2 + in the 3 A 2 g ground state. 2 0 ' 2 1 The Ni(II) centre in Ni[SbFe]2 has a d 8 configuration. For an octahedral geometry there should be two unpaired electrons (S = 1) and a 3 A 2 g ground state. Zero Field Splitting and TIP are also expected to be present (see Appendix 3 ) . 2 0 ' 2 1 The Zero Field Splitting Model for S = 1 2 2 with TIP Correction gives calculated magnetic moments higher than experimental values from 200 to 300 K but lower than the experimental data from 20 to 200 K. With the addition of the Molecular Field Correction, 2 2 the model f i t s the data much better (see Figure 5.7), as indicated by the decrease in F value from 0.0475 to 0.00322. The 0.2% error bars, determined from uncertainty in magnetic susceptibility data and from measurements of different samples, are about the same size as the data point and thus too small to be s h o w n on the plot (Figure 5.7). g and TIP were set as variables and fitted to be 2.29(2) and 6.(0)x 10"4 c n r V m o l respectively (see 164 Section 5.4.2 for calculations of g and TIP). The fitted zJ value of-1.88(2) indicates the presence of weak antiferromagnetic behaviour. 3.6 3.4 H 3.2 H 3.0 H CQ 2.8 c £ 2.6 -I 2.4 -co 2.2 H 2.0 1.8 1.6 H 0 O Experimental Magnetic Moment (Dashed) Calculated Moment for Zero Field Splitting with TIP (fitted: g = 2.12(1), D = 9.1(3), TIP = 0.0013(1) and F = 0.0475) (Solid) Calculated Moment for Zero Field Splitting with TIP and Molecular Field Correction (fitted: g = 2.29(2), D = 3.4(8), TIP = 0.0006(0), zJ = -1.88(2) and F = 0.00322) 50 100 150 200 Temperature (K) T 250 300 350 Figure 5.7 Experimental magnetic moments of Ni[SbF6]2 compared to the Zero Field Splitting Models with and without Molecular Field Correction. 165 The final fit of the model to the data suggests that both zero field splitting and weak antiferromagnetic interactions are present in Ni[SbFe]2 (see Figure 5.7). The experimental p.300 of 3.40 pe falls into the range of those for typical N i 2 + complexes (2.9 - 3.9 U , B ) - 2 1 It is also quite close to that of another high spin octahedral complex with weak field ligands, (NH 4)2Ni(S04)2 .6H 20, which has a u 3 0 0 of 3.23 u B 2 1 The comparison between the experimental magnetic moments of Ni[SbFe]2 found in the current work and those obtained from previous studies is shown in Figure 5.8. The magnetic moments in this work are close to the expected values of N i 2 + in the 3 A 2 g ground state. The previously published magnetic moments (prepared via solvolysis of Ni(S0 3 F) 3 in SbFs) 1 ' 2 6 are, however, lower than the expected values. These lowered moments were attributed in the earlier work to the presence of a spin free-spin paired (triplet-singlet) equilibrium, antiferromagnetism, or the presence of Ni of other oxidation states.1-26 Another set of data, obtained from a sample synthesized from the fluorination of Ni metal, 1 6 ' 2 6 gives magnetic moments that are higher than the values obtained here. The reported magnetic moment of 3.16 ± 0.06 u.B at 294 K of a sample prepared from reacting Ni and SbF5 in SO2 2 is, among all magnetic moment values of Ni[SbFe]2 in the literature, closest to the u.300 at 3.40 U.B found here. Hence, this survey of magnetic data collected for Ni[SbF6]2 samples suggests that the magnetic susceptibility depends on the method of preparation. In the present work, the method (reaction of N i F 2 and SbFs in HF) gives Ni[SbFe]2 samples with magnetic data that are consistent with pure Ni[SbF6]2. The other methods, according to previously reported results, gave atypical results (see Figure 5.8), and it is possible that these samples were contaminated by Ni in higher oxidation states than 2+, and / or other by-products. 166 4.5 CD o CD c= CD 4.0 H 3.5 H 3.0 2.5 H 2.0 H 1.5 1 1.0 • • • o ooooo o 0 • A 0 X P r e s e n t work: N i F 2 + S b F 5 + HF re fe rences 1, 26 : re fe rences 1, 26 : re fe rences 1, 26 : N i ( S 0 3 F ) 2 + S b F , Ni + F 2 + S b F 5 N i ( S 0 3 F ) 2 + S b F , re ference 2: Ni + S b F 5 + S 0 2 50 100 150 200 Tempera tu re (K) 250 300 350 Figure 5.8 A comparison of experimental magnetic moments obtained here for Ni[SbF6] with values from the literatureJ>2,26 167 6. Cu/SbF6/2 The experimental magnetic moment of Cu[SbF6]2 at 300 K of 2.07 p B (Figure 5.9) is higher than the spin-only value of 1.7 p B as expected for C u 2 + with a 2 E g ground state. 2 0 ' 2 1 This is to the presence of Temperature Independent Paramagnetism, TIP, and some orbital contributions from second order spin-orbit coupling causes mixing in of excited terms (see Appendix 3). 2 0> 2 1 0.040 0.035 0.030 o E i 0.025 Q. | 0.020 I 3 03 CD | 0.015 0.010 0.005 \ 0.000 O Experimental Magnetic Susceptibility (Dashed) Calculated Magnetic Susceptibility for Linear Chain Model (fitted: J = -3.18(4), g = 2.39(1), TIP = 0.0005(2) and F = 0.0181) • Experimental Magnetic Moment (with 2 % error bars) • (Dotted) Calculated Magnetic Moment for Curie Behaviour with TIP and Molecular Field Correction (fitted: g = 2.30(6), TIP = 0.0003(2) z J = -12(1) and F = 0.0947) (Solid) Calculated Magnetic Moment for Linear Chain Model (fitted: J =-3.18(4), g = 2.39(1), TIP = 0.0005(2) and F= 0.0181) 2.2 2.0 1.8 1.6'TS c co E o 1.4 1 .2 : 1.0 0.8 0.6 09 c CD CD 50 100 150 200 250 300 350 Temperature (K) Figure 5.9 Experimental magnetic data of Cu[SbF6j2 compared to the Curie Model with Molecular Field Correction and the Linear Chain Model. 168 The Cu(II) centre in Cu[SbFg]2 has a d 9 configuration. As for Cr(II), Cu(II) is subject to Jahn-Teller distortion. As a first approximation, however, an octahedral geometry is assumed in the discussion that follows. For an octahedral geometry, Cu(II) has one unpaired electron (S = 1/2) and a 2 E g ground state. It should be noted that for Cu 2 + , there is no zero field splitting as S = 1/2. The magnetic susceptibility data show an incipient maximum at 2K (Figure 5.9). This is indicative of antiferromagnetic behaviour. The values of g and TIP are calculated to be 2.31 and 9.7xl0"5 cm3/mol respectively (see Section 5.4.2 for calculation of g and TIP). The Linear Chain Model 2 3 was found to fit the data to within a 2% error (Figure 5.9). This is a better fit than the Curie Model with Molecular Field Correction,2 2 as indicated by the F values which decrease from 0.0947 for the Curie Model to 0.0138 for the Chain Model. Since the Molecular Field Model is only valid for extremely weak exchange, the improved fit using a Linear Chain Model indicates a significant amount of exchange in Cu[SbF6J2. This satisfactory fit to the Linear Chain Model, however, in no way proves that the structure of Cu[SbF 6] 2 is a linear chain. The exchange parameter, J, of the Linear Chain Model is fitted to be -2.94(6) while g and TIP are fitted to be 2.31(2) and 2.(0)x 10"4 cm3/mol respectively. The better fit of Linear Chain Model indicates that weak antiferromagnetic behaviour exchange is present in Cu[SbF6J2. The experimental u.300 at 2.07 U.B falls into the range of those for typical C u 2 + complexes (1.9 -2.1 U . B ) , 2 1 although it is slightly higher than that of similar high spin octahedral complexes with weak field ligands, like K2Cu(S04)2.6H20 (1.91 U.B) 2 1 A comparison of magnetic moment data from this work to those from previously published work 1 ' 2 6 is shown in Figure 5.10. The experimental magnetic moments here are close to the expected values for C u 2 + having a 2 E g ground state, while the moments of previous samples are much lower. 1 ' 2 6 The low values were 169 attributed in the earlier work to an equilibrium between C u 2 + and a mixture of Cu + and C u 3 + . 2 6 Our results here cast some doubt on the validity of the earlier data. 2.2 o o o o 0 o o o o o o 1.8 4 J£ 1.6 -c C D E o CD C CD CO 1.0 H 8 o o o o • o o o o O Present work: CuF 2 + S b F 5 + HF • references 1, 26: Cu(S0 3 F) 2 + SbF 5 0 50 100 150 200 250 300 350 Temperature (K) 0.8 Figure 5.10 A comparison of experimental magnetic moments of Cu[SbF6]2 obtained here with values in the literature. l>2f> 170 7. Pd[SbF6]2 The experimental magnetic moment at 300 K of 3.11 \XQ is slightly higher than the spin-only value of 2.8 U.B as expected for Pd 2 + with a 3 A 2 g term 2 0 ' 2 1 (Figure 5.11). As in the case of N i 2 + in Ni[SbF6]2, this is likely due to the presence of some orbital contribution which arises from the mixing in of excited terms as described in second order treatment2 0'2 1 (see Appendix 3). The Pd(II) centre in Pd[SbF 6] 2 has a d 8 configuration, and is found to have a distorted geometry surrounded by six F atoms.4 This gives two unpaired electrons (S = 1) and an 3 A 2 g ground state as for N i 2 + in Ni[SbFe]2. Both Zero Field Splitting and the presence of TIP are also expected to affect the moment in this case (see Appendix 3). When the Molecular Field Correction is also added, the Zero Field Splitting Model of S = 1 fits the data to within 0.2% error as shown on Figure 5.11, (the 0.2% error bars which are about the size of the data point are too small to be shown on the plot) with F = 1.6xl0"4. The parameters D, g, TIP and zJ are fitted to be 10.6(3) cm"1, 2.192(8), 2.(0)xl0"4 cm3/mol and -0.573(7) respectively. While there is a sudden decrease of magnetic susceptibilities with temperature observed in the previously published data, 1 ' 2 6 this behaviour is not observed in the current work. The accurate fit of the model to the data, however, shows the presence of both zero field splitting and weak antiferromagnetic interactions in Pd[SbFe]2 (Figure 5.12). It should be noted that the vibrational spectra of the sample show that there is an incomplete conversion of Pd(S03F) 2 into Pd[SbF 6] 2, and therefore there is a small amount of the fluorosulphate in addition to hexafluoroantimonate (see Section 5.3.2). Nonetheless, the magnetic behaviour of the Pd[SbF 6] 2 sample prepared here is consistent with those of octahedral Pd(II). 171 3.2 -3.0 -2.8 -2.6 -2.4 -m 3 2.2 -c C D I 2.0 H o 1 1.8 -O) CO 1.6 -1.4 1.2 H 1.0 0.8 O Experimental Magnetic Moment (Dashed) Calculated Moment for Zero Field Splitting with TIP (fitted g = 2.05(3), D = 28(1 )cm"1, TIP = 5(2)x10" 4 cm 3 /mol and F = 0.0679) (Solid) Calculated Moment for Zero Field Splitting with TIP and Molecular Field Correction (fitted: g = 2.192(8), D = 10.6(3)cm"1, TIP = 2(0) x 10"4 cm 3/mol, zJ = -0.573(7) and F = 1.60 x 10"4) ~~i 1 1 1 1 — 0 50 100 150 200 Temperature (K) 250 300 350 Figure 5.11 Experimental magnetic moments of Pd[SbF6]2 compared to the Zero Field Splitting Models. 172 CQ =1 3.2 3.0 2.8 2.6 2.4 2.2 c CD E ° 2.0 CD £ 1 - 8 -CO 1.6 1.4 -1.2 -1.0 0.8 • Q o 0 ° ° 0 0 ° ' oooooo' oooooooooo s o • O • o o o • o •° o LX) <§ % o B : o o O Present Work • references 1, 26 0 50 100 150 200 250 300 350 Temperature (K) Figure 5.12 A comparison of experimental magnetic moments of Pd[SbF6]2 obtained here with literature values J'26 173 8. Pd(S03F)2 The magnetic properties of Pd(S03F)2 were also studied in this work and compared to those of Pd[SbF6]2- The magnetic susceptibility and moment of Pd[SbF6]2 and of Pd(SC>3F)2 from the present work and a previous study 2 6 ' 2 7 are shown in Figure 5.13. The results obtained for Pd(S03F)2 here agree well with those obtained previously. 2 6 ' 2 7 Like the Pd(II) in Pd[SbF6]2, the metal centre in Pd(S03F3)2 has been shown to be in octahedral environment with a 3 A 2 g ground state.1 M 2 - 2 7 An interesting point to note is that octahedral Pd(II), found for these complexes, is formed only in very few other compounds such as P d F 2 , 2 8 > 2 9 and is rare compared to square planar Pd(II). The solid-state structures of both Pd(II) compounds can be described as the CdCb type structure, with alternating layers of metal atoms separated by the polyatomic anion [SbF6]~ or (SO3F)"; the ratio of Pd 2 + cation to polyatomic anion is one to two 4,11,12,27 The magnetic properties of the two compounds, however, are found to be distinctly different. Pd(S0 3F)2 is ferromagnetic while Pd[SbFe]2 has weak antiferromagnetic interactions at low temperatures, although the magnetic behaviour of the two compounds is quite similar at higher temperatures (100 - 300 K, see Figure 5.13). The magnetic properties of Pd(S03F)2 closely resemble those of the binary chlorides, FeCl 2 , C0CI2 and N i C l 2 , all of which have structures with layers of metal atoms separated by chloride ions . 2 7 ' 3 0 Based on results of neutron-diffraction studies of these binary, chlorides, it was reported that ferromagnetic coupling occurs between the ions within each layer, and weak antiferromagnetic coupling occurs between layers. 2 7 ' 3 1 The difference in the magnetic behaviours of Pd[SbF6]2 and Pd(S0 3 F 3 ) 2 at low temperatures may be due to the size of the anions. Since [SbFe]" is a larger anion, the distances between the Pd(II) centres both within and across the layers are larger in Pd[SbFe]2 than 174 Pd(S03F)2. Since ferromagnetism is by nature weaker than antiferromagnetism, the longer distances may be a reason for antiferromagnetic interactions dominating in Pd[SbF6]2- This assumption is consistent with the finding that, in changing from the smaller anion (S03F)~ to the larger (S0 3CF3)", Pd(S0 3CF 3)2 exhibits antiferromagnetism in contrast to ferromagnetism in Pd(S0 3 F) 2 . 2 6 1.0 0.8 =• 0.6 E o a. 0 4 o in -CO o "S c CO CD 5 0.2 0.0 \ O O A A O Magnetic Susceptibility of Pd[SbFg] 2 (Present Work) A Magnetic Susceptibility of P d ( S 0 3 F ) 2 (Present Work) • Magnetic Moment of Pd[SbFg] 2 (Present Work) ^ Magnetic Moment of P d ( S 0 3 F ) 2 (Present Work) O Magnetic Moment of P d ( S 0 3 F ) 2 (reference 26) c CO E o co CD 2 1 0 0 50 100 150 200 250 300 350 Temperature (K) Figure 5.13 The magnetic susceptibility and moment ofPd[SbF(,]2, Pd(SOiF)2 from present work and a previous study.2^ 175 The case is similar to the Ni compounds; Ni (S03CF 3 ) 2 exhibits antiferromagnetism while ferromagnetic interactions are observed in Ni(SC>3F)2 at low temperatures.26 Furthermore, the magnetic behaviours of Ni (S03CF 3 ) 2 and Ni(S03F) 2 are also very similar at 20 - 80 K, as for the Pd[SbF 6] 2 and Pd(S03F)2 above 100 K. For these Ni and Pd compounds, the magnetic moments are very similar at the higher temperature ranges, and become opposite in nature (increases or decreases) at the same temperatures (20 K for the Ni compounds and 100 K for the Pd compounds). It should be noted, though, that since geometry of the anions, octahedral [SbFe]" and tetrahedral SO3F" in the two compounds are different, the observed difference in magnetic behaviours is not surprising. 5.3.5 UV-VIS Spectroscopy It would have been interesting to compare the calculated lODq values (see Section 5.4.2) to experimental values obtained from the UV-VIS spectra of the metal(II) hexafluoroantimonate(V). Unfortunately, attempts to obtain UV-VIS-NIR spectra (employing a Cary 5 spectrometer) of the M[SbF 6 ] 2 compounds were unsuccessful. Mulls were made up using fluorocarbon grease and the powder samples; however, the samples decomposed too quickly to obtain any useful spectra. Therefore, this course of investigation was halted. 5.4 Discussion 5.4.1 The Relationship between the Isolable Metal Cations as [SbF6]~ Salts and the Transient Metal Cations that React with CO to form Carbonyl Cations /. The Isolable Metal Cations The M 2 + (M = Cr, Mn, Fe, Co, Ni, Cu and Zn), isolated as [SbF6]" salts are primarily from the first transition series. They are, according to the HSAB (hard - soft acid - base) concept. 176 classified as "hard" or borderline acids 3 2- 3 3 that will coordinate strongly to hard N - , 0-, F- type donors and exist in aqueous solutions as hydrated ions or in other coordinating solvents.34 A l l of these M 2 + have extensive coordination chemistry.34 Their d-electron configurations range from d 4 to d 1 0 . Except for d 1 0 (Zn 2 +), all of these cations have paramagnetic ground states. Their coordination geometries are mostly octahedral and occasionally tetrahedral or five-coordinate. They give rise to ligand field spectra which are well analysed.20 2. The Transient Metal Cations In this case, the M ' m + (M' = Os 2 + , Ir 3 +, Pt 2 + , A u + , Hg 2 + , R u 2 + and Pd 2 +) are from the 5d and 4d series. These cations are generated by reductive and solvolytic methods, exclusively in superacids. They are soft or borderline acids 3 2 - 3 3 that require stabilization by the soft base CO. Generally, they do not exist in aqueous solutions and, with the exception of Pd, have no coordination chemistry.34 These transient metal cations react with CO to form homoleptic carbonyl cations.7 The effective atomic numbers are 18 for [M' (CO )6 ] n + (n = 2, M ' = Ru, Os; n = 3, M ' = Ir), 16 for [M'(CO) 4 ] n + (n = 2, M ' = Pd, Pt) or 14 for [M'(CO) 2 ] n + (n = 1, M ' = Au; n = 2, M ' = Hg). 7 The resulting d configurations are d 6 (octahedral), d 8 (square planar) and d 1 0 (linear). Paramagnetic ground states are unknown among them, as discussed elsewhere in this thesis (see Chapter 1). There are two areas of overlap, namely Fe 2 + and Pd 2 + , for the isolable metal(II) hexafluoroantimonates(V) and the homoleptic metal carbonyl cations isolated as [SbiFn]" salts (see Figure 5.14). [Fe(CO)6][Sb2Fn]2 is obtained only by oxidative carbonylation of Fe(CO)s; it decomposes reversibly with loss of CO to Fe[SbFe]2 as discussed in Chapter 3. Pd , a soft acid, forms a very limited number of octahedral complexes which are paramagnetic (for 177 example, PdF 2 , 2 8 > 2 9 Pd(S0 3F) 2 2 6 and Pd[SbF 6] 2 discussed here). Square planar [Pd(CO) 4] 2 + is formed by reductive carbonylation of Pd(S0 3F) 3 or Pd[Pd(S0 3F) 6]. 3 5 [Pd(CO)4][Sb2F, , ] 2 decomposes with loss of COF 2 in a complicated manner.35 metals that form isolable metal cations as M[SbF6]2 salts 4 Ti Zr Hf V 7 3__ 9 Cr I Mn mtiym ,1 10V11 12 ] Fe * Co L L Ni i Cu Zn Nb E Mo !| Tc - J Ru |! Rh Ta I W Re Pd ] Ag Cd Os Ir Pt A metals that form carbonyl cations as -[SbjFJ" salts Au i j Hg 1 = metals that form transient metal cations stabilized by CO Figure 5. 14 The distribution of metals that form isolable M[SbF6]2 compounds, those that form carbonyl cations as -[Sb2Fi /]' salts and the transient metal cations stabilized by CO. 5.4.2 Parameters Parameters such as g and TIP generated in fitting magnetic data to the models of various magnetic behaviours are compared to the calculated values obtained using a set of equations. These equations are presented in the following section. The values of variables used in the calculations of these parameters as well as the calculated values of the parameters for each first row transition metal cation studied as a metal(II) hexafluoroantimonate(V) in this work are listed in Tables 5.12 and 5.13. 178 1. Calculation of X The spin-orbit coupling constant, X, of a complex can be calculated from the equation: 2 0 ' 2 1 ±X = ^„d I 2S (cm"1) (eq. 5.8) where X is positive for a less than half-filled shell and negative for more than a half-filled shell, E,„4 is the free ion single electron spin-orbit coupling parameter and S is the total spin on the metal centre. It should be noted that this X value is calculated for a free metal ion, and would be the maximum possible X for the metal centre in a complex. 2. Calculation of lODq The theoretical 1 ODq value for a complex can be calculated using an empirical relationship developed by C. K. Jorgenson:36 lODq = g x [(x/6)f ] x 103 (cm"1) (eq. 5.9) where the parameters g and f have been tabulated from empirical data, with g describing and metal and f a set of six ligands. In our case, the parameter f for F" ligands (0.90) is used in the calculations, since the closest atoms to the metal centre would be F from the counteranion SbF6~. If the predicted geometry is octahedral, x = 6 is used; x = 4 is used if the predicted geometry is tetrahedral. 3. Calculations of fie/f and g The effective magnetic moment, uefr, and the Lande splitting factor, g, can be calculated using the following equations: 2 0 ' 2 1 u €f f=u sx[l-(yA./10Dq)] (eq. 5.10) g = 2.0x[l-(yA./10Dq)] (eq. 5.11) 179 where y is equal to 4 for metal complexes with the A 2 ground term, and y is equal to 2 for those with the E ground term. Since X is positive for a less than half-filled shell and negative for more than half-filled shell, u.eir is smaller than p s for a less than half-filled shell and greater than p s for a more than half-filled shell. For the same reason, g is smaller than 2.0 for a less than half-filled shell and greater than 2.0 for a more than half-filled shell. Table 5.12 Variables used in the calculations of the parameters in equations 5.8 and 5.9 and the calculated values for each metal ion. Metal d" approximate ground S a Jergenson's k> X 10Dqb ion geometry term (cm1) (cm1) g for the (cm1) metal C r 2 + d4 octahedral S E c g 2 230 57.5 14.1 6 12700 M n 2 + d> octahedral 5/2 300 N / A 8.5 6 7650 Fe 2 + d6 octahedral ^ 2 400 -200 10.0 6 9000 cV + dV octahedral 4 T , g 3/2 515 -172 9.3 6 8370 N i 2 + d" octahedral - 'A 2 8 1 630 -315 8.9 6 8010 C u i + dy octahedral 1/2 830 -830 12.0 6 10800 a : The parameters S and £,nd are used in the calculation of X (see eq. 5.8). b : The parameters, J0rgenson's g and x, are used in the calculation of lODq (see eq. 5.9). 4. Calculations of TIP The Temperature Independent Paramagnetism, TIP, can be calculated from the following equation:21 TIP = z N p 2 / lODq (cm3 mole"1) (eq. 5.12) 180 where z is equal to 8 for metal complexes with the A 2 ground term, and 4 for those with the E ground term. N is the Avogadro's constant and B is the Bohr Magneton; N B 2 is equal to 0.260741 cm 3 mole"1 cm" 1 . 2 0 Table 5.13 Variables used in the calculations of parameters in equations 5.11 and 5.12, and the calculated values for each metal ion. Metal d" approximate ground S y c Lande d z TIP" ion geometry term Splitting (cm3mol1) Factor, g C r 2 + d 4 octahedral 5 E 2 2 1.98 4 8.2xl0"5 M n 2 + d> octahedral & A l g 5/2 N / A 2.0 N / A 0 Fe 2 + d 6 octahedral ^ g 2 N / A >2.0 N / A -C o 2 + A' octahedral 4 T l 8 3/2 N / A >2.0 N / A -N i 2 + d» octahedral J A 2 g 1 4 2.31 8 2.6xl0" 4 C u 2 + d 9 octahedral 2 E 1/2 2 2.31 4 9.7xl0"5 c: The parameters y as well as X and lODq from Table 5.12 are used in the calculation of the Lande splitting factor, g (eq. 5.11). d: The parameters z and lODq from Table 5.13 are used in the calculation of TIP (eq. 5.12). It should be noted that there are no second order effects for the A i g term. Thus, the Lande splitting factor, g, is 2.0 and there is no TIP. Also, the second order effects are much greater for metals with T ground states than those with A and E ground states. The parameters for metals with T ground states are much more difficult to predict. In general, the temperature dependence of the magnetic moments is high for systems with T ground terms. The magnetic moments are 181 usually higher than the spin-only values for a more than half-filled shell and smaller than the spin-only values for a less than half-filled shell. Correspondingly, the Lande splitting factor, g, is usually higher than 2.0 for a more than half-filled shell and smaller than 2.0 for a less than half-filled shell . 2 0 - 2 1 5.4.3 Magnetic properties of the metal(II) hexafluoroantimonates(V) The results of the magnetic susceptibility measurements on the M[SbF 6]2 compounds (M = Cr, Mn, Fe, Co, N i , Cu and Pd) are summarized in Tables 5.14 and 5.15. Table 5.14 Magnetic behaviours of the M[SbFe]2 compounds (M = Cr, Mn, and Fe). Complex d n S P-300 O B ) Proposed ground state Notes Cr[SbF 6] 2 d 4 2 4.38 5 E • experimental magnetic moments are much lower than the calculated values given by the Zero Field Splitting Model and the Linear Chain Model • possible impurities such as SbF3 and its adducts with SbF 5 in addition to 6SbF 3.5SbF 5 Mn[SbF 6] 2 d 5 5/2 5.93 «A,B • satisfactory fit of magnetic data to theory for octahedral geometry • characteristics of both zero field splitting and weak antiferromagnetic behaviour Fe[SbF 6] 2 d 6 2 5.38 • satisfactory fit of magnetic data to theory for octahedral geometry • pseudo-octahedral geometry • k value of 0.7 suggests electron derealization • A value of 1.5 is consistent with weak field approximation 182 Table 5.15 Magnetic behaviours of the MfSbFaj2 compounds (M = Co, Ni, Cu and Pd) . Complex d" S P300 (uB) Proposed ground state Notes Co[SbF 6] 2 d 7 3/2 5.28 4 T l g • satisfactory fit of magnetic data to theory for octahedral geometry • pseudo-octahedral geometry • X value of-140 cm"1 indicates electron derealization • A value of 1.5 is consistent with weak field approximation Ni[SbF 6] 2 d8 1 3.40 3 A 2 g • satisfactory fit of magnetic data to theory for octahedral geometry • characteristics of both zero field splitting and weak antiferromagnetic behaviour Cu[SbF 6] 2 d 9 1/2 2.07 2 E • weak antiferromagnetic behaviour Pd[SbF 6] 2 d 8 1 3.11 3 A 2 g • octahedral geometry • characteristics of both zero field splitting and weak antiferromagnetic behaviour • contains a small amount of fluorosulfate in addition to hexafluoroantimonate The analysis of the magnetic data for our samples of Cr[SbFe]2 was inconclusive. Attempts to fit the data to standard models failed, indicating that these samples may contain impurities such as SbF3 and its adducts with SbFs, in addition to 6SbF3.5SbFs. These results contrasted with those obtained for all of the other metals in this work, where satisfactory analyses of the magnetic data were obtained in all cases. 183 The magnetic data for the M[SbF,;]2 compounds studied (M = Mn, Fe, Co, Ni , Cu and Pd) were satisfactorily fit to models assuming octahedral (or slightly distorted octahedral) metal centres. Hence, the metals are certainly not "naked" but instead are coordinated by F" ligands in all cases. The M-F interactions are weak; this is shown by the fact that no low-spin behaviour was observed and also indicated by the results of the analysis for the Fe and Co compounds which required the parameter A to take on the weak field limit of 1.5. The question of whether the weak M-F interaction can be considered completely "ionic" has also been answered by this work. The two fitting parameters which model the degree of d-electron derealization onto the ligands are k and X. For a purely ionic interaction, k should be 1 and X should take on the free-ion value. The observation that the best-fit k for Fe[SbFe]2 is 0.7 and that X for Co[SbF6]2 is significantly less than the free ion value suggests that there is measurable covalency in the M - L bonds in these materials. Moreover, the presence of detectable antiferromagnetic coupling in the compounds of Mn, Ni, Cu and Pd supports the notion of covalency in the metal-ligand bonds, since such interactions most likely involve superexchange via the electrons in the bonds connecting the metals and ligands. The pathway of interactions through the F atom between two neighbouring metal centres is possible for these M[SbF 6 ] 2 complexes. The Ag-F distances in Ag[SbF 6] 2 are 2.132(4), 2.095(5) and 2.431(3) A . 3 These distances are comparable to the Mn-F distances of MruS (2.10 and 2.13 A), the Fe-F distances of FeF 2 (1.99 and 2.12 A) and the Co-F distances of CoF 2 (2.03 and 2.05 A ) . 3 7 These MF2 (M = Mn, Co and Fe) complexes have somewhat stronger antiferromagnetic interaction (zJ's at the order of 10), propagated via the "bridging" F " . 3 8 - 4 1 If the M-F distances in the M[SbFe]2 complexes (M = Mn, Ni, Cu and Pd) are assumed to be similar to those in Ag[SbFe]2, 184 then the propagation of weak antiferromagnetic interactions via the interionic contacts between F and neighbour metal centres in the these complexes is not surprising. Finally, attention should be drawn to the fact that, in general, anomalous variable temperature magnetic behaviours reported in the literature for the compounds Ni[SbF6]2, 1' 2' 2 6 Cu[SbF6]2 1 , 2 6 and Pd[SbFe]2 1 , 2 6 have not been supported in the present study. On the other hand, the previously reported ferromagnetic behaviour of Pd(S03F)2 2 6 has been confirmed here. 5.5 Summary and Conclusions The magnetic behaviours of the M[SbF6]2 compounds (M = Mn, Fe, Co, Ni, Cu and Pd), in general, are as predicted for metal ions in weak ligand fields of octahedral or pseudo-octahedral geometry. In some cases (M = Mn, Ni, Cu and Pd), clear evidence of weak antiferromagnetic coupling between metal centres was obtained. The experimental room temperature magnetic moments of the M[SbF6]2 compounds (M = Cr, Mn, Fe, Co, Ni, Cu and Pd) are much closer to ps, than P L + S or pj. This is an indication of quenching of orbital angular momentum by ligand fields, and is the clearest indication that the metals in these systems cannot be modeled as "free" or "naked" metal ions. The experimental P300 values of the M[SbF6]2 complexes (M = Cr, Mn, Fe, Co, Ni, Cu and Pd) fall into the accepted range for high spin octahedral (or pseudo-octahedral) complexes of the respective metals. The magnetic moments are also comparable to those published for complexes with weak field ligands. However, anomalous variable temperature magnetic behaviours reported in the literature for the compounds Ni[SbFe]2,1'2>26 Cu[SbF6]2 1 , 2 6 and Pd[SbF6J2 1 , 2 6 have not been supported in the current study. 185 5.6 References 1) Cader, M . S. R.; Aubke, F. Can. J. Chem. 1989, 6*7,.1700. 2) Dean, P. A. W. J. Fluorine Chem. 1975, 5, 499. 3) Gantar, D.; Leban, I.; Frlec, B.; Holloway, J. H . J. Chem. Soc. Dalton Trans. 1987, 1987, 2379. 4) Miiller, M . ; Miiller, B. G.; Abstract of 11th European Symposium on Fluorine Chemistry. Bled, Slovenia, September 17 - 22, 1995; pp 127. 5) Mazej, Z.; Benkic, P.; Lutar, K.; Zemva, B. Reactions between binary fluorides of manganese and Lewis acids in HF; Glavic, P. and Brodnjak-Voncina, D., Ed.: Univerza v Mariboru, Fakulteta za Kemijo in Kemijsko Tehnologijo: Maribor, Slovenia, 2000, pp 187. 6) Bernhardt, E.; Bley, B.; Wartchow, R.; Willner, H . ; Bill, E.; Kuhn, P.; Sham, I. H . T.; Bodenbinder, M . ; Brochler, R.; Aubke, ¥.J. Am. Chem. Soc. 1999, 121, 7188. 7) Willner, H. ; Aubke, F. Angew. Chem. Int. Ed. Engl. 2002, submitted. 8) Brochler, R.; Sham, I. H. T.; Bodenbinder, M . ; Schmitz, V.; Rettig, S. J.; Trotter, J.; Willner, H. ; Aubke, F. Inorg. Chem. 2000, 39, 2172. 9) Larson, E. M . ; Abney, K. D.; Larson, A. C ; Eller, P. G.; Acta Cryst. 1991, B47, 206. 10) Aynsley, E. E.; Peacock, R. D.; Robinson, P. L. Chem. Ind. 1951, 1117. 11) Lee, K. C ; Aubke, F. Can. J. Chem. 1977, 55, 2473. 12) Lee, K. C ; Aubke, F. Can . J. Chem. 1979, 57, 2058. 13) Leung, P. C ; Aubke, F. Inorg. Chem. 1978, 17, 1765. 14) Birchall, T.; Dean, P. A. W.; Gillespie, R. J. J. Chem. Soc. A. 1971, 1777. 15) Sorum, C. H. Introduction to Semimicro Qualitative Analysis; Prentice-Hall: New York, 1953. 186 16) Christe, K. O.; Wilson, W. W.; Bougon, R. A.; Charpin, P. J. Fluorine Chem. 1987, 34, 287. 17) Nandana, W. A. S.; Passmore, J.; White, P. S. J. J. Chem. Soc. Dalton. Trans. 1985, 1623. 18) Alleyne, C. S.; O'Sullivan Mailer, K.; Thompson, R. C. Can. J. Chem. 1974, 52, 336. 19) Skoog, D. A.; Leary, J. J. Principles of Instrumental Analysis; 4th ed.; Saunders College Publishing, 1992. 20) Figgis, B. N . ; Hitchman, M . A. Ligand Field Theory and Its Applications; Wiley-VCH, 2000. 21) Mabbs, F. E., Machin, D. J. Magnetism and Transition Metal Complexes; Chapman and Hall: London, 1973. 22) O'Connor, C. J. Magnetochemistry - Advances in Theory and Experimentation; Lippard, S. J., 1982; Vol. 29, pp 203. 23) Weng, C. H. , Ph. D. Thesis; Carnegie-Mellon University, 1968. 24) Lines, M . E. J. Phys. Chem. Solids 1970, 37, 101. 25) Wilson, E. B. J.; Decius, J. C ; Cross, P. C. Molecular Vibrations; McGraw-Hill: New York, 1955. 26) Cader, M . S. R., Ph. D. Thesis; The University of British Columbia: Vancouver, 1992. 27) Cader, M . S. R.; Thompson, R. C ; Aubke, F. Chem. Phys. Lett. 1989, 164, 438. 28) Rao, P. R.; Sherwood, R. C ; Bartlett, N . . / . Chem. Phys. 1968, 49, 3728. 29) Rao, P. R.; Bartlett, N . Chem. Soc. Proc. 1964, 393. 30) Starr, C ; Bitter, R.; Kaufmann, A. R. Phys. Rev. 1940, 58, 977. 31) Wilkinson, M . K.; Cable, J. W.; Woolan, E. O.; Koehler, W. C. Phys. Rev. 1959, 113, 497. 32) Huheey, J. E.; Keiter, E. A.; Keiter, R. L. Inorganic Chemistry, Principles of Structure and Reactivity; HarperCollins College Publishers: New York, 1993. 33) Pearson, R. G. J. Chem. Educ. 1987, 64, 561. 187 34) Cotton, F. A.; Wilkinson, G. Advanced Inorganic Chemistry; 5th ed.; Wiley: New York, 1988. 35) Willner, H . ; Bodenbinder, M . ; Brochler, R.; Hwang, G.; Rettig, S. J.; Trotter, J.; von Ahsen, B.; Westphal, U ; Jonas, V.; Thiel, W.; Aubke, F. J. Am. Chem. Soc. 2001, 123, 588. 36) J 0 r g e n s o n , C. K. Oxidation Numbers and Oxidation States; Springer: New York, 1969. 37) Peacock, R. D.; Progress in Inorganic Chemistry; Cotton, F. A., Ed.; Interscience Publishers, Inc.: New York, 1960; Vol. 2, pp 193. 38) Homma, A. J. Phys. Soc. Japan 1960, 75, 456. 39) Lines, M . E. Phys. Rev. 1965, 737, 982. 40) Moriya, T.; Motizuki, K.; Kanamori, J.; Nagamiya, T. J. Phys. Soc. Japan 1956, 77, 211. 41) Trapp, C ; Stourt, J. W. Phys. Rev. Letters 1963, 10, 157. 188 Chapter 6 Addition Reactions of the Metal(II) Hexafluoroantimonates(V) 6.1 In t roduct ion Addition reactions of the metal(II) hexafluoroantimonates(V), M[SbF6J2 (M = Cr, Mn, Fe, Co, Ni , Zn, Pd and Ag) with potentially bridging ligands in solution were studied and the results are presented in this chapter. The objective of this study was to generate extended one, two or three-dimensional cationic lattices. The study of the magnetic properties of such systems is a topic of considerable current interest, due to the potential for magnetic exchange interactions between metal ions propagated via the bridging systems. Thus, the new solid-state materials produced through this chemistry were investigated with special attention to their structure and magnetic properties. Potentially bridging ligands, pyrazine (1,4-C4F£4N2, denoted by pyz) and 4,4'-bipyridyl (4,4'-C8H10N2, denoted by 4,4'-bipy), were chosen for this work (see Figure 1.5 in Chapter 1). These ligands are good candidates for molecular building blocks due to their rod-like rigidity and length. Incomplete substitution leading to mixed ligand complexes was not a problem in the case of ligand addition to the M[SbF6J2 (M = Cr, Mn, Fe, Co, Ni , Zn, Pd and Ag) complexes. This is because the solvents used for these reactions, sulfur dioxide, ethanol and dichloromethane, are only very weakly coordinating and unlikely to coordinate to the metals in the presence of excess pyz and 4,4'-bipy. Furthermore, [SbFe]* is a very weak nucleophile and should not be in competition with the N , N ' donor ligands for coordination sites at the metal centre. 189 There is previous work in this area that will be briefly reviewed. Carlucci et al. reported a diamagnetic silver(I)-pyz compound (d1 0) prepared from Ag[SbF 6], formulated as [Ag(pyz)3][SbF6].1 The structure of [Ag(pyz)3][SbFe] is a simple cubic framework of octahedral A g + ions, surrounded by six bridging pyz ligands.1 The purpose of the present work was to produce this type of complex with paramagnetic metal centres. The reactions were expected to proceed as follows: M[SbF 6 ] 2 + nL -» [ML n][SbF 6] 2 (eq. 6.1) The systems investigated involved L = pyz for M = Cr, Mn, Fe, Co, Ni , Cu, Pd, Ag and Zn; and L = 4,4'-bipy for M = Cr, Mn, Fe, Co, Ni and Pd. These reactions were carried out by mixing the reagents in sulfur dioxide, or by layering an ethanol solution of the M[SbFg]2 compounds over a dichloromethane solution of the ligand. Mixtures of products were obtained in all cases. The composition of the isolated solid products (powder or crystalline materials) was investigated by microanalysis. The compounds were also characterized by vibrational spectroscopy and magnetic susceptibility measurements. 6.2 Experimental 6.2.1 Reaction of the M[SbF 6 ] 2 complexes ( M = Cr , M n , Fe, Co, Ni , Zn, Pd and Ag) with py z /. Reactions in SO2 In a typical experiment, 0.8 g (10 mmol) of white solid pyz was added to 0.3 - 0.5 g M[SbFe]2 (M = Cr, Fe, Zn, Pd and Ag) or 0.5 g M[SbF 6 ] 2 -yS0 2 (M = Mn or Ni) powder in a 100 mL one-part reactor under a dry nitrogen stream. Upon addition of pyz, some of the mixtures immediately changed colour (see Table 6.1 for a summary of the amount of reagents, the products collected, and the colour changes). The colour became more intense as the mixture 190 was left to sit for a few minutes. Excess solvent, 10 mL SO2, was then condensed onto the mixture, which was left to react for one week at room temperature. Upon removal of all volatiles in vacuo, the products were collected as powders. This powder was placed under dynamic vacuum to remove any excess pyz until a constant weight was obtained. Table 6.1 Experimental findings for reactions involving the M[SbF(,]2 (M = Cr, Mn, Fe, Ni, Zn, Pd and Ag) compounds and pyz in SO2. Amount of starting reagent / colour Colour of mixture Amount and colour of product 0.5 g (0.95 mmol) Cr[SbF 6] 2 / white no immediate colour change 0.5 g, pink powder 0.5 g (~1 mmol) Mn[SbF 6] 2-yS0 2 / white no immediate colour change 0.1 g, white powder 0.38 g (0.72 mmol) Fe[SbF 6] 2 / white yellow 0.48 g, light orange and dark orange powder 0.5 g (~1 mmol) Ni[SbF 6] 2-yS0 2 / light greyish green no immediate colour change observed 1.1 g, light grey powder 0.40 g (0.75 mmol) Zn[SbF 6] 2 / white or very light grey no immediate colour change observed 0.42 g, light grey powder 0.34 g (0.59 mmol) Pd[SbF 6] 2 / light purple brown 0.41 g, beige powder 0.39 g (0.67 mmol) Ag[SbF 6] 2 / white light brown 0.54 g, light brown powder In the experiments described, the products were collected upon removal of volatiles. Since pyz is volatile, any excess ligand in the reaction was expected to be removed in vacuo. It was suspected, however, that the powders obtained consisted of compound(s) that were soluble in S 0 2 in addition to the precipitates, since the volatiles are removed from the reaction mixture (precipitates in solutions) in vacuo. Therefore, the reactions for selected metals (Cr, Co and Fe) 191 were carried out in a one-part filtration reactor (with two 50 mL round bottom compartments separated by a frit), so that the precipitates could be separated from the solution before volatiles were removed. This also ensured that any excess pyz adhering to the precipitate was removed. In a typical experiment, excess (0.5 g, 6 mmol) pyz was added to 0.3 g (0.6 mmol) of Cr[SbF6]2, Co[SbFe]2, or 0.5 g (~ 1 mmol) Fe[SbF 6 ]2"yS02 powder in the same compartment of the reactor under dry nitrogen. As in the experiments just described, some of the mixtures immediately changed colour upon addition of pyz (see Table 6.1). Excess solvent, 10 mL S O 2 , was then condensed onto the mixture at -196 °C. The mixture was left to stir for one week at room temperature. On standing, a precipitate was observed, in addition to a clear solution for each metal complex, as described previously. The solution was filtered into the filtrate compartment through the frit, and the volatiles were pumped off while slowly warming from -50 °C up to room temperature. A very small amount of powder was left in the filtrate compartment. The product, in the form of powder, was left in the reaction compartment. Liquid S 0 2 was again condensed into the reaction compartment to wash off excess pyz; the mixture was again stirred and then filtered. Upon removal of volatiles, a powder remained in the reaction compartment; a smaller amount of powder was found in the filtrate compartment. The colours of the reagents, the precipitate, the filtrate and the powder collected from the filtrate are listed in Table 6.2. Table 6.2 Observations for reactions of CrfSbFej2, Co[SbF6]2 and Fe[SbF(,]2 yS02 with pyz in SO2, where the precipitates were filtered off before volatiles were removed. Reagent / colour Precipitate Solution or filtrate Powder from the filtrate Cr[SbF 6] 2 / white pink light pink light pink Co[SbF 6] 2 / violet dark pink red light pink F e [ S b F 6 ] 2 - y S 0 2 / white dark orange colourless or very pale yellow yellow 192 2. Synthesis involving layering of solutions In an attempt to obtain crystalline products suitable for single crystal X-ray diffraction studies, a synthetic method employing the layering of solutions was used. In a typical reaction, a solution of about 0.20 g (approximately 0.3 - 0.4 mmol) M[SbF 6]2 (M = Mn, Fe, Co, Ni and Cu) in 4 mL dry ethanol was layered onto a solution of 0.15 g (2 mmol) of pyz in 3 mL dichloromethane in a test tube. After three days, crystalline solid(s) was (were) produced in each case. Table 6.3 summarizes the colours of the layers, the interface between the layers and crystalline solids for each reaction. The two layers were completely mixed in three weeks; during this time, the size of the crystalline solids had increased. The crystalline solids were collected by filtration and washed with dry ethanol. These solids decomposed slowly in air. Table 6.3 Experimental findings for the layering reactions ofM[SbFe]2 (M = Cr, Mn, Fe, Co, Ni and Cu) with pyz. M in M[SbF 6 ] 2 Top layer* Interface Product(s) Mn colourless cloudy white white crystalline solids and brown powder Fe pale yellow bright yellow yellow and orange microcrystals Co pink dark pink pink microcrystals Ni pale yellow dark yellow Light green microcrystals and brown powder Cu light blue blue green and light blue microcrystals, royal blue crystals *The bottom layer (a solution of pyz in dichloromethane) was colourless in each case. 193 6.2.2 Reaction of Cr[SbF6]2 or M[SbF6]2-yS02 (M = Mn, Fe, Co or Ni) with 4,4'-bipy In a typical reaction, a one-part filtration reactor (with two 50 mL round bottom compartments separated by a frit bridge) was used. Excess (0.7 g, 5 mmol) 4,4'-bipy and 0.3 g (0.6 mmol) of Cr[SbF6]2 or 0.5 g (~1 mmol) M[SbFe] 2-yS0 2 (M = Mn, Fe, Co or Ni) powder were placed into the reaction compartment of the flask in the glove box. Excess solvent, 10 mL S0 2 , was then condensed onto the mixture. At room temperature, precipitation was observed from each solution. The mixture was left to react for five days. The solution with excess 4,4'-bipy dissolved was removed from the precipitate by filtering through the frit into the filtrate compartment. To wash away any excess 4,4'-bipy adhering to the precipitate, the S 0 2 liquid was condensed back into the reaction compartment after the filtration. After the condensation, the filtrate compartment contained a white solid, presumably 4,4'-bipy. The mixture of the precipitate and S 0 2 in the reaction compartment was stirred and warmed to room temperature. Afterwards, the colourless solution was filtered into the filtrate compartment again. This washing procedure was repeated 10 times. Afterwards, the volatiles were slowly removed in vacuo while warming the product from -50 °C to room temperature. The product was collected in the reaction compartment as powder. For all metals, the product was difficult to handle because of a build up of static charge. Table 6.4 lists the details of these experiments. Table 6.4 Observations for reactions of Cr[SbFe]2, and MfSbFtJ^ySOi (M = Mn, Fe, Co and Ni) with 4,4 -bipy in SO2-Reagent / colour Precipitate Solution or filtrate Powder from filtrate Cr[SbF 6] 2 / white greyish brown pale brown greyish brown Co[SbF 6] 2 / violet dark pink red light pink Mn[SbF 6] 2-yS0 2 / white white colourless white Fe[SbF 6] 2-yS0 2 / white bright yellow colourless bright yellow Ni[SbF6]2*yS02 / greyish green light green colourless light green 194 6.2.3 Study of by-products formed in reactions of M[SbF6]2 complexes with donor ligands pyz or 4,4'-bipy Since results of microanalysis showed that the products isolated were largely non-stoichiometric (see Section 6.3.1), it was suspected that the following side reactions had taken place: L + M[SbF 6 ] 2 -> L.mSbFj + M[Fm(SbF6)2.m] (eq. 6.2) M[F m(SbF 6) 2 . m] + nL -> [ML n][Fm(SbF 6) 2 .m] (eq. 6.3) (where L = pyz or 4,4'-bipy) To investigate the above hypothesis, the following test reactions were carried out: 1. Reaction of SbF5 with pyz or 4,4'-bipy 2. Reaction of K[SbFe] with pyz or 4,4'-bipy 3. Reaction of M F 2 with pyz (M = Fe and Ni) /. Reaction of SbFs with pyz or 4,4 -bipy In a typical experiment, a one-part filtration vessel with two 50 mL round bottom compartments was used. About 0.5 g pyz (6 mmol) or 4,4'-bipy (3 mmol) was loaded into one of the compartments in the vessel under dry nitrogen. SbFs (1.5 mL, 21 mmol) and 10 mL SO2 were then condensed onto the mixture at -196 °C. The mixture was left to stir overnight at room temperature. For the pyz reaction with SbFs, a white precipitate was observed in a clear, colourless solution. The solution collected in the filtrate compartment. The SO2 liquid was condensed into the reaction compartment to wash off any excess pyz. The mixture was again stirred and filtered, and then the volatiles were pumped off while the product was allowed to warm from -50 °C to room temperature. A white precipitate (3.2 g, 98% yield), identified as pyz.2SbF5 was obtained in the reaction compartment. The product, analyzed as pyz.2SbFs, was found to be thermally stable up to 200 °C when heated in a melting point capillary. Single 195 crystals, suitable for single crystal X-ray diffraction, were obtained by sublimation in vacuo at 200 °C. Crystals were collected on the cold finger and also in the colder walls of the sublimation apparatus. Vibrational spectra obtained on the crude product, the sublimed crystals and the yellowish residue were found to be identical. A small amount of another white powder, of limited thermal stability, was isolated from the filtrate and identified, by vibrational spectroscopy, as SCh.SbFs. 2' 3 For the 4,4'-bipy reaction with SbFs, no precipitation was observed in the clear, very pale yellow solution. Upon removal of volatiles in vacuo from -50 °C slowly up to room temperature, white solids, which consisted of a small amount of needles and a fine powder, were obtained. The needles appeared to be 4,4'-bipy, therefore the product is possibly a mixture of this and an adduct of 4,4'-bipy and SbFs (see Section 6.3.2). 2. Reaction ofK[SbF6] with pyz or 4,4-bipy In a typical experiment, a one-part filtration flask with two 50 mL round bottom compartments was used. About 0.5 g pyz (6 mmol) or 4,4'-bipy (3 mmol) was added to 0.25 g (0.9 mmol) K[SbFg] in one of the compartments under dry nitrogen. Excess solvent, 10 mL SO2, was then condensed onto the mixture at -196 °C. The mixture was left to stir overnight at room temperature. For the reaction of pyz with K[SbFe], white crystalline solids were observed in a clear, very pale yellow solution; while for the reaction of 4,4'-bipy with K[SbF 6], white crystalline solids were observed in a clear, colourless solution. In both cases, the solution was filtered into the other compartment through the frit, while the white crystalline solids remained in the reaction compartment. The SO2 liquid was condensed back into the reaction compartment to wash off any excess pyz or 4,4-bipy on the white crystalline solids. Once 196 again, the mixture was stirred and subsequently filtered. Upon removal of volatiles in vacuo, the white crystalline solids, identified by vibrational spectroscopy as K[SbF 6], were collected in the reaction compartment for both reactions. For the reaction involving pyz, only a small amount of fine white powder was obtained in the filtrate compartment. For that involving 4,4'-bipy, a fine white powder along with another flaky white powder, appearing to be excess 4,4'-bipy, were obtained from the filtrate. Since [SbF6]" was thought to react with pyz (see eq. 6.2) according to: M[SbF 6 ] 2 + pyz *5 pyz.2SbF5 + M F 2 (eq. 6.4) (M = Cr, Mn, Fe, Co, Ni , Zn, Pd and Ag) Excess KF (1 g, 17 mmol) was added to the reaction in an attempt to shift the equilibrium to the left. However, the product obtained has an identical vibrational spectrum to the products obtained without added KF. 3. Reaction of MF2 (M = Fe and Ni) with pyz Neat reaction In a typical experiment, 0.5 g (6 mmol) of pyz was added to about 0.3 - 0.4 g (3 - 4 mmol) yellow FeF 2 or greenish-yellow NiF 2 powder in a 50 mL one-part reactor under a dry nitrogen stream. The mixture was then placed in an oven at 70 °C. After 24 hours, the yellow FeF 2 turned brown in colourless liquid pyz, while greenish-yellow NiF 2 became more deeply green in colourless liquid pyz. Upon removal of the excess pyz in vacuo at room temperature, the product containing Fe was collected as light brown solids while the product containing Ni was collected as bright, light green solids. 197 Reaction in SO 2 solvent In a typical experiment, 0.5 g (6 mmol) of white pyz solids was added to about 0.3 - 0.4 g (3 - 4 mmol) yellow FeF2 or greenish-yellow NiF 2 powder in a 50 mL one-part reactor under a dry nitrogen stream. Excess solvent, 10 mL SO2, was then condensed onto the mixture. At room temperature, the reaction mixture of FeF 2 and pyz consisted of reddish brown solids in a clear, very pale red solution. For the reaction involving NiF 2 and pyz, a bright green solid was formed in a colourless solution. The mixtures were left to stir at room temperature for two weeks. Upon removal of all volatiles in vacuo, the Fe product was collected as brick-red solids while the Ni product was obtained as bright, light green solids. 6.3 Charac te r i za t ion , Resul ts and Discussion The addition reactions of pyz or 4,4'-bipy to the M[SbF 6 ] 2 salts clearly yield a mixture of products. The observation of products with more than a single shade of colour was the first clue to the generation of a mixture of complexes with cations, [ML„] 2 + (M = metal(II) centre, L = pyz or 4,4'-bipy), having different numbers of ligands. This was most obvious for Fe 2 + and C u 2 + , where different shades of orange and blue solids were obtained respectively (see Table 6.3). Support for this finding also came from the results of the microanalysis discussed in Section 6.3.1, where non-integral numbers of pyz or 4,4'-bipy ligands per metal were found for most metals. 6.3.1 Elemental Analysis The results of microanalysis for carbon, hydrogen and nitrogen in precipitates from the reactions of the M[SbF6]2 compounds (M = Cr, Mn, Fe, Co, Ni, Zn, Pd and Ag) with pyz or 4,4'-bipy in S 0 2 are given in Tables 6.5 and 6.6. These results are compared to the calculated 198 values for various values of n (in the formula [ML n][SbF6]2; M = Cr, Mn, Fe, Co, Ni, Zn, Pd and Ag; L = pyz or 4,4'-bipy). Most of the obtained values are closer to those with a non-integral n than to an integral n. Also, different samples of the same reaction give slightly different results. These results indicate the presence of a mixture of complexes with different number of ligands coordinated to the metal centre. Furthermore, the ratios between C, H, and N found are sometimes a little different from those in pyz or 4,4'-bipy, suggesting some decomposition of the ligands in these complexes (see Tables 6.5 and 6.6). Table 6.5 Results of elemental analysis for products, formulated as [M(pyz)<„]'[SbF\]'2, from the reactions involving the M[SbFt]2 (M = Cr, Mn, Fe, Co, Ni, Zn, Pd and Ag) compounds and pyz in SO2-M Found / Calculated %C % H % N n C :N mole ratio Cr Found 24.73 2.16 13.60 2.0 Calculated 24.46 2.05 14.26 4.5 2.0 Mn Found (sample 1) 20.4 1.66 11.53 2.1 Calculated 20.1 1.68 11.69 3.3 2.0 Found (sample 2) 22.04 1.93 12.79 2.0 Calculated 22.34 1.87 13.03 3.9 2.0 Fe Found (sample 1) 18.72 1.50 10.50 2.1 Calculated 18.78 1.58 10.95 3.0 2.0 Found (sample 2) 22.72 1.94 12.97 2.0 Calculated 22.67 1.90 13.22 4.0 2.0 Co Found 23.72 1.94 13.30 2.1 Calculated 23.27 1.95 13.58 4.2 2.0 Ni Found 22.43 1.78 12.36 2.1 Calculated 21.88 1.84 12.76 3.8 2.0 Zn Found 17.42 1.61 9.07 2.2 Calculated 17.22 1.44 10.04 2.7 2.0 Pd Found 25.69 2.09 12.45 2.4 Calculated 23.96 2.01 13.97 4.8 2.0 Ag Found 17.68 1.41 8.42 2.5 Calculated 17.58 1.48 10.3 3.0 2.0 199 Table 6.6 The obtained and calculated C, H and N contents in precipitates from the reaction of the M[SbF(]2 compounds with 4,4 -bipy in SO2, formulated as [M(4,4 -bipy)„][SbFs]2-M Found / Calculated %C % H % N n C :N mole ratio Cr Found 32.86 2.28 7.31 5.2 Calculated 32.85 2.21 7.66 2.5 5.0 Mn Found 39.13 2.51 8.58 5.3 Calculated 38.04 2.55 8.87 3.3 5.0 Fe Found 35.94 2.16 8.00 5.2 Calculated 34.86 2.64 8.13 2.8 5.0 Co Found 35.85 2.61 8.06 5.2 Calculated 35.42 2.38 8.26 2.9 5.0 Ni Found 35.33 2.40 8.00 5.1 Calculated 35.42 2.38 5.26 2.9 5.0 There was no trapped free pyz or 4,4'-bipy contributing to the non-integral ratios, as shown by vibrational analysis (see Section 6.3.2). Any excess pyz in the reaction mixtures was removed in vacuo, while excess 4,4'-bipy was removed from the precipitates in SO2 by filtration (see Section 6.2). The product that formed between 4,4'-bipy and SbFs is soluble in liquid SO2, as is free 4,4'-bipy; therefore it was difficult to isolate and identify this by-product. It is most probably an adduct of 4,4'-bipy and SbFs of composition 4,4'-bipy.mSbF5, similar to pyz.2SbF5. However, the solubility of this adduct was also an advantage since 4,4'-bipy.mSbFs, along with excess 4,4'-bipy, could be removed from the [MLn][SbF6]2 complexes by filtration. Hence, pyz.2SbFs was present in the products of the pyz reactions and affected the results of microanalysis. The products of the 4,4'-bipy reactions contained no adduct of 4,4'-bipy and SbF5, thus the non-integral ratio of M to L found for the 4,4'-bipy compounds is clear evidence for the presence of a mixture of complexes with different number of 4,4'-bipy ligands. 200 6.3.2 Vibrational Analysis The vibrational spectra of the precipitates from the reactions involving the M[SbF6]2 (M = Cr, Mn, Fe, Ni and Co) compounds and pyz were compared to the those of pyz.2SbFs and pyz alone. The analysis was, unfortunately, not very informative. Several bands were assigned to a symmetrically bridged pyz moiety for each of the precipitates. These bands might be consistent with the presence of pyz.2SbFs in the products. However, the pyz bands of pyz.2SbF5 could not be distinguished from those for a symmetrically bridged metal-pyz complex, such as M-pyz-M. Furthermore, the strong Sb-F bands of pyz.2SbFs, at around 650 - 700 cm"1, fall in a cluttered region where the Sb-F bands of the [SbFs]" are, so they were not very useful in terms of identification of pyz.2SbFs in the product. Although pyz.2SbF5 sublimes at 200 °C under vacuum, it could not be removed from the metal-pyz complexes by sublimation since the complexes are only thermally stable to approximately 30 °C. On the other hand, the vibrational analysis seemed to indicate the presence of little or no excess pyz in the products. The precipitates from the reactions involving the M[SbFg]2 salts (M = Cr, Mn, Fe, Ni and Co) and 4,4'-bipy should not contain excess 4,4'-bipy or the product from the reaction of 4,4'-bipy and SbF5. These compounds are soluble in SO2 and would be removed from the precipitates during filtration. There were near coincidences of the vibrational bands of the precipitates and those of free 4,4'-bipy, indicating that the coordinated 4,4-bipy ligands in the metal complexes have similar band positions to those of free 4,4'-bipy. There were also a large number of bands for the precipitates, which suggested that there were complexes with different numbers of 4,4'-bipy ligands, [M(4,4'-bipy)n]X2 (X = F" and / or [SbF6]") in the products. 201 The product from the reaction of 4,4'-bipy and SbFs consisted of crystalline needles and a fine powder (see Section 1 of 6.2.3). The needles appeared to be 4,4'-bipy while the fine powder might be an adduct of 4,4'-bipy and SbFs. There was, however, no clear evidence from the vibrational analysis that the product contained free 4,4'-bipy. Results of elemental analysis for this product (C = 10.43%, H = 0.76% and N = 2.26%) indicated an approximately 1 to 4 ratio of 4,4'-bipy to SbF5 (calculated for CioH8N2Sb4F2o: C = 11.74%, H = 0.79% and N = 2.74%). Although the vibrational spectra of the product did not indicate the presence of free SbF 5 , 4 it was possible that the product was a physical mixture of SbFs and an adduct of 4,4'-bipy and SbF5. To investigate whether pyz.2SbFs could be obtained, not just from the reaction of pyz and SbFs but also from [SbFc;]' and pyz, the IR spectra of the products of reactions between K[SbFe] and pyz in SO2 were examined in detail (see Table 6.7). The observation of bands at 1080 and 1169 cm"1 is consistent with the presence of pyz.2SbF5, which has intense bands at 1085 and 1170 cm"1. These bands appeared as weak bands in the product, indicating that there is only a small amount of pyz.2SbF5 in the product. Table 6.8 shows the same type of comparison for the 4,4'-bipy reactions. It is reasonable to assume that the reaction of 4,4'-bipy with SbFs also forms an adduct, as most bands observed in the product are shifted in comparison to the free 4,4'-bipy bands (see Table 6.8). Some of these bands (1292, 1516 and 1640 cm"1) are also present, in weaker intensities, in the product for the reaction of K[SbFe] and pyz. Therefore, a small amount of the "adduct" can also be obtained from the reaction of [SbF6]" and 4,4'-bipy in SO2. 202 Table 6.7 IR data*: Comparison of the precipitates from reaction of K[SbF6]with pyz in SO 2 with K[SbF6], pyz5 andpyz.2SbFs (500 - 1500 cm'1). Sample K[SbF 6 ] + pyz K[SbF 6 ] p y z 5 pyz.2SbF 5 IR (cm 1) 500 (s) 594 (m) 635 (m) 668 (vs) 668 (vs) 663 (vs) 684 (sh,vs) 683 (vs) 771 (w) 790 831 (m) 975 (vw) 988 (w) 1019 1037 (vw) 1080 (w) 1085 (s) 1065 1117 (s) 1128 1169 (w) 1170 (s) 1394 (w) 1441(m) 1339 1411 1485 *: Intensities are in parentheses: v = very, s = strong, m = medium, w = weak, sh = shoulder. Since our studies indicated that the reactions involving [SbFe]" and pyz produce some pyz.2SbF5, some F" should also be present in the SO2 solution. This could generate complexes of the form M[Fm(SbF6)2-m] (see eq. 6.2, 6.3 and 6.4). In the extreme case where m = 2, the complexes would be metal(II) fluorides, M F 2 . The vibrational spectra of the products obtained from the reactions of M F 2 (M = Ni or Fe) with pyz were compared to those of the reactions involving M[SbF6]2.yS02 (M = Ni or Fe) and pyz. There were several bands for the products of the M[SbF6]2.yS02 reaction which were also observed for the M F 2 products. These bands were assigned to the coordinated pyz vibrations for complex(es) of the type [M(pyz)„]X2 (M = Ni or Fe, X = F" and / or [SbFe]"). The intensities of these bands indicated that the(se) [M(pyz)„]X2 complex(es) was / were obtained in larger amount for the reaction involving M[SbFe]2.yS02 than those involving MF2. Table 6.8 Raman data*: Comparison of the precipitates from reaction of K[SbF6]with 4,4 -bipy in SO 2 with KfSbFsJ, 4,4 -bipy and the product from reaction of 4,4 '-bipy and SbF5 in SO 2 (200 -1650 cm'). Sample K[SbF6l + 4,4'-bipy K[SbF6] 4,4'-bipy SbF5 + 4,4' -bipy Raman (cm1) 262 (vw) 315 (w) 324 (w) 386 (w) 276 (m) 292 (m) 251 (w) 315 (m) 379 (w) 230 (w) 294 (w) 389 (w) 572 (w) 576 (mw) 571 (mw) 648 (m) 659 (m) 658 (vs) 659 (s) 684 (m) 755 (m) 752 (m) 790 (m) 853 (vw) 857 (w) 861 (w) 884 (vw) 999 (vs) 1020 (vs) 1032 (w) 1067(w) 1086 (w) 1084 (m) 1217 (m) 1218 (vs) 1223(w) 1297 (vs) 1300 (w) 1292 (vs) 1510 (m) 1510 (m) 1516 (s) 1595 (ms) 1606 (s) 1603 (s) 1620 (s) 1647 (vw) 1640 (vs) *: Intensities are given in parentheses: v = very, s = strong, m = medium, w = weak. 204 In summary, the vibrational analysis of the precipitates obtained from the reactions involving the M[SbF 6]2 salts (M = Cr, Mn, Fe, Co and Ni) were not very useful in terms of identification of the products. Results of the analysis only showed that the products consist of a mixture of [ M L n ] X 2 (L = pyz or 4,4'-bipy; X = F" and / or [SbFe]"), with different number of ligands, n, for each metal. There also seemed to be some disintegration of the anion, [SbFe]", to form an adduct with the ligands. 6.3.3 Magnetic Susceptibility Measurements Magnetic susceptibility measurements were performed on the precipitates from the reaction of the metal(II) hexafluoroantiinonates(V) with pyz or 4,4'-bipy in S 0 2 over the temperature range 2 - 300 K. For analysis of the magnetic data, the products are formulated as [ML n][SbF6]2 (where M= Cr, Mn, Fe, Co or Ni ; L = pyz or 4,4'-bipy), with n determined from results of elemental analysis (see Section 6.3.1). Results of the preliminary analysis showed that all of the complexes are paramagnetic, and the measured moments at 300 K (u.300) are generally consistent with high spin configurations. A high spin configuration for these metal-pyz or metal-4,4'-bipy complexes is not surprising, because the ligands are considered weak field.6 There was, however, no indication of strong interactions between the metal centres in these complexes. A sample of these results is shown for the Co and Ni complexes. Magnetic moment versus temperature plots for these products are shown in Figure 6.1. A comparison of the experimental u.300 for the [MLn][SbF6]2 (M = Co or Ni , L = pyz or 4,4'-bipy) complexes with corresponding spin-only moments (us) and experimental u.3oo values for typical metal coordination complexes7 is given in Table 6.9. The experimental u.300 values of the complexes are generally in the range expected for these metal systems (see Table 6.9). This 205 provides some support for the average stoichiometrics proposed for these complexes. Since the exact compositions of these [MLn][SbF6]2 complexes are unknown, further study on the magnetic behaviour of these products cannot be carried out at this time. 5.0 4.5 O „ v v O v V O v . v 4.0 '3.5 I c CD E o o aj 3.0 a> co 2.5 O V £ L A A A A < A i A A A A A A A A A A A A A A A A A A A ooooooooooooooooooooooooo 2.0 1.5 v [Co(4,4'-bipy) 2 9][SbF 6] 2 O [Ni(4,4 1-bipy) 2 9][SbF 6] 2 O [Co(pyz) 4 2 ] [SbF 6 ] 2 A [Ni(pyz) 3 8][SbF 6] 2 50 100 150 200 Temperature (K) 250 300 350 Figure 6.1 The experimental magnetic moment for products obtained from the reactions involving MfSbFa]2 complexes and pyz in SO2. 206 Table 6.9 Comparison of experimental fj.300 obtainedfor [ML„][SbF(]2 compounds with the range of p.300 for typical coordination metal complexes and jus. M / Magnetic Moment (U.B) Experimental u .300 (M-B) Range of p .300 for other high spin complexes of the metal7 *7 (M-B ) L = pyz(n) L = 4,4-bipy (n) Co(II) 4.92 (4.2) 4.77 (2.9) 4.3-5.3 3.9 Ni(II) 3.13 (3.8) 2.97 (2.9) 2.9-3.9 2.8 *: u.s calculated for high spin configurations. 6.3.4 Other Characterization Attempts Crystals of royal blue [Cu(pyz)4][SbF6]2 were produced from layering an ethanol solution of Cu[SbFe]2 over a dichloromethane solution of pyz. These crystals were physically separated from other green and blue crystalline solids with forceps. The formulation of the complex was determined from preliminary results of X-ray diffraction studies. Unfortunately, the crystals are twinned and the structure of the complex had not been solved at the time this thesis was prepared. In summary, the precipitates from the reactions of the M[SbFe]2 salts with pyz or 4,4'-bipy were found to consist of mixtures of [MLn][Fn i(SbF6)2-m] complexes (L = pyz, 4,4'-bipy,) with different numbers, n, of ligands. Results of vibrational analyses were also consistent with the formation of a small amount of pyz.2SbFs in the reactions involving pyz. 207 6.4 Isolation and Characterization of pyz.2SbF5 (C4H4N2.2SDF5) 6.4.1 Synthesis and Characterization For the experimental details on the preparation and characterization of pyz.2SbFs, see Section 1 of 6.2.3). 1. Molecular Structure Determination Results of single crystal x-ray diffraction studies show that this complex of pyz and SbF5 is a 1:2 adduct of composition l,4-C4H4N2.2SbF5 (see Figure 6.2). The crystal is orthorhombic and the space group is Pbca (for a list of structural parameters, see Appendix 1). The Sb-F4(eq) planes are eclipsed and the plane of the pyz ring forms a 45° angle to the trans-Feq-Sb-Feq vector (see Figure 6.2). The point group of pyz.2SbF5 is D2i,. For the pyz moiety, all bond angles are within error limits of 120° (see Table 6.11 as discussed in the next section and Appendix 1). The pyz ring in pyz.2SbFs is a near-perfect hexagon, only slightly elongated along C 2 h on account of the longer C - C bond lengths (1.383(8) vs. 1.340(7) A for C - N distances). H2* HI F5* H2 F5 HI * Figure 6.2 The molecular structure of pyz.2SbFs. 208 2. Vibrational Analysis The vibrational data of pyz.2SbFs are shown in Table 6.10. It can be seen the IR and Raman bands of the pyz moiety are mutually exclusive. This is consistent with the centrosymmetric geometry of pyz.2SbFs (see Figure 6.2). Table 6. 10 Vibrational data* of pyz. 2 SbFs. IR (cm 1) Raman (cm"1) 204 (s) 263 (w) 283 (w) 294 (w) 500 (s) 594 (m) 592 (w) 630 (vs) 635 (m) 663 (vs) 683 (vs) 680 (s) 691 (w) 715 (w) 831 (m) 988 (w) 1047 (m) 1085 (s) 1117 (s) 1170 (s) 1244(w) 1394 (w) 1441 (m) 1525(m) 1639 (w) *: Intensities are given in parentheses: v = very, s = strong, m = medium, w = weak. 3. Microanalysis Microanalysis of pyz.2SbFs powder gave carbon, hydrogen and nitrogen content results of C: 9.38%, H: 0.76%o, N : 5.14%. These results are consistent with the calculated contents for C 4 H4N2F 1 0 Sb2 of C: 9.35%, H: 0.79% and N : 5.45%. 209 6.4.2 Discussion /. Synthesis The formation of pyz.2SbFs in SO2 as solvent appears to be a simple addition reaction according to: pyz + 2SbF5 2 5 ° C ' S ° 2 pyz.2SbF5 (eq. 6.5) However, a displacement reaction is more likely since SO2 also forms a thermally unstable 1:1 adduct with SbFs. 2 ' 3 The weaker Lewis base, SO2, is likely to be replaced by a stronger one, pyz, which results in a 2:1 adduct. This adduct precipitates from solution and is isolated by filtration. Small amounts of SC^.SbFs 2>3 are obtained from the filtrate by evaporation of SO2. Thus, the adduct SC^.SbFs is both an intermediate and a by-product, because a slight excess of SbFs above the stoichiometric amount is employed in the formation reaction (see Section 6.4.1). Sublimation of the crude pyz.2SbFs in vacuo at about 200 °C produces a crystalline product, suitable for single crystal x-ray diffraction study, and ensures that any traces of S02.SbFs are removed on account of its limited thermal stability.3 The use of SO2 as solvent in the formation of Lewis acid-base adducts is unusual. These adducts are usually obtained in aprotic, weakly coordinating solvents.8"18 The reason for using SO2 in this case goes back to initial attempts to form metal-pyz complexes with weakly coordinating anions by the reaction of metal(II) hexafluoroantimonates(V) with pyz in liquid SO2 as discussed in the previous sections. These reactions are unexpectedly complex. Observations suggest the possibility of partial extrusion of SbFs from the [SbFg]" anion either by SO2 or pyz to form MF2 and probably a previously unknown adduct of pyz and SbFs. There is a remote precedent for this process in the extrusion of SbFs from [M(CO)6][Sb 2Fii] 2 (M = Fe, Ru, Os) by repeated washing with HF, HSO3F or SO2 at 25 °C to give the corresponding [SbFe]" (see 210 Chapter 3 for M = Fe) . 1 9 In order to illustrate the reaction and to identify the adduct of pyz with SbFs, the reaction of both in SO2 was undertaken. 2. Comparison of structural parameters ofpyz.2SbFs with calculated data and with experimental data of pyz. PCI5 A number of molecular 1:1 adducts of SbF5 with nitrogen donor ligands are known and have been studied by vibrational spectroscopy.8"11 While this pyz 1:2 adduct with SbFs appears not " to have been characterized and reported previously, a number of 1:2 adducts of pyz with various boranes are known. 1 2 " 1 6 Of these, the adduct pyz.2B(N3)3 has been structurally characterized.16 For group 15 pentahalides, only a 1:1 adduct of pyz and P C I 5 is known and has been characterized.1 7'1 8 Table 6.11 lists selected experimental and calculated structural parameters20 for pyz and some of its adducts with various Lewis acids. It can be seen from this table that the experimental structural parameters of pyz.2SbF5 are very close to the calculated parameters,20 and are similar to relevant structural parameters of pyz and pyz.PCU. It appears that coordination of a Lewis acid to one or both nitrogen atoms has very little effect on the C-N or C-C distances in the pyz ring, which reflects appreciable multiple bond character. For the adduct pyz.2SbF5, DFT calculations20 suggest a D211 ground state where the pyz ring is coplanar with two SbF2 groups on either side. In contrast, the pyz ring is tilted by 45° out of this plane in the crystal structure (see Figure 6.2). This gives a tilted form of D 2 i , symmetry which is calculated to lie +14.7 kJ mol"1 above the coplanar form of D2h state.20 Precedents for such a tilted conformation are found for the adducts pyz.PCb 1 8 and py.PFs (py = pyridine), 2 1 where dihedral angles of 44 and 41° respectively are found. It appears that crystal packing effects are stabilizing the tilted form over the calculated gas phase coplanar conformer for this 211 species. While the calculated C-C and C-N distances of the tilted and coplanar form agree well with observed distances, the compiled bond angles for the observed tilted conformer differ markedly from observations. Table 6.11 Selected experimental and calculated structural parameters for pyz and some of its adducts with various Lewis acids. 18,22,23 Compound / parameter (expt.)22*23 pyz.PCl5 (expt.)3'18 pyz.2SbF5 (expt.) pyz.2SbF5 (calc, tilted)b'20 pyz.2SbFs (calc, co-planar)b'20 Bond lengths (A) d C1-N1 1.334(15) 1.353(6) 1.340(7) 1.346 1.352 d C2-N1 1.334(15) 1.327(8) 1.339(7) 1.346 1.352 d C - C 1.378(15) 1.390(7) 1.383(8) 1.386 1.386 Bond angles (°) C-C-Ni 122.4(10) 120.2(5) 119.8(5) 118.32 119.38 C - C - N 2 122.4(10) 123.0(6) 119.8(5) 118.32 119.38 115.4(10) 116.8(5) 120.3(5) 123.35 121.28 Dihedral angle c 44.0 45.0 45.0 0 Abbreviations: expt.^ experimental value, calc. = calculated value a: nitrogen atom, N i , coordinated to P C I 5 1 8 b: calculated (DFT-B3LYP) data of pyz.2SbFs of tilted or coplanar form with D 2 h symmetry20 c: the dihedral angle between the plane of the pyz ring and the F eq-Sb-F e q vector for trans-equatorial fluorines 212 3. Comparison of structure parameters ofpyz.2SbF5 and [Sb2Fuf The [Sb2Fn]" anion in the metal carbonyl salts [Pvh(CO)4][Sb2Fn] 2 4 and [Au(CO) 2][Sb 2Fi i ] , 2 5 ' 2 6 with point group D 4h, is very similar to pyz.2SbF5 in terms of its high symmetry conformation. The metal centres in these a-carbonyl complexes are in the +1 oxidation state, and Ti-back bonding in this case is reduced to a lesser extent than for the metal centres in higher oxidation states. Therefore, significant interionic contacts between the C atom of the carbonyl ligand and the F atom of the [Sb2Fi i]" counteranion are absent, and none of the characteristic distortions (bending of the Sb-Fbridging-Sb bridge angle and rotation of the SbF 4 planes to a staggered conformation) of the [Sb2Fn]" are observed (see Chapter 1). 19,27,28 A similar anion conformation is also found for the structurally characterized compound, [H3F 2][Sb 2Fn], where the cation forms asymmetric hydrogen bonds to the axial F atoms of the anion, resulting in a chain with alternating cations and anions. 2 9 The similarity of the [Sb2Fn]~ anion in these complexes and pyz.2SbF5 is apparent by comparing Figures 6.2 and 6.3. Figure 6.3 The [Sb2Fnf anion in [Au(CO)2jfSb2Fu]P 213 In both molecules, the two SbF4 groups are eclipsed, such that pyz.2SbF5 has a D2h symmetry while [Sb2Fn]" has a D 4 h symmetry. This similarity is also reflected in the structural parameters of both complexes. The averaged bond parameters for the [Sb2Fn]" anion of [H3F 2][Sb 2Fn] 2 9 and the adduct pyz.2SbFs are listed in Table 6.12, where they are compared to calculated data.2 (For a list of all structural parameters of pyz.2SbFs, see Appendix 1.) Table 6.12 Experimental and calculated bond parameters for the anion [Sb2Fi/f (point group D4^29 and the adduct pyz.2SbF$ (tilted or coplanar form). Bond Parameters [SbzFn expt.3'29 ]" (D4h) calc.20 pyz.2SbF5 (D2h) expt.b calc.c'20 Bond lengths (A) d Sb-F eq avg. 1.848(6) 1.893 1.864(3) 1.878 d Sb-F a x . 1.905(6) 1.895 1.861(4) 1.901 d Sb-F b r . 2.012(2) 2.036 d Sb-N 2.172(5) 2.171) Bond angles (°) Fax." Sb-Feq. avg. 93.13(2) 93.45 92.8(2) 98.35) Feq."Sb-Feq. avg. 89.85 89.78 89.9(2) 88.79) Feq.-Sb-Fbr. avg. 86.87(2) 86.46 Feq-Sb-N 87.2(2) 81.65) Abbreviations: eq.= equatorial, avg. = average, ax. = axial and br. = bridging a : experimental data for [H3F2][Sb2Fn] 2 9 b : experimental data for the tilted form of pyz.2SbFs c: calculated values for the coplanar form of pyz.2SbFs Calculations at the B3LYP level confirm the D 4i, structure as the ground state for the unperturbed [Sb2Fn]", as opposed to the disagreement between the tilted ring conformation observed and the coplanar form calculated for the pyz.2SbFs adduct, although both forms have D 2h symmetry (see the previous section, Section 2 of 6.4.2).2 0 It can be seen from Table 6.12 214 that, in addition to the eclipsed SbF 4 planes, all of the Faxjai-Sb-Fbridging-Sb-Faxiai atoms in the anion and the F a x i a i -Sb -N--N -Sb -F a x i a i atoms in the adduct lie on the principal axes, C 4 in the case of [Sb2Fn]" and C2u for the adduct. These findings are consistent with the high symmetries of the two species. The bond parameters are rather similar for both species (see Table 6.12). Experimental Sb-F distances appear to be slightly longer for pyz.2SbFs than for [SbaFi i ] " . 2 9 Calculated Sb-F distances agree quite well for the most part; only the calculated bond angles for the adduct fit poorly. As is generally found in structurally characterized [SbiFn]" salts, experimentally observed Faxiai-Sb-Fequatoriai angles are wider than 90° by about 3°. 19,27,29-31 Correspondingly for [Sb2Fn]" and pyz.2SbF5, the FeqUatoriai-Sb-Fbridging 2 9 and F eqUatoriai-Sb-N bond angles are more acute than 90° by about 3°. In essence, the equatorial F atoms appear to lean slightly towards the bridging region in the [Sb2Fn]" salts and towards the pyz nitrogens in the adduct pyz.2SbFs. 1 9 ' 2 7> 2 9" 3 1 Differences between the [Sb2Fn]" anion and the pyz.2SbFs adduct arise from the nature of the two different bridging moieties between the two SbFs pyramidal groups, the fluoride bridge in [Sb2Fn]~ and the pyrazine ring in pyz.2SbFs. The bonding in the Sb-Fbridging-Sb bridge is viewed as a 3 centre-2-e" bridge, as in tetrameric (SbF5) 4 where comparable Sb-Fbndging (1.98 — 2.06 A) and Sb-F terminal (1.79 — 1.87 A) distances are found. 3 2 The Sb-Fbndging -Sb in [Sb2Fnr provides a flexible backbone, and along with its very low nucleophilicity, makes [Sb2Fn]" a versatile anion for the stabilization of highly electrophilic cations such as the a-carbonyl cations. 1 9 ' 2 6" 3 1 In contrast, the pyz ring in pyz.2SbF5 is, on account of strong bonding between the ring atoms (see Table 6.11), a rather rigid bridging unit between both pyramidal SbFs 215 groups. It is interesting to find that, despite the differences in the nature of the bridging moieties, the two species adopt such a similar conformation. 6.5 Summary and Conclusions Addition reactions involving pyz or 4,4'-bipy and several M[SbFe]2 (M = Cr, Mn, Fe, Co, Ni and Cu) salts yielded precipitates which were found to be mixtures of products. For the pyz reactions, the product mixtures included [M(pyz)„]X2 (X = F" and / or [SbF6]~); these complexes have different numbers, n, of pyz ligands. The formation of pyz.2SbF5 in these reactions was also indicated. The precipitates from the 4,4-bipy reactions were [M(4,4'-bipy)n]X2 (X = F" and / or [SbFe]"), and the complexes also have different numbers, n, of 4,4'-bipy ligands. The presence of these components in the product mixtures is supported by vibrational analysis and microanalysis. The product mixtures (for Cr, Mn, Fe, Co and Ni) are paramagnetic. The experimental P300 values of the Co(II) and Ni(II) complexes are generally in the range expected for these metal systems. This provides some support for the average stoichiometrics proposed for these complexes. The adduct, pyz.2SbFs was prepared from pyz and SbFs in SO2 as the solvent. Single crystals were obtained by sublimation, and the structure was determined by an X-ray diffraction study. The adduct has D21, symmetry, in which both Sb-F^q) planes are eclipsed and the plane of the pyz ring is at an 45° angle to the Feq-Sb-Feq vector for trans-equatorial fluorines. The structural parameters of pyz.2SbFs were compared to those of pyz.PCb; it was found that coordination of a Lewis acid to either one or both nitrogen atoms has very little effect on the C-N or C-C distances in the pyrazine ring, which reflect appreciable multiple bond character. A very similar species to pyz.2SbF5, the D 4 h [Sb2Fn]" anion in [Rh(CO)4][Sb2Fii], [Au(CO) 2][Sb 2Fn] and 216 [H3F2KSD2F11] was also discussed for comparison. In both of these dioctahedral species, the pyz.2SbF5 adduct and the [Sb2Fn]" anion, the equatorial SbF 4 planes are eclipsed; also, all of the Faxiai-Sb-Fbridging-Sb-F a Xi a| atoms in the anion and F a x j a i -Sb-N- -N-Sb-F a X i a i atoms in the adduct lie on the principal axes (C211 and C 4 respectively). 6.6 References 1) Carlucci, L. ; Ciani, G.; Proserpio, D. M . ; Sironi, A. Angew. Chem. Int. Ed. Engl. 1995, 34, 1895. 2) Moore, J. W.; Baird, H. W.; Miller, H . B. J. Am. Chem. Soc. 1968, 90, 1358. 3) Aynsley, E. E.; Peacock, R. D.; Robinson, P. L. Chem. Ind. 1951, 1117. 4) Gantar, D.; Leban, I.; Frlec, B.; Holloway, J. H. J. Chem. Soc. Dalton Trans. 1987, 1987, 2379. 5) Billes, F.; Mikosch, H. ; Holly, S. Theochem. 1998. 6) Figgis, B. N . ; Hitchman, M . A. Ligand Field Theory and Its Applications; Wiley-VCH, 2000. 7) Mabbs, F. E., Machin, D. J. Magnetism and Transition Metal Complexes; Chapman and Hall: London, 1973. 8) Kolditz, L. ; Rehak, W. Z. Anorg. Allg. Chem. 1966, 342, 34. 9) Byler, D. M . ; Shriver, D. F. Inorg. Chem. 1973, 12, 1412. 10) Byler, D. M . ; Shriver, D. F. Inorg. Chem. 1974, 13, 2697. 11) von Ahsen, B.; Bley, B.; Proemmel, S.; Wartchow, R.; Willner, H. ; Aubke, F. Z. Anorg. Allg. Chem. 1998, 624, 1225. 12) Martin, D. R.; Merkel, C. M . ; Mondal, J. U . ; Rushing Jr., C. R. Inorg. Chim. Acta 1985, 99, 81. 2 1 7 13) Martin, D. R.; Merkel, C. M . ; Drake, C. B.; Mondal, J. U . ; Iwamoto, J. B. Inorg. Chim. Acta 1985, 97, 189. 14) Caminati, W.; Damiani, D. J. Mol. Struct. 1986, 147, 389. 15) Lever, A. B. P.; Lewis, J.; Nyholm, R. S. J. Chem. Soc. 1963, 3165. 16) Fraenk, W.; Habeveder, T.; Hammerl, A.; Klapotke, B. ; Krumm, B.; Mayer, P.; Noth, M . ; Warchhold, M . Inorg. Chem. 2001, 40, 1334. 17) Ishley, J. N . ; Knachel, H. C. Inorg. Chem. 1975,14, 2558. 18) Meyer, B. N . ; Ishley, J. N . ; Fratini, A. V.; Knachel, H . C. Inorg. Chem. 1980, 19, 2324. 19) Willner, H. ; Aubke, F. Angew. Chem. Int. Ed. Engl. 2002, submitted. 20) von Ahsen, B. ; von Ahsen, S.; Willner, H. , Personal Communication. 21) Sheldrick, W. S. J. Chem. Soc. Dalton Trans. 1974, 1402. 22) DeWith, G.; Harkema, S.; Feil, D. Acta Cryst. 1976, B22, 3178. 23) Wheatley, P. J. Acta Cryst. 1957, 10, 182. 24) von Ahsen, B. ; Bach, C ; Kockerling, M . ; Willner, H. ; Aubke, F. Inorg. Chem., manuscript in preparation. 25) von Ahsen, B. Diploma Thesis; University of Hannover 1999. 26) Willner, H. ; Schaebs, J.; Hwang, G.; Mistry, F.; Jones, R.; Trotter, J.; Aubke, F. J. Am. Chem. Soc. 1992, 114, 8972. 27) Willner, H . ; Aubke, F.; Inorganic Chemistry Highlights; Meyer, G., Naumann, D. and Wesemann, L. , Ed.; Wiley-VCH Weinheim, 2002, Chapter 11, pp 195. 28) Willner, H. ; Aubke, F. Angew. Chem. Int. Ed. Engl. 1997, 36, 2402. 29) Mootz, D.; Bartmann, K. Angew, Chem. Int. Ed. Engl. 1988, 27, 39. 30) Willner, H. ; Bodenbinder, M . ; Brochler, R.; Hwang, G.; Rettig, S. J.; Trotter, J.; von Ahsen, B. ; Westphal, U . ; Jonas, V.; Thiel, W.; Aubke, F. J. Am. Chem. Soc. 2001, 123, 588. 218 31) Bernhardt, E.; Bley, B.; Wartchow, R.; Willner, H. ; Bill, E.; Kuhn, P.; Sham, I. H . T.; Bodenbinder, M . ; Brochler, R.; Aubke, F. J. Am. Chem. Soc. 1999, 727, 7188. 32) Edwards, A. J.; Taylor, P. J. Chem. Soc. Chem. Comm. 1971, 1376. 219 Chapter 7 General Summary and Suggestions for Future Work 7.1 G e n e r a l S u m m a r y This thesis describes the synthesis and characterization of potentially paramagnetic metal carbonyl cations as -[Sb2Fn]" and -[SbFe]" salts, and related metal(II) hexafluoroantimonates(V). The attempt to produce a paramagnetic carbonyl species by the preparation of a carbonyl cation of Fe(III), [Fe(CO)6] 3 +, has been unsuccessful. Fe(II) was found to be an extremely stable oxidation state in [Fe(CO)e]2+, as shown by the resistance of this cation to undergo further oxidation. Furthermore, thermal decomposition of [Fe(CO)6][Sb2Fn]2 and [Fe(CO)6][SbF6]2 at temperatures above 200 °C gave only the Fe(II) species, Fe[SbFe]2 and FeF2. The previous suggestion of formation of [Fe(CO)6] 3 + during the oxidation of Fe(CO)s with CI2 1 was shown to be in error. The ligand field strength of CO in a-carbonyl cations was expected to be reduced, due to a reduction in 71-backbonding, compared to the situation with other carbonyl complexes. However, the current work revealed that even in a-carbonyl cations, the ligand is strong field and the cations have low spin configurations. Hence, the Fe(II) centre in [Fe(CO)e]2+, having a d 6 configuration in pseudo-octahedral geometry, is diamagnetic. The source of observed paramagnetism in the samples of "[Fe(CO)6][Sb2Fn]3" and [Fe(CO)6][Sb2Fu]2 was identified as Fe[SbFe]2. Study of the thermolysis behaviours of [Fe(CO)6][Sb2Fn]2 and [Fe(CO)6][SbFe]2 showed that decarbonylation of these complexes does not occur in distinct steps. No carbonyl product, which could possibly be a high spin species (see eq. 1.20), was isolated on decomposition. It was also shown that the paramagnetic by-product, Fe[SbF6J2, could not be obtained by the decomposition of [Fe(CO)6] 2 + at temperatures at which the syntheses were carried out. A reaction pathway for the formation of [Fe(CO)6][Sb2Fn]2 and Fe[SbFe]2 via an intermediate of 220 the form Fe(CO) 4 X 2 (X = CI or F) was postulated. Oxidation of Fe(CO) 5 by XeF 2 in HF-SbF 5 was found to be the preferred route to the synthesis of [Fe(CO)e][Sb2F] i ] 2 ; the purity of the bulk sample was improved by treatment with F 2 in HF. Seven-coordinate cationic carbonyl derivatives of Mo and W, [{Mo(CO)4}2(cis-p-F 2SbF 4) 3] x[Sb 2Fi i ] x and [W(CO)6(FSbF5)][Sb2Fi i] were prepared from the oxidation of M(CO)6 by SbFs. These seven-coordinate complexes have complicated, low symmetry geometries with possibly low lying, thermally accessible excited states or even ground states that may be paramagnetic. Variable temperature magnetic susceptibility measurements (2-300 K), however, showed that the major contribution of the paramagnetism in the Mo(II) species (d4 configuration) is TIP. It should be noted that Mo and W, being 4d and 5d metals respectively, are expected to have larger ligand field splittings than 3d metals. Therefore, an analogous Cr carbonyl species to the Mo and W species would be a more promising candidate to exhibit paramagnetism. Vibrational analysis, however, showed that the Cr carbonyl species prepared from the oxidation of Cr(CO)6 in SbFs has a centre of symmetry, instead of the low symmetry expected for a seven-coordinate complex. Although the high symmetry of this Cr carbonyl species could lead to diamagnetism, this is not necessarily true. Nonetheless, the Cr carbonyl species is thermally unstable at room temperature, which made the study of this compound very difficult. Structural characterization of this Cr species is essential for a better understanding of this chemistry. Metal(II) hexafluoroantimonates(V), M[SbFe]2, are observed as thermal decomposition products or synthetic by-products of metal rj-carbonyl cations which exist as -[Sb 2Fn]" or -[SbFe]" salts. The magnetic behaviour of the M[SbFe]2 compounds (M = Mn, Fe, Co, N i , Cu and Pd), in 221 general, are as predicted for metal ions in a weak ligand field of octahedral or pseudo-octahedral geometry. The experimental magnetic moments of the M[SbF6]2 compounds (M = Cr, Mn, Fe, Co, N i , Cu and Pd) are much closer to ps, than pL+s or pj. This is an indication of quenching of orbital angular momentum by ligand fields, and is the clearest indication that the metals in these systems cannot be modeled as "free" or "naked" metal ions. Instead, they can be viewed as coordinated by F" ligands with weak M-F interactions. Anomalous variable temperature magnetic behaviours reported in the literature for the compounds M[SbFe]2 (M = N i , Cu and Pd) have not been supported in the current study. Addition reactions involving pyz or 4,4'-bipy and several M[SbFe]2 (M = Cr, Mn, Fe, Co, N i and Cu) complexes were carried out in the hope of generating extended cationic one-, two- or three-dimensional lattices with the potential for observing magnetic exchange interactions between metal ions propagated via the bridging systems. These reactions yielded mixtures of compounds. The product consisted of mixtures of [M(pyz)n]X2 or [M(4,4'-bipy)n]X2 (X = F" and / or [SbFe]") complexes; these complexes have different numbers, n, of pyz or 4,4'-bipy ligands. The product mixtures for M = Cr, Mn, Fe, Co and N i are paramagnetic. The formation of pyz.2SbF5 was observed in reactions involving the pyz ligand. The crystal structure of pyz.2SbF5 was determined; it has D211 symmetry in which both Sb -F^q ) planes are eclipsed and the plane of the pyz ring is at an angle of 45° to the F e q -Sb-F e q vector for trans-equatorial fluorines. 7.2 Suggestions for Future Work As described in the previous section, it is possible that a cationic Cr(II) carbonyl species would be paramagnetic. Further characterization of the transient Cr carbonyl species prepared in the 222 current work (see Chapter 4) would be worthwhile. This work should include (re)crystallization of the Cr carbonyl species and further structural studies. The current work revealed that crystalline solids in the product are stable only at low temperatures in HF or HF-SbFs solutions. This work will require the separation of these solids under these low temperature conditions. The carbonyl complex of the group 5 metal V , V(CO)6, remains the only paramagnetic carbonyl species isolated to date. If one sees V(CO)6 as a fragment of a species such as (CO)eV-V(CO)6, then binuclear carbonyl species of the form M(CO)6 for the other group 5 metals, M = Nb and Ta, may well exist. Therefore, attempts to prepare homoleptic carbonyl species of Nb and Ta, with odd numbers of d electrons may be worthwhile. Addition reactions of pyz or 4,4'-bipy and several M[SbF6]2 compounds studied in the current work yielded a complex mixture of products. It would be useful to carry out the reactions utilizing different stoichiometric amounts of reagents and to control the reaction times, such that the production of a particular metal-ligand complex is favoured. Also, when reactions were carried out in S O 2 in the present work, fine powders were obtained. When reactions were performed using layering techniques with ethanol and dichloromethane, the products obtained were crystalline solids. Therefore, the use of different solvents should be explored in an attempt to obtain crystalline products suitable for single crystal X-ray diffraction studies. It should also be noted that the formation of by-products such as pyz.2SbFs was avoided when liquid S O 2 was not involved. The use of other solvents may avoid the problem of obtaining a mixture of products. Structural information about these metal-ligand complexes would be valuable for an understanding of their magnetism. It is hoped that a single crystal X-ray structural analysis of [Cu(pyz)4][SbF6J2, obtained here as twinned crystals, can be completed. Finally, since the 223 [M(pyz)„]X2 or [M(4,4'-bipy)„]X2 (M = Cr, Mn, Fe, Co, Ni and Cu; X = F" and / or [SbF6]") complexes slowly decompose in air, they are suspected to be moisture sensitive. The use of anhydrous solvents is recommended, and the reaction would be better carried out under anhydrous conditions. 7.3 References 1) Bley, B. ; Willner, H . ; Aubke, F. Inorg. Chem. 1997, 36, 158. 224 Appendix 1 Crystallographic Data for the Structure Determination of Various Complexes 1. [Fe(CO)6] [ S b 2 F „ ] 2 and [Fe(CO)6] [SbF 6] 2 1 Table Al. 1 Crystal Data of [Fe(CO)6][Sb2F,,]2 and [Fe(CO)6][SbF6]2 [Fe(CO) 6 ][Sb 2 F„] 2 [Fe(CO) 6][SbF 6] 2 Formula C 6 F 2 2 Fe0 6 Sb 4 C 6 F 1 2 Fe0 6 Sb 2 Formula weight (g / mol) 1128.9 695.4 Crystal system monoclinic tetragonal Colour of crystal colourless colourless Lattice parameters a (A) 9.751(1) 8.258(1) b(A) 12.457(1) 83258 c(A) 10.542(1) 12.471(2) P O 110.63(1) 90 V ( A 3 ) 1198.4(2) 850.5(2) z 2 2 Space group P2,/n(No. 14) P4/mnc (No. 128) Goodness of fit on F 2 0.843 1.03 Ri [I>2a(I)] 0.0282 0.0259 W R 2 0.0532 0.0668 225 Table A1.2 Bond Lengths and Angles for [Fe(CO)e] [Sb2Fi /]2 (a) Bond Distances (A) atom atom distance atom atom distance Fe C(l)(2x) . 1.911(5) C(l) 0(1) (2x) 1.103(5) Fe C(2) (2x) 1.912(5) C(2) 0(2) (2x) 1.102(5) Fe C(3) (2x) 1.910(5) C(3) 0(3) (2x) 1.107(5) Sb(l) F(l) 2.053(3) Sb(2) F(l) 1.998(3) Sb(l) F(2) 1.850(3) Sb(2) F(7) 1.839(6) Sb(l) F(3) 1.835(3) Sb(2) F(8) 1.864(3) Sb(l) F(4) 1.842(3) Sb(2) F(9) 1.847(3) Sb(l) F(5) 1.824(3) Sb(2) F(10) 1.843(3) Sb(l) F(6) 1.869(3) Sb(2) F ( l l ) 1.810(3) (b) Bond Angles (degrees) atom atom atom angle atom atom atom angle C(l) Fe C(2) (2x) 89.2(2) C(l) Fe C(2)(2x) 90.8(2) C(2) Fe C(3) (2x) 88.9(2) C(2) Fe C(3) (2x) 91.1(2) C(3) Fe C(l)(2x) 89.8(2) C(3) Fe C(l)(2x) 90.2(2) C(l) Fe C(l) 180 C(2) Fe C(2) 180 C(3) Fe C(3) 180 Fe C(l) 0(1) (2x) 178.1(4) Fe C(2) 0(2) (2x) 177.5(4) Fe C(3) 0(3) (2x) 179.4(4) F(l) Sb(l) F(3) 86.2(2) F(l) Sb(2) F(8) 84.8(1) F(l) Sb(l) F(4) 86.2(1) F(l) Sb(2) F(9) 87.2(1) 226 F(l) Sb( l ) F(5) 84.3(2) F(l) Sb(2) F(10) 87.3(2) F(l) Sb( l ) F(6) 83.3(1) F(l) Sb(2) F ( l l ) 85.7(2) F(2) Sb( l ) F(3) 96.4(2) F(7) Sb(2) F(8) 92.0(2) F(2) Sb (1) F(4) 96.6(2) F(7) Sb(2) F(9) 91.4(2) F(2) Sb( l ) F(5) 93.1(2) F(7) Sb(2) F(10) 95.7(2) F(2) Sb( l ) F(6) 94.0(2) F(7) Sb(2) F ( l l ) 95.8(2) F(3) Sb( l ) F(4) 89.7(2) F(8) Sb(2) F(9) 88.4(2) F(4) Sb (1) F(5) 92.3(2) F(9) Sb(2) F(10) 86.8(2) F(5) Sb (1) F(6) 89.4(2) F(10) Sb(2) F ( l l ) 92.4(2) F(6) Sb (1) F(3) 87.0(2) F ( l l ) Sb(2) F(8) 91.4(2) F(l) Sb( l ) F(2) 176.2(1) F(l) Sb(2) F(7) 176.6(2) F(3) Sb (1) F(5) 170.1(2) F(8) Sb(2) F(10) 171.0(2) F(4) Sb( l ) F(6) 169.2(1) F(9) Sb(2) F ( l l ) 172.9(2). Sb(l) F(l) Sb(2) 148.5(2) 227 Table A1.3 Bond Lengths and Angles for [Fe(CO) g][SbF^J'2 (a) Bond Distances (A) atom atom distance atom atom distance Fe C(l) (2x) 1.917(7) C(l) 0(1) (2x) 1.096(9) Fe C(2)(4x) 1.903(6) C(2) 0(2) (4x) 1.114(8) Sb F(l) (4x) 1.852(3) Sb F(2)(2x) 1.832(3) (b) Bond Angles (degrees) atom atom atom angle atom atom atom angle C(l) Fe C(2)(8x) 90 C(2) Fe C(2)(4x) 90 C(l) Fe C(l) 180 C(2) Fe C(2) (2x) 180 Fe C(l) 0(1) (2x) 180 Fe C(2) 0(2) (4x) 178.9(5) F(l) Sb F(l)(2x) 88.3(1) F(l) Sb F(l)(2x) 91.9(1) F(l) Sb F(2)(4x) 87.2(1) F(l) Sb F(2)(4x) 92.8(1) F(l) Sb F(l)(2x) 174.4(2) F(2) Sb F(2) 180 228 2. [{Mo(CO)4}2(cis-u.-F 2SbF 4)3] x[Sb 2F 1 I] x 2 Table A1.4 Crystal Data of[{Mo(CO)4}2(cis-fi-F2SbF4)3JxfSb2FnJx Formula C 8 F29Mo 2 0 8 Sb 5 Formula weight 1575.67 g /mol Lattice parameters a 9.234(4) A b 13.858(3) A c 25.790(3) A P 90.532(2)° V 3300.1(12) A 3 z 4 Space group P2i / c (No. 14) Temperature -93 °C ^calculated 3.171 g / c m3 X 0.71069 A 49.5 cm"1 26 4 .0-60.2° Residuals (on F, I > 3a(I)) R a , R w b = 0.048, 0.047 a : R = Z | | F 0 2 | - | F c 2 | | / Z | F 0 2 | b : R w = Z W ( | F 0 2 | - | F c 2 | ) 2 / SwF 0 4)] 229 Table Al, 5 Bond Lengths and Angles Range of [{Mo(CO)4}2(cis-F2SbF4)3]'x[Sb2F) 1)\ Group Bond length range observed (A) Group Bond angles range observed (degrees) Mo-C 2.021(10)-2.052(10) C-Mo-C 70.2(4) and 75.3(4) 111.6(4) and 113.9(4) C-0 1.089(11)- 1.136(11) Mo-C-0 174.8(9)- 178.7(9) Mo-F 2.140(5)-2.196(5) F - M o - F 74.2(2) - 77.6(2) S b - F - ( S b ) 3 2.014(6) and 2.056(5) S b - F - S b a 150.0(3) Sb-F-(Mo) 1.933(6)- 1.988(6) S b - F - M o 142.1(3)- 162.8(3) Sb-Ftermina! -1.85 ±0.02 FterminarSb-Fterminal 91 ± 1 a : Data pertain to [Sb2Fn]" b : Denotes terminal F atom in cis-SbF6" and in [ S b 2 F n ] 230 3. [ W ( C O ) 6 ( F S b F 5 ) ] [ S b 2 F u ] 3 Table A1.6 Crystal Data of [W(CO)6(FSbF5)][Sb2Fn] Formula C 6 F 1 7 0 6 S b 3 W Formula weight 1040.14 g /mol Crystal system monoclinic Lattice parameters: a 8.2051(12) A b 16.511(3) A c 8.1432 A P 111.5967(6)° V 1025.8(2) A 3 z 2 Do 3.367 g/cm3 Fooo 928.00 Space group P 2 . (No. 4) Crystal dimensions 0.35 x 0.30 x 0.25 mm 20 max 60.0° no. of reflections measured 9112 Unique (including Friedel pairs) 4410 (R int = 0.029) Residuals (on F, I > 3rj(I)): R a , R w b = 0.023, 0.023 A : R = S | | F 0 2 | - | F C 2 | b : R w = E w ( | F 0 2 | - | F c 2 | ) 2 / £ M ; F 0 4 ) ] 1 / 2 231 Table A1.7 Bond Lengths and Angles for [W(CO)6(FSbF5)][Sb2Fn] (a) Bond Distances (A) atom atom distance atom atom distance W(l) F(l) 2.109(5) W(l) C(l) 1.996(9) W(l) C(2) 2.021(7) W(l) C(3) 2.139(9) W(l) C(4) 2.150(9) W(l) C(5) 2.130(10) W(l) C(6) 2.131(9) Sb(l) F(l) 2.010(4) Sb(l) F(2) 1.868(5) Sb(l) F(3) 1.839(6) Sb(l) F(4) 1.829(6) Sb(l) F(5) 1.812(7) Sb(l) F(6) 1.801(6) Sb(2) F(7) 2.034(5) Sb(2) F(8) 1.818(6) Sb(2) F(9) 1.841(5) Sb(2) F(10) 1.840(6) Sb(2) F ( l l ) 1.846(5) Sb(2) F(12) 1.864(6) Sb(3) F(7) 2.022(5) Sb(3) F(13) 1.851(6) Sb(3) F(14) 1.848(6) Sb(3) F(15) 1.833(7) Sb(3) F(16) 1.849(6) Sb(3) F(17) 1.811(6) 0(1) C(l) 1.160(10) 0(2) C(2) 1.155(8) 0(3) C(3) 1.128(10) 0(4) C(4) 1.109(10) 0(5) C(5) 1.118(10) 0(6) C(6) 1.111(10) 232 (b) Bond Angles (degrees) atom atom atom angle atom atom atom angle F(l) W(l) C(l) 143.3(3) F(l) W(l) C(2) 147.2(3) F(l) W(l) C(3) 87.7(3) F(l) W(l) C(4) 86.5(3) F(l) W(l) C(5) 79.4(3) F(l) W(l) C(6) 82.7(3) C(l) W(l) C(2) 69.3(4) C(l) W(l) C(3) 116.4(3) C(l) W(l) C(4) 73.6(3) t C(l) W(l) C(5) 74.9(3) C(l) W(l) C(6) 115.5(3) C(2) W(l) C(3) 74.3(3) C(2) W(l) C(4) 114.8(4) C(2) W(l) C(5) 117.1(3) C(2) W(l) C(6) 79.4(4) C(3) W(l) C(4) 77.2(3) C(3) W(l) C(5) 167.1(3) C(3) W(l) C(6) 102.3(3) C(4) W(l) C(5) 101.5(3) C(4) W(l) C(6) 169.2(3) C(5) W(10) C(6) 76.5(3) 233 4 . SbF 3 Table A1.8 Crystal Data of SbFs from determination by A. J. Edwards4 and the present work. A . J . Edwards 4 (1970) The present work (2001) Formula SbF 3 Formula Weight 178.75 g /mol Crystal system Orthorhombic Lattice parameters "c" - 7.26(9) A a = 7.2153(6) A "b" =7.46(6) A b = 7.4025(6) A "a" = 4.95(6) A c = 4.9104(9) A " U " = 268 A 3 V = 262.27(12) A 3 Z = 4 Z = 4 D c = 4.39 g/cm3 D c = 4.526 g/cm3 Space group C 2 c m (no. 40) Ama2 (No. 404) Crystal Dimensions 0.25 * 0.25 * 0.20 mm Temperature - 1 0 0 ± 1 ° C 29 58.3° Residuals (on F, I > 3a(I)) R a , R w b = 0.030, 0.040 *: Reference: Edwards, A. J. J. Chem. Soc. (A), 1970, 2751. a : R = s | | F021 - I F c 2 I | / 2 | F 0 2 | b : R w = S w ( | F 0 2 | - | F C 2 | ) 2 / 2 ^ F 0 4 ) ] 1 / 2 234 5. pyrazine.2SbF5 Table Al.9 Crystal Data of pyrazine. 2SbF5 (1,4-C4H4N2.2SbF5) Formula C 4H4N2F 1 0 Sb2 Formula weight 513.57 g /mol Crystal system orthorhombic Lattice parameters a 8.8052(8) A b 9.948(1)A c 12.297(2) A V 1077.1(2) A 3 Z 4 Space group P b c a (No.61) Temperature 173(1)K H 51.3 cm- 1 26max 225.8° Refl. collected/ unique / Rjnt 9501 / 1469/0.062 Residuals R i a , w R 2 b = 0.031, 0.089 a : R i = S I | F01 - | F c | | / S | F01 b : w R 2 = [ S ( | F 0 2 | - | F c 2 | ) 2 / S w ( F 0 2 ) 2 ) ] 235 Table ALIO Bond Lengths and Angles for pyrazine.2SbFs (l,4-C4H4N2-2SbFs) (a) Bond Lengths (A) atom atom distance atom atom distance Sb(l) F(l) 1.858(4) Sb(l) N( l ) 2.172(5) Sb(l) F(2) 1.863(3) N(l ) C(l) 1.340(7) Sb(l) F(3) 1.870(3) N(l ) C(2) 1.339(7) Sb(l) F(4) 1.866(3) C(l) C(2) 1.383(8) Sb(l) F(5) 1.861(3) (b) Bond Angles (degrees) atom atom atom angle atom atom atom angle F(l) Sb(l) F(2) 88.0(2) F(3) Sb(l) F(5) 93.4(2) F(l) Sb(l) F(3) 174.1(2) F(3) Sb(l) N( l ) 86.5(2) F(l) Sb(l) F(4) 90.5(2) F(4) Sb(l) F(5) 92.7(2) F(l) Sb(l) F(5) 92.6(2) F(4) Sb(l) F(5) 87.3(2) F(l) Sb(l) N( l ) 97.6(2) F(5) Sb(l) N( l ) 179.9(2) F(2) Sb(l) F(3) 91.7(2) Sb(l) N( l ) C(l) 119.5(4) F(2) Sb(l) F(4) 174.4(1) Sb(l) N( l ) C(2) 120.3(4) F(2) Sb(l) F(5) 92.7(2) C(l) N( l ) C(2) 120.3(5) F(2) Sb(l) N( l ) 87.3(2) N(l ) C(l) C(2) 119.8(5) F(3) Sb(l) F(4) 89.3(2) N(l ) C(2) C(l) 119.9(5) 236 Table ALU Atomic coordinates and equivalent isotropic thermal parameters for pyrazine. 2SbF5 (1,4-C4H4N2.2SbF5) atom X y z Sb(l) 0.63196(4) 0.20146(4) 0.37688(3) 0.830(8) F(l) 0.8284(4) 0.2567(4) 0.4093(3) 1.77(8) F(2) 0.6237(4) 0.1235(3) 0.5145(3) 1.89(8) F(3) 0.4274(4) 0.1624(4) 0.3510(3) 1.78(7) F(4) 0.6385(4) 0.2957(3) 0.2457(3) 1.62(7) F(5) 0.6994(4) 0.0432(4) 0.03123(3) 1.89(8) N(l ) 0.5527(5) 0.3861(5) 0.4522(4) 0.81(9) C(l) 0.4415(6) 0.4566(6) 0.4046(5) 1.1(1) C(2) 0.6126(7) 0.4277(6) 0.5465(4) 1.1(1) H(l) 0.398 0.4255 0.3357 1.5216 H(2) 0.6951 0.3765 0.5803 1.3402 *: B e q = 8/3TI 2 (Un(aa*)2 + U 2 2(bb*) 2 + U 3 3(cc*) 2 + 2U12(aa*bb*)cosy + 2Ui3(aa*cc*)cosp + 2U23(bb*cc*)cosa) 237 A l . l References 1) Bernhardt, E.; Bley, B.; Wartchow, R.; Willner, H. ; B i l l , E.; Kuhn, P.; Sham, I. H . T.; Bodenbinder, M . ; Brochler, R.; Aubke, F. J. Am. Chem. Soc. 1999,121, 7188. 2) Brochler, R.; Freidank, D.; Bodenbinder, M . ; Sham, I. H. T.; Willner, H. ; Rettig, S. J. Trotter, J.; Aubke, F. Inorg. Chem. 1999, 38, 3684. 3) Brochler, R.; Sham, I. H. T.; Bodenbinder, M . ; Schmitz, V. ; Rettig, S. J.; Trotter, J.; Willner, FL; Aubke, F. Inorg. Chem. 2000, 39, 2172. 4) Edwards, A. J. J. Chem. Soc. (A) 1970, 2751. 238 Appendix 2 Experimental Details for Differential Scanning Calorimetric (DSC) Studies on [Fe(CO)6][Sb2Fn]2 and [Fe(CO)6][SbF6]2 A2.1 [Fe(CO)6][Sb2F„]2 Stainless steel, gold-plated sealed crucibles were used for all DSC studies of [Fe(CO)6][Sb2Fn]2. The samples were accurately weighed. After heating, the content of the gas phase was checked by IR spectroscopy and that of the solid residue was checked by Raman spectroscopy. The experimental details are shown in Table A2.1. Table A2.1 DSC studies on [Fe(CO)6][Sb2F,,J2. Experiment Weight of sample Temperature Sequence Heating Rate (K/min) 1 50.77 mg (0.04497 mmol) 2 5 - 2 5 0 ° C 10 2 43.54 mg (0.03857 mmol) 2 5 - 2 5 0 ° C 10 11 26.01 mg (0.02304 mmol) 2 5 - 1 3 0 - 8 0 - 190 °C. 10 From 25 - 250 °C, there are two events with onset at about 100 °C and 185 °C (with a shoulder onset at 170 °C) respectively. Experiments 1 and 2 (25 - 250 °C, see Figures A2.1 and A2.2.) Experiments 1 and 2 were performed to confirm the reproducibility of results. After the sample 1 was heated to 250 °C, the gas phase (12 mbar) expanded into a 200 mL vacuum line was found to have about 2 mbar (0.02 mmol) C 0 2 and about 6 mbar CO (0.05 mmol, 16% of CO 239 content in sample.) Other components in the gas phase included trace amounts of COF2 and CF4. nW/mg) i:Exo _ , , , , , _ _ , , , , , , , , , _ - , , , , , Figure A2. J DSC Plot of [Fe(CO)6J'[S02F11yY Experiment 1. 50 100 150 ' 200 250 Figure A2.2 DSC Plot of [Fe(CO)e][Sb2Fi/yV Experiment 2. 240 The events are assigned as described previously (see chapter 3 for justification of assignments): • The small peak with onset at about 100 °C is tentatively assigned to liberation of HF. • The shoulder with onset at about 170 °C is assigned to the decomposition of the anion: [Sb 2 F 1 1 ] -^ [SbF 6 ] - + SbF5(,) • The main event, which was the large peak with onset at about 185 °C, is assigned to the decarbonylation of [Fe(CO)6J and the disproportionation reaction of CO: 2CO(g) ^ » C0 2( g) + C(S). There was a problem with the large amount of sample used for these two experiments. About 40 - 50 mg of [Fe(C0)6][Sb 2Fn] 2 was used in each experiment and the resulting pressure inside the crucible after heating was close to 100 atm. As a result, the lid of the crucible deformed. Therefore, experiments were scaled down subsequently due to limitations of the apparatus. Experiment 11 (25 - 130 - 80 - 190 °C, see Figure A2.3) Experiment 11 was used to check the reversibility of the events between 8 0 - 130 °C. It was found that none of the events are reversible. The irreversibility of the events is consistent with the above assignments. After the sample has been heated to 190 °C, the Raman spectrum shows strong fluorescence with a broad, weak peak at 653 cm"1. The peak corresponds to the A i g Sb-F stretch of [SbFe]". This confirms that decomposition of [Sb2Fi i]" and that decarbonylation of [Fe(CO) 6] 2 + begins before reaching 190 °C. 241 A2.2 [Fe(CO)6][SbF6]2 Both stainless steel, gold-plated sealed crucibles and aluminum pans with pierced lids were used for DSC studies of [Fe(CO)6][Sb2Fn]2. The samples were accurately weighed. For the samples in the crucibles, the content of the gas phase was checked by IR and that of the solid residue was checked by Raman after heating. The experimental details are shown in Table A2.2. Due to the weight difference of the two kinds of sample holders, the A l pan runs have much higher sensitivity than the crucible runs, as evident in the DSC plots. Experiments 3 and 4 confirmed the reproducibility of the runs (Figures A2.4 and A2.5). 242 Table A2.2 DSC Studies on [Fe(CO)6][SbF6]2 Name of Experiment Sample holder Weight of sample Temperature Sequence Heating Rate (K / min) 3 A l pan 12.61 mg (0.01813 mmol) 25 - 300 °C 10 4 A l pan 10.47 mg (0.01506 mmol) 25 - 400 °C 10 5 A l pan 10.70 mg (0.01539 mmol) 2 0 - 1 1 0 - 8 0 - 170- 130-270 °C 10 6 Crucible 39.10 mg (0.05623 mmol) 30-270 °C 10 7 A l pan 8.12 mg (0.01168 mmol) 3 0 - 3 3 0 ° C 2 8 A l pan 10.35 mg (0.01488 mmol) 3 0 - 3 3 0 ° C 20 9 A l pan 11.37mg (0.01635 mmol) 3 0 - 3 3 0 ° C 5 10 A l pan 8.94 mg (0.01286 mmol) 30 - 275 - 180 - 340 - 270 - 350 °C 10 12 A l pan 11.05 mg (0.01590 mmol) 30-330 °C 2 13 Crucible 23.77 mg (0.03418 mmol) 2 5 - 160 °C 10 14 Crucible 11.22 mg (0.01613 mmol) 25 - 204 °C 10 15 Crucible 11.51 mg (0.01655 mmol) 25 - 220 °C 10 16 Crucible 9.97 mg (0.01434 mmol) 25 - 240 °C 10 243 I Era Area: 0.3438 J/g Peak: 96.6 °C Onset: 95.3 °C End: 98.8 °C Height: 0.03216 mW/mg Area: 241.8 J/g Peak: 231.1 "C Onset: 199.9 °C End: 243.6 "C Height: 1.383 mW/mg Area: 7.011 J/g Peak: 156.0 °C Onset: 151.7 °C End: 158.8 °C Height: 0.209 mW/m Figure A2.4 Temperature (°C) DSC Plot of[Fe(CO)6][SbF6]2: Experiment 3. kiW/ttn) J. E * Area: 201.6 J/g Peak: 230.3 °C Onset: 208.2 °C End: 240.7 °C Height: 1.272 mW/m, Area: 0.4307 J/g Peak: 96.6 °C Onset: 95.4 "C End: 98.7 °C Height:0.03683 mW/mg Area: 7.456 J/g Peak: 155.6 °C Onset: 151.6°C End: 158.7 °C Height:0.2217mW/mg Area: 43.47 J/g Peak: 304.8 °C Onset: 292.6 °C End: 308.3 °C Height:0.5817 mW/mg Area: 4.663 J/g Peak: 364.7 °C Onset: 352.5 °C End: 373.3 °C Height:0.1642 mW/mg Area: 15.81 J/g Peak: 389.0 °C Onset: 378.4 °C End: 393.0 °C Height:0.2774 mW/mg 1. Temperature (°C) Figure A2.5 DSC Plot of [Fe(CO)6][SbF6]2: Experiment 4. 244 A summary of the thermal events for [Fe(CO)6][Sb2Fn]2 is given in Table A2.4. Table A2.3 DSC Studies on [Fe(CO)6][SbF6]2: the events from 25 - 400 °C. Event # Approximate onset temp. (°C) Peak size Notes 1 95 very small • liberation of HF from lattice • irreversible 2 150 small • phase transition • reversible 3 180 large (~ 200 J/g) • overlap of at least 2 peaks • irreversible 4 240 medium (~ 40 J/g) • irreversible 5 280 medium • irreversible 6 340 medium • overlap of at least 2 peaks Experiment 5 (20 - 110 - 80 - 170 - 130 - 270 °C, see Figure A2.6) Experiment 5 examined the reversibility of events # 1 (onset: 95 °C) and 2 (onset: 150 °C). Event # 1 is irreversible. Due to its irreversibility, event # 1 is again attributed to liberation of HF from the lattice as for [Fe(CO)6][Sb2Fn]2. However, event # 2 is found to be reversible. Due to its reversibility, it can be certain that this event is not related to the decarbonylation of [Fe(CO)6] 2 + . The reversibility of the event suggests that it may be due to a structural phase transition of [Fe(CO)6J[SbF6]2 as follows: <150°C: a phase 150 - 170 °C: a phase ±5 B phase >170°C: p phase 245 fcWJmg) •i' E i o Area: 5.653 J/g Peak: 155.2 °C Onset: 151.0 "C End: 158.1 °C Height:0.2027 mW/mg Area: 1.149 J/g Peak: 96.3 °C Onset: 98.8 °C End: 98.7 °C Height:0.07053 mW/mg Area: 6.151 J/g Peak: 154.2"C Onsel: 150.5 "C End: 157.3 "C Heighl:0.2248 mW/mg Area:6.419.I/g Peak: 147.3 "C Onset: 143.2 "C End: 151.1 "C Height:0.2006 mW/mg Ml Area: 207.6 J/g Peak: 231.5"C Onset: 209.5 "C End: 242.9 °C Height: 1.423 mW/mg .H •;/5v. "V \ ..n / j \ Temperature (°C) Figure A2.6 DSC plot of [Fe(CO)(][SbF(l]2: Experiment 5. Experiments 6, 7, 8, 9 and 12 (30 - 270 or 330 °C, see Figure A2.7) Experiments 6, 7, 8, 9 and 12 studied the events # 1 - 4 at different heating rates. These runs confirmed the reproducibility of the thermal events. In general, the corresponding events show onset at about the same temperatures. Experiment 10 (30 - 275 - 180 - 340 - 270 - 350 °C, see Figure A2.8) Experiment 10 confirmed that events # 3 (onset: 180 °C), 4 (onset: 240 °C) and 5 (onset: 280 °C) are irreversible. The onset temperatures and irreversibility, along with the comparison with 2_)_ the thermal events of [Fe(CO)6J[Sb2Fi i ] 2 , suggest that decarbonylation of [Fe(CO)6] in [Fe(CO)6][SbF6]2 begins between 180 - 240 °C. 246 -i-Exo Area: 5.224 J/g Peak: 154.3 "C Onset: 151.4"C End: 156.4 "C HeiKht:0.054S2 n>W/nm Area: 4.402 J/g Peak: 151.1 °C Onset: 151.7 °C End: 157.6 °C Height:0.1048 mW/mg Area: 6.713 J/g Peak: 156.0 °C Onset: 151.9 °C End: 158.7 °C Height:0.02062 mW/mg Area: 203.6 J/g Peak: 222.5 °C Onset: 195.0 "C End: 27.3 °C Height:0.6473 mW/mg Area: 230.1 J/g Peak: 241.5 T Onset: 211.9"C End: 254.8 °C Height:2.51 mW/mg Temperature (°C) Figure A2.7 DSC Plot of [Fe(CO)>6jfSbF\)V Experiments 3, 7, 8 and 9. Experiment 3 Experiment 7 Experiment 8 = P1 Experiment 9 = / • / / I Eio Area: 6.207 J/g Peak: 155.2 °C Onset: 151.0 °C End: 158.4 °C Height:0.1832 mW/mg Area: 174.1 J/g Peak: 231.5 °C Onset: 207.5 °C End: 241.5 °C Height: 1.18 mW/m; Area: 1.15 J/g Peak: 96.5 °C Onset: 94.3 °C End: 98.7 "C Height:0.07025 mW/m: Temperature (°C) Figure A2.8 DSC Plot of [Fe(CO)6][SbF6]2: Experiment 10. 247 Experiments 13, 14, 15 and 16 (see Figure A2.9) Experiments 13, 14, 15 and 16 were performed using crucibles, which enabled the gas phase and solid residue content to be studied after the heating process. Each of these samples was heated to a specific temperature to enable the study of a particular event. The results are summarized in Table A2.5. Experiment 13 = [1] Experiment 14 - [2] Experiment 15 = Experiment 16 •-• [4] i I / To"1 ' ' ' ' 100 ' ' ' ' 150 ' ' ' ' 200 ' Temperature (°C) Figure A2.9 DSC Plot of [Fe(CO)6] [SbF6]2: Experiments 13, 14, 15 and 16. From Table A2.5, it can be seen that event # 2 is unrelated to the decarbonylation of the Fe carbonyl cation while event # 3 can be assigned to the loss of CO. At 204 °C, the amount of CO given off might have been too small to be detected. From 220 °C up, CO evolution was observed. From 240 °C up, disproportionation reaction of 2CO(g) ^ C02(g) + C(S) was evident. 248 Table A2.4 Summary of results from DSC Experiments 13, 14, 15 and 16 Events being studied Maximum temp. (°C) Information about gas phase content (from IR) Information about solid residue content (from Raman) Phase #2 160 no sign of CO (6 x 10"3 mmol gas) • vc-o = 2244, 2220 cm"1 • vsb-F = 653 cm"1 a* + p #3 204 no sign of CO (2 x 10"3 mmol gas) • vc-o =2241, 2219 cm"1 • vSb-F = 650 cm"1 P #3 220 presence of CO (0.03 mmol gas, equivalent to -25% of total CO content) • strong fluorescence • no vc-o identified • only observable band: VSD-F = 666 cm"1 (weak) P #3 240 presence of CO and C 0 2 (0.034mmol gas, equivalent to ~60% of total CO content) • Strong fluorescence • no vc-o identified • only observable band: vsb-F = 656 cm"1 (weak) • dark solids present P : vc-o = 2241.8, 2219.0 cm"1 (Raman) at room temperature for a phase solid (see chapter 3). 249 Appendix 3 Magnetic Properties of Free Metal(II) Cations and Weak Field Metal(II) Coordination Complexes A3.1 Free Metal Ion Terms The terms or energy levels of a free ion arise from the electronic configuration.1'2 They are characterized by the total orbital angular momentum quantum number, L, and the total spin angular momentum quantum number, S; these are the free ion equivalents to the single electron angular momentum quantum number, 1, and the spin quantum number, s. A free ion term is labeled by means of its values of L and S as ( 2 S + 1 ) X , where (2S+1) is known as the multiplicity of the term and X is the letter S, P, D, F etc. when L is 0, 1,2,3.. . as for the single electron case.1 When there is more than one d-electron, their mutual repulsions will also contribute to the possible energies, resulting in having more than one term for each configuration. The ground term, which lies at lowest energy of all, is determined by Hund's rules:1 1. Of the terms of a configuration, those with the maximum multiplicity lie lowest. 2. Of the terms with maximum multiplicity, that with the largest value of L lies lowest. Table A3.1 shows the maximum values of Ms and M L for a particular configuration, which give the largest possible values of S and L respectively, and hence the ground terms for free transition metal ions from d 4 to d 9 . The effects of spin-orbit coupling upon the free ion terms can be described as functions of the quantum numbers L and S; 2 each of the terms is (2L+l)(2S+l)-fold degenerate. The Russell-Saunders coupling scheme, in which spin-orbit coupling is considered as a small perturbation on the terms derived from electron repulsions, is a reasonable model for most lighter elements.1 Under this scheme, each term is split up into a number of states that are specified by the total 250 angular momentum quantum number J that runs in intergral steps from | L+S | to | L-S | of that term.1 Each state is (2J+l)-fold degenerate. For heavier elements, the spin-orbit coupling is much more important than electron repulsions, and the free ion terms split into levels determined by spin-orbit coupling. This is referred to as the j-j coupling scheme. Table A3.1 Maximum values of Ms and Mi as well as ground terms for theoretical free metal ions from d4 to cf. l>2 Configuration M s M L Free Ion Ground Term d 4 2 2 >D d> 5/2 0 bS d b 2 2 3 D d 7 . 3/2 3 4 F d 8 1 3 J F d y 1/2 2 2 D Since most naked metal(II) hexafluoroantimonates(V) studied in this project are of first row transition metals, the Russell-Saunders coupling scheme is used throughout. When spin-orbit coupling is considered, the full symbol for a state is ( 2 S + I ) X j . The energy difference between the J and J+l state is A,(J+1), where X is the spin-orbit coupling constant.1 For a less than half-filled shell, X is positive and the lowest value of J (| L-S |) lies lowest in energy. For a more than half-filled shell, X is negative because the system may be regarded as a set of positively charged holes in a filled shell. In this case, the highest value of J (| L+S |) lies lowest in energy. 251 A3.2 Crystal Field (Weak Field) Terms When a free metal ion is placed in a crystal field, the energies of the free ion terms change. Due to the "naked" nature of the metal(II) hexafluoroantimonates(V), the weak crystal field case is considered. In the weak field approximation, the effects of inter-electron repulsion exceed those of the crystal fields, and the separation between free ion terms is large compared to the crystal field splittings of the terms.2 The crystal field potential is an electrostatic potential that interacts with the orbital angular momentum of the electrons, thus the weak field terms have the same spin-degeneracy as their parent free ion term.2 Table A3.2 summarizes the splitting of the free ion S, P, D and F terms into weak field terms.1'2 Table A3.2 Summary of the splitting offree ion terms into weak field terms for an octahedral geometry. Free ion term Weak field terms S A l g P T.g D T2g, Eg F Tig, T2g, A2g On forming a complex, the orbital contribution in a free ion is quenched. This is reflected in most transition metal complexes, as opposed to the free ion, having an experimental magnetic moment much closer to the spin-only value. A generalized qualitative prediction as to when orbital contribution will appear is: 2 For an electron in a particular orbital to have orbital angular momentum about an axis, it must be possible to convert the orbital into another identical, degenerate orbital which can accommodate an electron of the same spin by simple rotation about that axis. Thus in a free ion 252 the orbital d x y can be converted in dx2-y2 by a 45° rotation and d x z into d y z by a 90° rotation about the z-axis. In order to be valid, this must be accomplished without having two electrons of the same spin in a single orbital. When these conditions are met, orbital contributions may arise from having electrons in these orbitals. On the other hand, the d x y , d y z and d^ orbitals of t2(g> symmetry in an octahedral or tetrahedral crystal field are no longer degenerate with the dx2-y2 and dz2 orbitals of e(g) symmetry. Therefore the orbital contribution about the z-axis arising from the d x y , and dx2-y2 pair of orbitals is removed. Moreover, the energetically degenerate dx2-y2 and dz2 orbitals cannot be interconverted by rotation. Thus there is no orbital contribution associated with these two orbitals. By contrast, the t2(g) orbitals may be interconverted by rotation about suitable axes, so orbital contributions can be expected from the t2(g> electrons. Therefore, orbital contributions are not completely quenched, and for example, d 6 as well as d 7 systems would have orbital contributions in octahedral symmetry due to degeneracy in the t2g orbitals. Orbital contributions can be expected for the metal ions with T ground terms, since all configurations having degeneracy in the t2g set have either Ti or T2 ground terms. There is also a lot more temperature dependence of the magnetic moment for the metal ions with T ground terms than A or E ground terms. The energy gaps between the states split by the spin-orbit coupling for metal ions with T ground terms are of the order of kT, where k is the Boltzman constant and T is the temperature in K. For the A and E ground terms, a very small amount of orbital contribution can sometimes still be present although there is no degeneracy in the t2g set. This orbital contribution comes from mixing in of excited terms with the required symmetry. As a result of the partial quenching of orbital contributions in the complex, the effective magnetic moment of a metal complex is much closer to the spin-only value, ps (ps = 2V[S(S+1)]),1 than that of a free metal ion, which is close to U L + S or uj ( U L + S = V[4S(S+1)+L(L+1)], uj = gV[J(J+l)] where g is the Lande Splitting Factor).1 A summary of the 253 free ion ground terms and the weak field ground terms, as well as the comments on orbital contributions for d 4 to d 9 configurations is presented in Tables A3.3 to A3.4. Table A3.3 Summary offree ion ground terms, weak field ground terms and comments on orbital contributions for d4 to d5 configurations. Configuration Free ion ground term Weak field ground term Comments on orbital contributions1'2 d4 5D • no orbital angular momentum, thus no orbital contribution in first order approximation • can get mixing through the action of spin-orbit coupling between the 5 E g ground term and the higher lying 5 T 2 g term in second order approximation, which results in a very small amount of orbital contributions • the experimental moment is usually less than ps d5 6S 6 A , g • no orbital angular momentum, thus no orbital contribution in first order approximation • no orbital contribution from second order mixing, since there are no excited terms of same multiplicity • the experimental moment is usually equal to ps d6 5 D 5 T 2 g • A significant amount of orbital contributions is expected from first order approximation • the experimental moment is usually greater than ps 254 Table A3.4 Summary of free ion ground terms, weak field ground terms and comments on orbital contributions for d6 to d3 configurations. Configuration Free ion ground term Weak field ground term Comments on orbital contributions1'2 d7 4 F 4 T Mg • A significant amount of orbital contributions is expected from first order approximation • the experimental moment is usually greater than ps d8 3 F 3 A 2 g • no orbital angular momentum thus no orbital contributions from first order approximation • can get mixing through the action of spin-orbit coupling between the 3 A 2 ground term and the higher lying 3 T terms in second order approximation, which results in a very small amount of orbital contributions • the experimental moment is usually greater than ps d9 2D 2 E c g • no orbital angular momentum, thus no orbital contributions associated with first order approximation • can get mixing through the action of spin-orbit coupling between the 2 E g ground term and the higher lying 2 T 2 g term in second order approximation • the experimental moment is usually greater than ps 255 A3.3 References 1) Figgis, B. N . ; Hitchman, M . A . Ligand Field Theory and Its Applications; Wiley-VCH, 2000. 2) Mabbs, F. E., Machin, D. J. Magnetism and Transition Metal Complexes; Chapman and Hall: London, 1973. 256 Appendix 4 Magnetic Data for Various Complexes 1. [Fe(CO) 6 ] [Sb 2 F„] 2 and [Fe(CO) 6][SbF 6] 2 Table A4.1 Magnetic Data of [Fe(CO)g][Sb2FjiJ2 collected at 10000G. The sample was purified by fluorination of the compound in HF solution. Temperature X M Temperature X M (K) (x 10"4 cm 3 mol 1 ) 0*B) (K) (x 10"4 cm 3 mol *) ( U B ) 300 3.79 0.95 35 8.47 0.49 2 8 0 3.87 0.93 3 0 9.20 0.47 260 3.95 0.91 25 10.20 0.45 240 4 .00 0.88 22 11.01 0.44 2 2 0 4 .13 0.85 2 0 11.67 0.43 2 0 0 4.25 0.82 17 12,92 0.42 180 4 .34 0.79 15 14.00 0.41 160 4.46 0.76 13 15.38 0.40 140 4 .62 0.72 12 16.24 0.39 120 4.84 0.68 11 17.26 0.39 100 5.14 0.64 10 18.43 0.38 9 0 5.33 0.62 9 19.85 0.38 80 5.58 0.60 8 21 .59 0.37 70 5.90 0.57 7 23 .69 0.36 60 6.33 0.55 6 26 .34 0.36 50 6.91 0.53 5 29 .99 0.35 45 7.48 0.52 4 35 .36 0.34 4 0 7.91 0.50 3 43 .50 0.32 257 Table A4.2 Magnetic Data of [Fe(CO)[SbF^J'2 collected at 10000G. The sample was prepared from purified [Fe(CO) (J [SbjFu] 2-Temperature Temperature X M V-(K) (x 10 4 cm 3 mol"1) (UB) (K) (x 10"4 cm 3 mol"1) (Wl) 300 6.15 1.22 30 17.14 0.64 280 6.31 1.19 25 19.30 0.62 260 6.44 1.16 22 20.97 0.61 240 6.58 1.12 20 22.34 0.60 220 6.72 1.09 17 24.90 0.58 200 6.94 1.05 15 27.08 0.57 180 7.14 1.01 13 29.76 0.56 160 7.41 0.97 12 31.36 0.55 140 7.76 0.93 11 33.27 0.54 120 8.21 0.89 10 35.40 0.53 100 8.86 0.84 9 37.88 0.52 90 9.28 0.82 8 40.77 0.51 80 9.80 0.79 7 44.14 0.50 70 10.45 0.77 6 48.27 0.48 60 11.29 0.74 5 53.35 0.46 50 12.45 0.71 4 59.85 0.44 45 13.39 0.69 3 67.97 0.40 40 14.37 0.68 2 78.02 0.35 35 15.58 0.66 258 2. [{Mo(CO)4}2(cis-p-F2SbF4)3]x[Sb2Fn]x and [W(CO)6(FSbF5)][Sb2Fn] Table A4.3 Magnetic Data of [{Mo(CO)4}2(cis-/u-F2SbF4)3Jx[Sb2F//Jx collected at 10000G. Temperature X M Temperature X M V-( K ) (x 10"4 cm3 mol1) ( K ) (x 10 4 cm3 mol"1) (HB) 300 2.42 0.76 60 3.33 0.40 290 2.42 0.75 50 3.54 0.38 280 2.44 0.74 45 3.48 0.35 270 2.46 0.73 40 3.54. 0.34 260 2.51 0.72 35 3.61 0.32 250 2.53 0.71 30 3.72 0.30 240 2.55 0.70 25 3.87 0.28 230 2.60 0.69 22 4.00 0.27 220 2.64 0.68 20 4.10 0.26 210 2.58 0.66 17 4.29 0.24 200 2.72 0.66 15 4.48 0.23 190 2.76 0.65 13 4.70 0.22 180 2.79 0.63 12 4.85 0.22 170 2.83 0.62 11 5.02 0.21 160 2.88 0.61 10 5.21 0.20 150 2.92 0.59 9 5.47 0.20 140 2.97 0.58 8 5.77 0.19 130 3.00 0.56 7 6.15 0.19 120 3.03 0.54 6 6.64 0.18 110 3.07 0.52 5 7.34 0.17 100 3.11 0.50 4 7.79 0.16 90 3.16 0.48 3 9.19 0.15 80 3.22 0.45 2 11.85 0.14 70 3.27 0.43 259 Table A4.4 Magnetic Data of [{Mo(CO)4}2(cis-/u-F2SbF4)3]x[Sb2Fi\]x collected at 10000G. Temperature X M V- Temperature X M (K) (x 10 4 cm3 mol"1) ( M (K) (x 10"4 cm 3 mol"1) (11B) 3 0 0 9.90 1 . 5 4 60 41 .14 1.41 2 9 0 10.16 1.53 50 48 .94 1.40 2 8 0 10.46 1.53 45 54 .16 1.40 2 7 0 10.74 1.52 40 60.68 1.39 260 11.11 1.52 35 69 .18 1.39 2 5 0 11.45 1.51 3 0 80.33 1.39 240 11.84 1.51 25 95 .74 1.38 2 3 0 12.28 1.50 22 108.37 1.38 2 2 0 12.75 1.50 20 118.99 1.38 2 1 0 13.26 1.49 17 139.59 1.38 2 0 0 13.82 1.49 15 157.73 1.38 190 14.43 1.48 13 181.49 1.37 180 15.11 1.48 12 196.30 1.37 170 15.82 1.47 11 213 .59 1.37 160 16.72 1.46 10 234 .30 1.37 150 17.70 1.46 9 259.41 1.37 140 18.83 1.45 8 290 .15 1.36 130 20 .12 1.45 7 329.53 1.36 120 21.63 1.44 6 380.08 1.35 110 23 .39 1.43 5 450 .30 1.34 100 25 .52 1.43 4 548 .24 1.32 90 28.11 1.42 3 696.53 1.29 80 31 .37 1.42 2 912 .62 1.21 7 0 35.58 1.41 260 3. The M[SbF6]2 complexes (M = Cr, Mn, Fe, Co, Ni, Cu, Pd) and Pd(S03F)3 Table A4.5 Magnetic Data ofCr[SbF6]2 collected at 10000G. Temperature (K) X M (x IO"3 cm3 mol"1) U ( U B ) Temperature (K) X M (x 10 3 cm3 mol') ( U B ) 300.0 8.03 4.39 45.0 49.26 4.21 290.0 8.31 4.39 40.0 54.97 4.19 280.0 8.60 4.39 35.0 62.31 4.18 270.0 8.88 4.38 30.0 71.77 4.15 260.0 9.22 4.38 25.0 84.34 4.11 250.0 9.57 4.37 22.0 94.46 4.08 240.0 9.96 4.37 20.0 102.94 4.06 230.0 10.37 4.37 17.0 118.12 4.01 220.0 10.83 4.36 15.0 131.33 3.97 210.0 11.33 4.36 13.0 147.32 3.91 200.0 11.87 4.36 12.0 156.79 3.88 190.0 12.46 4.35 11.0 166.74 3.83 180.0 13.13 4.35 10.0 179.79 3.79 170.0 13.88 4.34 9.0 192.85 3.72 160.0 14.71 4.34 8.0 209.16 3.66 150.0 15.64 4.33 7.0 225.48 3.55 140.0 16.79 4.34 6.0 245.06 3.43 130.0 17.93 4.32 5.0 266.27 3.26 120.0 19.40 4.31 4.0 284.21 3.02 110.0 21.03 4.30 3.0 295.61 2.66 100.0 22.98 4.29 2.8 292.35 2.56 90.0 25.59 4.29 2.6 284.19 2.43 80.0 28.53 4.27 2.4 272.75 2.29 70.0 32.45 4.26 2.2 261.33 2.14 60.0 37.51 4.24 2.0 254.80 2.02 50.0 44.52 4.22 261 Table A4.6 Magnetic Data of Mn[SbF6]2 collected at 10000G. Temperature X M V- Temperature X M V-(K) (x 10 2 cm 3 mol"1) ( H B ) (K) (x 10 2 cm 3 mol 1 ) ( H B ) 300 1.47 5.93 30 12.43 5.46 280 1.58 5.95 25 14.48 5.38 260 1.70 5.94 22 16.04 5.31 240 1.83 5.93 20 17.29 5.26 220 1.99 5.92 17 19.58 5.16 200 2.19 5.91 15 21.47 5.07 180 2.42 5.90 13 23.69 4.97 160 2.71 5.89 12 25.00 4.90 140 3.09 5.88 11 26.48 4.83 120 3.57 5.85 10 28.12 4.75 100 4.25 5.83 9 29.97 4.65 90 4.67 5.80 8 32.07 4.53 80 5.21 5.77 7 34.44 4.39 70 5.90 5.75 6 37.15 4.23 60 6.76 5.70 5 40.16 4.01 50 7.98 5.65 4 43.82 3.74 45 8.76 5.61 3 47.67 3.38 40 9.71 5.57 2 51.05 2.86 35 10.91 5.53 35 10.91 5.53 262 Table A4.7 Magnetic Data ofFe[SbF6]2 collected at 10000G. Temperature X M M- Temperature X M u (K) (x 10 2 cm 3 mol 1 ) ( U B ) (K) (x 10 2 cm 3 mol 1 ) ( U B ) 300 1.21 5.38 60 5.61 5.19 290 1.25 5.38 50 6.59 5.13 280 1.29 5.38 45 7.26 5.11 270 1.34 5.37 40 8.07 5.08 260 1.39 5.37 35 9.08 5.04 250 1.44 5.37 30 10.35 4.98 240 1.50 5.36 25 12.04 4.91 230 1.56 5.36 22 13.39 4.85 220 1.63 5.35 20 14.47 4.81 210 1.70 5.35 17 16.44 4.73 200 1.78 5.34 15 18.08 4.66 190 1.87 5.33 13 20.10 4.57 180 1.97 5.33 12 21.30 4.52 170 2.09 5.33 11 22.65 4.46 160 2.21 5.32 10 24.19 4.40 150 2.36 5.32 9 25.88 4.32 140 2.52 5.31 8 27.94 4.23 130 2.70 5.30 7 30.33 4.12 120 2.91 5.29 6 33.06 3.99 110 3.17 5.28 5 36.47 3.82 100 3.47 5.27 4 40.35 3.59 90 3.84 5.25 3 45.13 3.29 80 4.29 5.24 2 50.12 2.83 70 4.86 5.22 263 Table A4.8 Magnetic Data of Co[SbFe]2 collected at 10000G. Temperature X M u Temperature X M ( K ) (x 10 2 cm 3 mol 1 ) ( U B ) ( K ) (x 10 2 cm 3 mol"1) . ( H B ) 300 1.16 5.28 30 6.94 4.08 280 1.25 5.29 25 8.05 4.01 260 1.34 5.28 22 8.96 3.97 240 1.44 5.25 20 9.71 3.94 220 1.55 5.22 17 11.13 3.89 200 1.68 5.18 15 12.37 3.85 180 1.83 5.14 13 13.94 3.81 160 2.01 5.08 12 14.91 3.78 140 2.23 5.00 11 16.04 3.76 120 2.49 4.89 10 17.36 3.73 100 2.84 4.77 9 18.95 3.70 90 3.05 4.69 8 20.88 3.66 80 3.31 4.60 7 23.23 3.61 70 3.63 4.51 6 26.20 3.55 60 4.06 4.41 5 30.06 3.47 50 4.64 4.31 4 35.62 3.38 45 5.03 4.25 3 43.18 3.22 40 5.51 4.20 2 54.34 2.95 35 6.13 4.14 264 Table A4.9 Magnetic Data ofNi[SbF6]2 collected at 10000G. Temperature X M u Temperature X M u (K) (x IO"3 cm 3 mol"1) (uB) (K) (x 10 3 cm 3 mol 1 ) ( U B ) 299.99 4.82 3.40 60 21.11 3.18 289.95 5.00 3.41 50 24.88 3.15 279.96 5.17 • 3.40 45 27.46 3.14 269.97 5.34 3.40 40 30.46 3.12 259.97 5.53 3.39 35 34.36 3.10 249.97 5.73 3.39 30 39.43 3.08 239.97 5.95 3.38 24.98 46.23 3.04 229.98 6.18 3.37 22.03 51.49 3.01 219.97 6.44 3.37 19.99 55.82 2.99 209.97 6.68 3.35 17 64.04 2.95 199.98 7.00 3.35 15 70.76 2.91 189.98 7.35 3.34 13.02 79.10 2.87 179.98 7.73 3.33 12.02 83.98 2.84 169.98 8.14 3.33 11.02 89.83 2.81 159.99 8.60 3.32 10.02 96.15 2.78 149.99 9.13 3.31 9.02 103.53 2.73 139.99 9.73 3.30 8.02 112.15 2.68 129.99 10.41 3.29 7.02 122.03 2.62 119.99 11.19 3.28 6.01 133.79 2.54 110 12.12 3.27 5 147.63 2.43 99.99 13.27 3.26 3.99 164.62 2.29 90 14.61 3.24 3 184.34 2.10 79.99 16.25 3.22 2 206.01 1.82 70 18.40 3.21 265 Table A4.10 Magnetic Data of Cu[SbF6J2 collected at 1OOOOG. Temperature X M M- Temperature X M (K) (x IO"3 cm3 mol"1) (11B) (K) (x 10 3 cm3 mol1) ( U B ) 300 1.78 2.07 35 12.94 1.90 2 8 0 1.91 2.07 3 0 14.71 1.88 2 6 0 2.05 2 .06 25 17.02 1.84 240 2.21 2 .06 2 2 18.75 1.82 2 2 0 2 .40 2.05 2 0 20 .09 1.79 2 0 0 2.62 2.05 17 22 .44 1.75 180 2 .90 2.04 15 24 .27 1.71 160 3.24 2.04 !3 26 .29 1.66 140 3.66 2.03 12 27 .39 1.62 120 4 .26 2 .02 11 28.54 1.59 100 5.07 2.01 10 29 .69 1.54 9 0 5.57 2 .00 9 30.81 1.49 80 6.20 1.99 8 31 .84 1.43 7 0 7.01 1.98 7 32.71 1.35 60 8.07 1.97 6 33.31 1.27 50 9.50 1.95 5 33 .57 1.16 45 10.43 1.94 4 33 .52 1.03 4 0 11.55 1.92 3 33.68 0.90 266 Table A4.ll Magnetic Data ofPd[SbF6]2 collected at 10000G. Temperature X M M- Temperature X M V-(K) (x IO"3 cm 3 mol"1) 0 * B ) (K) (x IO"3 cm 3 mol 1 ) 0 * B ) 300 4.02 3.11 60 16.67 2.83 290 4.14 3.10 50 19.40 2.79 280 4.27 3.09 45 21.07 2.75 270 4.41 3.09 40 23.12 2.72 260 4.57 3.08 35 25.55 2.67 250 4.73 3.08 30 28.41 2.61 240 4.91 3.07 25 32.07 2.53 230 5.10 3.06 22 34.81 2.48 220 5.31 3.06 20 36.87 2.43 210 5.55 3.05 17 40.25 2.34 200 5.79 3.04 15 42.84 2.27 190 6.06 3.04 13 45.71 2.18 180 6.37 3.03 12 47.15 2.13 170 6.70 3.02 11 48.74 2.07 160 7.07 3.01 10 50.30 2.01 150 7.49 3.00 9 51.91 1.93 140 7.96 2.99 8 53.57 1.85 130 8.51 2.97 7 55.08 1.76 120 9.15 2.96 6 56.47 1.65 110 9.88 2.95 5 57.94 1.52 100 10.77 2.93 4 59.11 1.37 90 11.80 2.91 3 60.99 1.21 80 13.08 2.89 2 63.14 1.00 70 14.66 2.86 267 Table A4.12 Magnetic Data ofPd(S03F)3 collected at 1OOOOG. Temperature X M u Temperature X M u (K) (x 10 3 cm 3 mol"1) ( U B ) (K) (x IO"3 cm 3 mol"1) ( U B ) 300 4.99 3.46 60 28.26 3.68 290 5.17 3.46 50 35.41 3.76 280 5.35 3.46 45 40.48 3.82 270 5.54 3.46 40 47.39 3.89 260 5.75 3.46 35 57.07 4.00 250 5.98 3.46 30 71.71 4.15 240 6.22 3.45 25 96.19 4.38 230 6.48 3.45 22 119.93 4.59 220 6.79 3.46 20 143.05 4.78 210 7.14 3.46 17 198.15 5.19 200 7.52 3.47 15 261.43 5.60 190 7.92 3.47 13 363.25 6.15 180 8.37 3.47 12 436.23 6.47 170 8.88 3.47 11 526.11 6.81 160 9.46 3.48 10 620.01 7.05 150 10.13 3.49 9 706.42 7.14 140 10.92 3.50 8 775.32 7.05 130 11.82 3.51 7 824.03 6.80 120 12.88 3.52 6 856.73 6.42 110 14.16 3.53 5 879.52 5.93 100 15.75 3.55 4 896.45 5.36 90 17.66 3.57 3 906.00 4.66 80 20.16 3.59 2 912.15 3.82 70 23.54 3.63 268 4. The [MLn][SbF6]2 complexes (M = Cr, Mn, Fe, Co and Ni; L = pyrazine or 4,4'-bipyrdine) Table A4.13 Magnetic Data of [Cr(pyz)45][SbF6]2 collected at 10000G. Temperature X M u Temperature X M u (K) (x IO' 3 cm3 mol"1) (UB) (K) (x IO"3 cm3 mol1) (UB) 300 6.92 4.07 60 25.91 3.53 290 7.11 4.06 50 30.18 3.47 280 7.31 4.05 45 32.93 3.44 270 7.55 4.04 40 36.37 3.41 260 7.79 4.02 35 40.66 3.37 250 8.06 4.01 30 46.52 3.34 240 8.33 4.00 25 54.30 3.29 230 8.61 3.98 22 60.36 3.26 220 8.91 3.96 20 65.34 3.23 210 9.25 3.94 17 74.92 3.19 200 9.63 3.92 15 83.19 3.16 190 10.02 3.90 13 93.76 3.12 180 10.47 3.88 12 100.12 3.10 170 10.99 3.86 11 107.96 3.08 160 11.58 3.85 10 117.02 3.06 150 12.22 3.83 9 128.13 3.04 140 12.90 3.80 8 141.87 3.02 130 13.70 3.77 7 159.19 2.99 120 14.63 3.75 6 181.73 2.96 110 15.72 3.72 5 213.05 2.92 100 17.00 3.69 4 260.16 2.88 90 18.52 3.65 3 332.24 2.82 80 20.40 3.61 2 453.79 2.69 70 22.77 3.57 269 Table A4.14 Magnetic Data of [Mn(pyz)3.9][SbF6]2 collected at 10000G. Temperature X M M- Temperature X M (K) (x 10 2 cm3 mol"1) (HB) (K) (x 10 2 cm3 mol"1) (U-B) 300 1.52 6.05 60 7.92 6.17 290 1.58 6.06 50 9.53 6.17 280 1.64 6.06 45 10.61 6.18 270 " 1.70 6.07 40 11.95 6.18 260 1.77 6.07 35 13.69 6.19 250 1.84 6.07 30 15.99 6.20 240 1.92 6.07 25 19.25 6.20 230 2.01 6.08 22 21.90 6.21 220 2.10 6.09 20 24.08 6.21 210 2.21 6.09 17 28.34 6.21 200 2.33 6.10 15 32.07 6.20 190 2.45 6.10 13 36.73 6.18 180 2.59 6.11 12 39.62 6.17 170 2.75 6.11 11 42.93 6.15 160 2.93 6.12 10 46.84 6.12 150 3.13 6.13 9 51.38 6.09 140 3.36 6.13 8 56.87 6.04 130 3.63 6.14 7 63.42 5.96 120 3.94 6.15 6 71.21 5.85 110 4.31 6.16 5 80.98 5.69 100 4.74 6.16 4 94.06 5.49 90 5.28 6.16 3 110.57 5.15 80 5.94 6.16 2 132.46 4.60 70 6.79 6.17 270 Table A4.15 Magnetic Data of [Fe(pyz)40] [SbF6] 2 collected at 10000G. Temperature X M V- Temperature X M (K) (x IO"2 cm 3 mol 1 ) (U-B) (K) (x IO"2 cm 3 mol"1) (M*) 300 1.53 6.07 60 6.95 5.77 290 1.59 6.07 50 8.14 5.71 280 1.64 6.07 45 8.89 5.66 270 1.70 6.06 40 9.79 5.60 260 1.77 6.06 35 10.89 5.52 250 1.83 6.05 30 12.22 5.41 240 1.91 6.05 25 13.91 5.27 230 1.98 6.04 22 15.16 5.16 220 2.07 6.03 20 16.09 5.07 210 2.16 6.03 17 17.66 4.90 200 2.27 6.02 15 18.82 4.75 190 2.38 6.01 13 20.04 4.57 180 2.51 6.01 12 20.70 4.46 170 2.65 6.00 11 21.36 4.34 160 2.80 5.99 10 21.90 4.19 150 2.98 5.98 9 22.46 4.02 140 3.18 5.97 8 22.98 3.84 130 3.41 5.96 7 23.48 3.63 120 3.67 5.94 6 23.94 3.39 110 3.99 5.93 5 24.54 3.13 100 4.36 5.90 4 25.45 2.85 90 4.80 5.88 3 26.76 2.53 80 5.35 5.85 2 28.56 2.14 70 6.05 5.82 271 Table A4.16 Magnetic Data of [Co(pyz)42] [SbF6]2 collected at 10000G. Temperature X M Temperature X M (K) (x 10"2 cm3 mol"1) 0 * ) (K) (x 10 2 cm3 mol"1) (ue) 300 1.01 4.92 60 3.55 4.13 290 1.04 4.92 50 3.97 3.98 280 1.08 4.91 45 4.22 3.90 270 1.12 4.91 40 4.50 3.80 260 1.16 4.90 35 4.83 3.68 250 1.20 4.89 30 5.21 3.54 240 1.24 4.88 25 5.64 3.36 230 1.29 4.87 22 5.92 3.23 220 1.34 4.86 20 6.10 3.12 210 1.40 4.84 17 6.37 2.94 200 1.46 4.83 15 6.52 2.80 190 1.53 4.82 13 6.63 2.63 180 1.60 4.80 12 6.67 2.53 170 1.68 4.78 11. 6.70 2.43 160 1.77 4.75 10 6.71 2.32 150 1.86 4.72 9 6.72 2.20 140 1.97 4.69 8 6.71 2.07 130 2.09 4.66 7 6.71 1.94 120 2.22 4.61 6 6.76 1.80 110 2.37 4.56 5 6.88 1.66 100 2.53 4.50 4 7.17 1.51 90 2.73 4.43 3 7.74 1.36 80 2.96 4.35 2 8.76 1.18 70 3.22 4.25 272 Table A4.17 Magnetic Data of [Ni(pyz)3,8J[SbF6J2 collected at 10000G. Temperature X M u Temperature X M u ( K ) ( x 10 3 c m 3 mol"1) (UB) ( K ) ( x 10 3 c m 3 mol"1) (UB) 3 0 0 4 .07 3.13 60 19.87 3.09 2 9 0 4.23 3.13 50 23 .70 3.08 2 8 0 4 .37 3.13 45 2 6 . 1 9 3.07 2 7 0 4.53 3.13 40 29.23 3.06 2 6 0 4 .69 3 .12 35 33 .09 3.04 2 5 0 4.87 3 .12 3 0 38 .04 3.02 240 5.07 3 .12 25 44 .67 2 .99 2 3 0 5.28 3 .12 22 49 .69 2.96 2 2 0 5.52 3 .12 2 0 53 .67 2.93 2 1 0 5.78 3 .12 17 60 .79 2 .87 2 0 0 6.07 3 .12 15 66 .47 2.82 190 6.39 3 .12 13 72 .79 2.75 180 6.74 3.11 12 76 .27 2.71 170 7.13 3.11 11 79.98 2.66 160 7.58 3.11 10 83 .74 2 .59 150 8.07 3.11 9 87 .56 2.51 140 8.65 3.11 8 91 .34 2.42 130 9.31 3.11 7 94 .90 2.31 120 10.09 3.11 6 98 .15 2 .17 110 10.99 3.11 5 100.97 2.01 100 12.08 3.11 4 103.57 1.82 9 0 13.39 3 .10 3 105.88 1.59 80 14.99 3 .10 2 108.25 1.32 70 17.11 3.09 273 Table A4.18 Magnetic Data of [Cr(4,4'-bipy)2.5][SbF6]2 collected at 10000G. Temperature X M u Temperature X M u ( K ) (x 10 2 cm 3 mol"1) (UB) (K) (x IO"2 cm 3 mol 1 ) (UB) 3 0 0 1.08 5.08 60 3.61 4 .16 2 9 0 1.10 5.06 50 4 .16 4.08 2 8 0 1.13 5.03 45 4.53 4 .04 2 7 0 1.16 5.00 40 4.98 3.99 2 6 0 1.19 4.97 35 5.54 3.94 2 5 0 1.22 4 .94 3 0 6.25 3.87 240 1.26 4.91 25 7.22 3.80 2 3 0 1.30 4.88 22 7.97 3.74 2 2 0 1.34 4.85 20 8.58 3.71 2 1 0 1.38 4 .82 17 9.75 3.64 2 0 0 1.43 4 .79 15 10.75 3.59 190 1.49 4.75 13 11.98 3.53 180 1.55 4 .72 12 12.73 3.50 170 1.62 4 .69 11 13.60 3.46 160 1.69 4.65 10 14.61 3.42 150 1.78 4.62 9 15.77 3.37 140 1.87 4 .58 8 17.20 3.32 130 1.98 4 .54 7 18.89 3.26 120 2.11 4 .50 6 21 .02 3.18 110 2.26 4.46 5 23.65 3.08 100 2.43 4.41 4 27 .34 2.95 9 0 2.64 4 .36 3 32 .27 2.78 80 2.89 4 .30 2 38 .92 2 .50 7 0 3.20 4.23 274 Table A4.19 Magnetic Data of[Mn(4,4'-bipy)3.3j[SbF6J2 collected at 10000G. Temperature X M u Temperature X M u (K) (x IO"2 cm 3 mol"1) ( U B ) (K) (x 10"2 cm 3 mol"1) ( U B ) 300 2.26 7.37 60 8.85 6.52 290 2.35 7.38 50 10.16 6.37 280 2.43 7.37 45 10.95 6.28 270 2.51 7.36 40 11.91 6.17 260 2.59 7.34 35 13.07 6.05 250 2.68 7.32 30 14.57 5.91 240 2.78 7.31 25 16.54 5.75 230 2.89 7.29 22 18.11 5.64 220 3.00 7.27 20 19.38 5.57 210 3.12 7.24 17 21.84 5.45 200 3.25 7.21 15 23.96 5.36 190 3.39 7.18 13 26.68 5.27 180 3.55 7.14 12 28.35 5.22 170 3.73 7.12 11 30.35 5.17 160 3.93 7.09 10 32.72 5.12 150 4.15 7.06 9 35.54 5.06 140 4.41 7.03 8 39.06 5.00 130 4.70 6.99 7 43.43 4.93 120 5.04 6.96 6 48.93 4.85 110 5.40 6.89 5 56.38 4.75 100 5.85 6.84 4 67.15 4.63 90 6.35 6.76 3 82.11 4.44 80 6.99 6.69 2 104.27 4.07 70 7.80 6.61 275 Table A4.20 Magnetic Data of [Fe(4,4'-bipy)28] [SbF6] 2 collected at 10000G. Temperature X M Temperature X M M-(K) (x 10 2 cm 3 mol"1) (UB) (K) (x 10 2 cm 3 mol"1) (HB) 300 1.24 5.46 60 5.12 4 .96 2 9 0 1.28 5.46 50 5.96 4.88 2 8 0 1.32 5.44 45 6.48 4.83 2 7 0 1.37 5.43 40 7 .14 4.78 2 6 0 1.41 5.42 35 7.96 4 .72 2 5 0 1.46 5.41 3 0 9.02 4.65 240 1.52 5.40 25 10.46 4 .57 2 3 0 1.58 5.38 22 11.60 4 .52 2 2 0 1.64 5.37 2 0 12.54 4.48 2 1 0 1.71 5.36 17 14.27 4.41 2 0 0 1.78 5.34 15 15.82 4 .36 190 1.87 5.32 13 17.72 4 .30 180 1.96 5.31 12 18.87 4 .26 170 2 .06 5.29 11 20 .20 4 .22 160 2 .17 5.27 10 21 .74 4 .17 150 2 .30 5.25 9 23 .57 4 . 1 2 140 2.44 5.23 8 25.73 4 .06 130 2.61 5.21 7 28 .33 3.99 120 2.80 5.19 6 31 .56 3.89 110 3.03 5.16 5 35.65 3.78 100 3.29 5.13 4 41 .15 3.62 9 0 3.60 5.09 3 48 .09 3.40 80 3.99 5.05 2 56 .82 3.01 7 0 4.48 5.01 276 Table A4.21 Magnetic Data of [Co(4,4'-bipy)2.9]'[SbF6]2 collected at 10000G. Temperature X M H- Temperature X M (K) (x 10 3 cm 3 mol 1 ) (UB) (K) (x 10"3 cm 3 mol 1 ) (UB) 300 9.48 4.77 60 33.27 4.00 290 9.82 4.77 50 37.52 3.87 280 10.16 4.77 45 40.19 3.80 270 10.50 4.76 40 43.3.6 3.72 260 10.87 4.75 35 47.27 3.64 250 11.25 • 4.74 30 52.26 3.54 240 11.66 4.73 25 59.02 3.43 230 12.09 4.71 22 64.36 3.36 220 12.57 4.70 20 68.68 3.31 210 13.08 4.69 17 77.05 3.24 200 13.65 4.67 15 84.26 3.18 190 14.27 4.66 13 93.30 3.12 180 14.89 4.63 12 98.94 3.08 170 15.60 4.61 11 105.63 3.05 160 16.38 4.58 10 113.28 3.01 150 17.27 4.55 9 122.64 2.97 140 18.23 4.52 8 133.91 2.93 130 19.32 4.48 7 148.18 2.88 120 20.54 4.44 6 166.16 2.83 110 21.92 4.39 5 190.36 2.76 100 23.49 4.33 4 225.71 2.69 90 25.30 4.27 3 276.89 2.58 80 27.45 4.19 2 355.18 2.38 70 30.03 4.10 277 Table A4.22 Magnetic Data of [Ni(4,4'-bipy) 2.9]'[SbF6]2 collected at 10000G Temperature X M M- Temperature X M (K) (x IO"3 cm 3 mol 1 ) (UB) (K) (x 10 3 cm 3 mol"1) (UB) 300 3.68 2.97 60 17.47 2.89 290 3.82 2.98 50 20.79 2.88 280 3.96 2.98 45 22.98 2.88 270 4.11 2.98 40 25.76 2.87 260 4.26 2.97 35 29.22 2.86 250 4.42 2.97 30 33.76 2.85 240 4.60 2.97 25 39.95 2.83 230 4.78 2.97 22 44.77 2.81 220 5.00 2.97 20 48.68 2.79 210 5.22 2.96 17 56.02 2.76 200 5.48 2.96 15 62.24 2.73 190 5.75 2.95 13 69.46 2.69 180 6.05 2.95 12 73.67 2.66 170 6.39 2.95 11 78.99 2.64 160 6.81 2.95 10 84.48 2.60 150 7.23 2.94 9 90.68 2.56 140 7.71 2.94 8 97.37 2.50 130 8.31 2.94 7 105.04 2.43 120 8.96 2.93 6 113.84 2.34 110 9.77 2.93 5 124.29 2.23 100 10.70 2.93 4 135.98 2.09 90 11.85 2.92 3 149.20 1.89 80 13.24 2.91 2 163.53 1.62 70 15.08 2.91 278 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0061269/manifest

Comment

Related Items