Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A synthetic, spectroscopic and magnetic susceptibility study of selected main group and transition metal.. 1992

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


ubc_1992_fall_cader_m_shah_roshan.pdf [ 4.71MB ]
JSON: 1.0060513.json
JSON-LD: 1.0060513+ld.json
RDF/XML (Pretty): 1.0060513.xml
RDF/JSON: 1.0060513+rdf.json
Turtle: 1.0060513+rdf-turtle.txt
N-Triples: 1.0060513+rdf-ntriples.txt

Full Text

A SYNTHETIC, SPECTROSCOPIC AND MAGNETIC SUSCEPTIBILITY STUDY OF SELECTED MAIN GROUP AND TRANSITION METAL FLUORO COMPOUNDS By M. SHAH ROSHAN CADER M.Sc., The University of British Columbia, 1986 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Department of Chemistry) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April, 1992 © M.S.R. CADER, 1992 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. _____________________ Department of CItE WI/c The University of British Columbia Vancouver, Canada Date ft TC)(36rQj , 1di9J DE-6 (2/88) ABSTRACT This study was initiated in order to synthesize, and where appropriate, to investigate the magnetic properties of selected main group and transition metal cationic complexes, all stabilized by weakly basic fluoro anions derived either from the Brönsted superacids HSO3Fand HSO3CF,or the Lewis acids SbF5 and AsF5. Of the preparative reactions, the solvolysis of metal(ll) fluorosulfates in excess SbF5 according to: 25-60°C M(SO3F)2 + 6SbF5 > M(SbF6)+ 2SbF9( O3) with M=Ni, Pd, Cu, Ag or Sn, is found to be a useful synthetic route to the corresponding divalent hexafluoro antimonates. The products, as their precursors, are characterized as CdC12- type layered polymeric compounds. Relevant vibrational (Raman and IR), electronic and 119Sn-Mössbauer spectra as well as magnetic susceptibility measurements and X-ray powder data are reported. Several compounds prepared by this method display unusual features: Pd(SbF6)2is, like its fluorosulfate precursor, paramagnetic with the Pd2 ion in a 3A2g ground state. Ag(SbF6)2,unlike its paramagnetic blue valence isomer, is diamagnetic and nearly white in color, and is formulated as the mixed valency complex Ag(I)[Ag(III)(SbF64]. The Cu(SbF6)2compound also contains, in addition to Cu2 ions, small quantities of Cu(I) and Cu(Ill) ions. Both Ni(SbF6)2 and Pd(SbF6)2exhibit temperature dependent low magnetic moments, indicative of antiferromagnetic exchange. Pd(SbF6)2also displays very weak ferro magnetism below —lOK. 11 The Sn(SbF6)2product from the above synthesis, and its precursor Sn(SO3F)2,react with excess 1,3,5-trimethylbenzene(mesitylene or mes) to give the it-arene adducts Sn(SbF62mes and Sn(SOF)2mes in high yield. The adducts are characterized by elemental analysis and infrared spectra. The adduct formation is followed by 119Sn Mössbauer spectroscopy. It is found that only tin(ll) compounds with large, weakly nucleophilic anions are capable of forming mesitylene complexes, while SnC12,SnF2, and stannocene do not give any indication of adduct formation under similar reaction conditions. The divalent fluorosulfates Ni(SO3F)2,Pd(SO3F)2and Ag(SO3F)2,precursors to the M(SbF6)2 compounds, and the mixed valency Pd(II)[Pd(IV)(SO6], as well as their corresponding trifluoromethylsulfate derivatives Ni(SO3CF)2, Pd(SO3CF)2 and Ag(SO3CF)2,investigated for their magnetic behavior by susceptibility studies down to -•4 K, show significant magnetic exchange, and except in Ag(SO3CF)2,the onset of magnetic exchange becomes observable at low temperatures. The fluorosulfates are found to exhibit strong ferromagnetism below —11 K, whereas the trifluoromethylsulfates behave as anti ferromagnets with the spin interactions noted over a wider temperature range. The maximum magnetic susceptibilities of Ni(SO3F)2, Pd(SO3F)2 and Ag(SO3F)2 indicate saturation magnetization, and hence for these compounds field dependent maximum magnetic moments are obtained in the temperature range —5 to 10.5 K. Maxima in the susceptibility vs. tempera ture plots are noted for the antiferromagnets Pd(SO3CF)2and Ag(SO3CF)2at —4 and —13 K respectively. Unlike in the corresponding divalent antiferromagnetic fluorides, no spin canting is detected in the trifluoromethylsulfates at lower temperatures. Magnetic susceptibility measurements to —4 K are also carried out for the main group molecular cations withinO2[AsF6],Br2[Sb316] andI2[SbF1i]• The data are interpreted utilizing previous results from photoelectron spectroscopy, known crystal structures, magnetic studies on the superoxide ion and the ozonide ion, and in the case of Oj[AsF6],previous ESR studies. 111 The magnetic properties of the three materials are quite different. Br2[Sb3F16]obeys Curie- Weiss law between 80 and 4 K. The magnetic moment decreases slightly from 2.04 B at room temperature to 1.93 B at 4 K. I2[SbF11] exhibits relatively strong antiferromagnetic coupling with a maximum in XM observed at -54 K. The magnetic moment (corrected for TIP) decreases from 1.92 B at 124 K to 0.41 B at 4 K. Experimental susceptibilities for this compound over the temperature range 300-4 K have been compared to values calculated using three different theoretical models for extended chains of antiferromagnetically coupled paramagnetic species. 02+[AsF6] exhibits Curie-Weiss behavior over the temperature range 60-2 K. The magnetic moment, uncorrected for TIP, varies from 1.63 B at 80 K to 1.17 I.LB at 2 K, and the presence of weak antiferromagnetic coupling in this material is suggested. iv TABLE OF CONTENTS Page ABSTRACT ii LIST OF TABLES ix LIST OF FIGURES xi LIST OF SYMBOLS xiii ACKNOWLEDGEMENTS xiv CHAPTER 1 INTRODUCTION 1 1.1 General Introduction 1 1.2 The Lewis Acid SbF5 3 1.2.1 Physical and Chemical Properties 3 1.2.2 Cationic Derivatives 4 1.3 The Transition Metal Fluorosulfates and Trifluoromethyl sulfates 9 1.4 Post-Transition Metal Arene fl-Compounds 13 1.5 Magnetic Measurements 16 1.6 Objectives of this Study 20 References 22 CHAPTER 2 GENERAL EXPERIMENTAL 27 2.1 Introduction 27 2.2 Chemicals 27 2.2.1 Purification Methods 29 2.2.2 Synthetic Methods 29 V Page 2.3 Apparatus 31 2.3.1 Reaction Vessels 33 2.3.2. SbF5 Storage-bridge Vessel 33 2.3.3 S206F- Addition Trap 33 2.3.4 Pyrex Vacuum Line 37 2.3.5 Metal Vacuum Line 37 2.3.6 Dry Atmosphere Box 37 2.4 Instrumentation and Methods 38 2.4.1 Infrared Spectroscopy 38 2.4.2 Raman Spectroscopy 38 2.4.3 Nuclear Magnetic Resonance Spectroscopy 38 2.4.4 Electronic Spectroscopy 39 2.4.5 Mössbauer Spectroscopy 39 2.4.6 X-ray Photoelectron Spectroscopy 39 2.4.7 Magnetic Susceptibility Measurements 40 2.4.8 X-ray Powder Diffractometry 41 2.4.9 Differential Scanning Calorimetry (DSC) 41 2.4.10 Elemental Analyses 42 References 42 CHAPTER 3 METAL(II) HEXAFLUORO ANTIMONATES M(SbF6)2, M(ll) = Sn(II), Ni(II), Pd(ll), Cu(ll) AND Ag(ll) 43 3.1 Introduction 43 3.2 Experimental 46 3.2.1 General Synthetic Scheme to M(SbF6)2 46 3.2.2 Physical Properties and Analyses 47 3.2.2a Cu(SbF6)2 48 vi Page 3.2.2b Pd(SbF6)2 48 3.2.2c f3-Ag(SbF6) 48 3.2.3 Alternate Synthetic Route to 3-Ag(SbF6)2 48 3.3 Results and Discussion 49 3.3.1 Synthesis 49 3.3.2 Vibrational Spectra 55 3.3.3 Electronic Spectra 62 3.3.4 Magnetic Susceptibility Measurements 66 3.3.5 X-ray Photoelectron Spectra 82 3.3.6 Attempted Synthesis of Au(SbF6)2 83 3.4 Conclusion 85 References 87 CHAPTER 4 MESITYLENE ADDUCTS OF TIN(II) FLUORO COMPOUNDS, Sn(SO3F)2C9H12and Sn(SbF6)2C9H12 90 4.1 Introduction 90 4.2 Experimental 93 4.2.1 Synthesis 93 4.2.1 a Synthesis of Sn(SO3F)21 ,3,5-(CH)C6H 93 4.2.1 b Synthesis of Sn(SbF6)2[1 ,3,5(CH] 94 4.3 Results and Discussion 95 4.3.1 Synthesis 95 4.3.2 119Sn Mössbauer Spectra 99 4.3.3 Infrared spectra 101 4.4 Conclusion 108 References 108 vii Page CHAPTER 5 A LOW TEMPERATURE MAGNETIC STUDY OF THE MOLECULAR CATIONS O2, Br2 AND I2 111 5.1 Introduction 111 5.2 Experimental 112 5.3 Results and Discussion 113 5.3.1 Synthesis 114 5.3.2 Magnetic Measurements 118 5.3.3 Br2[SbF16] 118 5.3.4 O[AsF6] 126 5•3•5 I2[SbF11f 130 5.4 Conclusion 136 References 137 CHAPTER 6 MAGNETIC EXCHANGE IN M(II) SULFONATES, M(ll) = Ni(II), Pd(I1) AND Ag(II) 141 6.1 Introduction 141 6.2 Experimental 144 6.3 Results and Discussion 144 6.3.1 Ferromagnetism of M(II) fluorosulfates Ni(SO3F)2, Pd(SO3F)2,Pd(II)[Pd(IV)(SO3F6]and Ag(SOF) 145 6.3.2 Antiferromagnetism of M(ll) thfluoromethylsulfates Ni(SO3CF),Pd(SO3CF)2and Ag(SO3CF)2 163 6.4 Conclusion 176 References 178 CHAPTER 7 SUMMARY AND GENERAL CONCLUSIONS 181 APPENDIX A 185 APPENDIX B 194 yin LIST OF TABLES Table 1.1: Table 1.2: Table 2.1: Table 3.1: Table 3.2: Table 3.3: Table 3.4: Table 3.5: Table 3.6: Table 3.7: Table 4.1: Table 4.2: Table 5.1: Table 5.2: Table 5.3: Table 6.1: Table 6.2: Table 6.3: Table 6.4: Table 6.5: Page 5 6 28 56 57 64 67 68 69 75 100 104 106 119 120 121 146 147 148 149 150 Solid Polyhalogen Cationic Derivatives of SbF5 Solid Binary Transition Metal Derivatives of SbF5 Chemicals Used Without Purification Vibrational Spectra of Ni(SbF6)2,Pd(SbF6)2and Cu(SbF6)2 Vibrational Spectra of the Two Valence Isomers of Ag(SbF6)2 Electronic Transitions and Ligand Field Parameters for Ni(SbF6)2,Pd(SbF6)2and Related Compounds Low Temperature Magnetic Data of Ni(SbF6)2 Low Temperature Magnetic Data of Pd(SbF&2 Low Temperature Magnetic Data of Cu(SbF6)2 Magnetic Moment Data of Ni(SbF6)2for the Temperature Range —80 to 295 K 119Sn Mössbauer Parameters of Relevant Tin(H) Compounds at 80 K Infrared Bands of Liquid Mesitylene and Bands Attributed to Mesitylene in the Adducts Sn(SO3F)2mes and Sn(SbF6)2mes Table 4.3: Infrared Frequencies for Sn(SO3F)2and Sn(SnF6)2and Bands Attributed to the Anions in the Mesitylene Adducts Sn(SO3)2mes and Sn(SbF6)2mes Magnetic Data ofBr2[Sb3F16j Magnetic Data ofI2[SbF1j andO2[AsF6] Structural and Spectroscopic Information onO2[AsF6i,I2[SbF11], Br2[Sb3F16]and the corresponding Ions O2, Br2 and I2 Low Temperature Magnetic Data of Ni(SO3F)2 Low Temperature Magnetic Data of Pd(SO3F)2 Low Temperature Magnetic Data of Pd(II)[Pd(IV)(SOF6] Low Temperature Magnetic Data of Ag(SO3F)2 Magnetic Data of Ni(SO3F)2for the Temperature Range 291 to 79K x Page Table 6.6: Magnetic Parameters of Ni(SO3F)2,Pd(SO3F)2, Pd(II)[Pd(IV)(SO3F]and Ag(SOF) 151 Table 6.7: Experimental and Calculated Saturation Magnetic Susceptibilities of Ni(SO3F)2,Pd(SO3F)2,Pd(ll)[Pd(IV)(SO3F fJand Ag(SO3F)2 157 Table 6.8: Magnetic Data of Ag(SO3CF)2for the Temperature Range 304 to 4 K 164 Table 6.9: Low Temperature Magnetic Data of Pd(SO3CF)2 165 Table 6.10: Low Temperature Magnetic Data of Ni(SO3CF)2 166 Table 6.11: Magnetic Data of Ni(SO3CF)2for the Temperature Range 292 to 80 K 167 x LIST OF FIGURES Page Figure 2.1: Apparatus for PreparingS206F 30 Figure 2.2: Typical Pyrex Reaction Vessels 32 Figure 2.3: Vacuum Filtration Apparatus 34 Figure 2.4: Kel-F Tubular Reactor 35 Figure 2.5: SbF5 Storage-bridge Vessel 36 Figure 3.1: Raman Spectrum of -Ag(SbF6)2 58 Figure 3.2: Crystal Structure of cz-Ag(SbF6)2 60 Figure 3.3: Spin Allowed Electronic Transitions from 3A Ground Term for Pd2 and Ni2 (d8) in Octahedral Ligand 1e1d 63 Figure 3.4: Magnetic Moment vs. Temperature of M(SbF&2,M = Ni, Pd and Cu 70 Figure 3.5: Magnetic Moment vs. Temperature of Ni(SbF6)2 74 Figure 3.6: Magnetic Moment vs. Temperature of HF-Treated Cu(SbF6)2 80 Figure 3.7: Magnetic Moment vs. Temperature of SbF5-Treated Cu(SbF6)2 81 Figure 4.1: 119Sn Mössbauer Spectrum of Sn(SbF6)2mes at 77 K 102 Figure 5.1: Magnetic Moment vs. Temperature ofBr2[Sb316], 12[51J andO2[AsF6J 122 Figure 5.2: Inverse Susceptibility vs. Temperature ofO2[AsF6j andBr2[SbF3F16J 123 Figure 5.3: Crystal Structure of Br2[Sb3F16] 125 Figure 5.4: Energy Level Diagram of the Dioxygenyl Ion with the a and it-Bonding 2p Orbitals 127 Figure 5.5: Crystal Structure ofI2[SbF11J 131 Figure 5.6: Magnetic Susceptibility vs. Temperature ofI2[SbF1] 132 Figure 6.1: Proposed Structure of Pd(SO3F)2 143 Figure 6.2: Inverse Susceptibility vs. Temperature of Ni(SO3F)2 152 Figure 6.3: Inverse Susceptibility vs. Temperature of Pd(SO3F)2, Pd(II)[Pd(IV)(SOF6}and Ag(SO3F)2 154 Figure 6.4: Magnetic Susceptibility vs. Temperature of Pd(SO3F)2at 7501 and 9625 G 155 xi Page Figure 6.5: Magnetic Susceptibility vs. Temperature of Pd(H)[Pd(IV)(SO3F6] at750land9625G 156 Figure 6.6: Magnetic Moment vs. Temperature of Pd(SO3F)2and Ag(SO3F)2 159 Figure 6.7: Magnetic Moment vs. Temperature of Pd(SO3CF)2 170 Figure 6.8: Magnetic Susceptibility vs. Temperature of Pd(SO3CF)2 171 Figure 6.9: Magnetic Moment vs. Temperature of Ni(SO3CF)2and Ni(SO3F)2 173 xii LIST OF SYMBOLS N = AvogaLiro’s number k = Boltzmann’s constant T = Absolute temperature g = Lande splitting factor 3 = Exchange coupling constant D = Zero-field splitting parameter = Spin-orbit coupling constant Dq = Ligand field splitting parameter B = Racah parameter (interelectronic repulsion parameter) or Bohr magneton Spin only magnetic moment eff = Effective magnetic moment XM = Molar magnetic susceptibility XMCOff = Molar magnetic susceptibility corrected for diamagnetism Cm = Curie constant e = Weiss constant G = Magnetic field in Gauss = Quadrupole splitting = Isomer shift xlii ACKNOWLEDGEMENTS It has been rewarding experience to work under the supervision of Professor F. Aubke. His guidance and encouragement were invaluable during the course of this work, and for this I extend my sincere gratitude and thanks to him. I also wish to thank Professor R.C. Thompson for his constructive and valuable collaboration in this study over the past few years. Many thanks are also due to my co-workers Dingliang Zhang, Germaine Hwang, Fred Mistry and Walter Cicha for their pleasant friendship and enlightening discussions. Tom Otieno and Martin Ehiert are thanked for their kind assistance in obtaining the magnetic measurements. Thanks are also due to the members of the mechanical, glassblowing, and electronic shops for their technical expertise. I am also indebted to Mr. P. Borda for his microanalytical services. Special thanks are extended to Germaine Hwang for proof-reading this thesis and Rani Theeparajah for the excellent work done in typing the manuscript. Finally I would like to express my sincere thanks to my family for their kind understand ing and words of encouragement throughout my years of graduate study. xiv DEDICATED TO MY MOTHER J.BA. CADER xv CHAPTER 1 INTRODUCTION 1.1 GeneraJ Introduction The synthesis and physical study of solid compounds where unusual metal and non metal cations are stabilized by wealdy coordinating fluoro anions is an area of research which has grown steadily over the last two decades, and is the primary focus of this dissertation. In recent years, there has been an increasing interest shown both by chemists and physicists in the synthetic and solid state properties of these rare cationic complexes, often obtainable only as derivatives of strong Brönsted (protonic) fluoroacids or Lewis acids (1, 2). The use of a wide array of physical methods including various magneto-chemical techniques to characterize these compounds has permitted structure-property relationships to be understood in detail and has also provided rational approaches to the synthesis of new and unusual materials. The types of compounds which have been studied in this thesis are varied, and range from main group non-metallic molecular cationic complexes to inorganic and organometallic coordination polymers, where the organometallic polymers are t-arene adducts of the post- transitional metal Sn(ll) fluoro derivatives. The molecular cations and the transition metal coordination polymers have been investigated for their magnetic properties to obtain informa tion on the ground state electronic structure as well as to detect any magnetic exchange inter actions between the paramagnetic centers (particularly at low temperature) in the respective compounds. Furthermore, spectroscopic and structural data (where available) have been utilized in the interpretation of the magnetic behavior observed for the compounds. 1 The inorganic and organometallic polymers synthesized in this study contain fluoroanions such as SbF6 and S03F, which are usually generated in superacidic media. Strong protonic fluoroacids and superacids have been used extensively as reaction media, solvents and synthetic reagents in both inorganic and organic syntheses (1). The role of these acids and their anions in the synthesis and stabilization of unusual cations is of significance to this work, since all the fluoro compounds studied here have anions which are derived from either the protonic fluoroacids HSO3Fand HSO3CF or the Lewis acids SbF5 and AsF5. Conse quently, the corresponding anions of interest S03F, S03CF, SbF6,Sb2F11,Sb3F16 and AsF6 are all poorly coordinating, weakly nucleophilic anions, well capable of stabilizing a variety of electrophiic cationic centers either in solid compounds or in solutions. These fluoro anions are in general non-oxidizable and are reasonably resistant toward reduction and, when coordinating to transition metals, act as monodentate as well as bidentate or tridentate ligands, usually with bridging, rather than chelating configurations. The coordinating ability of the fluorosulfate SO3F, and the trifluoromethylsulfate SO3CF groups has been discussed relatively recently by Lawrence (3), and the generation and stabilization of various halogen and interhalogen cations in protonic fluoroacids and superacids has been reviewed in the past by Gillespie and Morton (4), and many years later by Shamir (5). It is generally agreed that the anions mentioned above are all very weakly basic and have high group electronegativities. For a number of dimethyltin(IV) compounds of the type (CH3)2SnX (X = a fluoro or fluoroxy anion) with linear or near linear C-Sn-C groupings and bidentate bridging anions, the ‘19Sn Mössbauer parameters suggest the following order of anion basicity: F- > S03CH > SOCF3 Z SOF > AsF6 > SbF Z Sb2F1f(6). These general concepts mentioned briefly above will now be discussed in some detail in the following sections to provide the necessary background information for the extended study. 2 1.2 The Lewis Acid SbF5 1.2.1 Physical and Chemical Properties Antimony(V) fluoride, SbF5 is generally regarded as the strongest molecular Lewis acid (7). It is a very viscous (460 cP at 20°C) (8), colorless liquid, with a specific gravity of 3.145 g/cm3 at 15°C. It has a relatively high melting point (8.3°C) and a high boiling point (142°C) (la), compared to the Lewis acid AsF5 (mp = -80°C, bp = -53°C). This suggests a considerable degree of association for the molecule, and vapor density measurements indicate aggregates corresponding to (SbF5)3 at 150°C and (SbF5)2at 250°C (9). The polymeric structure of the liquid SbF5 has been established by 19F-NMR spectroscopy (10), and is found to have a cis fluorine bridged framework in which each antimony atom is surrounded by six fluorine atoms in an octahedral arrangement. In the solid state, SbF5 is tetrameric with octahedral coordination achieved again by cis-bridging fluorine atoms (11). Antimony pentafluoride is a good oxidizing and a moderately strong fluorinating agent. As shown by conductomethc, cryoscopic, and related acidity measurements, it appears that SbF5 is by far the strongest Lewis acid known, and hence is preferentially used in preparing stable metallic as well as non-metallic cationic derivatives. For the purpose of comparison, the following order of acidity can be assigned for a series of Lewis acids (la): SbF5 > AsF5 > TaF5 > BF3 > NbF5 > PF5. For their anions, the basicity is expected to increase in nearly the same order from SbF6 to PF6. Moreover, SbF5 also shows a remarkable ability to coordinate to protonic fluoroacids such as HF, HSO3Fand HSO3CF,resulting in vastly enhanced acidity for the conjugate superacid systems, where many types of unusual cations have been observed as stable species. 3 1.2.2 Cationic Derivatives Cation formation in SbF5 or in its conjugate superacid solutions is not resthcted to metallic species only, but occurs with equal frequency among non-metals such as halogen and polyhalogen derivatives (4,5). In most instances SbF5 has proven to be an excellent fluoride-ion acceptor, and the hexafluoro anion SbF6 or polyanions of the type SbF6. (SbF5), n = 1 or 2, are readily formed. Some selected solid polyhalogen and binary transition metal cationic species stabilized by such anions are listed in Tables 1.1 and 1.2. Extensive research has been carried out in our laboratory to study the possible extension of this fluoride-ion acceptor ability of SbF5 to include a similar “fluorosulfate-ion transfer” via the solvolysis of fluorosulfate compounds in SbF5 (12-15). The solvolysis process was initially used in our group to synthesize non-metallic cations (12,13,15), and later for the preparation of the dimethyltin(IV) cation (CH3)2Sn(16): C1O2S3F + 4SbF5 — C1O2[SbF1] + Sb2F9(SO3) [1.1] IX2SO3F + 4SbF5 — IX2[SbF11J + Sb2F9(SO3) [1.2] (X = Cl or Br) Br(SO3F) + 7SbF5 — BrF2[Sb6] + 3Sb2F9( O) [1.3] 2Br + 5206F + 1OSbF5 — 2Br[Sb3F16] + 2SbF9( O3) [1.4] 212 + S206F + 8SbF5 .—> 2I[SbF11] + 2SbF9( O3) [1.5] CH3Sn(SOF)2+ 8SbF5 —> (CH3)2Sn[SbF11]+ 2SbF9( O3) [1.6] 4 Table 1.1: Solid Polyhalogen Cationic Derivatives of SbF5 Compound Synthesis Reference C13[SbF6] Cl2 + CIF + SbF5 in HF 4 Br2[Sb3F16J Br-t-B F5+Sb 17 Br2 +S206F+ SbF5 15 12[2111] + SbF5 in SO2 18 12+S06FSbF 15 C1F2[Sb6j CW3 + SbF5 19 BrF2[Sb6] BrF3 + SbF5 20 Br(SO3F)+ SbF5 15 IF2[SbF6] IF3 + SbF5 21 ICl2[SbF11] IC12SO3F+ SbF5 13 IBrj[Sb2F11] IBr2SO3F+ SbF5 13 ClF4[Sb6j C1F5 + SbF5 in HF 22 BrF4[Sb2F1] BrF5 + SbF5 22 IF4[SbF61 IF5 + SbF5 8 BrF6[Sb2F1i] BrF5 + Kr2F3SbF6 23 IF6[Sb2F11 IF+SbF5 24 5 Table 1.2: Solid Binary Transition Metal Derivatives of SbF5 Compound Synthesis Reference Cr(SbF6)2 CrF2 + SbF5 in HF 25 Mn(SbF6)2 Mn + SbF5 in SO2 26 Fe(SbF6)2 FeF2+ SbF5 in HF 25 Fe + SbF5 in SO2 26 Co(SbF6)2 CoF2+ SbF5 in HF or SO2 25 Co + SbF5 in SO2 26 Ni(SbF6)2 NiF2 + SbF5 in HF 25 Ni+SbF5inSO2 26 Ni+SbF+F2 27 Ag(SbF6)2 AgF2+ SbF5 in HF 25 Zn(SbF6)2 ZnF2 + SbF5 in HF 25 Cd(SbF6)2 ‘2 + SbF5 in HF 25 Hg3(Sb2F11) Hg + SbF5 in SO2 28 6 Two of the above halogen cations synthesized by the solvolysis route, i.e.Br2[Sb3Fi& andI2[SbF1J, are investigated in this work for their magnetic properties by low temperature magnetic susceptibility measurements. This improved preparative method (15) provides a simple and straightforward route to sufficiently large quantities of very pure paramagnetic Brj and 12+ compounds, and hence is of value in magnetic studies where highly pure materials are desired. It is interesting to note that in Table 1.1, Br2[Sb3F16] andI2[SbF11} are the only two solid polyhalogen cation derivatives that are paramagnetic. The volatile components formed in these solvolysis reactions (SbF4O3F) and more frequently Sb2F9(SO3) where SbF5 is in an excess) can be removed easily in a dynamic vacuum (14), facilitating the isolation of pure solid products in very high yield. In this study, the solvolysis method is extended to divalent transition metal fluorosulfates as well. The polyhalogen derivatives of SbF5 listed in Table 1.1 are generally accepted to be of ionic structure, although there is evidence that in most of them some secondary cation-anion interaction of various degrees does exist. Single crystal X-ray studies performed on two compounds of interest to this study, i.e. Br2[Sb3F16] andI2[SbF1J, support an ionic for mulation since the cation-anion contacts are rather long, although shorter than the sum of the van der Waals radii, indicating a very small cation-anion interaction (17,18). Interestingly, the shortest FF contact distance inI2[SbF1i] is almost the largest cation-anion contact observed in a series of Sb2F1 derivatives with the following cations of decreasing contacts: SbCl4 > 12+ > XeF3 > Br2+ > BrF4. This decreasing distance indicates an increase in the acidity for the cations (18). The Sb-F (terminal) distances in Br2[Sb3Fi& andI2[SbF11J (1.83 A and 1.85 A) are similar to those observed in SbF6 derivatives like CW2[SbF6]and BrF2[Sb6j,namely 1.84 A and 1.835 A respectively (19,20). 7 The transition metal hexafluoroantimonates listed in Table 1.2 are, in most reported instances, formulated as metal difluoride adducts of the parent acid SbF5, and are of the type (25). This formulation is derived from a common synthetic route where the metal difluoride is reacted with SbF5, usually in the presence of anhydrous l{F or SO2 to yield the desired product. However, alternative structural forms like M(SbF6)2and MF(Sb211)are also possible for these products. When other preparative methods such as the oxidation of metals by SbF5 in SO2 (26) or metal fluorination with F2 in the presence of SbF5 (27) are used, the result ing products are conveniently formulated as M(SbF6): so2 M + 4SbF5 > M(SbF6)2 + SbF35 [1.7] where M = Mn, Fe or Ni 270°C Ni + 2SbF5 + F2 > Ni(SbF [1.8] 250 atm In the case of cobalt, reaction [1.7] reportedly leads to the ternary CoF(SbF6)2(26), whereas mercury is converted toHg3(Sb2F11)(28). It is significant to note here that all the above routes to transition metal hexafluoro antimonates have various limitations and complications. Metal oxidation by F2 at elevated temperature may lead to higher oxidation state compounds, whereas oxidation by SbF5 may not be a suitable method for metals with higher oxidation potentials than provided for by the Sb(V)/Sb(llI) couple. In addition, the quantitative separation of the solid byproduct SbF35 from the main product may prove to be the difficult (29), and consequently, impure materials are isolated as reaction products. Even the more versatile synthetic method of fluoride abstraction from MF2 by SbF5 (25) could lead to compounds of the type MF(Sb21 a structural isomer of the binary M(SbF6)2,due to an incomplete breakup of the MF2 lattice. 8 Some structural information on the transition metal derivatives of SbF5 (Table 1.2) has been reported. For the Mn, Fe and Ni compounds obtained from SbF5 in SO2 (26), magnetic moment values at room temperature appear to indicate octahedral coordination for the metal centers. Based on vibrational and X-ray powder data, Ni(SbF6)2which is synthesized from the high temperature fluorination of nickel in SbF5 (27), is shown to be related to the LiSbF6 struc ture by the occupation of only every second octahedral Li site with Ni2, leading to a layer type structure. Furthermore, a crystal structure has been reported for the paramagnetic (vide infra) Ag(SbF6)2(25), where the Ag2 ion is located in a tetragonally elongated octahedral environ ment, which in turn implies a layer structure with iridentate bridging SbF6 moiety for the compound. Except for the observed distortion due to the Jahn-Teller effect, the reported structure appears to be consistent with the proposed structure of the above mentioned Ni(SbF6)2 compound (27). 1.3 Transition Metal - Fluorosulfates and Trifluoromethylsulfates The transition metal sulfonates studied in this work can be considered as derivatives of the strong protonic fluoroacids fluorosulfuric acid (HSO3F) and trifluoromethylsulfuric acid (HSO3CF). The two acids rank among the strongest known protonic acids (1,3,7). Consequently, the corresponding anions S03F and SO3CF behave as weakly coordinating ligands, and are well suited to stabilize many unusual transition metal (as well as main group) cations. Compared to other poorly coordinating anions like perchiorate (C104j or tetra fluoroborate (BF4j, the SO3F and S03CF ions are arguably the most stable and non-oxidizing species available for synthetic purposes (3). Interestingly, almost all the transition metal trifluoromethylsulfate compounds are made by the solvolysis of suitable metal salts in an excess of HSO3CF, and include many metal fluorosulfates as precursors (30,3 1). Therefore, it is appropriate to examine first the syntheses and properties of transition metal fluorosulfates (particularly the derivatives of electron rich metals) in some detail. 9 Fluorosulfate chemistry displays many parallels to halogen chemistry, and the fluoro sulfate group may be viewed as a pseudohalide. Hence, the synthetic methods used in the preparation of the flouorosulfates have striking parallels to those used in the synthesis of halides, with the necessary modifications. Two synthetic methods in general have been used to prepare a large number of transition metal fluorosulfate derivatives: (a) solvolysis of a corresponding metal salt such as MCi2, MSO4 or M(RCOO)2 in excess HSO3F, and (b) the oxidation of a metal with the strongly oxidizing and fluorosulfonating reagent bis(fluorosulfuryl)peroxide, S206F (32,33) in the presence or absence of HSO3F. Solvolysis is almost exclusively the route of choice for the synthesis of 3d-block metal(ll) fluorosulfates (34,35), whereas metal oxidation byS206Fyields a variety of electron rich 4d- and 5d-metal fluorosulfates, in particular Pd(U)[Pd(IV)(SO3F6](36), Ag(SO3F)2(37), Pt(SO3F)4(38) and Au(SO3F) (39) according to: HSO3F M + x12S06F > M(SOF), [1.9] where x 2, 3 or 4 A major advantage to the use of theS2O6FJHSO3reagent combination in metal oxida tion is that precursors, in the form of fine metal powders, are available in high purity for all the transition metals, and consequently, very pure products can be isolated after the removal of the excess reagents in a dynamic vacuum. Furthermore, some noble metal fluorosulfates such as Pd(SO3F)2(36), Pt(SO3F)4(40) and Au(SO3F) (40) are also prepared by the oxidation of the respective metals by bromine monofluorosulfate, BrSO3F (41). However, the use of BrSO3F 10 instead ofS206F/HSO3in the synthesis of binary metal fluorosulfates offers no real advantage, primarily asS206Fis initially required to synthesize BrSO3F. For these transition metal fluorosulfates, only a single molecular structure, that of gold(llI) fluorosulfate, which was obtained by single crystal X-ray diffraction, has been reported so far (42). The compound is a dimer and contains both monodentate and symmetrically bridging bidentate fluorosulfate groups. Unfortunately, the polymeric nature of most transition metal fluorosulfates and their resulting lack of volatility and solubility in HSO3F or other suitable solvents have prevented the formation of single crystals, and hence structural evidence rests largely on vibrational spectra and magnetic properties. Fluorosulfates of the type M(SO3F)2appear to belong to a single structural type, derived from the CdC12 layer structure, with one exception - the mixed valency Au(I)[Au(Ill)(SO3F4] (43). The 0-tridentate bridging fluorosulfate group in M(SO3F)2results in M06-coordination polyhedra within the layer structure. Regular octahedra are found for M = Fe, Co, Ni, Pd, Zn, Cd and Hg (35,36,44), while the symmetry of the SO3F group appears to be reduced for M = Mn, Cu and Ag (35,37), with Jahn-Teller distortions expected for Cu2 and Ag2. This structure type is also postulated for the divalent metal trifluoromethylsulfates Fe(SO3CF)2(45), Co(SO3F)2(46), Pd(SO3CF2(30) and Cu(SO3F)2(46), and also extends to divalent pre-transition and post-transition metal S03F, S03CF and SO3CH derivatives as well. The only two molecular structures reported so far for this type of compounds are that of Sn(SO3F)2 (47a) and Ca(SO3H)2(47b). For the transition metal M(SO3R)2type compounds with R = F, CF3 or CH3, electronic spectra and magnetic data, where reported, confirm the structural conclusions reached. A greater structural diversity is encountered for binary fluorosulfates of the general composition M(SO3F). The dimeric structure of [Au(SO3F)]2(42) and the mixed-valency 11 Pd(II)[Pd(IV)(SO3F6](36) are clear exceptions. The mixed valency formulation follows the precedent of PdF3 (48) and is supported by the magnetic behavior and the synthesis and struc tural characterization of bimetallic compounds of the types Pd(II)[M(IV)(SO3F6](M = Pt or Sn) and Ba[Pd(IV)(SO3F6j(49). The fluorosulfate group appears to bond strongly to the M(IV) metal center and coordinates weakly to M(II), in an “anisobidentate” bonding mode. Therefore, it seems that where the SO3F (and the SO3CF or SO3CH)group functions as a polydentate ligand, a bridging configuration is observed, resulting in stable solids, often viewed as coordina tion polymers. Moreover, the versatile coordinating ability of the fluorosulfate group, which may function as a mono-, hi-, or tn-dentate ligand leads to the stabilization of metal ions in high, intermediate, and low oxidation states. In contrast to the large number of binary transition metal fluorosulfate compounds reported, only a small number of trifluoromethylsulfate derivatives are known. Almost all the M(SO3CF) species are made by the solvolysis of suitable metal salts in an excess of HSO3CF. In the solvolysis of M(SO3F) in HSO3CF (30,31), the reaction initially proceeds according to: HSO3CF M(SO3F) + x HSO3CF > M(SO3CF)+ x HSO3F [1.10] where x = 2 or 3, M = Mn, Pd, Ag or Au However, the by-product HSO3F and the reactant HSO3CF undergo a degradation reaction and produce a series of products (50,51) which do not appear to interfere in the isolation of the trifluoromethylsulfates (30). As mentioned above, for the divalent iron, cobalt and copper derivatives a layered lattice structure involving hexacoordinated metal centers has been suggested on the basis of vibrational and electronic spectra, as well as magnetic and Mössbauer data (45,46). In a manner similar to the SO3F groups discussed previously, the 12 SO3CF groups act as bridging tridentate ligands in these compounds. Although the solvolysis of transition metal fluorosulfates in excess trifluoromethyl sulfuric acid, which is a generally applicable method, should allow the synthesis of a compara tively large number of SO3CF compounds (30), only limited use has been made of this synthetic possibility. The principal reasons for this are threefold: (a) As discussed in Section 1.1, the SO3CF group has a basicity comparable to that of the SO3F group. An extended synthetic approach is not fruitful where compounds with similar properties and molecular struc tures result, as is often the case. (b) With vibrational spectroscopy used as a principal method of structural analysis, the coincidence of SO3 and CF3 stretching modes in S03CF causes a greater complexity, making vibrational assignments and structural conclusions frequently uncertain and ambiguous. (c) The S—C linkage in HSO3CF and its derivatives is sensitive to oxidative cleavage (30), and hence the acid’s use in the synthesis of high-valent metal deriva tives with a good oxidizing potential, is not advisable. However, the S-C linkage, unlike the S-F bond, is hydrolytically stable and hence chemistry in aqueous medium is possible with HSO3CF. There are nevertheless a number of interesting cases where metal trifluoromethylsulfates display fundamentally different magnetic properties from the coffesponding fluorosulfates, as will be discussed in Chapter 6 of the text. 1.4 Post-Transition Metai Arene it-Compounds Both main group and transition metal organometallic n-complexes have been studied extensively, and a very large number of compounds have appeared in the chemical literature in this field over the past few years alone. In this discussion, however, the emphasis is placed on the relatively rare post-transition metal arene n-complexes. In terms of bonding, these 13 compounds differ significantly from most of the transition metal it-compounds for three principal reasons (52): (a) Metals of the post-transition series contain no partially filled or empty valence d subshells of comparable energy to those of the it orbitals of arene ligands. As a result, the it electron density of the ligand can be transferred only to a very small extent to the s or p type orbitals of the metal. (b) The filled d subshells of the metals are usually of much lower energy than the antibond ing it’ orbitals of the arene ligand. Consequently, unlike the transition metal counter parts, back-donation from the d electrons to effect synergic bonding is not observed in post transition it-complexes. (c) The effective atomic number rule, which is so useful in transition metal carbonyls and arene and Cp derivatives, is rendered ineffective and inapplicable in the case of post transition metal derivatives. The consequence of factors (a) and (b) is a weak metal ligand interaction, leading to compounds which are not as varied as the transition metal n-complexes. The study of neutral arene-metal complexes with the post transition metals in recent times has included the elements Ga, In, Tl, Sn, Pb, and Bi (53). However, these studies indicate that the it-interactions with the elements above the fourth period are very weak, resulting in reduced hapticity from a desired 6 to i3 or 2 (53) Studies by Schmidbaur et al. (54) have concentrated on neutral arene complexes of univalent gallium, indium and thallium, which directly follow the d block elements. Both mono and bis(arene) complexes have been synthesized and characterized, and among these is the 14 mixed mono and bis(arene) thallium complex, [(Mes)6T14][GaBr whose crystal structure was also solved. This complex is actually a skeletal framework of tetrameric T14(GaBr),with one or two mesitylene molecules alternately coordinated to the Tl+ cations. In the MesTl+ unit, the Tl is located directly above the face of the ring, implying 6 coordination. A similar structure was also found for the gallium analogue (55) [Mes)4Ga]GaC1]indicating that, although weak, it-interactions do exist with univalent Group 13 metals. It is significant to note that in the above studies it was found that the complex salts T1A1C14,TIA1Br4 and T1GaC14all dissolved in hot benzene to give colorless crystals upon cool ing. However, when dried under a stream of dry N2, the benzene content of the crystals steadily decreased, implying very weak it coordination of the benzene ligand to the Ti center. Apparently, only the mesitylene ligand afforded a dry, isolable product. Strauss et al. (56) have also synthesized TI-Mes complexes such as [Tl(OTeF5)(Mes)2},but found that this compound also loses all traces of the weakly bound arene ligands under N2 atmosphere or in vacuo. There fore, it appears that the high lability of these arene ligands makes synthesis and purification a very delicate task, even though the synthesis itself is relatively simple (i.e., arene addition to a suitable substrate). It is apparent from the above studies that the electronic configurationnd10(n+1)s2plays a particularly important role in the metal’s affinity for arene ligands. In view of this, the next logical step is to explore the divalent Group 14 elements (which have the same electronic configuration) to investigate their affinities for arene ligands. In fact, the synthesis and structure of the arene M(II) complexes(C6H)M(AlCl42(C6H),M=Pb(1I) or Sn(II), have been reported by Amma and co-workers (57,58). The X-ray crystallographic study of the compounds reveals a polymeric chain structure, with a benzene ring wealdy bound in a manner to the M(II) center. In this thesis, the synthesis and characterization of arene adducts of divalent tin fluoro deriva tives which contain the weakly coordinating SbF6 and SO3F ligands will be considered. 15 1.5 Magnetic Measurements Magnetic susceptibility measurements, when recorded as a function of temperature, provide primarily information regarding the nature of the ground state of the paramagnetic ions. The extent of the information that can be obtained is related to the accuracy of the measure ments, the range of temperature over which the measurements are carried out, as well as to the purity of the samples used. The magnitude and temperature dependence of the magnetic moment data of magnetically dilute systems are determined by several factors such as ground state occupancy and degeneracy, crystal field symmetry, spin-orbit coupling and electron delocalization effects. The theory of magnetic susceptibilities of paramagnetic molecular species and transition metal complexes is well covered in several texts (59). In magnetically concentrated systems, the factors mentioned above are also present, and any magnetic inter action is superimposed upon these single-ion phenomena. Furthermore, it is now recognized that magnetic exchange interactions are not at all uncommon in inorganic compounds, even where magnetic dilution (usually at room temperature) appears at first sight to be dominant. It is for this reason that magnetic susceptibility measurements should be performed, where possible, over as wide a temperature range as possible. Qualitatively, magnetic exchange interactions may be thought of as arising from unpaired spin densities on neighbouring paramagnetic centers, being aligned either parallel or anti-parallel to each other, resulting in ferromagnetism or antiferromagnetism respectively. Magneto-structural relationships emphasize the importance of such factors as the stereochemistry around the paramagnetic center, the efficiency of orbital overlap which may be direct or via a superexchange pathway, the geometry of the bridging anions, the types of substituents on the bridging group and the nature of any nonbridging groups. The theoretical aspects of magnetic exchange phenomena and the models used to interpret the empirical data have been the subject of extensive investigations (60). 16 The main group molecular complexes and several of the transition metal derivatives studied in this work for their magnetic properties contain unusual paramagnetic cations in the solid state, which are stabilized by weakly basic fluoroanions. Nearly all the previous magnetic measurements on the molecular cations and the divalent sulfonates have been carried out at higher temperatures only, where in most instances, magnetic exchange interactions are not detected. The halogen cations I2 and Br2 are shown to have 2H3g ground states with the first excited states21I2g at approximately 5100 and 2800 cmt above the ground state, respectively, both in the gas phase (61) and in solution (15). The magnetic moment in both cases is expected to be independent of temperature due to the absence of thermally accessible excited states, and predicted to have a value of 2.0 B at higher temperatures, in analogy to the 1’eff of the NO molecule, discussed by van Vieck several years ago (62): Peff = 2[(1 — e + xe”)/(x + xe)]1/2 [1.11] where x = VkT. = Spin-orbit coupling constant k = Boltzmann’s constant T = Absolute temperature In an earlier report, Gillespie and Mime (63) showed the magnetic moment of in a solution of HSO3F to be 2.0 ± 0.1 Kemmitt et al. (64) reported a magnetic susceptibility study on (SbF5)21, a material thought to contain the ‘2 cation, and found magnetic moments that ranged from 2.25 B at room temperature to 2.05 B at 100 K. Considering the uncertainty in the chemical composition of the material, not much significance could be attached to these values. A later work from our group, where theI2[SbF1i] and [Br2][Sb3F16Jcompounds 17 were synthesized by an improved method (15), indicated the possible existence of antiferro magnetic interaction between the ‘2 cations, with jj values of 2.15 and 1.68 B at 295 and 81 K respectively. Furthermore, the Br2 species was found to be magnetically dilute in the temperature range 297 to 80 K, and a magnetic moment close to 2.0 B was obtained for the compound (15). This is in contrast to a previous study, where a room tempreature value of 1.6 11B had been reported for the Br2 derivative (17). In contrast to the and Br2 cations, the dioxygenyl cation 02 (with ground state 21f2g and first excited state2113ag at —1480 cmt above ground state) investigated in this work has been the subject of several previous magnetic studies (65-68). The 02+ cation can be prepared by a variety of methods, and is stabilized in the solid state by various fluoroanions, leading to complexes like02[PtF6] (69), O2[BF4] (70), O2[AsF6] (71), and O2[SbF6] (71). The magnetic behavior ofO2[PtF6} over the temperature range 77-298 K has been postu lated to be similar to that of NO, and the magnetic moment was reported as 1.57 1.LB at room temperature (65). A magnetic moment of 1.66 B has been found for theO2[SbF6]-compound (68b), and a value of 1.70 B forO2[BF4J (68a). Two previous studies onO2[AsF6] down to 4 K give contradictory results. The study by Grill et al. (67) indicates no magnetic ordering of the 02+ cations down to 4 K, whereas weak O2O2+ interaction is suggested in the work of DiSalvo et al. (66). Even though in all the above studies the eff values reported for the O2 salts fall well below the spin only value of 1.73 B’ no satisfactory reason has been given so far to explain this curious phenomenon. It appears, however, that sample purity and identity play an important role in the interpretation of magnetic data of the O2 salts. Several transition metal sulfonate derivatives studied here for their magnetic properties contain paramagnetic cations in unusual coordination environments. As in PdF2 (48), in 18 Pd(SO3F)2(36) and Pd(SO3CF)2(30) compounds, the Pd(ll) ions (d8) with 3A2g ground states are located in octahedral environments, a situation found only in some of their cationic and anionic derivatives. Similarly, the two silver derivatives Ag(SO3F)2(37) and Ag(SO3CF)2 (31) investigated in this work have remained, in addition to the AgF2 compound (72), the only simple binary compounds of Ag(I1) with a d9 configuration. Previous magnetic susceptibility measurements down to —80 K on Pd(SO3F)2, “Pd(SO3F)” and Ag(SO3F)2 indicated that these fluorosulfate derivatives were relatively magnetically unconcentrated in that temperature range with Peff values of 3.34, 3.45 and 1.92 B respectively (36,37). Furthermore, their susceptibilities followed the Curie-Weiss law with positive Weiss constants. The trifluoromethylsulfate derivative Ag(SO3CF)2was found to be an antiferromagnetic compound, with Xmax at —138 K (31). Only two other binary fluorosulfate and trifluoromethylsulfate derivatives, Ir(SO3F)4 (73) and Fe(SO3CF)(45), are known up to now to be magnetically concentrated. However, unlike in Ag(SO3CF)2,the magnetic moments obtained for Ir(SO3F)4are only slightly below the calculated values, suggesting weak antiferromagnetic coupling down to —80 K (73). The Fe(SO3CF)compound is magnetically more concentrated than the iridium species, exhibiting magnetic moments which are significantly less than the expected values, and which also decrease with decreasing temperature, although no maximum is detected in the susceptibility curve down to —80 K (45). As in the cases of palladium and silver derivatives mentioned above, no low temperature magnetic data are available for these complexes. The antiferromagnetic coupling observed in all the above mentioned sulfonates is not unusual for magnetically concentrated transition metal fluoro compounds, since antiferro magnetism, rather than ferromagnetism, appears to be the more common type of magnetic ex change interactjon among the majority of these compounds (59). Moreover, in most instances 19 these antiferromagnetic transition metal fluoro derivatives contain small monoatomic ligands (59b,c), whereas the sulfonate coordination polymers studied in this work are composed of the much larger polyatomic S03F and SO3CFç ligands. 1.6 Objectives of this Study The research work presented in this thesis can be categorized into two general, but inter related sections: (A) Synthesis and characterization of divalent metal coordination polymers, and (B) Magnetic susceptibility studies on unusual main group and transition metal cations. The specific types of research relevant to these two sections are summarized below. (A) Synthesis and Characterization (i) The solvolysis of main group fluorosulfate derivatives in the Lewis acid SbF5 was extended in the present study to include the transition metal fluorosulfates according to: 25-60°C M(SO3F)2 + 6SbF5 > M(SbF + 2SbF9( O3) [1.12] SbF5 Where M = Ni, Pd, Cu, Ag or (Sn) The ready availability of the M(SO3F)2precursors, the low oxidation potential of SbF5, and the easy removal of the volatile byproduct Sb2F9(SO3), as well as mild reaction conditions and the possibility of using glass vessels as reactors were all positive factors in choosing the above solvolysis preparation. The divalent metal hexauluoro anlimonates obtained were characterized, where appropriate, by elemental analysis, vibrational, electronic and 119Sn Mössbauer spectra, and magnetic susceptibility measurements. 20 (ii) The favorable properties of 1 ,3,5-trimethylbenzene (mesitylene) as a ii-adduct ligand and the possible participation of the lone pair electrons of the Mössbauer nuclide Sn in the +2 state in bonding by being donated to the antibonding it’ ligand orbitals are contributing factors in the syntheses of the post-transition metal Sn(ll) arene adducts according to: 25°C Sn(SO3F)2+ mes > Sn(SO3F)2mes [1.131 mesitylene 25°C Sn(SbF6)2+ 2 mes > Sn(SbF6)2mes [1.14] mesitylene (mes = mesitylene) The adducts isolated were characterized by elemental analysis, infrared and tt9Sn- Mössbauer spectra. (B) Magnetic Susceptibility Studies (i) Low temperature magnetic susceptibility measurements down to —4 K were performed on02[AsF6j,Br2[SbFi& andI2[SbF1i]’ in order to compare the magnetic behavior of the three cations. These paramagnetic homonuclear derivatives seem to be the only three stable and isolable cations formed by non-metals that are suitable for solid state magnetic studies. The two dihalogen cations 12+ and Br2+ were investigated for possible exchange interactions, and the dioxygenyl cation 02+ was studied here to explain its observed low magnetic moment values. Furthermore, the application of van Vleck’s theory of molecular paramagnetism to solid state cations was also examined. (ii) Several Group 10 and Group 11 divalent transition metal sulfonate compounds were considered in this study as likely materials to exhibit magnetic exchange interations. 21 The metal fluorosulfates Pd(SO3F)2,“Pd(SO3F)”and Ag(SO3F)2,previously described as relatively magnetically dilute down to —80 K, were re-investigated down to —4 K for their low temperature magnetic properties. The nickel(ll) fluorosulfate, Ni(SO3F)2was, however, measured in the temperature range —291 to 2 K. The corresponding trifluoromethylsulfate derivatives Ni(SO3CF)2,Pd(SO3CF)2and Ag(SO3CF)2were similarly studied for their variable temperature magnetic susceptibilities. The last two compounds were measured at low temperatures, whereas the Ni(SO3CF)2complex was studied in the extended temperature range —292 to 2 K. References 1.a) G.A. Olah, G.K.S. Prakash, and J. Sommer, “Superacids”, John Wiley and Sons, New York, 1985. b) T.A. O’Donnell, Chem. Soc. Rev., i. 1 (1987). 2. P. Hagenmuller (Ed.), “Inorganic Solid State Fluorides”, Academic Press, New York, 1985. 3. G.A. Lawrence, Chem. Rev., , 17 (1986). 4. R.J. Gillespie and M.J. Morton, Ouart. Rev. Chem. Soc., 25, 553 (1971); and M.T.P. International Review of Science, Inorg. Chem., Ser. 1, Vol. 3, Butterworths, London, 1972. 5. J. Shamir in “Structure and Bonding”, Eds. J.D. Dunitz, J.B. Goodenough, P. Hem merich, J.A. Ibers, C.K. Jorgensen, J.B. Neilands, D. Reinen, and R.J.P. Williams, vol. 37, Springer-Verlag, Berlin, pp 141-210, 1979. 6. S.P. Mallela, S. Yap, J.R. Sams, and F. Aubke, Inor. Chem., 25, 4327 (1986). 7. P.-L. Fabre, J. Devynek, and B. Tremillon, Chem. Rev., , 591 (1982). 8. A.A. Woolf and N.N. Greenwood, J. Chem. Soc., 2200 (1950). 22 9. E.E. Ainsley, R.D. Peacock, andP.L. Robinson, Chem. md., 1117 (1951). 10.a) E.L. Muetterties and W.D. Phillips, Adv. Inorg. Chem. Radiochem., 4, 234 (1962). b) G.J. Hoffman, B.F. Holder, and W.L. Jolly, 3. Phys. Chem., 62, 364 (1958). 11. A.J. Edwards and P. Taylor, Chem. Comm., 1376 (1971). 12. P.A. Yeats and F. Aubke, J. Fluorine Chem., 4, 243 (1974). 13. W.W. Wilson, J.R. Dalziel, and F. Aubke, J. Inorg. Nucl. Chem., 2. 665 (1975). 14. W.W. Wilson and F. Aubke, J. Fluorine Chem., fl, 431 (1979). 15. W.W. Wilson, R.C. Thompson, and F. Aubke, Inor. Chem., j, 1489 (1980). 16. S.P. Mallela, S. Yap, J.R. Sams, and F. Aubke, Rev. Chim. Minerale, 2, 572 (1986). 17.a) A.J. Edwards, G.R. Jones, and R.J.C. Sills, Chem. Comm., 1527 (1968). b) A.J. Edwards and G.R. Jones, J. Chem. Soc., A2318 (1971). 18. C.G. Davies, R.J. Gillespie, S.R. Ireland, and J.M. Sowa, Can. 3. Chem., 2048 (1974). 19.a) F. Seel and 0. Detmer, Z. Anorg. Aug. Chem., , 113 (1959). b) A.J. Edwards and R.J.C. Sills, J. Chem. Soc., A2697 (1970). 20.a) A.J. Edwards and G.R. Jones, Chem. Comm., 1304 (1967). b) A.J. Edwards and G.R. Jones, 3. Chem. Soc., A 1467 (1969). c) K.0. Christe and C.J. Schack, Inorg. Chem., 2 2296 (1970). 21.a) M. Schmeisser and W. Ludovici, Z. Naturforsch., 602 (1965). b) M. Schmeisser, W. Ludovici, D. Naumann, P. Sartori, and E. Scharf, Chem. Ber., 101, 4214 (1968). 22. K.O. Christe and W. Sawodny, Inorg. Chem., j.2. 2879 (1973). 23.a) R.J. Gillespie and G.J. Schrobilgen, Inorg. Chem., j3, 1230 (1974). b) K.O. Christe and R.D. Wilson, Inorg. Chem., .1.4, 694 (1975). 24. F.A. Hohorst, L. Stein, and E. Gebert, Inorg. Chem., 14, 2233 (1975). 25. D. Gantar, I. Leban, B. Friec, and J.H. Holloway, J. Chem. Soc. Dalton Trans., 2379 (1987). 23 26. P.A.W. Dean, J. Fluorine Chem., , 499 (1975). 27. K.O. Christe, W.W. Wilson, R.A. Bougon, and P. Charpin, J. Fluorine Chem., 34, 287 (1987). 28. B.D. Cutforth, C.G. Davies, P.A.W. Dean, R.J. Gillespie, P.R. Ireland, and P.K. Ummat, Inorg. Chem., 12, 1343 (1973). 29. L. Kolditz, Adv. Inorg. Chem. Radiochem, , 1 (1965). 30. S.P. Mallela, J.R. Sams, and F. Aubke, Can. J. Chem., 3, 3305 (1985). 31. P.C. Leung, K.C. Lee, and F. Aubke, Can. J. Chem., 2. 326 (1979). 32. F.B. Dudley and G.H. Cady, J. Am. Chem. Soc., 2. 513 (1957). 33. R.A. DeMarco and J.M. Shreeve, Adv. Inorg. Chem. Radiochem., j., 109 (1974). 34. A.A. Woolf, J. Chem. Soc., A355, (1967). 35. C.S. Alleyne, K.O. Mailer, and R.C. Thompson, Can. J. Chem., , 336(1974). 36. K.C. Lee and F. Aubke, Can. 3. Chem., , 2473 (1977). 37. P.C. Leung and F. Aubke, Inorg. Chem., 1.2, 1765 (1978). 38. K.C. Lee and F. Aubke, Inorg. Chem., 3, 2124 (1984). 39. K.C. Lee and F. Aubke, Inorg. Chem., 31. 389 (1979). 40. W.M. Johnson, R. Dev, and 0. Cady, Inorg. Chem., 11, 2260 (1972). 41. F. Aubke and R.J. Gillespie, Inorg., Chem., 7, 599 (1968). 42. H. Wiliner, S.J. Rettig, J. Trotter, and F. Aubke, Can. J. Chem., 69, 391 (1991). 43. H. Wiliner, F. Mistry, 0. Hwang, F.G. Herring, M.S.R. Cader, and F. Aubke, 3. Fluorine Chem., 5, 13 (1991). 44. S.P. Mallela and F. Aubke, Can. J. Chem., , 382 (1984). 45. J.S. Haynes, J.R. Sams, and R.C. Thompson, Can. J. Chem., 59, 669 (1981). 46. A.L. Arduini, M. Garnett, R.C. Thompson, and T.C.T. Wong, Can. J. Chem., 53, 3812 (1975). 47.a) D.C. Adams, T. Birchall, R. Faggiani, R.J. Gillespie, and J.E. Vekris, Can. J. Chem., , 2122 (1991). 24 b) F. Charbonnier, R. Faure, and H. Loiseleur, Acta Crvst, 1478 (1977). 48. N. Bartlett and P.R. Rao, Proc. Chem. Soc., 393 (1964). 49. K.C. Lee and F. Aubke, Can. J. Chem., , 2058 (1979). 50. G.A. Olah and T. Ohyama, Synthesis, 5, 319 (1976). 51. R.E. Noftie, Inorg. Nuci. Chem. Lett., j, 195 (1980). 52. H. Schmidbaur, Angew. Chem. mt. Ed. Engi., 24, 893 (1985). 53. H. Schmidbaur, W. Bublak, B. Huber, and G. Muller, Angew. Chem. mt. Ed. Engi., ., 26 (1987). 54. H. Schmidbaur, W. Bublak, I. Riede, and G. Muller, Angew. Chem. mt. Ed. Engi., 5, 24, (1985). 55. H. Schmidbaur, U. Thewalt, and T. Zafiropoulos, Z. Naturforsch, 1642 (1984). 56. S.H. Strauss, M.D. Nairot, and O.P. Andersen, Inorg. Chem., 25, 3850 (1986). 57. E.L. Amma, P.F. Rodesiler, and M.S. Weininger, Inorg. Chem., jj, 751 (1979). 58. P.F. Rodesiler, Th. Auel, and E.L. Amma, J. Am. Chem. Soc., 97,7405 (1975). 59.a) R.J. Myers, “Molecular Magnetism and Magnetic Resonance Spectroscopy”, Prentice Hall Inc., Englewood Cliffs, New Jersey, 1973. b) E.A. Boudreaux and L.N. Mulay (Ed.), “Theory and Applications of Molecular Paramagnetism”, John Wiley and Sons, New York, 1976. c) R.L. Carlin, “Magnetochemistry”, Springer-Verlag, Berlin, 1986. 60.a) W.E. Hatfield, W.E. Estes, W.E. Marsh, M.W. Pickens, L.W. ter Haar, and R.R. Weller in “Extended Linear Chain Compounds”, Ed. J.S. Miller, Vol. 3, Plenum Press, New York, pp. 43-50, 1983. b) A.P. Ginsberg, Inorg. Chim. Acta. Rev., 5, 45 (1971). c) C.J. O’Connor, Prog. Inorg. Chem., 22. 203 (1982). 61. A.B. Cornford, D.C. Frost, C.A. McDowell, J.L. Ragle, and l.A. Stenhouse, J. Chem. Phys., M 2651 (1971). 25 62. J.H. van Vieck, “Electric and Magnetic Susceptibilities”, Oxford University Press, London, 1932. 63. R.J. Gillespie and J.B. Mime, Inorg. Chem., , 1577 (1966). 64. R.D.W. Kemmitt M. Murray, V.M. McRae, R.D. Peacock, M.C.R. Symons, and T.A. O’Donnell, I. Chem. Soc., A862 (1968). 65. N. Bartlett and S.P. Beaton, Chem. Comm., 167 (1966). 66. F.J. DiSalvo, W.E. Falconer, R.S. Hutton, A. Rodriguez, and J.V. Waszczack, J. Cheni Phys., 2575 (1975). 67. A. Grill, M. Schieber, and 3. Shamir, Phys. Rev. Lett., 747 (1970). 68.a) V.1. Belova, V. Ya. K. Syrkin, D.V. Bantov, and V.F. Sukhoverkhov, Russ. J. Inorg. Chem., (Engi. Trans.), j.., 772 (1971). b) V.1. Belova, V. Ya. Rosolovskii, and E.K. Nikitina, Russ. 3. Inorg. Chem., (Engi. Trans.), .j, 772 (1971). 69. N. Bartlett and D.H. Lohmann, J. Chem. Soc., 5253 (1962). 70. J.N. Keith, 1.3. Solomon, I. Sheft, and H.H. Hyman, Inorg. Chem., 7, 230 (1968). 71. D.E. McKee and N. Bartlett, Inorg. Chem., U 2738 (1973). 72. P. Fischer, G. Roult, and D. Schwarzenbach, 3. Phys. Chem. solids, 32, 1641 (1971). 73. K.C. Lee and F. Aubke, J. Fluorine Chem., i, 501 (1982). 26 CHAPTER 2 GENERAL EXPERIMENTAL 2.1 Introduction This chapter will deal with general experimental techniques, chemical sources, purifica tion procedures and the syntheses of starting materials used in this study. Specific synthetic procedures will be described in the appropriate chapters. Since most of the compounds involved in this work are extremely hygroscopic, they had to be handled in an environment free of moisture. Hence, standard vacuum line techniques were employed for the transfer of volatile liquids, and less volatile liquids and solids were manipulated inside an inert atmosphere dry box. All reactions were performed inside well ventilated fumehoods. Reactions were monitored by weight where possible, and the removal of volatile materials in vacuo was usually done at room temperature. However, where liquids with low vapor pressures like SbF5 or HSO3Fwere involved, elevated temperatures had to be used even under vacuum conditions. Fluorolube grease type 25-1OM (Halocarbon Corporation) was used to lubricate ground glass connections to maintain vacuum tight conditions. 2.2 Chemicals Some chemicals were used without purification as received, and these are listed in Table 27 2.1, along with their sources and purities. The other chemicals used were purified or synthesized according to the methods described below. Table 2.1: Chemicals Used Without Purification Chemical Source Purity (%) Ag, -100 mesh Alfa 99.95 Au, -20 mesh Alfa 99.99 Pd, -60 mesh Alfa 99.95 Pt, -60 mesh Alfa 99.90 Sn, -100 mesh Alfa 99.99 ‘2 Fisher 99.9 CuF2 Alfa 99.5 AgF2 Aldrich 98.0 SnF2 Matheson 98.0 AgSbF6 Aldrich 98.0 AsF5 Ozark Mahoning reagent grade HF Matheson reagent grade SnC12 BDH reagent grade KJ Fisher 99.95 CaH2 BDH reagent grade NaC5H Aldrich 2.0 M solution in THF 28 2.2.1 Purification Methods (a) SbF5, obtained from Ozark Mahoning, was purified (1) first by purging the crude liquid of most HF by bubbling dry N2 through in a 500 mL two-necked Pyrex flask fitted with a gas inlet tube and Drierite guard tube. Subsequent purification was done by repeated distillation, first at atmospheric pressure in a stream of dry N2 and later in vacuo. (b) HSO3F, from Orange County Chemicals, was purified by double distillation in a Pyrex apparatus under a counter flow of dry N2 at atmospheric pressure (2). The constant boiling fraction at 162-163°C was collected into Pyrex storage vessels. (c) HSO3CF, from Aldrich, was purified by repeated vacuum distillation and stored in Pyrex vessels for synthetic use. (d) Br2, from Aldrich, was stored in a Pyrex vessel containing P205 to exclude moisture and KBr to remove Cl2. (e) Mesitylene (1 ,3,5-Trimethylbenzene), obtained from Aldrich, was dried over CaH2 for one week and distilled in vacuo prior to use. 2.2.2 Synthetic Methods (a) Bis(fluorosulfuryl) peroxide,S206Fwas prepared in one to two kilogram quantities by the reaction of SO3 and F2, using AgF2 as catalyst and N2 as carrier gas (3). The synthesis was carried out at —180°C in a flow reactor made of Monel metal (Figure 2.1). The crude liquid product was condensed into Pyrex vessels, cooled to -78°C with dry ice. 29 CTo F2 c yl in de r C ro sb y P re ss ur e G uo ge F 2 O ut le t To Fl ow m .t. r C op pe r 6 W hl t.y - f3- H ok . 41 3 V ol v. A ut oc la v. E ng ln ..r ln g V al v. s R ea ct or T ra p Oi l Tu be A B 25 C — 7C — 7W C CO LL EC TI ON V ES SE LS Fi gu re 2. 1: A pp ar at us fo r Pr ep ar in g S206F Most of the potentially explosive byproduct FSO3 was removed by intemittently warming the product to room temperature and cooling down to -78°C. Further purifica tion was achieved by pumping on the product overnight at -7 8°C to remove any residual FSO3. Unreacted SO3 was removed by extracting the crudeS206Fwith concentrated H2S04 in a separatory funnel. A product obtained by this route may contain a small amount of disulfuryl difluoride,S205Fwhich has no effect on the synthetic reactions, except where a stoichiometric amount of S206F is required. The purified colorless liquid was vacuum distilled into large (500-1000 mL) one-part Pyrex storage vessels equipped with Kontes Teflon stem stopcocks. The purity of the reagent was checked by both JR and 19F-NMR spectroscopy. (b) BrSO3F was synthesized by reacting Br2 with a slight excess ofS206F (4) in a long stem one-part Pyrex reactor. The excess S206Fwas required to remove any unreacted Br2 from the liquid product. The BrSO3F obtained by this method can be vacuum distilled directly from the Pyrex vessel when required. 2.3 Apparatus 2.3.1 Reaction Vessels One part glass Pyrex reactors of 25-100 mL capacity were used when solid products could be isolated by removing the volatiles in vacuo. These round bottom reactors were fitted with Kontes Teflon stem stopcocks and had side arms extending to B 10 ground glass cones (Figure 2.2(a)). If high pressures were anticipated during reactions, 3 mm thick-wall rather than 2 mm normal wall vessels were employed. To facilitate the isolation of products by filtration, two-part reactors made from 50-100 31 KO NT ES TE FL ON ST EM tJ (b ) Fi gu re 2. 2: Ty pi ca lP yr ex R ea ct io n V es se ls (a) O ne Pa rt R ea ct or (b) Tw o Pa rt R ea ct or B— 1O GR OU ND GL AS S CO NE B— 19 GR OU ND GL AS S JO IN T B— 1O GR OU ND GL AS S CO NE 50 — lO O m L BO 1I OM (a ) FL AS KS mL round bottom flasks were utilized (Figure 2.2(b)). A typical reactor consisted of a round bottom flask with a B 19 ground glass cone fitted with a “drip lip” to trap possible grease- contaminated liquids. The corresponding adaptor top had a Kontes Teflon stem stopcock between the B 19 socket and a B 10 cone. During reaction work-up, the adaptor could be substituted by an appropriate equipment - such as a vacuum filtration apparatus, seen in Figure 2.3. The design of the filter was adopted from the one described by Shriver (5). A Kel-F tubular reactor (Figure 2.4) was employed for syntheses involving liquid HF. The Kel-F tube (2 cm o.d. and 1.2 cm i.d.; obtained from Argonne National Laboratory, USA) was held by a Monel top adaptor equipped with a Whitey valve (type 1KS4-316) which could be fitted to a metal vacuum line. 2.3.2. SbF5 - Storage-bridge Vessel The purified liquid SbF5 was stored in a dual purpose one-part Pyrex container as shown in Figure 2.5. The two Kontes Teflon stem stopcocks and BlO ground glass joints made it feasible for the vessel to be attached to a glass vacuum line and a Pyrex reactor simultaneously. SbF5 could then be distilled directly from the storage vessel into the reactor via the side arm extension. 2.3.3 S206F- Addition Trap Where exact amounts of S206F had to be used, a 1.00 or 4.00 mL graduated pipet equipped with an overflow bulb and fitted on top with a Kontes Teflon stem stopcock was employed. The side arm of the trap ended in a BlO ground glass cone, which could be connected to a Pyrex T-shaped bridge for the transfer ofS206F. Determination of the precise amount ofS206Fused was obtained by weight difference. 33 Figure 2.3: Vacuum Filtration Apparatus B—1O GROUND GLASS CONE 50—lOOmI ROUND BOflOM F1.ASK KONTES TEFLON STEM STOPCOCK GLASS FRIT (MEDIUM POROSITY) B—19 GROUND GLASS JOiNT 34 C’2 S C.) S cp Figure 2.4: Kel-F Tubular Reactor Whitey Valve Monel Alloy Top Kel—F Tube Copper Ferrule Brass Nut m 2cm 35 Figure 2.5: SbF-Storage-bridge Vessel B—iC GROUND GLASS CONE KONTES TEFLON STEM STOPCOCK 8cm E 1) Lfl B—iC GROUND GLASS SOCKET 36 2.3.4 Pyrex Vacuum Line A general purpose glass vacuum line, consisting of five Kontes Teflon stem stopcocks with B 10 ground glass sockets, was used. The manifold was approximately 60 cm long, and a detachable safety trap, cooled with liquid N2, was placed between the manifold and the vacuum pump to protect the pump from volatile corrosive materials. Typical vacuum generated on this line was about 10.2 torr. 2.3.5 Metal Vacuum Line For the reactions involving liquid HF, a metal vacuum line was used. It was operated in a manner similar to that of the Pyrex line. The metal manifold was constructed using 6 cm o.d. Monel tubing equipped with Whitey valves (type 1KS4-316), and connected to a liquid N2 cooled safety trap via a Teflon adaptor. 2.3.6 Dry Atmosphere Box Hygroscopic solids and low volatility liquids were manipulated inside a Vacuum Atmos phere Corporation ?DriLabH Model DL-001 SG dry box, filled with K-grade N2 gas. The removal of moisture inside the dry box was accomplished by circulating the nitrogen over molecular sieves located in the “Dri-Train” Model HE-493. The molecular sieves were periodi cally regenerated by heating in a stream of 10% H2 mixed with N2. Fresh P205 was also kept inside the dry box to exclude any residual moisture. 37 2.4 Instrumentation and Methods 2.4.1 Infrared Spectroscopy Room temperature IR spectra were recorded using three types of spectrometers: (a) Nicolet 5DX Fr-IR, (b) Perkin Elmer 1710 FT-IR, and (c) Perkin Elmer 598. Samples were run on thin solid films pressed between AgCl or AgBr windows, with transmission ranges down to —400 and —250 cmt respectively. The high reactivity of the compounds studied precluded the use of mulling agents or other window materials. Samples were prepared inside the dry box and the spectra were recorded immediately after removing the samples from the box. Spectra of gaseous samples were recorded using a glass cell of 10 cm path length, fitted with AgBr windows and a Kontes Teflon stem stopcock. All spectra were calibrated with a polystyrene reference. 2.4.2 Raman Spectroscopy Raman spectra were obtained with a Spex Ramalog 5 Spectrophotometer equipped with a Spectra Physics 164 argon ion laser, using the 514.4 nm green line as the excitation wavelength. Solid samples were packed in the dry box into melting point capillaries, temporarily sealed with Fluorolube grease and then immediately flame-sealed. 2.4.3 Nuclear Magnetic Resonance Spectroscopy The FT-NMR spectra were obtained on a Varian XL-300 multinuclear spectrometer, with the following operating frequencies and external references: (a) 1H = 300 MHz, TMS; (b) = 282.23 1 MHz, CFC13. The solutions were either loaded into 5 mm NMR tubes inside the 38 dry box or, in the case of volatile liquids, transferred via a static vacuum into NMR tubes fitted with B 10 ground glass cones and then flame sealed. Low temperature spectra were recorded by cooling the probe with liquid N2 and controlling the temperature with a high precision thermo couple. 2.4.4 Electronic Spectroscopy Solid state electronic spectra in the near infrared and visible regions (4,000 to 30,000 cm-1) were recorded on a Cary 14 spectrophotometer. Samples were prepared in the dry box, and run as mulls in either SbF5 or Fluorolube oil as mulling agent in a Teflon cell fitted with quartz windows of 2.5 cm diameter. The absorbance level was adjusted with neutral density filters. Nujol mull spectra in the UV-visible range (12,000 to 50,000 cm4) were run on a Hewlett Packard 8452-A diode array spectrophotometer using the same Teflon Cell described above. 2.4.5 Mössbauer Spectroscopy The 1195n Mössbauer spectra were recorded on a constant acceleration spectrometer (6). Data counts were accumulated on a Tracer-Northern TN-1706 multichannel analyzer, linked to the mainframe computer via an IBM-PC. The y-ray source used was Ba9SnO3,and the Doppler Velocity Scale was calibrated with an iron foil absorber and a 57Co source. The chemi cal shifts were measured relative to Sn02. 2.4.6 X-ray Photoelectron Spectroscopy X-ray photoelectron spectra were recorded with a Varian IEE-15 spectrometer using Al Ka X-rays of 1486.6 eV energy. The detachable XPS sample probe was taken inside the dry 39 box where powdered samples were thinly dusted onto 2 cm length 3M Scotch tapes, which were wrapped around the sample slugs. The carbon 1S binding energy peak at 284.0 eV was used as a calibrant in all the measurements. The chemical shift is taken as the difference between the measured binding energy of a peak and that of the chosen standard atomic peak for a given energy level. The accuracy of the measurements is estimated to be ±0.1 eV. 2.4.7 Magnetic Susceptibility Measurements Variable temperature magnetic susceptibility measurements from —2.0-124 K were made using a Princeton Applied Research Model 155 Vibrating Sample Magnetometer, internally calibrated with ultrapure nickel (7). Temperature equilibration was obtained using a Janis Research Company Model 153 Cryostat and a Princeton Applied Research Model 152 Cryogenic temperature controller. Accurately weighed samples of --200-300 mg were loaded into Kel-F capsules inside the dry box and sealed with epoxy resin. Temperature measurements were taken with a chromel vs. Au-0.02% Fe thermocouple, located in the sample holder immediately above the sample. The thermocouple was calibrated using the known susceptibility vs. temperature behavior of tetramethylethylenecliammonium tetrachiorocuprate (II) and checked with mercury tetrathiocyanatocobaltate (II). From the scatter in data points of four different calibrations, the temperatures are estimated to be accurate to ±1% over the range studied. Thermocouple potentials were measured using a Fluke 8200A digital voltmeter. Magnetic fields of 7501, 9225, and 9625 G were employed, and set to an accuracy of 0.5%, and measured with a F.W. Bell Inc. Model 620 Gaussmeter. The accuracy of the susceptibility values is estimated to be ±1%. A Gouy balance, equipped with a Mettler AE 163 balance was used in the temperature range —80-300 K to measure the susceptibility of some compounds. Samples were packed in a 40 Pyrex tube containing an air-tight Teflon cap. Measurements were made in a nitrogen atmosphere at a field strength of 8000G and HgCo (NCS)4 was again used as a calibrant. The accuracy of the susceptibility values obtained by this method is estimated as ±5%. Corrections were made for the diamagnetic contribution of the holders and the molar susceptibilities were corrected for diamagnetism using the following values (8) (units of 10-6 cm3 mol-1): SO3F 40; S03CF 46; SbF6 80; Sb3F16-218; Sb2F11 149; AsF6 72; O2 7; Br2 61; 89; Ni2 11; Pd2 25; Pd4 18; Cu2 11; Ag2 24. 2.4.8 X-ray Powder Diffractometry X-ray powder diffraction patterns were obtained using a RU 200B series Rigaku Rotating Anode Diffractometer operating at 12 KW maximum power output. The Diffracto meter detected Cu Ka target radiation through a 200 urn nickel filter with a horizontal-type Nal Scintillator probe. A horizontal goniometer was used for the rotating anode. The diffractometer was interfaced with a DMaxIB computer system driven by an IBM PS/2. The peak-finding program was provided by Rigaku. Finely powdered samples were put on two sided 3M Scotch tapes attached to glass slides and protected from moisture by sealing the samples with plastic film. All preparations were carried out inside the dry box and samples were analyzed as soon as they were removed from the dry box. 2.4.9 Differential Scanning Calorimetry (DSC) DSC studies were performed using a Mettler DSC-20 cell interfaced with a Mettler TC1O TA processor and a Swiss Matrix printer. Finely powdered samples of approximately 2-4 g were sealed into aluminum pans and mounted on the measuring cell, under a dry N2 flow 41 rate of 50 mI/mm. The samples were scanned through a temperature range of 35 to 450°C with 4°C per minute increments. 2.4.10 Elemental knalyses Carbon, hydrogen and some of the sulfur analyses were performed by Mr. Peter Borda of this Department. All other elemental analyses were carried out by the Analytische Laboratorien, Gummersbach, Germany. References 1.a) W.W. Wilson and F. Aubke, J. Fluorine Chem., 13, 431 (1979). b) W.W. Wilson, R.C. Thompson, and F. Aubke, Inorg. Chem., j, 1489 (1980). 2. J. Barr, R.J. Gillespie, and R.C. Thompson, Inor. Chem., , 1149 (1964). 3.a) J.M. Shreeve and G.H. Cady, Inorg. Synth., 2, 124 (1963). b) G.H. Cady, Inorg. Synth., 11, 155 (1967). 4. F. Aubke and R.J. Gillespie, Inorg. Chem., 2, 599 (1968). 5. D.F. Shriver, “The Manipulation of Air Sensitive Compounds”, McGraw Hill Book Co., New York, 1969. 6. J.R. Sams and T.B. Tsin, Inorg. Chem., .14, 1573 (1975). 7. J.S. Haynes, K.W. Oliver, S.J. Rettig, R.C. Thompson, and J. Trotter, Can. J. Chem., 62, 891 (1984). 8. Landolt-Börnstein, Numerical data and functional relations in Science and Technology, Vol. 2, Magnetic properties of coordination and organometallic compounds, Springer- Verlag, Berlin, 1966. 42 CHAPTER 3 METAL(1I) HEXAFLUORO ANTIMONATES M(SbF6)2, M(ll) = Sn(J1), Ni(I1), Pd(II), Cu(ll) AND Ag(H) 3.1 Introduction Antimony(V) fluoride, SbF5, is commonly regarded as the strongest molecular Lewis acid (1). Conversely the anions SbF6 and the related Sb2F11-are extremely weak nucleophiles (2), capable of stabilizing a wide range of electrophilic cations, both in solid compounds and in HF-SbF5superacid solution (3). The objective of this study was the synthesis and characterization of SbF6 salts formed by divalent metal cations. The synthetic route chosen was the solvolysis of metal fluorosulfates in liquid SbF5 according to the general route: 25-60°C M(SO3F)2 + 6SbF5 > M(SbF + 2SbF9( O3) [3.1] with M = Sn, Ni, Pd, Cu, or Ag. Early work by Gillespie and Rothenbury (4), and a subsequent investigation by our group (5) into the antimony-fluoride-fluorosulfate system, form the basis for this synthetic reaction. The byproduct Sb2F9O3which is the most volatile component in this system, facilitates the ready removal of the SO3F- group from the reaction mixture. As mentioned previously in Chapter 1, initial uses of the solvolysis reaction in SbF5 have involved the stabilization of non-metallic cations, such as C102+ (6), a series of triatomic inter- 43 halogen cations (7), or the dihalogen cations Br2 and I2 (8) (see Chapter 4) in solid compounds. Use of this method in the synthesis of the first stage graphite salt C8SbF6 (9) and stabilization of the dimethykin(IV) cation as (CH3)2Sn(SbF11)2 (10) illustrate its versatility. There has been a variety of alternative methods used for the synthesis of hexafluoro antimonates of divalent metals, and two of the compounds discussed here (Ni(SbF6)2(11,12) and Sn(SbF6)2(13)) had been prepared when this study was started. All methods have the use of SbF5 in common, even though the routes vary. Metal oxidation either by elemental fluorine or by SbF5 in SO2 solution has allowed the synthesis of Ni(SbF&2(11,12) as well as Fe(SbF6)2 and Mn(SbF6)2(12). However, there are obvious limitations to this approach. Where higher oxidation states are accessible, direct fluorination may result in oxidation beyond the +2 state, unless the reaction conditions are well understood and carefully controlled. Oxidation by SbF5 fails for metal with higher oxidation potentials than provided for by the Sb(V)/Sb(III) couple (Appendix A-i). Quantitative separation of the reduced product, solid SbF35,may prove to be difficult, and mixed fluoride-hexafluoroantimonates like CoF(SbF6)(12) may form instead. Fluoride abstraction from MF2 by SbF5 in SO2 or anhydrous HF may be a more versatile synthetic method (13,14), but here another problem surfaces: the MF2 lattice could be incompletely broken up under the reaction conditions. This can lead to the formation of MF(Sb21i)’ a structural isomer of M(SbF6)2,and consequently chemical analysis becomes inconclusive. When moving from ionic SbF6 to coordinated SbF6 in these compounds, vibra tional spectra become more complex and are less readily interpreted. As a consequence, formulation as MF2SbF5 is employed in a recent publication for materials obtained in this manner (14). There is a fair amount of evidence for the existence of cations of the type [MFI+, [M2F3j, and [MF1 when AsF6 is used as a counter anion (15,16), and a crystal structure 44 obtained on AgF2s5(17) shows a chain-type cation [AgF] with AsF6 as counter anion. In AgF2Sb5however, a true Ag2 appears to be present and the crystal structure supports its formulation as Ag(SbF6)2(14). The advantages of using the M(SO3F)2-excess SbF5 solvolysis method as a viable synthetic route to metal(II) hexafluoro antimonates are summarized below: (i) M(SO3F)2precursors are readily available either from the solvolysis of SnC12 or NiC12, or the corresponding Ni(ll) or Cu(II) benzoates or other carboxylates in HSO3F(18), or by the use ofS206Fin the case of Pd(SO3F)2(19) and Ag(SO3F)2(20). (ii) In all instances further oxidation by SbF5 appears unlikely, since higher oxidation states for the metals cannot be achieved with the potential provided for only by the Sb(V)/Sb(ffl) couple. (iii) Vibrational spectra allow monitoring of the reaction by probing the absence of SO3F- vibrations. All the vibrational bands due to SbF6 appear below —750 cm1, whereas SO3Fstretching vibrations are present well above this value (see Appendix A-2). (iv) Reactions can be carried out under mild conditions in glass vessels and followed by weight. This also facilitates detection of any color changes that may occur during the reaction process. The reasons for the preparation and selection of these five M(SbF6)2compounds for this study, with M = Sn, Ni, Pd, Cu, or Ag are two-fold: (i) To avoid structural ambiguities of the type observed above, it became necessary to 45 obtain a 119Sn Mössbauer spectrum on a sample of Sn(SbF6)2prepared by a different route and to compare it to the spectrum reported previously (13). Furthermore, it was found previously that the SbF6 and Sb2F11 anions allow a close approach to the true (CH3)2Sn cation (10) on account of their low nucleophilicity. The same anions, it is felt, also permit the closest approach to a true Sn2 cation. In both instances, 119Sn- Mdssbauer spectroscopy can be effectively employed. (ii) Significant ferromagnetism was observed at low temperature in the fluorosulfates of Pd2 and Ag2 (Chapter 6, also ref. 21) with a weaker manifestation of this interesting coupling phenomenon seen also in the analogous Ni(SO3F)2 compound. However, surprisingly the Cu(SO3F)2compound is magnetically dilute to low temperatures (see Chapter 6). Therefore it is of interest to study the magnetic properties of the four cor responding transition-metal SbF6 derivatives at temperatures below 80 K. Since Ni(SbF6)2had been obtained previously by all three methods (11,12,14) discussed above, its formation by solvolysis in liquid SbF5 is viewed as a test case. A final point of inter est concerns the nickel(II) and palladium(ll) compounds. The Pd2 and Ni2 ions in the fluorosulfate are located in an octahedral coordination environment which is unusual for Pd2+, and as a result paramagnetic Pd(ll) and Ni(11) with 3A2g ground state is found in both Ni(SO3F)2 and Pd(SO3F)2. It is expected that similar paramagnetic ions are present in the corresponding Ni(SbF6)2and Pd(SbF6)2compounds as well. 3.2 Experimental 3.2.1 General Synthetic Scheme to M(SbF6)2 The fluorosulfates Ni(503F)2(18), Cu(SO3F)2(18), Sn(SO3F)2(18), Pd(SO3F)2(19), 46 and Ag(SO3F)2(20) were all synthesized according to published methods. A general synthetic method was applied to the solvolysis reactions in liquid SbF5. Approximately 500 mg of the M(SO3F)2compound was transferred inside the inert-atmosphere box into a preweighed reaction vessel, and -P10 mL of freshly purified SbF5 was added subsequently by vacuum distillation. The reaction vessel was warmed up, initially to room temperature, and subsequently to 50-60°C in an oil bath. Detailed temperatures and reaction times are given below. Only the reactions of Pd(SO3F)2 (purple —> light-grey) and Ag(SO3F)2(black-brown — yellow-green —> off-white) involved perceptible color changes of the solid reactant and the mixture remained heterogeneous throughout. The initially very viscous antimony(V) fluoride became less viscous after about 24 h and the mixture could be stirred effectively with a magnetic stirrer. Volatiles were removed in a dynamic vacuum. A sample of Ni(SbF6)2,made from Ni and SbF5 by fluorinating with F2 (11), was obtained from Dr. Karl 0. Christe of Rocketdyne, U.S.A. 3.2.2 Physical Properties and Analyses Both Ni(SbF6)2(11,12,14) and Sn(SbF6)2(13) are known compounds. Their identities were ascertained by weight and by infrared spectroscopy. Formation of both Ni(SbF6)2and Sn(SbF6)2required reaction times of 14 and 2 days at 60 and 50°C, respectively. The JR frequencies observed for Ni(SbF6)2are listed in Table 3.1. For Sn(SbF6)2the following JR bands were found (estimated intensities are in parentheses): 700 (s, sh), 678 (s), 648 (s), 620 (m), 595 (ms), 571 (ms), 520 (m, sh), 477 (m), 432 (w). 47 3.2.2a Cu(SbF6)2 Reaction time 10 days, reaction temperature 50°C, white hygroscopic solid that is thermally stable up to 210°C. Anal. Calcd. for CuSb2F12: Cu, 11.88; Sb, 45.51; F, 42.61%. Found: Cu, 11.65; Sb, 45.75; F, 42.47. Total: 99.87%. 3.2.2b Pd(SbF6)2 Reaction time 14 days, reaction temperature 50°C, light-grey solid, very hygroscopic and thermally stable up to 250°C. Anal. Calcd. for PdSb2F12:Pd, 18.62; Sb, 42.14; F, 39.45%. Found: Pd, 18.35; Sb, 41.90; F, 39.18. Total: 99.43%. 3.2.2c f-Ag(SbF6) Reaction time 10 days, reaction temperature 25°C, creamy white, hygroscopic solid, melts at 180-182°C to clear liquid. Anal. Calcd. for AgSb2F11:Ag, 19.25; Sb, 43.45; F, 37.25%. Calcd. for AgSb2F12:Ag, 18.62; Sb, 42.03; F, 39.35%. Found: Ag, 18.65; Sb, 42.30; F, 39.06. Total: 100.01%. 3.2.3 Alternate Synthetic Route to 3-Ag(SbF6)2 Two other preparative methods for 3-Ag(SbF6)2using the silver compounds AgSbF6and AgF2 as precursors are shown below: 48 a) 0.323 g of AgSbF6was allowed to react with an excess ofS206Fat room temperature. The color of AgSbF6 changed immediately from white to black-brown. The excess S206F was removed in vacuo and subsequently replaced by an excess of SbF5. The mixture was kept at room temperature for 2 days, then heated at 60°C for 2 h. A white solid of the composition Ag(SbF6)2was isolated by removing the excess SbF5 in vacuo and identified by its JR spectrum. b) 0.436 g (2.99 mmol) of AgF2 was loaded inside the dry box into a Kel-F reactor together with 4.50 g (20.8 mmol) of SbF5. About 6.5 mL anhydrous HF was distilled into this mixture in vacuo. After warming to room temperature under magnetic stirring, a light- yellow solid formed immediately. The solution’s initial color was light blue, which faded quickly. Removal of all volatiles yielded again a compound corresponding to the composition Ag(SbF6)2as a cream-colored solid. 3.3 Results and Discussion 3.3.1 Synthesis As noted previously (6-10), solvolysis reactions of fluorosulfates in a large excess of SbF5 proceed smoothly and frequently without any color change of the solid reactant. Hence, rather long reaction times are chosen to ensure complete conversion. Two indications that a reaction takes place are a noticeable decrease in the viscosity of SbF5 after about one day, and a slight increase in the vapor pressure above the reaction mixture. Both observations may be attributed to the formation ofSb2F9(SO3)(7). To reduce the viscosity even further in order to stir the heterogeneous mixture more effectively, slightly elevated reaction temperatures are chosen. Removal of all volatiles 49 proceeds easily in a dynamic vacuum, with the reaction flask at room temperature. It is noteworthy that after pumping overnight, the correct weight for M(SbF6)2 is obtained. Previous use of this synthetic method (6-8,10) had in all instances led to products where molecular cations are stabilized by Sb2F1f,and in one case (8) even by Sb3F16. For the spherical, less electrophilic M2+ cations, SbF6 appears to be the more suitable counter anion allowing formation of layered materials, as discussed below. In any event, neither product weights on isolation, nor chemical analyses, nor vibrational spectra give any indication of Sb2F1 containing intermediates or by-products. Of the resulting compounds, Ni(SbF6)2(11) and Sn(SbF6)2(13) are identified by their weights and their previously reported IR spectra. The 119Sn Mössbauer spectrum of Sn(SbF6)2 shows a single broad line (r’ = 1.35 mm s1) caused by unresolved quadrupole splitting. The isomer shift is found at 4.39 mm s1 relative to Sn02, in excellent agreement with the previously reported value (13). The solvolysis of Ag(SO3F)2is expected to lead to the recently reported blue form of Ag(SbF6)2since solutions of Ag2 in HF are deep blue in color. The X-ray diffraction study had revealed a true Ag2 ion in a distorted octahedral environment (14). However, in this study the course of the solvolysis and the final product obtained are unanticipated. The initial black brown color of Ag(SO3F)2quickly disappears and a greenish-blue solid slowly changes to yel low in color. Ultimately a cream-colored, diamagnetic solid is isolated, which melts at -480°C without decomposition to a clear, colorless liquid. The magnetic behavior, the color, and the observed lack of solubility in anhydrous HF (22) suggest the possible presence of univalent silver in the resulting yellow compound. An alternative formulation of the reaction product as AgSb2F11 is, however, not supported by the 50 weight change during reaction, a complete chemical analysis (see experimental section), the vibrational spectrum, or the chemical behavior of the material. Furthermore, with paramagnetic divalent silver well established in Ag(SO3F)2(20), the other reactant, the Lewis acid SbF5, can hardly be regarded as a reducing agent. A possible reductive decomposition of the reaction product appears unlikely, since besides excess SbF5,Sb29(SO3F)is the only volatile product formed during the reaction process. The chemical behavior of the white Ag(SbF6)2is not consistent with the presence of Ag in the reaction product. Unlike other silver(I) salts, the material will not undergo further oxidation by S206F in the presence or the absence of HSO3F. In contrast, AgSbF6 is immediately oxidized by S206Falone, to give a black-brown solid, with an increase in weight consistent with the composition Ag(SbF6)(SO3F). Subsequent solvolysis of this material in SbF5 proceeds in an identical manner to the solvolysis of Ag(SO3F)2,suggesting a viable alter native synthetic route to the white diamagnetic Ag(SbF6)2. Strong evidence for the presence of silver in an oxidation state higher than +1 comes from the hydrolysis of Ag(SbF6)2 in aqueous KI solution. The reaction proceeds very vigorously and ‘2 and 02 (where the latter evolves with rapid bubbling) are produced. Forma tion of 2 may in part be due to oxidation by Sb(V) in an acidic environment (23), but evolution of 02 can only be caused by Ag(II) or Ag(llI) ((24), also Appendix A-i): 2Ag + H20 —÷ 2Ag + 2W + 1/202 [3.2] and Ag2 + 2K1 — AgI + 1/212 + 2K [3.3] In summary, all evidence suggests that the material obtained in the solvolysis of 51 Ag(SO3F)2in SbF5 is J3-Ag(SbF6)2,a valence isomer of the previously reported a-Ag(SbF6)2 (14). Attempts were made to synthesize the a-form using the published method (14), where AgF2was reacted with SbF5 in anhydrous HF according to: HF AgF2 + 2SbF5 > Ag(SbF [3.41 These syntheses are only partialiy successful if a slight excess of SbF5 (ratio SbF5:Ag2= 2.32) over the stoichiometric amount is used. Even here a fair amount of solid -Ag(SbF6)2forms in addition to a blue solution, from which crystalline a-Ag(SbF6)2is obtained by slow evaporation of the volatiles. When a larger excess of SbF5 is used (SbF5:Ag2= 6.01), -Ag(SbF62is the only product. The blue color of the HF solution quickly fades within an hour. Removal of the HF and the excess SbF5 yields pure 3-Ag(SbF6)2. It seems that formation of a-Ag(SbF6)2in HF represents an acid-base titration with both the cation Ag2(50l)and the final product a-Ag(SbF2stable in, and isolable from, anhydrous HF. Similar behavior is reported for all other transition-metal hexafluoro antimonates of the M(SbF6)2or MF2SbF5type (14). On the other hand, p-Ag(SbF6)2is insoluble in anhydrous HF, suggesting structural differences. The observed diamagnetism of the compound is best explained by assuming a mixed- oxidation-state compound of the composition Ag(I)Ag(Ill)(SbF64.To account for the diamag netism, Ag3 should be in a square planar, or at least in a tetragonally elongated, octahedral environment while Ag+ would be located in a tetragonally compressed, nearly linear coordina tion environment. Argentic oxide, AgO, represents a precedent, according to neutron diffraction studies on this compound (25). The black-brown color, also found for the recently reported 52 oxides Ag203 (26) and Ag304 (24), is seemingly common to binary oxides and oxysalts of di- or trivalent silver, e.g. the sulfonates Ag(SO3F)2(20) or Ag(SO3CF)2(27), and may possibly be due to a charge transfer transition (28). In contrast, alkali metal salts containing the [AgF4] ion are reported to be yellow (25,29), while AgF2 and many Ag(ffl) fluoro derivatives are blue (30), just like a-Ag(SbF6)2(14). It appears, then, that all the observations mentioned above support formulation of the [3-form as Ag(I)Ag(Ill)(SbF64;however, the inability to oxidize Ag further withS206Fis not consistent with an ionic formulation as silver(I) tetrakis(hexafluoroantimonato)argentate(Ill), Ag[Ag(SbF6)4]. No report has been published so far of any other valence isomeric pair of the Ag(II) vs. Ag(I)Ag(ITI) type. The above mentioned AgO is reported to be Ag(I)Ag(llI)02(25), but a true Ag(ll)O appears to be still missing. However, the crystal structure of Ag304 (24) shows both Ag(ll) and Ag(III) in square planar coordination sites, and mixed oxidation-state compounds of the type Pd(II)Pd(IV)X6,with X F (31) or SO3F (19), are known for palladium, the neighbor ing element in the 4d series, with Pd(llI)X3so far unknown. Little is known, at least with regard to structure, about the paramagnetic red-brown compound AgF3 (32). The mixed valency for mula Ag(II)Ag(IV)F6has been proposed but the product may still contain some impurities (33). Observations made during the course of this study suggest that a-Ag(SbF6)2may be irreversibly converted to the [3-valence isomer. This conversion occurs at room temperature, if SbF5 is present in an excess, either in the presence or absence of HF. It is therefore not surpris ing that all the attempts made to synthesize pure a-Ag(SbF6)2from AgF2 and SbF5 in anhydrous HF (14) produce [3-Ag(SbF6)2as well, as an insoluble precipitate in addition to the blue solution. Removal of the solvent at -7 8°C in a dynamic vacuum affords a mixture of the two valence isomers. 53 Heating the isomeric mixture allows transformation to the pure [3-form. Following this conversion by differential scanning calorimetry indicates two endothermic events, a sharp peak at 100°C, and a broad peak at 139-140°C. The latter peak coincides with the melting point, where a colorless liquid forms. The [3-form melts at —180°C, consistent with differential scan ning calorimetry. Absence of the peak at 100°C suggests interpretation of this thermal event as a phase transition from a-Ag(SbF6)2to [3-Ag(SbF6)2. It seems likely that electron transfer between two Ag2 ions to give Ag and Ag3 is mediated by an antimony(V) fluoro species (SbF5 or SbF6). Considering the observations made regarding the relative stability of the two valence isomers, it is surprising that [3-Ag(SbF6)2has not been reported while the a-form has. The solvolysis of Pd(SO3F)2in SbF5, on the other hand, does not appear to change the electronic structure of the Pd2 ion (31,34). The electronic spectra and magnetic susceptibility data clearly show that in Pd(SbF6)2,the palladium ion, like in its parent compound, remains as a paramagnetic Pd2 species. However, in Cu(SbF6)2,the solvolysis product of Cu(SO3F)2, identification of all the copper ions as purely divalent may not be possible. Magnetic measure ments taken on Cu(SbF6)2also indicate, as in the case of [3-Ag(SbF6)2,mixed valency although to a more limited extent. This is discussed in more detail in Section 3.3.4, which deals with magnetic studies. It is of interest to note here that this solvolysis behavior leading to unexpected mixed valent products is exhibited by the two transition metal precursor compounds Cu(SO3F)2and Ag(SO3F)2with HJahnTeller ions”, i.e. Cu2 and Ag2 (d9). 54 3.3.2 Vibrational Spectra The infrared spectra obtained for Ni(SbF6)2,Pd(SbF6)2,Cu(SbF6)2,and -Ag(SbF6)2, together with the Raman spectra for the last two compounds, are summarized in Tables 3.1 and 3.2. Also included are the previously reported Raman spectra for Ni(SbF6)2 (11) and a-Ag(SbF6)2(14) as well as an approximate band description, also proposed previously (11). However, this description pertains only to the Ni2,Pd2, and Cu2 compounds in Table 3.1. The Raman spectrum of [3-Ag(SbF6)2is illustrated in Figure 3.1. Attempts to obtain a Raman spectrum of Pd(SbF6)2result in partial sample decomposition. Apparently the excitation line of the Ar laser falls within ii2 of the electronic spectrum of Pd(SbF6)2. Agreement with the JR and Raman spectra previously reported for Ni(SbF6)2(11) is very good with only a minor exception: an JR band at 585 cm1 attributed to out of phase for bridging SbF3 group is not observed in this study. This band does not have a counterpart in the Raman spectrum, and may be spurious. The band description in Table 3.1 has been amended slightly. There is only a partial comparison possible with the vibrational spectra reported for NiF2SbF5(14). The limited number of bands listed, six Raman and four JR bands with five non-coincidences, suggests an incomplete listing. Nevertheless, the Raman bands reported for NiF2•2SbF5 (14) are all observed with similar relative intensities in the Raman spectrum of Ni(SbF6)2(11). Since the reported X-ray powder data have the principle lines in common (11,14; see also Appendix A-3), it is reasonable to conclude that the two compounds are identical. This is possibly not the case for Cu(SbF6)2reported here, and CuF2Sb5(14). Very intense Raman and IR bands at 725 cm1 (see Table 3.1) are apparently not observed for CuF2Sb5 (14). However, the case for polymorphism as evidenced by magnetic measure ments (see Section 3.3.4) is by no means as strong as in the case of the two forms of Ag(SbF6)2. 55 Ta bl e 3. 1: V ib ra tio na l S pe ct ra o fN i( Sb F 6)2, P d (S bF 6)2, a n d C u (S bF 6)2 N i( S bF 6)2a ) Ra .b ) N i(S bF 6)2 IR P d (S bF 6)2 IR C u( Sb F 6)2 IR C u( Sb F 6)2 Ra A pp ro xi m at e Ba nd D es cr ip tio n Al .) [c m 1] Tn t. 1) [c m 1] m t. 1) [c m 1 Tn t. 1) [c m - 1] Tn t. A l) [c m - 1] m t. fo r M (S bF 6)2; M = N i, Pd o r Cu 74 2 (1) 73 8 v s 73 0 s, sh 72 9 s 72 5 (65 ) l)a sS b1 3’ o u to fp ha se 71 7 (12 ) 72 1 s, sh 71 6 v s 70 7 v s, b 72 1 (70 ) lJ as Sb F 3t in ph as e 71 0 (44 ) 71 0 v s 69 8 s 67 2 s 70 9 (18 ) lJ S b F 3t o u to fp ha se 67 4 (10 0) 67 2 m 66 7 m s 66 5 m 67 1 (10 0) 1 S b F 3t in ph as e 65 7 (22 ) 63 1 m 63 0 m 63 3 m a s S t 3b o u to fp ha se 61 8 (5) 62 1 m ,s h 61 5m s 61 7m 62 0 (4) l ) S b F 3’o u to f p h a s e 56 8 (2) 57 3 s 57 0 m 55 6 m s 59 2 (8) l . ) 3S S b F b in ph as e 56 0 w ,s h 55 0 w ,s h 51 1 (2) 52 1 m w 52 2 m s 49 6 m l) S b F 31’ in ph as e 46 5 m 45 0 (8) 34 8 (0+ ) 35 0 w , sh 34 5 m 39 5 (7) M ”F -S b st re tc hi ng 32 2 (5) 33 0w 33 5m ,s h 33 5 (11 ) 30 8 (24 ) 31 0w — 30 0w 30 5 (19 ) 29 9 (25 ) 29 5 w 28 5 w 29 5 (26 ) Sb -F de fo rm at io n m o de s 27 2 (3) 27 0 (29 ) A bb re vi at io ns : s = st ro ng , m = m ed iu m , w = w ea k, v = v er y, sh ,= sh ou ld er , b = br oa d, as = as ym m et hc , sy m = sy m m et ric , t = te rm in al , b = br id gi ng a) A dd iti on al ba nd s a t2 46 , 2 20 an d 19 8 c m1 an d sh ou ld er s a t 17 2, 14 6, an d 13 0c m b) Re f. 14 Ta bl e 3. 2: V ib ra tio na l Sp ec tra o ft he Tw o V al en ce Is om er s o fA g (S bF 6)2 - A g (S bF 6)2 Ra - A g (S bF 6)2 JR (X -A g( Sb F 6)2 Ra a) C -A g( Sb F 6)2 IR a) A pp ro xi m at e Ba nd D es cr ip tio n t\ D [c m 1] m t. )[ cm 1] m t. M )[ cm 1j m t. 1) [c m m t. 72 0s h (12 ) 71 9 s, sh 70 0s h (33 ) 69 2v s 69 2 vs 68 2 (10 0) 67 3 s Sb F’ st re tc hi ng 65 9 (93 ) 65 4 s 66 8 (10 0) 66 8 m s 65 1 sh (30 ) 65 8 (75 ) 64 0 m 60 0 (16 ) 60 2m 60 0 (37 ) 1 58 0 (31 ) f 58 3w Sb Fb st re tc hi ng 52 6 (4) 51 5v w 55 2 (20 ) 55 4w 49 0m s 49 6w 36 5 (1) 36 8m w M FS b st re tc hi ng 35 7w ,sh 31 2s h (13 ) 30 3 (32 ) 30 1 sh Sb F de fo rm at io n 29 0• (18 ) 28 3s 29 0 (37 ) 27 8 s a) Re f. 14 . Figure 3.1: Raman Spectrum of -Ag(SbF6)2 750 682 700 650 600 550 500 450 400 350 300 250 v, cm1 659 700 65! 600 303 312 290 58 The available structural information on M(SbF6)2compounds facilitates a discussion of the vibrational data shown in Table 3.1. A common CdC12-type layer structure is proposed for Ni(SbF6)2,based on X-ray powder diffraction data (11), and evident for a-Ag(SbF6)2from the X-ray single crystal diffraction study (14) (Figure 3.2). The structure for Ni(SbF6)2is deduced, starting from the well known LiSbF6 structure (35) by placing Ni2 into every second, nearly octahedral Li site (11). In a-Ag(SbF6)2,Ag2 is situated in a tetragonally elongated octahedral site (Figure 3.2(c)), resulting in almost square planar coordination, similar to the coordination in AgF2 (36). SbF6 acts as a thdentate bridging group (Figure 3.2(b)), with the bridging fluorines, Fb, and the terminal fluorines, Ft, in fac-octahedral positions. Unlike the SbF31’ unit, the corresponding SbF3t trigonal pyramid is remarkably regular with two dsb..Ft --1.836 A and one slightly longer at 1.846 A. The vibrational assignment for Ni(SbF6)2by Christe et al. (11), adopted here in this study, suggests a useful subdivision into SbF3tstretching modes found above —660 cm1 and SbF3t) below —635 cm1. Each set is further divided into symmethc and asymmetric in-phase and out-of-phase modes. The basic difference between both sets, SbF3b and SbF3t, is that the former belong to a filled, and the latter to an unoccupied octahedral hole in the M(SbF6)2-layer structure. The data summarized in Table 3.1 for M(SbF6)2with M = Ni, Pd, and Cu, show indeed a common group of four strong vibrations, some small band splittings for Cu(SbF6)2 not withstanding, which are assigned as SbFt vibrations. Most prominent among them is a Raman band at —670 cm1, which is clearly the strongest band in the respective spectra. Interestingly, this band is found as well for a number of additional M(SbF6)2compounds with M = Mg, Zn, Fe, Co, and Cu (14), in the same position, and always of the highest intensity. This is not unexpected, since the reported X-ray powder diffraction data had indicated the existence of two structurally related triads, consisting of the hexafluoro antimonates of Mg, Zn, and Ni, and those 59 Figure 3.2: Crystal Structure of cz-Ag(SbF6)2(Ref. 14) (a) The Unit Cell (b) ORTEP view of [SbF6] anion (c) ORTEP view of Ag2 environment F(3) 60 of Fe, Co, and Cu (14). However, the structural differences between the two triads appear to be so small that they do not seriously affect the positions of the USbF3tbands. In the Raman data, the 670 cm1 band is attributed to a symmetric in-phase SbF3tstretching mode, which suggests a common SbF3L grouping in both triads as well as in Cu(SbF6)2reported in this work. A corresponding IR band of medium intensity is observed in all instances as well. This band is found for Pd(SbF6)2 at 667 cm1. In addition, there is a close correspondence between Ni(SbF6)2and Pd(SbF6)2in both JR band positions and intensities to suggest isostructural compounds. The remaining M(SbF6)2compounds reported by Gantar et al. (14) with M = Cr, Pb, and Cd, differ slightly in both vibrational spectra and X-ray powder diffraction data. The highest intensity Raman band is now found at about 650 cm1. Even for these compounds, as for all the others (14), formulation as M(SbF6)2is suggested. None of the vibrational data reported here or published previously (11,14) suggest the presence of the anion Sb2F11, by comparison to published precedents (13,37,38). The vibrational data for both forms of Ag(SbF6)2 in Table 3.2 show interesting differences. Both have two very intense Raman bands, rather than one, in the uSbF3tregion, at 668 and 658 cm’ for a-Ag(SbF6)2and at 682 and 659 cm1 for the f3 form (see Figure 3.1). In each case the band with the largest Raman shift also has the highest intensity, and yet in spite of the incomplete band listing for a-Ag(SbF6)2,there appears to be some structural similarity between the two forms. Only a rather general band description is suggested, because there may be some band overlap for 3-Ag(SbF6)2,while the listing for the a-form appears, as stated above, to be incomplete. The distorted layer structure formed for a-Ag(SbF6)2with all tetragonally elongated octahedral holes occupied by Ag2, and concommitant tetragonally compressed, vacant sites 61 (14) provide a suitable model for the valence isomer as well. Regular occupation of half of the compressed holes by Ag and half of the elongated holes by Ag3 would allow retention of the layer structure, where Ag achieves linear and Ag3+ square planar coordination. Conversely, two types of vacant sites are formed with two sets of SbFt vibrations and different band positions for the SbFb stretching modes as well, consistent with observations. 3.3.3 Electronic Spectra Electronic mull spectra were recorded for the nickel and palladium hexafluoro antimonates, and it appears based on these spectra that the solvolysis of Ni(SO3F)2 and Pd(SO3F)2in SbF5 does not lead to a change in the electronic structures of Ni2 and Pd2 ions. The resulting compounds Ni(SbF6)2and Pd(SbF6)2,like their precursors, are paramagnetic (Section 3.3.4). It is therefore not surprising that the magnetic results and the vibrational spectra point to octahedrally coordinated Ni(ll) and Pd(II) in Ni(SbF6)2and Pd(SbF6)2. These two compounds show similar three-band electronic spectra, which can be attributed to d-d transitions. Although no extinction coefficients were obtained to support the assignment due to the insolubility of the compounds in a suitable solvent, it seems that the ligand field parameters Dq and B derived from such an assignment are very reasonable, particularly in comparison to the reported Dq and B values of the two parent compounds Ni(SO3F)2(39) and Pd(SO3F)2(19). The electronic ground term for a d8 ion in an octahedral ligand field is 3A2g and three spin allowed d-d transitions to the excited triplet-terms are expected. The energy level diagram for a d8 ion in an octahedral field is given in Fig. 3.3. The band positions of the observed electronic spectra and the calculated ligand field parameters of Ni(SbF6)2, Pd(SbF6)2, Ni(SO3F)2(39) and Pd(SO3F)2(19) are listed in Table 3.3. It seems that the agreement in band 62 Figure 3.3: Spin Allowed Electronic Transitions from 3A25 Grou nd Term for Pd2 and Ni2 (d8) in Octahedral Ligand Field \ \ \ ____ ____ ____ ____ ____ ____ __ __ __ __ _ \ 3P / / / / / / / / / V;3 V2 11 \ \ \ ig ig 2g 2g \ \ \ \ free—ion \ ‘ terms weak flgand field terms 63 Table 3.3: Electronic Transitions and Ligand Field Parameters for Ni(SbF6)2, Pd(SbF6)2and Related Compounds Compound Electronic Transition Energy lODq B B/B° d Reference (cm-1) (cm-’) 1a 3c Ni(SbF6)2 6740 11300 22400 6740 899 0.832 This work Ni(SO3F)2 - 12400 23300 7340 912 0.844 39 Pd(SbF6)2 11900 18200 26700 11900 613 0.739 This work Pd(SO3F)2 11800 17400 27000 11800 606 0.730 19 a i: 3A2g __> 3T2g b 2: 3A2g 3Tig (F) C 1)3: 3A2g 3Tig(P) d B/B° = f3, where B° is the free ion value obtained from Ref. 34(a). 64 positions for Ni(SO3F)2/Ni bF6)and Pd(SO3F)2IPd bF6)pairs suggests similar coordina tion environments for the Ni2+ and Pd2+ in the fluorosuifate and hexafluoroantimonate derivatives. In addition, the Dq and B values reported here also point to a close structural similarity between the two types of compounds. For Pd(SO3F)2(19), like for most fluorosulfates of divalent metals (18), a layer structure based on the CdC12 prototype is postulated. As discussed in Section 3.2, the same structural type is reported for a-Ag(SbF6)2(14) and implicitly also for Ni(SbF6)2(1 1). Paramagnetic Pd2+ complexes with fluorometallate anions have precedents. Complexes of the general type Pd[MF6], with M = Pd, Pt, Ge, or Sn, have been known for over 25 years now (31), and their magnetic susceptibility measurements have been reported down to 80 K. The presence of MF62 anion would suggest a different structural type, but the coordination environment of Pd2 should again be octahedral. The ligand field parameters, the octahedral splitting Dq and the interelectronic repulsion term B are obtained by using the appropriate equations (see Appendix A-4) as suggested by Lever (34(b)). The increase in Dq and the decrease in j (defined as B/B°, where B° is the free ion value) when moving from Ni(ll), (3d8) to Pd(II), (4d8) are not unexpected in view of the higher nuclear charge and more spacially diffuse 4d orbitals for palladium (40). Although no splitting of any of the bands is observable, the broad nature of the bands makes it difficult to confinn or deny the possible existence of distortion in the octahedral coordination sphere of the respective metal centers. 65 3.3.4 Magnetic Susceptibility Measurements Magnetic susceptibilities over the temperature range of —2 to 80 K are recorded for Ni(SbF6)2,Pd(Sb13)2,and Cu(SbF&2on a P.A.R. vibrating sample magnetometer. Relevant data are summarized in Tables 3.4, 3.5 and 3.6 respectively, and the plot of the magnetic moments vs. temperature for all three compounds is given in Figure 3.4. As discussed in the experimental section, attempts to obtain pure a-Ag(SbF6)2in quantities large enough for a bulk magnetic measurement were not successful. The results of measurements made, on what is obviously a mixture of the two valence isomeric forms, allow only limited conclusions, since the small amount of paramagnetic material in the sample did not permit the extension of the measurements to temperature higher than 65 K. Generally, for paramagnetic materials with one or two unpaired electrons, the vibrating sample magnetometer used in this work is useful to —90 K. In the temperature range of 65-3 K the magnetic moment calculated for cx-Ag(SbF6)2appears to be independent of temperature with a shallow maximum at --6 K before falling off. It is unclear whether the observed dilute magnetic behavior is due to a-Ag(SbF6)2,or caused by 3-Ag(SbF6)2acting as a diluent. The diamagnetism of -Ag(SbF6)2 is confirmed by measurements made at room temperature, using the Gouy technique. The measured susceptibility of -68 x 10 cm3 mo11 is less than the sum of Pascal constants of -128 x 10-6 cm3 mol1;however, the Gouy balance used is insufficiently sensitive for a more accurate determination of diamagnetic susceptibilities. In addition, trace amounts of a-Ag(SbF6)2could be responsible for the slight discrepancy, as well as temperature independent paramagnetism (TIP). As seen in Figure 3.4, the magnetic moment decreases gradually with decreasing temperature for Ni(SbF6)2and Pd(SbF6)2,before a steep decline becomes apparent at —10 K, 66 Table 3.4: Low Temperature Magnetic Data of Ni(SbF6)2 Temperature [K] XMCOff x i05 [cm3mol1] Peff B1 81.67 750 2.21 77.78 780 2.20 74.08 820 2.20 69.72 870 2.20 65.31 920 2.19 60.15 1000 2.19 54.20 1100 2.18 47.60 1240 2.17 40.05 1450 2.15 30.80 1850 2.14 26.28 2150 2.13 21.17 2610 2.10 16.20 3280 2.06 11.00 4560 2.00 7.72 6060 1.93 5.34 8020 1.85 4.78 8660 1.82 4.32 8990 1.76 3.70 9770 1.70 2.99 10340 1.57 2.50 10730 1.47 2.00 11160 1.34 67 Table 3.5: Low Temperature Magnetic Data of Pd(SbF6)2 Temperature [KJ XMCOIT x105 [cm3moP1] J4ff [PB] 81.50 1460 3.09 77.55 1530 3.08 73.91 1600 3.07 69.77 1680 3.06 65.19 1770 3.04 60.03 1900 3.02 54.00 2060 2.99 47.15 2300 2.95 39.80 2630 2,89 31.10 3210 2.83 30.45 3260 2.82 26.55 3620 2.77 21.60 4200 2.69 20.80 4310 2.68 16.50 5220 2.62 10.80 8410 2.70 9.94 9170 2.70 7.12 11270 2.53 5.84 11720 2.34 5.26 11830 2.23 4.76 11910 2.13 4.36 11910 2.04 4.00 11950 1.96 3.64 11950 1.87 3.37 11990 1.80 2.99 11990 1.69 2.29 11990 1.45 2.10 12020 1.42 68 Table 3.6: Low Temperature Magnetic Data of Cu(SbF6)2 Temperature [K] XMCOff x [cm3mol1] eff [ILBI 81.83 370 1.56 78.06 390 1.56 74.24 410 1.57 70.11 440 1.57 65.65 470 1.57 60.44 510 1.57 54.45 560 1.56 51.40 590 1.56 47.95 640 1.56 40.55 750 1.56 30.70 980 1.55 26.00 1150 1.55 21.22 1410 1.55 16.25 1820 1.54 11.00 2690 1.54 7.46 3970 1.54 5.56 5420 1.55 4.24 7290 1.57 4.04 7250 1.49 3.02 8890 1.47 2.40 10750 1.44 1.97 12480 1.40 1.86 13230 1.40 69 Figure 3.4: Magnetic Moment vs. Temperature of M(SbF6)2,M=Ni, Pd and Cu f2K<T<82K1 2.:• ‘ Pd(SbF6)2 2- .. .-o-.-o.-.o.-.o 0 Ni(SbF6)2 1.5—i 4 •...+....•.--4...•..•-.•..• • Cu(SbF6)2 1- 0 20 40 60 80 100 TEMPERATURE,K 70 indicative of possible antiferromagnetic ordering. An additional contributing factor to the sharp drop in the magnetic moment values at very low temperatures could come from zero-field splitting of the triplet spin states in the nickel and palladium compounds. It was shown in Section 3.3.3 wher’ the electronic spectra of both Ni(SbF6)2and Pd(SbF6)2were discussed that the ground term of Ni2 and Pd2 ions in these two compounds is3A2g. For ad8-ion in an octahedral ligand field with 3A2g ground term, the magnetic moment can be expressed as follows (41): 4?. 4 ji = (8)112 (1— ) = (1— ) [35]e lODq S.O. lODq where ? = spin-orbit coupling constant = J2 for d8. lODq = ligand field splitting parameter = spin-only magnetic moment = [4S(S+l)]112 Since the ground term discussed here is 3A2g a first-order orbital contribution to the magnetic moment is not expected. However, through spin-orbit coupling, which is expected to be quite large for second and third row transition metals, the observed magnetic moment is enhanced beyond the spin-only value. In the absence of any magnetic exchange between the paramagnetic centers, the moment calculated is independent of temperature and should depend only on . and lODq. The sign of is negative for transition metals where the d-shell is more than half full, and therefore according to equation [3.5], for Ni2 and Pd2 ions Peff greater than is expected. For Pd(SbF6)2,using equation [3.5] with lODq = 11,900 cm (Table 3.3) and the estimated value of ?. for Pd2 = 1600/2 cm4 (42), a temperature independent-moment of 3.59 B is obtained. The observed magnetic moment of 3.09 B at —82 K is reasonable when compared 71 to the predicted value, since the magnetic moment in Pd(SbF6)2has been lowered by anti- ferromagnetic exchange, as evident from the plot in Figure 3.4. A slight increase in the magnetic moment is observed in the Pd(SbF6)2plot (also to a very small extent in Cu(SbF6)2,just before dropping off at very low temperatures. The observed weak effect is reproducible and two possible interpretations are suggested: (a) a phase transition occurs at --10 K and (b) very weak ferromagnetism is responsible for the small effect. The latter explanation is favored here because there are a number of precedents for simultaneous ferromagnetic and antiferromagnetic ordering in fluoro derivatives of palladium(ll). Very weak ferromagnetism is observed in PdF2 (43), which has a rutile structure (44), and in the ternary fluorides Pd(ll)Pd(IV)F6and Pd(U)Pt(IV)F6(45). According to a recent crystal structure reported for the latter compound (46), there may even be a close structural relationship: Pd(ll)Pt(IV)F6has a LiSbF6 structure. For Pd(ll)(SbF6)2a layer structure is suggested, which, as discussed above, is also derived from the LiSbF6prototype, with half of the octahedral holes occupied by Pd2 compared to all holes filled for Pd(ll)Pt(IV)F (46). Some similarity in magnetic behavior is also evident from the I’80eff values, which are for the ternary palladium(ll) fluorides in the range of 2.7 to —3.0 B for the series Pd(II)M(IV)F6,with M = Pd, Pt, Ge or Sn (31a), while for Pd(SbF6)2,where Pd2 is slightly more dilute, a value of 3.09 B is found. Substantially higher magnetic moments (3.30-3.60 B) are found for the corresponding fluorosulfato complexes of palladium(II) (19b), where the bulkier fluorosulfate groups appear to prevent antiferromagnetic exchange. It is interesting to note that at low temperatures (—20 K) significant ferromagnetism has been observed for two members of this group Pd(SO3F)2and Pd(ll)Pd(IV)(SO3F6(see Chapter 6). The former, like Pd(SbF6)2,has a layer structure, derived 72 from the CdC12 type, and both appear to have the Pd2 ion in very similar electronic environ ments at room temperature, as reflected in their respective electronic mull spectra (Table 3.3) discussed in Section 3.3.3. The magnetic moments observed for Ni(SbF6)2are rather low (g = 2.21 B at —82 K), even though the moments are reduced by possible antiferromagnetic exchange (see Figure 3.4 and Table 3.7). If there is no magnetic exchange between the paramagnetic Ni2+ centers, equation 3.5 can again be utilized to calculate the temperature independent moment. The lODq value of the compound is 6740 cm1 (Table 3.3), and the estimated value of . for Ni2 is 644/2 cm1 (42) which when substituted in equation 3.5 yields al1eff value of 3.37 B• In order to understand the unusual magnetic results of Ni(SbF6)2,two other Ni(SbF6)2 samples, made by different methods were also investigated for their magnetic properties. A sample was obtained for this study from Dr. Karl 0. Christe of Rocketdyne, U.S.A., which was prepared from Ni, SbF5 and F2 under high temperature and pressure (11). The second sample was synthesized for this study from NiF2 and SbF5 in anhydrous HF according to the published method (14). The low temperature magnetic moment vs. temperature plot of these two samples is given in Figure 3.5 (see Appendices A-5 and A-6 for data). In addition, magnetic measurements were also taken using a Gouy balance on the solvolysis sample as well as on the above mentioned high temperature fluorination product (11). Results of this high temperature study are given in Table 3.7. Although all the three samples of Ni(SbF6)2display the common behavior of tempera ture dependent low magnetic moments, it is rather puzzling to note that their moments differ so substantially. The plots of ieff vs. T display very similar slopes which suggests the possible presence of a common magnetic substance at different concentrations in these samples. The 73 Figure 3.5: Magnetic Moment vs. Temperature of Ni(SbF6)2 3.4 - , o o [A] 0 ° 0 0 0 0 0 0 0 - C’ . • . . . . . [B] C • • 2.6- g • . . 2.4- o . . C C L.L I 0 20 40 60 80 100 TEMPERATURE,K [A] Gift sample, from Ni + F2 + SbF5, Ref. 11 [B] Sample from NiF2 + SbF5 in HF, Ref. 14 74 Table 3.7: Magnetic Moment Data of Ni(SbF6)2for the Temperature Range 8O to 295 K Ni(SbF6)2a Ni(SbF6)2b [Ni(SO3F)2+ SbF5] [Ni + F2 + SbF5J Temperature (K) ILeff (PB)C Temperature (K) p. (JtB)’ 290.5 2.64 293.5 4.25 268.5 2.61 268.8 4.16 251.0 2.61 251.0 4.11 234.5 2.61 234.5 4.07 218.0 2.60 218.0 4.02 200.3 2.58 201.0 3.97 176.0 2.58 177.0 3.90 151.0 2.56 151.5 3.81 126.3 2.45 127.5 3.73 101.8 2.42 103.0 3.61 86.5 2.31 86.5 3.53 78.0 2.26 78.0 3.46 a This work b Gift sample, made according to Ref. 11 C eff = 2.828 [(Xmoi - TIP)T]1/2; = 2 320 x 10-6 cm3 mo11 75 solvolysis product from this study appears to have the lowest magnetic moments, whereas the gift fluorination sample exhibits moments which are unexpectedly high for an octahedrally coordinated Ni(I1) species (Table 3.7). The sample made using HF as solvent medium has moments (measured up to —80 K only, Figure 3.5) which fall between the above two sets of values. Furthermore, a Ni(SbF6)2sample made from the oxidation of Ni by SbF5 in SO2 solution is reported as having a magnetic moment of 3.16 1B at 294 K (12). It is also important to note here that if the symmeuy of the ligand field acting on a d8 ion like Ni(II) is allowed to be lower than cubic and this low symmetty component is large enough, then a spin free-spin paired (triplet-singlet) equilibria may occur in the paramagnetic ions of the compound (47,48). This could reduce the magnetic susceptibility of the system, yielding magnetic moments lower than expected. Although this situation is observed mostly in Ni(II) complexes, it is reported to occur almost exclusively in solutions and is dependent on the solvent, concentration and temperature (47). Therefore, this unusual behavior may not be a significant factor in this solid state magnetic study of Ni(SbF6)2. However, the existence of such systems in the solid state cannot be ruled out completely. The temperature dependent low moments obtained for the solvolysis sample may be due primarily to the following factors: (a) possible antiferromagnetic interaction between the Ni(II) centers, which will lead to lower than expected moments; (b) small amounts of nickel ions in oxidation states other than +2 acting as cliluents to predominantly octahedral Ni(ll) ions in the lattice; and (c) a combination of factors (a) and (b) which will result in lower magnetic moments for the compound. The contribution of antiferromagnetism toward lower moments can be seen clearly in both Figure 3.5 and Table 3.7, where in the latter tabulation the small contribution (320 x 10-6 cm3 mo11) from the temperature independent paramagnetism (TIP) to the susceptibility has 76 been removed, since this contribution is relatively significant at higher temperatures. However, as no Xmax is observed in the susceptibility data, antiferromagnetism alone cannot account for such low moments in the solvolysis product. Therefore, the existence of small quantities of other nickel ion species than octahedral Ni(II) in the lattice may have to be considered as a possibility. Interestingly, two octahedrally coordinated spin paired Ni(ffl) fluoro compounds, K3NiF6 and Na3iF6 have been reported in the literature (49), where the elongation of the [NiF6]3 octahedron is ascribed to the Jahn-Teller effect to be expected for the t2g6eg’ configuration. The magnetic moment ofK3NiF6is given as 2.51 and 2.12 B at 295 and 90 K respectively (50), although the large temperature dependence of the moments is surprising for an 2Eg ground term, unless thermal equilibrium between low and high spin configurations (49) and/or antiferromagnetism is taken into consideration. It is however still rather difficult to rationalize the generation and hence the presence in the solvolysis product of nickel ions in oxidation states other than +2, since the parent compound Ni(SO3F)2is well characterized as a Ni(ll) octahedral complex, both by electronic spectra (39) and magnetic studies (Chapter 6). The magnetic moments calculated at higher temperature for the fluorination product (Table 3.7) in contrast, are significantly above the values expected for a d8 Ni(II) complex. It was shown earlier that equation [3.5] predicts a temperature independent moment of 3.39 1B’ whereas the moment obtained at room temperature of this compound exceeds this value by nearly one Bohr magneton. As discussed in Section 3.1, the obvious limitation in the high temperature synthesis of Ni(SbF6)2is that direct fluorination may lead to oxidation of the metal beyond the +2 state. This is possible in nickel where higher oxidation states are accessible and the metal can be oxidized under severe conditions up to the +4 state (40). However, as pointed out above, the known octahedral Ni(ffl) fluoro compounds are of the spin-paired type (49) and 77 Ni(IV) derivatives like (CIF2O)Ni6and K2NiF6 (5 1,40) are diamagnetic with the low-spin t2g6 configuration. Therefore, the origin of the high magnetic moments calculated for this compound is not clear, especially as the compound has been sufficiently characterized using microanalysis, vibrational spectroscopy and X-ray powder diffraction (11). However, a very likely cause of the high magnetic moments for this sample can be Ni metal impurities. Since Ni metal is ferromagnetic, trace amounts of it would cause significantly larger ).teff values for the sample. Furthermore, the metal saturates at relatively low magnetic fields, and as ‘eff is propor tional to (XMT)”2,this will result in decreasing moments with decreasing temperatures for the compound. Field dependent magnetic susceptibility studies are required to verify this possible ferromagnetic contamination of the samples. The magnetic moments obtained for Cu(SbF6)2are given in Table 3.6, and appear to be independent of temperature down to —4 K. It is also of interest to note here that of the four transition metal fluorosulfate precursors used for this study, only Cu(SO3F)2is magnetically dilute to low temperatures (Chapter 6). The moments of Cu(SbF6)2are lower than expected for a Cu(II) (d9) ion in an octahedral ligand field (52). Since the ground term 2Eg has no orbital angular momentum associated with it the moments should be close to the spin only value of 1.73 B in the first approximation. However, spin-orbit coupling can occur between the 2E ground term and the higher lying 2T2g term, leading to slightly higher moments predicted by the follow ing expression (41): 2?. 2A Peff (3) ( 1 lODq = (1 lODq [3.6] where the terms used are as in equation [3.5]. It was noted earlier in Section 3.3.2 that the vibrational spectrum of Cu(SbF6)2made via the solvolysis method is slightly different when compared to that of CuF2Sb5(14). This was 78 taken as a clue to the possible existence of polymorphism in the two forms of Cu(SbF6)2, although to a very limited extent than in the case of Ag(SbF6)2. Magnetic measurements obtained on several samples of Cu(SbF6)2also indicate this possibility (Figures 3.6 and 3.7), with Cu(I), Cu(II), and Cu(III) present in an equilibrium of the type: k1 2 Cu(II) Cu(I) Cu(Ill), k2 >> k1 [3.7] k2 with the ratio k2/k1 differing very slightly relative to the experimental conditions of the formation reaction and the subsequent treatment of the reaction product. Solvolysis of Cu(SO3F)2in SbF5 appears to generate small quantities of Cu(I) and Cu(III) ions, in addition to the relatively larger number of Cu(ll) ions (k2 >> k1). The diamagnetic Cu(I) and Cu(llI) ions seem to act as diluents to the paramagnetic Cu(ll) centers, thereby lowering the observed magnetic moments of the compound below the expected values. This could also account for the temperature independent magnetic moment behavior of the compound, where the Cu(II) centers are now more dilute than in the corresponding Pd(II) or Ni(ll) derivatives. When the product is treated with anhydrous HF for 3 days, equilibrium [3.7] is shifted further to the left (k2 >>> k1), resulting in an increase in the Cu(II) ions present in the compound. Consequently, the magnetic susceptibility of the sample is enhanced to a small but significant extent leading to correspondingly higher magnetic moments (Figure 3.6). The opposite effect is found when Cu(SbF6)2is re-reacted with an excess of SbF5 for a prolonged period of time, —6 to 8 weeks (Figure 3.7). In all instances, no weight changes were observed in the vacuum dried products. 79 Figure 3.6 Magnetic Moment vs. Temperature of HF Treated Cu(SbF6)2 1.8- aD — Cu(sbF6)2 1.6- 0 + excess HF0 Z I I — Cu(SbF6)2 0 1.4- C-) LI z o 1.2- 1— 0.8- 0 20 40 60 80 100 TEMPERATURE,K 80 Figure 3.7: Magnetic Moment vs. Temperature of SbF5 Treated Cu(SbF6)2 1.8- 1.6- ‘ • • . . . • • . — Cu(SbF6)2 z I I ° o o o o o ° ° ° — Cu(SbF6)20 I 0 0 0 0 + excess SbF5 C-) LJ zç 1.2- 1— 0.8- I I 0 20 40 60 80 100 TEMPERATURE,K 81 The presence of Cu(I) and Cu(IH) ions, in addition to Cu(H) in Cu(SbF6)2is not totally unexpected, since both Ag(II)(SO3F2and Cu(H)(SO3F2which contain “Jahn-Teller ions” (d9), could behave in a similar manner toward SbF5, generating Ag(I), Ag(llI) and to a lesser degree Cu(I), Cu(llI) ions in their respective reaction products. As in the case of 3-Ag(SbF6)2,the Cu(I) and Cu(llI) centers could be accommodated in the layer structure of Cu(SbF6)2,with Cu(I) in a linear and Cu(ffl) in a square planar environment respectively. 3.3.5 X-ray Photoelectron Spectra Attempts were made to characterize the f3-Ag(SbF6)2and Cu(SbF6)2compounds using X-ray photoelectron measurements. It was hoped that the difference in Binding Energy (BE) between the cations in oxidation states +1, +2, and ÷3 could be utilized to identify these ions present in the silver compound as Ag(I), Ag(III) and in the copper compound as Cu(I), Cu(II), and Cu(Ill) respectively. The reported BE values corresponding to these oxidation states (for a given energy level) were reconfirmed by a series of measurements carried out with relevant silver and copper fluoro complexes. For the XPS study of -Ag(SbF6)2compound, the 3d512 energy level is chosen, which is found at BE = 366.8 eV in Ag(0). The Ag(I) and Ag(llI) BE peaks for the same energy level are expected at —367.7 and —371.0 eV respectively (53,54). The sample was scanned in the BE range of 357 to 377 eV. Unfortunately, for the 3d52 level, only a broad band spanning a BE range of —367.5 to 370 eV is obtainable, with no resolution of the relevant peaks expected for the Ag(I) and Ag(Ifl) species. It is however possible that the two peaks corresponding to the two silver ions are hidden under the broad band centered at —369 eV. A similar situation is observed when the alternate energy level 3d2 is chosen for the measurements. 82 The XPS spectra obtained for Cu(SbF6)2show a more complicated spectral pattern. The energy level 2P3 located at BE = 932.2 eV for Cu(O) is chosen, and BE peaks of Cu(I), Cu(II) and Cu(Ill) corresponding to this energy level are expected at 932.6, 936.1 and 938.0 eV respectively (53,54,55). However, in this compound, only small amounts of Cu(I) and Cu(ffl) ions are expected to be present (see Section 3.3.4), in contrast to the silver compound, where Ag(1) and Ag(ffl) ions are present in equimolar quantities. The BE range scanned for Cu(SbF6)2is from 921 to 941 eV, and as in the silver compound, no distinct peaks corresponding to the three oxidation states are observed. The broad band obtained covers an energy range of —932.5 to 938 eV, and is split by several intense satellite lines. This phenomenon is commonly seen in the XPS spectra of cupric compounds, and the number and the splitting of the peaks are found to be sensitive to the chemical environ ment of the ions (55). This situation is further complicated by the reduction of Cu(ll) ions when subjected to X-rays, and consequently, additional satellite peaks may appear in the spectra (55). Although no individual peaks at energy level 2P312 are seen in the spectra, it is again possible to have the less intense Cu(I) and Cu(ffl) peaks hidden under the broad band, since the energy range of the band covers the three BE values of the copper ions Cu(I), Cu(II), and Cu(III) expected in Cu(SbF6)2. 3.3.6 Attempted Synthesis of Au(SbF&2 Gold(lI) fluorosulfate, which was synthesized recently in our group by the reduction of Au(SO3F) with either gold powder or CO (56), was selected as the precursor to react with excess SbF5 in the attempted preparation of the corresponding gold(ll) hexafluoro antimonate, Au(SbF6)2. The gold compound has been formulated as a mixed valency, diamagnetic Au(I)[Au(III)(SO3F41complex, based on magnetic measurements and vibrational spectra (56). 83 Therefore, it was expected that the resulting binary compound may also have the composition Au(I) [Au(III)(SbF64],analogous to the diamagnetic -Ag(SbF6)2compound. However, in contrast to the other four transition metal fluorosulfate precursors, the reaction of Au(SO3F)2with excess SbF5, carried out in a manner similar to the other prepara tions discussed in this chapter (equation 3.1), does not yield the anticipated binary product Au(SbF6)2. InsteaLt, the synthesis follows the reaction scheme shown below, yielding a ternary compound: 60°C Au(SO3F)2+ excess SbF5 > Au(SO3F)(SbF6 [3.81 21 days The values of x and y (typically x = 1 to 1.5, and y = 2-x), calculated from microanalytical data, seem to vary slightly and appear to depend on the reaction conditions. The bright yellow Au(SO3F)2gradually turns color to give a dark brown-green powder. When the reaction is performed at elevated temperatures, a dark brown, very viscous liquid which may form due to the melting or decomposition of the lower temperature product, is isolated. Infrared spectra run on several samples clearly show the presence of both SO3F and SbF6 groups, and magnetic measurements indicate a diamagnetic compound, as anticipated. Interestingly, when diamagnetic gold(III) fluorosulfate Au(SO3F),which is synthesized by the oxidation of gold powder withS206Fin HSO3F(57), is reacted with excess SbF5 under the same experimental conditions (equation [3.8]), a dark blue-green powder is isolated, again with composition Au(SO3F)(SbF6where the values of x and y (typically x =2 to 2.3 and y =3-x), are dependent on the reaction conditions. However, this ternary compound, in contrast, is weakly paramagnetic. The low temperature magnetic data, tabulated in Appendix A-7, yield moments that range from 0.52 to 0.72 B at 82 and 2.7 K respectively. These values are 84 tentative, since the composition of the compound (and hence the molar mass) cannot be deter mined accurately. It is important to note here that like in the fluorides, where both AuF and AuF2 are unknown (33), examples for lower valent fluoro derivatives of gold are so far lacking. It is conceivable that Au(SO3F),which is dimeric in the solid state (58) when reacted with SbF5, may undergo a reduction process to a veiy slight extent. This reduction of the gold compound may be due to the inability of substituent SbF6 anions to suitably stabilize the higher oxidation state and hence strongly oxidizing Au(Ill) species, and consequently, may lead to the formation of small quantities of paramagnetic Au(II) ions in the solid lattice. This observation appears to be valid also for the solvolysis reaction of Sn(IV)(SO3F4in SbF5. When about 0.679 g of Sn(SO3F)4,made according to a published method (59) is reacted with an excess of SbF5 for two weeks at 60°C, a white solid that corresponds to the composition Sn(SbF6)2is isolated. The reduction of the Sn(IV) species to Sn(II) could occur according to: 60°C Sn(SO3F)4+ 6SbF5 > Sn(SbF + 2SbF9( O3) + S206F [3.91 and 2S06F + SbF5 —* 2S0F + 2S03 + 02 [3.10] The decomposition of S206F in the presence of SbF5 (5) leads to byproducts which are all removable in a dynamic vacuum and consequently solid Sn(SbF6)2is obtained in high yield. 3.4 Conclusion Solvolysis of M(ll) fluorosulfates with M = Sn, Ni, Pd, Cu or Ag, in antimony(V) fluoride is found to be a useful synthetic route to the corresponding metal(ll) hexafluoro 85 antimonate compounds. Even though alternate synthetic routes to compounds of the type M(SbF6)2are known, only the solvolysis method leads to the Pd(SbF6)2and the f3-Ag(SbF6)2 compounds. The latter complex, unlike a recently reported paramagnetic, blue valence isomer is diamagnetic and based on its chemical and magnetic behavior, formulated as a mixed valency Ag(I)[Ag(III)(SbF64]species. Additionally, the copper compound synthesized by this route also appears to be unique, with Cu(ll) and, to a lesser extent Cu(I) and Cu(llI) ions, all present simultaneously in the lattice of Cu(SbF6)2, as confirmed by low temperature magnetic susceptibility measurements. The Ni(SbF6)2compound prepared via solvolysis appears to be structurally similar to the compound obtained by the high temperature fluorination method. However, in the solvolysis sample the nickel centers are predominantly found as Ni(II) ions, whereas in the fluorination product trace Ni metal impurities seem to be present, which is evident from their respective magnetic studies. Ligand field analyses of the electronic spectra of Ni(SbF6)2and Pd(SbF6)2indicate that, like in their fluorosulfate precursors, the Ni(ll) and Pd(ll) ions are located in approximate octahedral ligand fields with 3A2g ground terms. Both compounds exhibit temperature dependent low magnetic moments, most likely due to antiferromagnetic exchange. Addition ally, Pd(SbF6)2shows very weak ferromagnetism below -40 K. It appears, based on X-ray powder data and vibrational spectra that a common CdC12- type layer structure is present in all the M(SbF6)2compounds synthesized in this study, although ultimate structural proof will have to come from single crystal X-ray diffraction studies. The lack of solubiity of the compounds in suitable solvents like anhydrous HF and their high reactivity toward many of the organic solvents present substantial obstacles to the crystal growth process. 86 References 1. P.L. Fabre, J. Devynck, and B. Tremillon, Chem. Rev., , 591 (1982), and references therein. 2. S.P. Mallela, S. Yap, J.R. Sams, and F. Aubke, Inor. Chem., , 4327 (1986). 3. G.A. Olah, G.K.S. Prakash, and J. Sommer, “Superacids”, John Wiley and Sons, New York, 1985, and references therein. 4. R.J. Gillespie and R.A. Rothenbury, Can. J. Chem., 42,416 (1964). 5. W.W. Wilson and F. Aubke, J. Fluorine Chem., j.j 431 (1979). 6. P.A. Yeats and F. Aubke, J. Fluorine Chem., 4, 243 (1974). 7. W.W. Wilson, J.R. Dalziel, and F. Aubke, J. Inor. Nuci. Chem., 37, 665 (1975). 8. W.W. Wilson, R.C. Thompson, and F. Aubke, Inorg. Chem., i, 1489 (1980). 9. S. Karunanithy and F. Aubke, J. Fluorine Chem., fl 541 (1983). 10. S.P. Mallela, S. Yap, J.R. Sams, and F. Aubke, Rev. Chim. Minerale, j 572 (1986). 11. K.O. Christe, W.W. Wilson, R. Bougon, and P. Charpin, J. Fluorine Chem., 3.4, 287 (1987). 12. P.A.W. Dean, J. Fluorine Chem., 5, 499 (1975). 13. T. Birchall, P.A.W. Dean, and R.J. Gillespie, J. Chem. Soc. A, 1777 (1971). 14. D. Gantar, I. Leban, B. Friec, and J.H. Holloway, J. Chem. Soc. Dalton Trans., 2379 (1987). 15. B. Friec, D. Gantar, and J.H. Holloway, J. Fluorine Chem., 20, 385 (1982). 16. B. Frlec, D. Gantar, and J.H. Holloway, J. Fluorine Chem., 1.2. 485 (1982). 17. D. Gantar, B. Frlec, D.R. Russell, and J.H. Holloway, Acta Crvst.. C, 4j 618 (1987). 18. C.A. Alleyne, K. O’Sullivan-Mailer, and R.C. Thompson, Can. J. Chem., 52, 336 (1974). 19.a) K.C. Lee and F. Aubke, Can. J. Chem., , 2473 (1977). b) K.C. Lee and F. Aubke, Can. J. Chem., 2. 2085 (1979). 87 20. P.C. Leung and F. Aubke, Inorg. Chem., ii, 1765 (1978). 21. M.S.R. Cader, R.C. Thompson, and F. Aubke, Chem. Phys. Lett., j.4, 438 (1989). 22. A.F. Clifford, H.C. Beachell, and W.M. Jack, S. Inorg. Nuci. Chem., , 57 (1957). 23. A.I. Vogel, “Quantitative Inorganic Analysis”, 3rd Edition, John Wiley and Sons, New York, 1961. 24. B. Standke and M. Jansen, Angew. Chem. mt. Ed., , 77 (1986). 25. J.A. McMillan, Chem. Rev., 2. 65 (1962) and references therein. 26. B. Standke and M. Jansen, Angew. Chem. mt. Ed., 24, 118 (1985). 27. P.C. Leung, K.C. Lee, and F. Aubke, Can. J. Chem., 7, 326 (1979). 28. L. Helmholtz and R. Levine, J. Am. Chem. Soc., 4, 354 (1942). 29. W. Klemm, Angew. Chem., , 468 (1954). 30. B.G. Muller, Z. Anorg. Aug. Chem., 3, 196 (1987) and references therein. 31.a) N. Bartlett and R.P. Rao, Proc. Chem. Soc., 393 (1964). b) R.P. Rao, Ph.D. Thesis, University of British Columbia, 1965. 32. R. Bougon, T.B. Huy, M. Lance, and H. Abazli, Inorg. Chem., 23, 3667 (1984). 33. B.G. Muller, Angew. Chem. Tnt. Ed. Engi., , 1081 (1987). 34.a) B.N. Figgis, “Introduction to Ligand Fields”, John Wiley and Sons, New York, 1966. b) A.B.P. Lever, I. Chem Educ., 4, 711 (1968). 35. J.H. Burns, Acta Crvst., .15, 1098 (1962). 36. P. Fischer, 0. Roult, and D. Schwarzenbach, J. Phys. Chem. Solids, 32, 1641(1971). 37. B. Friec and J.H. Holloway, J. Chem. Soc. Dalton Trans., 535 (1975). 38. J. Fawcett, J.H. Holloway, D. Laycock, and D.R. Russell, J. Chem. Soc. Dalton Trans., 1355 (1982). 39. D.A. Edwards, M.J. Stiff, and A.A. Woolf, Inorg. Nuci. Chem. Letters., 3, 427 (1967). 40. F.A. Cotton and G. Wilkinson, “Advanced Inorganic Chemistry”, 5th Edition, John Wiley and Sons, New York, 1989. 88 41. F.E. Mabbs and D.J. Machin, “Magnetism and Transition Metal Complexes”, Chapman and Hall, London, 1973. 42. R.L. Carlin, “Magnetochemistry”, Springer-Verlag, Berlin, 1986. 43. R.P. Rao, R.C. Sherwood, and N. Bartlett, J. Chem. Phys., 42. 3728 (1968). 44. N. Bartlett and R. Maitland, Acta. Cryst., .IQ, 63 (1957). 45. J.-M. Dance and A. Tressaud in “Inorganic Solid Fluorides”, Ed. P. Hagenmuller, Academic Press, New York, 1985, and references therein. 46. B.G. Muller, Z. Anor. Aug. Chem., 55, 79 (1988). 47. B.N. Figgis and 3. Lewis, Prog. Inorg. Chem., , 37 (1964). 48. G. Maid, I. Chem. Phys., 2. 651 (1958). 49. G.C. Allen and K.D. Warren, Inorg. Chem., 8, 1895 (1969). 50. W. Klemm, W. Brandt, and R. Hoppe, Z. Anorg. Aug. Chem., 3, 179 (1961). 51.a) W.W. Wilson and K.O. Christe, Inorg. Chem., , 3261 (1984). b) R. Hoppe in “Inorganic Solid Fluorides”, Ed. P. Hagenmuller, Academic Press, New York, 1985. 52. Landolt-Börnstein, Numerical Data and Functional Relationships in Science and Technology, Vol. 2, Magnetic Properties of Coordination and Organometallic Transition Metal Compounds, Springer-Verlag, Berlin, 1966. 53. T.A. Carlson, “Photoelectron and Auger Spectroscopy”, Plenum Press, New York, 1975. 54. C.D. Wagner in “Practical Surface Analysis by Auger and X-ray Photoelectron Spectroscopy”, Ed. D. Briggs and M.P. Seth, John Wiley and Sons, Newe York, 1983. 55. D.C. Frost, A. Ishitani, and C.A. McDowell, Mol. Phys., 24, 861 (1972). 56. H. Willner, F. Mistry, G. Hwang, F.G. Herring, M.S.R. Cader, and F. Aubke, J. Fluorine Chem., , 13 (1991). 57. K.C. Lee and F. Aubke, Inorg. Chem., ., 389 (1979). 58. H. Wiliner, S.J. Rettig, J. Trotter, and F. Aubke, Can. J. Chem., 69, 391 (1991). 59. S.P. Mallela, K.C. Lee, and F. Aubke, Inorg. Chem., 2, 653 (1984). 89 CHAFFER 4 MESITYLENE ADDUCTS OF TIN(II) FLUORO COMPOUNDS, Sn(SO3F)2C9H12and Sn(SbF62C9H12 4.1 Introduction Arene ic-complexes, where low valent post-transition metal centers act as acceptors, and benzene or various alkyl benzenes, function as ic-donors, form a small but interesting group of weakly bound complexes. Structural and bonding aspects of these compounds have become more widely known through the work of Amma et al. (1), and Schmidbaur et al. (2). A review on the more general topic of ic-complexes of main group elements has also appeared recently (3). Arene complexes of the post-transition metals are best regarded as donor-acceptor adducts, with the arene ic-system being the donor component. For Group 13 metal derivatives, an increased arene donor strength through electron releasing substituents leads to enhanced stability of the complexes. However, as observed in X-ray crystal structure studies, the metal ligand interactions are not strong enough in these complexes to induce significant distortions of the arene rings (2). Consequently, complexes with low thermal stability are usually isolated (1,2). Within this group of it-complexes, interest has focussed mainly on arene complexes with univalent metal centers such as Ag(I), Cu(I), Hg(I), Ga(I), In(I) and Tl(I) and comparatively little is known about similar compounds with the Group 14 centers like Sn(II) (4-7) and Pb(II) (1,8). There are commonly two reasons given for the paucity of arene H-complexes in group 14: 90 a) as exemplified by the structural chemistry of divalent tin (9), the 52 pair is stereochemically far more active than in compounds of isoelectronic In(I) and consequently the Sn(II) derivatives are apparently less inclined to form it-complexes with arenes (2); b) lattice energies for Sn(II) and Pb(II)-salts tend to be considerably higher than those of Ga(I), In(I), or Tl(I) salts. The non-polar, it-donating arenes are normally incapable of breaking up such lattices effectively. To reduce the lattice energy and to facilitate complex formation, salts with large, univalent anions are chosen and the majority of arene it-complexes reported so far feature tetrahalometallate (III) anions of the type MX4 with M=B, Al, Ga, In or Ti, and X=Cl or Br (1-8) as counter anions; however weakly basic anions like OTeF5 appear to be suitable as well (10). The tin(II)-ic-arene complexes reported so far are limited largely to mono-arene complexes with either Sn2 (4) or the chloride bridged moieties such as (Sn2C1) (5) or (Sn4Ci)(7) as acceptors, but recently the first bis(arene) complex has been reported as well, [(‘6-CH2SnC1(AlC1)J(6). Benzene functions primarily as donor, bonded to tin in the rj6- mode, and A1C14 is the counter anion, with Cl atoms bridging to tin. Frequently, however, arenes are also found in the lattice, without being coordinated to tin (4,6,7). Detailed crystal structures (see Appendix A-8) are reported for all compounds (4-7) and have been used in the development of bonding concepts based on the Molecular Orbital theory ((1, 4, 5), Appendix A-9). The interest in the chemistry of tin for this work stems from the following two specific related objectives: a) the use of 119Sn Mössbauer spectroscopy as a structural tool (this spectro scopic technique has not been applied so far to this small group of tin complexes) and, b) the stabilization of “bare”, non - or very weakly coordinated cations like Sn2 (11) orR2Sn (12). As discussed in Chapter 3, two compounds of interest, tin(II)bis(fluorosulfate), Sn(SO3F)2and tin(II)bis(hexafluoroantimonate), Sn(SbF62 (13), both originally reported by Gillespie and coworkers (14), appear to be capable of functioning as potential acceptors in arene it-complexes. 91 Both compounds give rise to broad single line 119Sn Mössbauer spectra with large positive isomer shifts relative to Sn02, and, in the case of Sn(SO3F)2,a small resolvable quadrupole splitting (11, 14) suggests non-directional, nearly spherical distribution of the 5s2 electron pair. In addition, it had been shown previously, based on 119Sn Mössbauer parameters, that the anions SbF6- and Sb2F1fare extremely wealdy basic and the least nucleophilic anions when stabilizing the cation (CH3)2Sn with a linear C-Sn-C group (12). Therefore, it was expected the SbF6 anion to be ideally suited to stabilize the “bare” Sn2 ion equally well. The single line 119Sn Mössbauer spectrum with an isomer shift of 4.39 mm s1 relative to Sn02 obtained for Sn(SbF6)2(see Chapter 3), which is a shift of about 0.45 mm s’ lower than suggested by calculations for Sn2 (15), indicate that steric factors rather than low nucleophilicity of the counter anion may be important. The list of stannous compounds with single-line 119Sn Mössbauer spectra and isomer shifts higher than the ones observed for Sn(SbF6)2provides a clue. Sn[Sn(SO3F)6],4.48 (11); Sn(ClO42(15-crown5),4.53 (16); Sn(SbF2As3,4.66 (17); and Sn[Sn(SOCF14.69 mm s1 relative to Sn02 (18) all approach a true Sn2 more closely than Sn(SbF6)2and all involve sterically more hindered ligands than SbF6. A crystal structure for Sn(SbF2As3(17), where the Lewis base AsF3 is coordinated to tin as well, indicates 9-coordinate tin. The use of crown ether ligands (16) also supports the view that steric factors play an important role in stabilizing the “bare” Sn2 ion. Furthermore, the two anions of the compounds, SO3F and SbF6, as seen earlier in Chapter 3, function frequently as weakly coordinating, tridentate ligands towards divalent metal ions giving rise to layer structures, based on the CdCl2-prototype. This structural type is commonly found for most sulfonates of divalent metals, and the recently reported structure of Sn(SO3F)2confirms this structural type, where a three-dimensional framework of SO3F groups linked by O-Sn-O bridges with the two crystallographically independent fluorosulfates acting as tridentate bridging ligands between the tin atoms is indicated (19). The structure of Ag(SbF6)2 (20) (Chapter 3, Figure 3.2) may serve as an example for the layer structure formed by 92 M(SbF6)2. Both SO3F and SbF6 are very weakly basic, and covalent contributions to the lattice energies of their tin(J1) salts should be small. Alternatively, both anions may stabilize arene complexes of tin with the help of oxygen or fluorine bridges. As donor, 1,3,5- trimethylbenzene (mesitylene) is chosen, because it is expected and found (2) to be a better donor than benzene. 119Sn Mössbauer spectroscopy is proven to be a useful technique for this investigation. Its application to the study of tin(II) compounds has been reviewed previously (21). 4.2 Experimental 4.2.1 Synthesis Sn(SO3F)2was obtained from SnCI2 and HSO3F(22) as described (23), and Sn(SbF6)2 was prepared from the solvolysis of Sn(SO3F)2in an excess of SbF5 as pointed out in Chapter 3, Section 3.2. Stannocene,(5-CH)2Sn, was synthesized from SnCI2 and NaC5H. The crude product was purified by sublimation and its purity checked by microanalysis. Synthesis ofSn(SO3F)21,3,5-(CH)C6H To a one part reactor a sample of Sn(SO3F)2(1.06 g, 3.35 mmol) was added, and after evacuation of the reactor, dry mesitylene (—7 mL, —50 mmol) was transferred into the reaction vessel via a distillation bridge. The white suspension was stirred in vacuo at room temperature for three days. Excess mesitylene was removed in vacuo at -20 to -15°C (ice/NaCl bath) over a period of three days yielding 1.39 g of a white powder, of the composition Sn(SO3F)2mes, with a decomposition point of 97.2-98°C. 93 Anal calcd. forC9H12F2O6Sn: C, 24.74; H, 2.77; F, 8.70; S, 14.67; and Sn, 27.16%. Found: C, 25.77; H, 2.87; F, 8.53; S, 14.84; and Sn, 27.10%. An alternate preparation with mesitylene removed in vacuo at 50C gave the following analysis: Anal. calcd.: C, 24.74; H, 2.77; and S, 14.67%. Found: C, 24.97; H, 2.80; and S, 14.55%. JR bands and estimated intensities [cm1]: 2910w,b, 2860vw, 1760vw, 1604m, 1575w,sh, 1446w, 1337w, ll9lvs, 1127w,sh, 1054vs, 888s, 786w, 667w, 589ms, 544m, 507vw, and 405m. Synthesis ofSn(SbF6)2[1,3,5(CH3CH] To the bottom portion of a two part reactor, freshly prepared Sn(SbF6)2(3.11 g, 5.27 mmol) was added. The flask was then fitted to the lower end of a filtration apparatus (24). On the top end of the filtration apparatus a seond 100 mL flask was fitted and the entire apparatus was then evacuated. Through a Teflon valve in the bottom chamber of the apparatus mesitylene (15 mL, 108 mmol) was transferred via a distillation bridge. Upon exposure to the mesitylene vapor, the white powder immediately turned yellow and then during the filtration, pale, off white. When the mesitylene distillation was complete, the suspension was stirred in vacuo for 1.5 hours. The apparatus was inverted, and the mixture filtered. The filtrate flask was cooled to -10°C (ice/NaC1 bath) and the mixture was left to filter for 17 hours. When gravity filtration was complete, the filtrate chamber of the apparatus was evacuated to further dry the solid. The apparatus was refilled with dry nitrogen inside the dry box, and the filtrate chamber was evacuated again. At the end of this drying procedure, a pale, 94 off-white solid (4.18 g, 5.04 mmol, 95.5% yield) was obtained together with a yellow-green filtrate, which was discarded. The product was transferred to a two part reactor and was further dried in vacuo for 24 hours at room temperature. No further weight change was observed. The solid decomposed at 69.5-71.5°C to a greenish yellow liquid. Anal. calcd. forC18HF2Sb2n: C, 26.03; H, 2.91; F, 27.45; Sb, 29.32; Sn, 14.29%. Found: C, 26.87; H, 3.00; F, 27.31; Sb, 29.05; Sn, 14.20; Total 100.43%. JR bands and estimated intensities [cm-1]: 2920mw, 2860w, 1601m, 1580m, 469w, 1384m, 1168vw, lO8Ovw, lO4Ovw, 1007w, 960vw,b, 860w, 687s, 671vs, 635s, sh, 580m,b, 544m, 5 12w, 440w, and 395w. 4.3 Results and Discussion 4.3.1 Synthesis The previously reported synthesis of Sn(II) containing benzene it-complexes (4-7) involves either the reaction of SnC12 and A1C13 at different mole ratios and temperatures (4,5), or of molten Sn(A1C14)2(6, 7) with the arene to give either Sn2, (Sn2Cl)or (Sn4Cl)’ containing complexes as colorless crystals. Crystalline material suitable for single crystal X-ray diffraction studies are obtained (4-7) in all instances. The course of the reaction appears to be influenced in part by the mole ratios of the reactants used and in part by the reaction tempera ture. The different products obtained, with Sn2 or (SnCl)’, n = 2 or 4, are not totally surprising since in the binary system SnC12-A1C3,phase studies (9) have revealed the existence of two distinct phases: SnC12A13 and SnC12AlCl3. It appears that the same two phases crystallize now in the form of ,t-arene complexes with A1C14 as a coordinated counter ion and in several instances with the lattice arene present (4, 6, 7). 95 The synthetic reactions reported here differ in a number of ways. Binary rather than tertiary reaction mixtures are used, the reaction temperature is room temperature and the tin(U) substrates are synthesized and characterized prior to their reaction with mesitylene. Addition of mesitylene occurs smoothly according to: 250C Sn(SO3F)2+ mes > Sn(SO3F)2mes [4.1] mesitylene or 25°C Sn(SbF6)2+ 2 mes > Sn(SbF6)2mes [4.2] mesitylene mes = l,3,5-trimethylbenzene (mesitylene), and the two complexes are formed in nearly quantitative yield. Solid material remains throughout the reaction and in case of Sn(SO3F)2as acceptor, there is no visible sign of a reaction, except for an increase in the bulk of the solid residue. Complete removal of excess mesitylene in vacuo is difficult to judge and the conditions quoted in the experimental section were arrived at by trial and error. It does appear that samples submitted for microanalysis (C, H, and S content) immediately after isolation had small amounts of residual mesitylene trapped while samples sent to Germany for further analysis (S, Sn, Sb, and F content) seemed to have lost the excess reactant. The reaction of mesitylene with Sn(SbF6)2differs even further. A 2:1 complex forms and a color change to bright yellow is noted as soon as the tin salt is exposed to mesitylene vapor. However the color of the solid quickly fades during product isolation and an off-white solid is obtained in very high yield together with a yellow-green filtrate. The origin of the initially observed color is unclear. A UV-visible spectrum taken as Nujol mull on a solid sample at this stage shows an intense UV-absorption at —240 nm tailing off in the visible region 96 with weak shoulders at 486, 580 and 656 nm respectively. Observations made during product isolation suggested that the intense color seen was due to a by-produc., which was soluble in excess mesitylene. A complex formed by mesitylene and small, residual amounts of SbF5 becomes a distinct possibility. Furthermore, the addition of mesitylene to pure SbF5 resulted in an intensely colored dark brown solid, which was not investigated further. There is little doubt that the final product, like other arene complexes of Group 13 and 14 metals (1-8), is a white solid. The use of gravity filtration as a means of product isolation is important in the case of the Sn(SbF6)2mes synthesis. Attempts to isolate this compound by slow, careful removal of excess mesitylene in vacuo or in a stream of dry N2 at room temperature and at -5°C lead to partial decomposition and materials that analyze as Sn(SbF6)21 .5mes. In addition, the product has a slight yellow tint. The thermal stability of both mesitylene complexes described here is limited. Sn(SO3F)2mes melts with decomposition at 97-98°C while Sn(SbF6)2mes has a decomposi tion range of 69.5-71.5°C, where a greenish yellow liquid forms. Attempts to extend the mesitylene addition reaction to three other tin(ll) compounds were unsuccessful. Both SnC12 and SnF2 were recovered unchanged. Both salts have Sn2+ in distorted octahedral environments (7) and the structures of SnC12 (25) and orthorhombic SnF2 (14,26), one of the two polymorphic forms, may be regarded as layer structures. It can be concluded that the lattice energies of both compounds are too high and adduct formation does not occur in these two compounds. 97 Stannocene,(5-CH)2Sn, has a molecular structure with a tilted arrangement of the C5H-rings (27) and presumably a stereochemically active electron pair. Its 119Sn Mössbauer spectrum shows a quadrupole splitting of 0.86 mms (28). The compound dissolves in mesitylene, but stnocene is again recovered unchanged after removal of the arene in vacuo. The 1H NMR spectrum of stannocene in mesitylene is studied down to -40°C, but only rather subtle changes are noted. The principle resonance due to stannocene (29) shifts from 5.96 ppm in CDC13 to 5.91 ppm relative to TMS in the presence of mesitylene while J1H - 119/7Sn changes from 15.76 Hz to 15.16 Hz. Beginning at -20°C and becoming more pronounced at -40°C, a second, smaller and rather broad resonance at 5.88 ppm emerges with a similar separa tion of the satellite peaks. Mesitylene resonances appear almost in the same position as in pure mesitylene as single lines at 2.28 (CH3) and 6.78 (C-H) ppm and do not shift with decreasing temperature. While there may be a weak interaction between stannocene and mesitylene at reduced temperature, all attempts to isolate a reaction product by low temperature filtration have failed. Therefore, it appears that tin(ll) is sufficiently acidic only when bulky and very weakly nucleophilic anions like A1C14, S03F and SbF6 are present, and when lattice energies are reduced in the resulting layer structures. From a correlation of 119Sn Mössbauer data of dimethyltin(IV) salts (12) it seems that SbF6 is the weaker nucleophile of the two fluoroanions, and formation of a 2:1 complex with mesitylene is not unexpected, while Sn(SO3F)2forms a 1:1 complex only. Complex formation by addition of mesitylene to a solid tin(ll) salt as described here does not lead to the formation of single crystals suitable for structural studies. Hence evidence for the presence of n-arene complexes has to come from two other sources, i.e. 119Sn Mdssbauer and vibrational spectra, to be discussed subsequently. 98 4.3.2 119Sn Mössbauer Spectra Isomer shifts, relative to Sn02 for tin(1I) compounds, fall within the range of —2.1 mm s1, viewed as the borderline between Sn(II) and Sn(IV), and values of about 5.1-7.7 mm s1 for a “bare’t Sn2 have been calculated where a spherically distributed 5S2 electron pair (9,15) is considered. Experimental values obtained have not reached the ionic limit. The highest values reported so far are 4.69 mm s1 for the tin(II) moiety in Sn(II)[Sn(IV)(SO3CF)&(18) and 4.66 mm s1 for Sn(SbF6)2As3(17), where an X-ray diffraction study reveals a distorted nine- coordinated environment for Sn2+. In these close approaches to a Sn2+ ion no quadrupole splitting is observed(17,18); however, with decreasing isomer shift, first line broadening already evident for Sn(SbF6)2,is noted and eventually small but well resolved quadrupole splittings are found (30). Hence 119Sn Mössbauer spectroscopy provides two criteria which determine the ionic character of Sn(II) compounds: (a) the isomer shift which should be high, close to or in excess of 4.69 mm s’ relative to Sn02,and (b) the quadrupole splitting which should be zero. Some 119Sn Mössbauer data of Sn(II) compounds which are of interest to this study are listed in Table 4.1. Of the five tin(II) compounds studied here Sn(SO3F)2(4.18 mm s1) and Sn(SbF6)2(4.44 mm c1) have the highest isomer shifts of those listed and only the former has a small, resolvable quadrupole splitting. Not unexpectedly, the more acidic Sn(SbF6)2is the better acceptor, seemingly capable of coordinating two mesitylene donors, while Sn(SO3F)2 forms a 1:1 complex only. Upon arene uptake the isomer shift is reduced by 0.17 mm s1 for Sn(SO3F)2and by 0.40 mm s for Sn(SbF6)2. This reduction suggests that mesitylene is interacting with Sn(ll) and, judging by the magnitude of the shift, that in the case of Sn(SbF6)2mes both mesitylene 99 Table 4.1: 119Sn Mössbauer Parameters of Relevant Tin(I1) Compounds at 80K Isomer Shift Quadrupole Splitting Compound 8 [mm s1] rel. to Sn02 LLEQ [mm s1] Reference SnF2 (orthorhombic) 3.30 2.15 14 SnF2 (tnonoclinic) 3.49 1.61 14 (r5-CH)2Sn 3.74 0.86 28a (n-C6)S (A1C14 3.93 0 4c SnC12 4.08 0.66 15c Sn(SO3F)2 4.18 0.68 14 Sn(SOF)mes 4.01 1.04 This work Sn(SbF6)2 4.44 0 14 Sn(SbF2mes 4.04 1.13 This work Accuracy limits for 6 and AEQ are ± 0.03 mm 54 100 molecules are probably coordinated to tin. A similar isomer shift of 3.93 mm s1 is reported for (-CH)Sn(AlCl4)26(4c) but no Mössbauer data appear to have been reported for the two SnC12-A1C3phases (9), which would allow an estimation of the decrease in the isomer shift upon arene addition. The reduction in the isomer shift upon binding to mesitylene is accompanied by an increase in quadrupole splitting to 1.04 mmst for Sn(SO3F)2mes whereas for Sn(SbF6)2mes a slightly asymmetrical doublet (see Figure 4.1) with a £S.EQ value of 1.13 mm s1 is obtained. These findings are consistent with trends on 6 and summarized above. The occurrence of quadrupole splitting suggests an increase in asymmetry in the coordination environment of tin(1I) when coordination to only oxygen or fluorine changes to include coordination to the arene as well. For both mesitylene adducts the 1195n Mössbauer parameters obtained remain well within the range reported for typical tin(II) compounds (21). The addition of mesitylene to Sn2, as suggested by the 119Sn Mössbauer data, is expected to cause a change in the manner in which the anions SO3F and SbF6 coordinate to tin. This change in anion denticity is probed by infrared spectroscopy and discussed in the subsequent section. 4.3.3 Infrared Spectra Attempts to obtain Raman spectra of both mesitylene adducts are unsuccessful and quick darkening of areas exposed to laser light indicates thermal degradation is occurring even at very low laser power. Hence evidence rests on infrared spectra obtained on solid films between silver halide windows. Both infrared and Raman spectra for Sn(SO3F)2(14, 23) and Sn(SbF6)2(14) have been reported previously, and for tridentate SbF6-groups present in M(SbF6)2compounds a useful vibrational assignment has been presented (13, 31). As discussed previously in Chapter 3, 101 Figure 4.1: c 0 Cl) Cl) E (I) C 0 H 119Sn Mössbauer Spectrum of Sn(SbF)2mes at 77K 0 1.5 3.0 4.5 0 I0& 98 96 C 0 -75 -6.0 -4.5 -3.0 -1.5 Velocity (mms1) 102 Section 3.3.2, to account for the observed spectral complexity in the SbF-stretching region, a subdivision into metal coordinated, through fluorine bridges SbFb, and non-coordinated or terminal SbFt stretches is suggested with the former usually between 550 and 650 cm1 and the latter above 650 cm1 (31). Vibrational bands reported for Sn(SbF6)2(14) allow such a division as well, even though a distorted environment around tin introduces additional complexity. Assignments for Sn(SO3F)2have suggested the presence of ionic SO3F with the symmetry of the anion reduced from C3,, for the free ion to Cs due to coordination to tin (14, 23), but there are still unresolved problems, best reflected in the occurrence of two equally intense bands at --770 and 830 cm1 possibly due to iSF(A1)and a splitting found for 3S0(A1). It seems more appropriate here to view the spectrum as being due to two non-equivalent SO3F-groups with approximately C3,, symmetry with some band overlap in the areas of deformation modes. This implies either ionic groups or more likely 0-tridentate groups. As mentioned previously, these conclusions are further confirmed by the recently published crystal structure study of Sn(SO3F)2 by Adams et al. (19). Addition of mesitylene to both Sn(II) compounds has two general effects: a) the infrared spectrum is dominated in both instances by bands due to the SO3F and SbF6 groups, respec tively, with all mesitylene bands of very low intensity, and b) the vibrational spectra of the anions appear to change, and at least the SbF6 bands reflect a change in anion coordination. Low intensity of vibrational bands due to organic groups is commonly found for organotin(IV) salts (32), and this feature appears to extend to the mesitylene adducts of tin(U) salts as well. The observed infrared frequencies for the two mesitylene adducts are listed in the experimental section together with estimated relative intensities. The discussion in this section will center around two aspects, (a) the mesitylene bands in neat mesitylene and in the complexes, and (b) the “anion” bands before and after mesitylene addition to Sn(SO3F)2and Sn(SbF6)2,respectively. Vibrational bands attributed to mesitylene are listed in Table 4.2 for 103 Table 4.2: Infrared Bands of Liquid Mesitylene and Bands Attributed to Mesitylene in the Adducts Sn(SO3F)2mesand Sn(SbF6)2mes Mesitylene Sn(SO3F)2mes Sn(SbF6)2mes [cm1]Tnt. [cm1]Tnt. [cm-1]mt. 3018 s 2921 vs 2910w,b 2920mw 2860 s 2860 vw 2860 vw 2730m 1760 m-w 1760 vw 1715 m-w 1608vs 1604m,1575w,sh 1601m,1580m 1512w 1473s 1446w 1469w 1442 m, sh 1375m-s 1384m ll7Ovw 1168vw 1037 s 1080 vw, 1040 vw, 1007 w 932vw 960vw,b 882vw 860w 835 vs 786w 687 vs 667 w 687 ms 561w,b 544m 515w 507vw 512w 443m 440w 421 m 411m 395w See Table 4.3 for abbreviations 104 both adducts and compared to bands observed for liquid mesitylene. The listing of bands for the adducts is incomplete because of the very low intensity mentioned above which causes medium and weak bands to be unobservable and in case of Sn(SO3F)2mes, partial overlap of mesitylene bands with more intense anion bands, in particular in the region of 800-1300 cm’. In regions where no anion bands are expected, small band shifts are observed, mostly to lower frequencies. The aromatic C=C stretch at 1608 cm1 in free mesitylene is a good example. This peak is shifted to 1604 and 1601 cm1 for Sn(SO3F)2mes and Sn(SbF6)2mes respectively. The lower frequency indicates a very slight withdrawal of it electron density of the aromatic ring towards tin. Interestingly, some bands assigned to bonded mesitylene appear to be split compared to those of free mesitylene. This is noticeable in the 1570-1610 cm1 region of both spectra, and especially in the 1000-1200 cm’ region for Sn(SbF6)2mes. The splitting can be attributed to a reduction in symmetry of the bonded mesitylene, and/or the presence of two non-equivalent mesitylenes, in particular in Sn(SO3F)2mes. However some solid state splitting is possible as well. All features affecting bands due to mesitylene are subtle, and the observed features are in general consistent with a very weak interaction between the arene and the two Sn(11) salts. The infrared bands attributed to the anion are summarized in Table 4.3 and are compared to the reported values for the parent compounds Sn(SO3F)2(14, 23) and Sn(SbF6)2(14). For Sn(SO3F)2mes the JR spectrum is dominated by an intense broad band centered at 1054 cm1 and two bands at 1191 and 1127 cm1. These are assigned collectively as SO3 stretching modes, probably masking some of the less intense motions of the mesitylene component in the adduct. 105 T ab le 4. 3: In fr ar ed Fr eq ue nc ie s fo r S n (S O 3F ) 2 a n d S n (S bF 6)2 a n d B an ds A ttr ib ut ed to th e A ni on s in th e M es ity le ne A dd uc ts S n (S O 3F ) 2m e s a n d S n (S b F 6)2 m e s S n (S O 3F ) 2 a S n (S O 3F ) 2m e s A pp ro xi m at e S n (S bF 6)2 b S n (S b F 6)2 m e s [c m ] m t. [c m 1 hi t. Ba nd D es cr ip tio n [c m ] Tn t. [c m ] m t. fo r C3, , A ss ig nm en tC A ss ig nm en tC 12 90 s, sh 13 37 w u SO3 (E ) 74 1m , s h 12 40 vs ll 9l vs 71 3s Sb Ft 67 1v s u Sb F 6 11 80 s, sh 11 27 w ,s h 67 6m s in an d o u t 63 5s ,s h as . (E i) 65 0s o fp ha se 10 62 s 10 54 vs u S O 3 (A1) 62 9s Sb .F b - . 61 3s h in an do ut 83 3s 88 8s o SF (A1) — 57 5s h o fp ha se 58 0m , b U Sb F 6 (E 2g ) 77 2s 78 6w 60 6m S O 3 (E ) 47 7w S b F 6- 59 2m 58 9m s 31 3s h de fo rm at io ns 29 2m 57 3s 54 4m ö S O 3( A 1) 27 1m 55 4s 40 3m 40 5m ö S O 3F (E ) 39 5s h a Re f. 23 b Re f. 14 C Re fs . 13 an d 31 A bb re vi at io ns : s = st ro ng , v = v er y, as = as ym m et ric , m = m ed iu m , sh sh ou ld er , t = te rm in al , w = w ea k b = br oa d b = br id gi ng C The most noticeable change appears to have occurred in the S-F stretching region. The two intense bands at 772 cm1 and 833 cm1 in Sn(SO3F)2 (Table 4.3) are now replaced by a single strong band at 888 cm1,with a weak band at 786 cm1 (possibly due to mesitylene). From this overall band distribution and the increased wave number of the US-F band, it appears that a change in denticity of the anion has taken place due to mesitylene adduct formation. The findings are consistent with a bidentate, possibly bridging configuration tentatively formulated as [(ri6-mes)Sn(SO3F)2]rather than an ionic (perturbed) or tridentate grouping (33). Compared to the rather complicated pattern observed for Sn(SbF6)2in the region of 520-750 cm1 (Table 4.3), the band pattern displayed by the anion in the adduct Sn(SbF6)2mes is very simple: a very strong band at 671 cm1, a sharp, medium band at 635 cm1, and a weak band at 580 cm1 are found. For a free SbF6 anion, a peak at —660 cm’ is attributed to the asymmetric stretch with Raman active bands observed at —660 again and at 575 cm1. The pattern observed for Sn(SbF6)2mes suggests the presence of a wealdy coordinated SbF6 anion. Peak splitting in the SbF6 stretching region produces bands at 671 and 635 crn1, with the loss of a symmetry center allowing detection of a Raman active fundamental at 580 cm1 (34). It appears then that addition of mesitylene has a different effect on the anion bands in Sn(SO3F)2mes and Sn(SbF6)2mes, respectively. In the former adduct the SO3F group seems to be still coordinated to tin, but more likely in a bidentate, possibly bridging manner. In Sn(SbF6)2mes only weak coordination of the SbF6 group is evident and any departure from an ionic SbF6 with °h symmetry is rather slight. It is unfortunate that in particular for this compound support from Raman spectroscopy is lacking due to sample decomposition in the laser light, as mentioned earlier. 107 The findings for Sn(SbF6)2mes are consistent with the view that both mesitylene groups are coordinated to tin, which in turn is supported also by the 119Sn M6ssbauer spectra. 4.4 Conclusion Direct addition of an arene, in this case mesitylene, to suitable Sn(II) salts like Sn(SO3F)2 and Sn(SbF&2 at room temperature, and product isolation well below room temperature, allow the high yield synthesis of veiy wealdy bound mesitylene complexes. However, only microcrystalline materials result which precludes structural studies by single crystal X-ray diffraction. The adducts are characterized by chemical analysis and infrared spectra and their formation is followed by 119Sn Mössbauer spectroscopy. This spectroscopic method allows not only product characterization, but also, using the isomer shift and the absence of quadrupole splitting as criteria, the identification of other suitable substrates for complex for mation. It is observed that only tin(II) compounds with large, weakly nucleophilic anions are capable of forming mesitylene adducts, while SnC12, SnF2 and stannocene do not give any indication of adduct formation under similar conditions. References 1. A.G. Gash, P.F. Rodsiler, and E.L. Amma, Inorg. Chem., j., 2429 (1974). 2.a) H. Schmidbaur, Angew. Chem. mt. Ed. (English), 4, 893 (1985). b) H. Schmidbaur, W. Bublak, B. Haber, and G. Muller, Angew. Chem. mt. Ed. (English), , 234 (1987). 3. P. Jutzi, Adv. Organomet. Chem. , 217 (1986). 4.a) Th. Auel and E.L. Amma, J. Am. Chem. Soc., 90, 5941 (1968). b) H. Luth and E.L. Amma, J. Am. Chem. Soc., 91,7515 (1969). 108 c) P.F. Rodsiler, Th. Auel, and E.L. Amma, J. Am. Chem. Soc., 97, 7405 (1975). 5.a) M.S. Weininger, P.F. Rodsiler, A.G. Gash, and E.L. Amma, J. Am. Chem. Soc., 4, 2135 (1972). b) M.S. Weininger, P.F. Rodsiler, and E.L. Amma, Inorg. Chem., Ia. 751 (1979). 6. H. Schmidbaur, T. Probst, B. Huber, 0. Steigelmann, and G. Muller, Organomet. , 1567 (1989). 7. H. Schmidbaur, T. Probst, B. Huber, G. Muller, and C. Kruger, 3. Organomet. Chem., , 53 (1989). 8. Th. Auel and E.L. Amma, 3. Am. Chem. Soc., 9Q, 5941 (1968). 9. J.D. Donaldson, Prog. Inorg. Chem., , 287 (1967) and references herein. 10. S.H. Strauss, M.D. Noirot, and 0.P. Anderson, Inorg. Chem., 2, 3851 (1986). 11. S.P. Mallela, S.T. Tomic, K. Lee, J.R. Sams, and F. Aubke, Inorg. Chem., , 2939 (1986). 12. S.P. Mallela, S. Yap, J.R. Sams, and F. Aubke, Inorg. Chem., 5, 4327 (1986). 13. M.S.R. Cader and F. Aubke, Can. J. Chem., 2, 1700 (1989). 14. T. Birchall, P.A.W. Dean, and R. 3. Gillespie, J. Chem. Soc. A, 1777 (1971). 15.a) J.K. Lees and P.A. Flinn, Phys. Lett., j, 186 (1965); and J. Chem. Phys., 4, 882 (1968). b) P.A. Flinn, G.K. Shennoy, F.E. Wagner, Eds., HMössbauer Isomer Shifts”, North- Holland Publishing Co., Amsterdam, 1978, p. 593 ff. c) J.D. Donaldson and B.J. Senior, 3. Chem. Soc. (M, 1821 (1967). 16. R.H. Herber and G. Carrasquilo, Inorg. Chem., , 3693 (1981). 17. A.J. Edwards and K.I. Khallow, Chem. Comm., 50 (1984). 18. R.J. Batchelor, J.N.R. Ruddick, J.R. Sams, and F. Aubke, Inorg. Chem., .j, 1414 (1977). 19. D.C. Adams, T. Birchall, R. Faggiani, R.J. Gillespie, and J.E. Vekris, Can. J. Chem., 9, 2122 (1991). 109 20. D. Gantar, I. Leban, B. Friec, and J.H. Holloway, 3. Chem. Soc. Dalton Trans., 2379 (1987). 21. R.V. Parish in “Mössbauer Spectroscopy Applied to Inorganic Chemistry”, Vol. I, Ed. G.J. Long. Dienum Press, New York, 1984. 22. J. Barr, R.J. Gillespie, and R.C. Thompson, Inorg. Chem., a, 1149 (1964). 23. C.S. Alleyne, K. O’Sullivan Mailer, and R.C. Thompson, Can. 3. Chem., , 336 (1974). 24. D.F. Shriver, “The Manipulation of Air-Sensitive Compounds”, McGraw-Hill, New York, 1969. 25. R.E. Rundle and D.H. Olson, Inorg. Chem., 3, 596 (1964). 26. J.D. Donaldson, R. Oteng, and B.J. Senior, Chem. Comm., 618 (1965). 27. A. Almenningen, A. Haaland, and T. Motzfeld, 3. Organomet. Chem., 7., 97 (1967). 28.a) P.G. Harrison and J.J. Zuckerman, 3. Am. Chem. Soc., 9, 2577 (1970). b) T.S. Dory and J.J. Zuckerman, J. Organomet. Chem., 264, 295 (1984). 29. L.D. Dave, D.F. Evans, and G. Wilkinson, 3. Chem. Soc., 3684 (1959). 30.a) J.G. Stevens and V.E. Stevens in “Mössbauer Effect Data Index”, Plenum Press, New York, 1958. b) J.J. Zuckerman in “Chemical Mössbauer Spectroscopy”, Ed. R.H. Herber, Plenum Press, New York, 1984. 31. K.O. Christe, W.W. Wilson, R. Bougon, and P. Charpin, J. Fluorine Chem., , 385 (1982). 32. P.A. Yeats, J.R. Sams, and F. Aubke, Inorg. Chem., ii, 2634 (1972). 33. W.W. Wilson and F. Aubke, Inorg. Chem., fl, 326 (1974). 34. A.M. Qureshi, A.H. Hardin, and F. Aubke, Can. 3. Chem., 4, 816 (1971). 110 CHAPTER 5 A LOW TEMPERATURE MAGNETIC STUDY OF THE MOLECULAR CATIONS O2, Br2 AND I2 5.1 Introduction Only a limited number of compounds with paramagnetic homonuclear ions of non metallic main group elements are known. Of these, compounds with diatomic cations stabilized by very weakly basic fluoroanions are the subject of this study. The cations in this group consist of the dioxygenyl cation Oj’, first discovered in O2[PtF6] by Bartlett and Lohmann (1), and the two dihalogen cations I2 and Br2, first identified in solutions of strong acids and super- acids by Gillespie et al. (2,3). Single crystal X-ray diffraction studies were subsequently reported forBr2[Sb3F16] (4) andI2[SbF1J (5), while powder diffraction studies have afforded a more limited structural insight into a number of dioxygenyl salts (6). For the dihalogen cations, electronic spectra of the solvated species (2,3), resonance Raman spectra (7), and magnetic measurements at room temperature (2,4(a)) have allowed some information on the electronic structure, suggesting a 2fl3jg ground state (10). In addition, the study by Herring and McLean (8) has pointed out the close similarity between the solvated cations I2+(1v) and Br2(SOlV) and their gaseous counter parts as studied by photoelectron spectroscopy (9) or molecular spectroscopy (10). This strong resemblance should permit at least an approximation of the energy separation to the nearest excited states, not only for the dihalogen cations, but also for 02+, where a21,’2g ground state is indicated (10). 111 The objectives of the present study which involves low-temperature magnetic suscep tibility measurements on suitable compounds containing the three homonuclear diatomic cations can be summarized as follows: (i) To investigate the magnetic behavior of what appear to be the only three suitable paramagnetic molecular cations formed by non-metals. Structural and spectroscopic information mentioned previously should help in the interpretation. (ii) To explore whether, and to what extent, van Vleck’s theory of molecular paramagnetism (11), developed 60 years ago, can be applied to solid-state cations. So far, nitric oxide, NO, has served as the best example; however, useful experimental data for this molecule do not extend into the condensed phase due to intermolecular association. (iii) To examine the reasons why the magnetic moments reported so far for O2 salts fall well below the spin-only value of 1.73 B• Measurements down to very low temperatures should allow the detection of magnetic behavior consistent with antiferromagnetic exchange, which is a possible cause for reduced magnetic susceptibilities. (iv) To compare the magnetic susceptibilities ofI2[SbF11J and Br2[Sb3F16J below —80K, since earlier measurements above that temperature (12) had suggested anti- ferromagnetic coupling forI2[SbF11],but not forBr2[Sb3F16j. 5.2 Experimental BothBr2[Sb3Fi& andI2[SbF1i] were synthesized according to the method reported by Wilson et al. (12). This method involved the oxidation of previously purified Br2 and resublimed 12 by bis(fluorosulfuryl) peroxide, S206F at a precise 2:1 mole ratio, and the 112 subsequent solvolysis of the product mixture in an excess of freshly distilled antimony(V) fluoride. The reactions were followed by weight and the purity checked by melting points. Br2[Sb3F16] (4.777 g, 5.762 mmol) was obtained from 0.924 g (5.782 mmol) of Br2 and 0.5724 g (2.89 1 mmol) ofS206Fafter solvolysis in —10 g of SbF5 at room temperature and removal of all volatiles in vacuo. The bright red solid melted at 84 ± 1°C (lit. 85.5°C) (12). I2[SbF1J- (2.775 g or 3.928 mmol) was synthesized from 0.9986 g (3.934 mmol) ‘2 and 0.3895 g (1.967 mmol) ofS206F and subsequent solvolysis in —12 g of SbF5 at 50°C. The black-blue solid isolated after the removal of all volatiles melted at 129 ± 1°C (lit. mp. 127°C) (12). A sample of02[AsF6j was obtained from Dr. Karl 0. Christe of Rocketdyne. The compound was synthesized by UV photolysis of 02, F2, and AsF5 in quartz. Magnetic field strengths of 7501, 9225, and 9625 G and a temperature range of —2-124K were used in this study. Molar magnetic susceptibilities, XM, were corrected for diamagnetism and for Br2[Sb3F16]-andI2[SbF1], slightly larger X values were used than reported previously (12). The magnetic moments of the Br2 and compounds were corrected for temperature- independent paramagnetism (TIP = 120 and 68 x 10 cm3 mol1 respectively, calculated as described in the text). The Curie-Weiss law in the form XM = CmITO is used throughout. 5.3 Results and Discussion The discussion of the magnetic behavior of the three cationic species will center around three aspects: (i) The selection of suitable compounds for this study in the light of previous magnetic susceptibility studies. 113 (ii) The magnetic measurements onO2[AsF6],I2[SbF1], andBr2[Sb3F16J. (iii) An attempted interpretation of these results. 5.3.1 Synthesis Magnetic measurements on a single compound in the current study require —300-500 mg of high-purity sample for duplicate measurements when the vibrating sample magnetometer is used, while the Gouy measurements require between 1 and 2 g. All cations are extremely reactive and only ‘2 is sufficiently stable in HSO3F- bF5 solution (2,12), not undergoing disproportionation or further oxidation. This rules out purification by recrystallization as a general procedure for all the three salts. The limited thermal stability reported for all materials has so far precluded purification by sublimation. Hence, synthesis on the desired scale from pure starting materials, with little or no chance for side-reactions and facile product isolation, becomes the only route to the compounds of this study. There appears to be no real choice among the Br2 containing compounds, since structural and spectroscopic details on Br2[AsF6] (14), the only other Br2 salt, are not readily available. The original synthesis ofBr2[Sb3F16} (4), from BrF5,Br2, and SbF5, did produce a sample suitable for a single-crystal X-ray diffraction study, but judging by the reported physical data, in particular the melting point and a magnetic moment of Peff — 1.6 B’ the product obtained from this method appears to be unsuited for this study. A similar problem arises in the case of the ‘2 species as well. The method used for the preparation of single crystals ofI2[SbF11] (5), the oxidation of 12 by SbF5 in SO2, is not suited for the preparation on a larger scale, as is evident from the reported analytical data (5). The problem appears to be the quantitative separation of the by-products SbF3 or SbF35. 114 Formation of anions of the type Sb3F14 becomes a possibility and, in addition, a diamagnetic compound of the compositionI2[(SbF6) b14] has recently been obtained by this synthetic route (15) with reaction conditions only slightly different. Substitution of SO2 by AsF3 as solvent appears to lead toI2[SbF1] of a higher purity, butI2[Sb3F16] may form as a by-product as well (16). Any uncertainty in the molar mass of the product will limit the usefulness of a magnetic study. Other materials considered unsuitable for this study, as mentioned in Chapter 1, include the substances formulated as (SbF5)21(17) and the insufficiently characterized12[’aF6J (17). An alternate synthetic route to bothI2[SbF1i] andBr2[Sb3F16],the oxidation of 12 or Br2 by stoichiometric amounts of bis(fluorosulfuryl) peroxide and subsequent solvolysis in an excess of SbF5 according to: 50°C 212 + S206F + 8SbF5 > 2I[SbF11] + 2SbF9SO3, [5.1] 25°C 2Br + S206F + 1OSbF5 > 2Br1Sb3F6] + 2SbF9SO3, [5.2] has produced sufficiently large quantities of both compounds for a magnetic study between 298 and 80 K using the Gouy technique (12). With all starting materials readily purified and the volatile byproduct Sb2F9O3(18) easily separated from the product, this method, employed in the present study, was chosen as the most appropriate route to the synthesis of bothI2[SbF1i] andBr2[Sb3F16]. With well over a dozen different O2 salts reported so far, this group of compounds appears to offer a wider choice, but here additional problems surface, which seem to have adversely affected earlier magnetic studies. Some of the reported compounds, O2[PtF6] (1), 115 O2[PdF& (19),02[RuF6j (20), orO2[RhF6] (20) have two different paramagnetic centers in the same molecular unit. The initial approach by Bartlett and Beaton (21) involving the subtraction of the molar susceptibilities of N0[PtF&- from those obtained for02[PtF6] has clearly allowed th identification of 02+ as a paramagnetic ion; however, magnetic exchange between 02 and Pt(V) (d5) is a plausible contributing factor for the observed decreasing trend of the magnetic moments attributed to 02+ with decreasing temperature. Magnetic ordering due to 02 Pt(V) exchange has subsequently been reported for this compound (23). In addition, the high-pressure synthesis of dioxygenyl salts from 02/F mixtures and either metal fluorides or the metal itself (20,22,24) in metallic reactors has often resulted in materials with ferromagnetic contaminants (19,23). Although monel impurities can be recog nized by the characteristic Curie temperature of 335 K and corrections have been attempted (23), such materials remain suspect in magnetic susceptibility studies. This is due to the possible existence of weak antiferromagnetic ordering in such materials contaminated with monel. UV photolysis in quartz vessels (25) represents a more suitable synthetic method, but a number of 02 salts such as02[BF4] (26) and02[GeF5] (27) prepared in this manner show rather limited thermal stability. Nevertheless, a magnetic study of02[BF4J between 293 and 85 K has been reported (28a), but only limited conclusions were reached. For the products obtained from the reaction of 02 and F2 with SbF5 by either UV photolysis or high-pressure synthesis, some structural ambiguity has been noticed. This problem arises since in these reactions, a variety of [SbF5+iJ type anions may be formed to stabilize the 02 species. Both 02[SbF6j and 02[SbF1i1 are well characterized (6,24), and O2[Sb3F16] has been postulated as well (29). The problems caused by this structural 116 ambiguity are apparent from magnetic susceptibility studies on “O2[SbF6j”reported in the Soviet literature (28). To avoid similar problemsO2[AsF6] obtained by UV photolysis (25) was chosen for this study. An additional reason for this choice is the availability of structural information on this compound. A phase transition at 255 ± 3 K (30) results in a rhombohedral distortion of the cubic structure found at room temperature (6) This distortion is also evident from ESR studies down to 4 K (23,3 1). Conclusions reached regarding the ground state of 02 in these ESR studies are useful in the interpretation of our low temperature susceptibility results. There have been two previous magnetic susceptibility studies of02[AsF6] down to 4 K, with rather contradictory results. In a study by Grill et al. (32), no magnetic ordering of the 02+ cations is observed down to 4 K, while weak 0202+ exchange is suggested in another study (23) based on the small Weiss constant obtained. However, these conclusions are rendered somewhat tenuous because of ferromagnetic monel impurities in the sample, requiring correction of the susceptibility data (23). In all previous magnetic susceptibility studies of 02k, the magnetic moments obtained over the whole temperature range are well below the spin-only value of 1.73 B’ and explana tions have varied from suggesting “van-Vleck behavior” (22) in analogy to the behavior of NO (11), to invoking the presence of 17% of an unidentified, magnetically inert impurity in 02[AsF6] (32). Likewise in some studies down to 80 K (28), values for O2[SbF& appear to decrease with decreasing temperature, while for the structurally related02’IAsF6], I1eff is said to be invariant with temperature to 4 K (32). It is unclear why there are so many discrepancies and contradictions in previous studies. It is felt, however, that sample identity and purity play a major role here. 117 5.3.2 Magnetic Measurements The results of our magnetic susceptibility studies onBr2[Sb3Fid below 80 K, together with results obtained in an earlier study (12) at higher temperatures, are summarized in Table 5.1. Both sets of XM and eff data have been calculated using the same diamagnetic correction and TIP values (279 and 120 x 10-6 cm3 mo11,respectively). The agreement between the data sets in the overlap region is excellent for this compound. The data obtained here forI2[SbF11] andO2[AsF6} are given in Table 5.2. Because of poor agreement in the 100 K region with the earlier Gouy data (12), the current study was extended with the vibrating sample magnetometer up to 124 K for the I2 compound. Addition ally, the measurements for the 02+ compound was also extended down to 2 K in the hope (not realized, unfortunately) of observing a maximum in the susceptibility data of this compound at very low temperatures. Pertinent structural and spectroscopic features of the three compounds studied and of the three cations, 02+, Br2+, and 12+, either as solvated or gaseous species, are summarized in Table 5.3. The contrast in the magnetic properties of the three paramagnetic cationic species studied in this work is clearly evident from the plots of magnetic moment versus temperature shown in Figure 5.1. Two of the compounds,02[AsF6] and Br2[Sb3F16],exhibit Curie Weiss behavior over a wide temperature range, as is seen by the plots of 1/XM vs. T given in Figure 5.2. 5.3.3 Br2[SbF16] The magnetic behavior ofBr2[Sb3F16}appears to be rather straightforward and will be discussed in detail first. The 1/XM versus T plot of this compound indicates Curie-Weiss 118 Table 5.1: Magnetic Data ofBr2[Sb3F16] Temperaturea XM X 106 L’eff (JIB)b (K) (cm3mol1) 297 1870 2.04 275 2010 2.04 255 2160 2.04 235 2330 2.03 213 2530 2.03 193 2780 2.03 172 3070 2.02 151 3430 2.00 131 3900 1.99 116 4500 2.02 104 4910 2.00 80.0 6130 1.96 81.1 6070 1.96 73.5 6720 1.97 65.7 7440 1.96 55.0 8870 1.96 43.7 10900 1.94 31.4 15000 1.93 21.4 21900 1.93 10.2 45700 1.93 9.88 45900 1.90 4.20 111300 1.93 a First twelve data points from Ref. 12. b Corrected for TIP using 11eff = 2.828 ((XM - TIP)T)1t with TIP = 120 x 10-6 cm3 mo11. 119 Table 5.2: Magnetic Data ofI2[SbF1] and O2[AsF6i I2[SbF1J- O2[AsF6] Temperature XM X 106 ffa (RB) Temperature XM x 106 ffb (iB) (K) (cm3mo11) (K) (cm3mo11) 124.1 3770 1.92 80.1 4170 1.63 117.5 3860 1.89 76.3 4410 1.64 106.5 4290 1.90 72.7 4650 1.64 98.6 4510 1.87 68.5 4950 1.65 90.9 4730 1.84 64.3 5480 1.65 84.3 4900 1.81 59.2 5600 1.63 81.2 5030 1.80 53.2 6170 1.62 76.8 5120 1.76 46.5 6990 1.61 71.8 5300 1.73 42.9 7520 1.61 66.6 5380 1.68 39.1 8200 1.60 63.5 5470 1.66 30.3 10530 1.60 58.4 5560 1.60 25.3 12430 1.59 51.9 5560 1.51 20.5 15100 1.57 44.9 5090 1.34 15.3 19700 1.55 36.9 4910 1.20 10.1 28800 1.52 31.6 4700 1.08 7.12 37500 1.46 27.4 4490 0.98 5.92 44500 1.45 27.1 4520 0.98 5.28 48300 1.43 21.8 4300 0.86 4.78 52200 1.41 11.6 4270 0.62 4.62 52700 1.40 8.64 4290 0.54 4.40 54800 1.39 6.30 4490 0.47 4.04 59400 1.39 4.54 4790 0.41 2.70 70400 1.23 2.40 75500 1.20 2.10 81300 1.17 a Corrected as in Table 3.1 with TIP = 68 x 106 cm3 mo11 b Not corrected for TIP 120 Table 5.3: Structural and Spectroscopic Information on O2IAsF6],I2[SbF1i]’ Brj’1Sb3F16] ,and the Corresponding Ions O2, Br2and I2 Data, X Denotes 0, Br, or I O2[AsF6J Br2[Sb3F16J I2[SbF11] or °2g) or Br2+(g) or12g) Shortest XX non-bonding 4 to4•05a 6.445 4.29 interaction [A] (ref. 28b,30) (ref. 12, 4b) (ref. 12, 6) van der Waals Radii 1.52 1.85 1.98 for X [A] (ref. 34) Ground state of X2(g) 2hhl/2g 23f2g 23/2g Spin orbit coupling 185 (ref. 52) 2820 ± 40 5125 ± 40 (ref. 53) (ref. 53) const. [cm1]for X2(g) 195 (ref. 10) 3141 ± 160 5081 ± 60 (ref. 55) (ref. 55) tS.E(2fl1/2g2h13/2 ) As above forX2(l) [cm1] --- 2890 (ref. 12) 5190 (ref. 2) Diamagnetic Correction (ref. 56) 79 279 238 [10-6 cm3 mo11] a Estimated from powder diffraction data 121 Figure 5.1: Magnetic Moment vs. Temperature of Br2[Sb3F16],I21SbF1], and O[AsF6] 2- __ - - - 0— - --CD 0- 8 - — .-. — _ . ..._..- —.— 1.5- F , //z Lii 1 / /0 4 / 1 4 Z / ° Br2[Sb3F16j o / / .I[sb11]— • O2[AsF6 0.5- w / 0- 0 20 40 60 80 100 120 140 TEMPERATURE,K 122 -J 0 () -J 0 w 0 U) D U) C-) z C, 0 TEMPERATURE,K 100 Figure 5.2: Inverse Susceptibility vs. Temperature of O2[AsFj andBr2[Sb3F16] 0 0 250 200 150 100 50 0 0 0 ,0 ‘Pt / A’ 0’ / 0’ / 0’ / / / / / / / 0’ / .— .-. 0 O2[AsF6J . Br[Sb3F16 0 20 40 60 80 123 behavior over the temperature range from 80-4 K (Cm = 0.49 ± 0.01 cm3 mo11 K, 9 = -0.74 ± 0.01 K). The magnetic moments (even when corrected for TIP) decrease slightly with temperature. A TIP correction had been arbitrarily assigned in the earlier work (12) in order to bring the magnetic moments in the high-temperature region into agreement with the value of 2.0 B predicted by theory (11) for a species with a2H3g ground state and no thermally acces sible excited state (see Table 5.3). The correction for TiP in the present study was taken as 4N32/3zE (33), where sE is the energy separation between the ground 2fl3g and the next excited state with which it is mixing. The AE value for Br2 is the spin-orbit coupling constant, , and for this species, estimated at 2890 cm1 for the solvated ion (Table 5.3), and using this value a TIP of 120 x 10-6 cm3 mo11 is calculated. Magnetic moments calculated using this TIP correction are within ±2% of the theoretical value of 2.0 B over the wide temperature range of 297-55 K (Table 5.1). The cause of the very small drop in the moment, particularly at very low temperatures (corresponding to the effect of a Curie-Weiss rather than Curie law behavior) is not certain. Although the shortest non-bonding BrBr contact at 6.445 A (4b) is too long to invoke significant direct magnetic interaction (Figure 5.3), F”Br contacts ranging from 2.86-3.34 A, comparable to or shorter than the sum of the van der Waals radii of 3.32 A (34), suggest the possibility of very weak exchange via bridging anions. However, the effect is very minor and it is reasonable to conclude that inBr2[Sb3F16] the Br2 cation is in a thermally isolated ground state and there is no conclusive evidence for any magnetic exchange between para magnetic centers down to 4.2 K. In this regardBr2[Sb3F16] is somewhat unique, since for all other paramagnetic main group species, including 02 (35), the recently studied ozonides, K03 Rb03,and Cs03 (36), and solid 02 (37), as well as ‘2 and O2 (discussed below), at least weak antiferromagnetic coupling has been suggested. 124 Figure 5.3: [Br] [Br] Crystal Structure of Br2ISb3F16] (redrawn using data from ref. 4b) a b C V z X—X, bond length X ... F, closest contact tSb..F (terminal) Sb...F(btidIng) R tbrid,Irter B4Sb3F1I Menoclinic 13.58 i 0.02 A 7.71 * 0.01 A 14.33 ±0.02 A 93.7 a 0.2 1497 A3 3.68 g cm3 4 2.15 A 2.86 A 1.83 A 2.10; 1.97 A 1,11 0 r Br—Br2 = 6.445A 125 5.3.4 O[MF(] The magnetic properties of02[AsF6] are similar to those of the Br2 compound in that Curie-Weiss behavior is followed (Figure 5.2), although over a more limited temperature range of 60-2 K. Here the Curie constant Cm, is significantly smaller and the absolute value of the Weiss constant, 101, greater (C = 0.34 ± 0.01 cm3 moi1 K, 0 = -1.90 ± 0.01 K). Consequently, the effective magnetic moments are significantly smaller for02+[A5F6J and show a stronger temperature dependence, particularly in the low temperature region (Figure 5.1). A slight departure from linear behavior in the Curie-Weiss plot is noted (Figure 5.2) above 60 K. This may be significant, as it coincides with the observed broadening of the ESR line (23), possibly caused by tumbling motions of 02 in the crystal lattice. These findings are at variance with a previous report by Grill et al. (32), who observed Curie-Weiss behavior for 02+ with a positive 0 value (according to our formulation of the law) of 0.7 K. The results of our study are in reasonable agreement with those of DiSalvo et al. (23), who obtained Cm = 0.309 cm3 mol-1 K and 9 = -0.8 K (sign changed from (23) to conform to our formulation of the law), although it must be taken into consideration that a temperature- independent paramagnetism contribution as well as corrections for ferromagnetic impurities were made in the earlier work. In this work the susceptibilities of 02 were not corrected for TIP. A crystal-field analysis of ESR data for02[AsF6]- suggests the separation between the ground and first electronic excited state to be 1480 cm1 (Figure 5.4), caused by both spin-orbit coupling and crystal-field splitting due to 02 being in a site of orthorhombic or lower symmetry (31). In this case one is clearly not dealing with pure21i2g and 2113flg states and it is not clear what the TIP correction should be. In any event, the large separation between the states means such a correc tion will be small, and hence ignoring it should not significantly affect the conclusions. 126 Figure 5.4: Energy Level Diagram of the Dioxygenyl Ion with the a and it-Bonding 2p Orbitals (Ref. 31) 2pJ* _ __ T 2pit (2÷2)j- 2p ii 2p H (a) (b) (c) free ion spin—orbit spin—orbit coupling. ? coupling + orthorhombic field, 127 When considering the magnitude of the magnetic moment, it is observed that forO2[AsFj the measured magnetic moments in the range of 1.6 B at the higher temperature region agree well with values predicted using the average g values, gay of 1.89 (23) and 1.905 (31) obtained by ESR. Using the equation for the magnetic moment of a single electron B = g[S(S+1)11/2,values of 1.64 and 1.65 B are obtained respectively for the above two gay data. As described by Goldberg et al. (31), both crystal-field splitting and spin-orbit coupling contribute to the ground state inO2[AsF6]-. In a pure2fllag ground state the spin and orbital contributions cancel, and there is no first-order paramagnetism associated with this state (33). If this were the case for O2 with negligible thermal popuation of excited states, only temperature-independent paramagnetism would be present. This is clearly not the case here. The crystal-field interactions have the effect of quenching the orbital contribution, although not completely, resulting in a ieff value slightly below the spin-only value. Quite the opposite effect is observed for the alkali metal superoxides, MO2 with M = Na, K, Rb, Cs (35). Here the ground state is and in all instances the .teff values are well above the spin-only value of 1.73 B’ as was found forI2[SbF11] andBr2[Sb3F16Y(12). The decrease in neff values with decreasing temperature forO2[AsF6], particularly significant below —20 K, requires further consideration. With a splitting of 1480 cm1 separat ing the ground state from the nearest excited state, the variation in the effective magnetic moment with temperature observed here cannot be accounted for by the sort of van Vieck behavior observed for gaseous NO (33). The existence of antifeffomagnetic exchange between paramagnetic centers must be considered as a possibility. The limited structural information available on O2[AsF6} (25,31,38) derived from powder diffraction data does not permit an accurate determination of the shortest OO non bonding contact distance. Estimates of 4.00 (30b) and 4.05 A (32) respectively, which were 128 used to argue against 0”{) interaction, are simply half the unit cell length ad2 for the room- temperature phase (25,31,38). The distance ad2 describes the separation between the mass centers of the 02+ cations, which are assumed to be in the most probable sites. The low- temperature phase, which is directly relevant to this study, has a powder diffraction pattern that has not been successfully indexed (30); hence the conclusion (32) that there can be no antiferro magnetic exchange in02[AsF61-because the 02 ions are too far apart is most likely invalid. For02[MnF9],where a single crystal X-ray diffraction study at room temperature and at -150°C is reported (39), non-bonding 00 contacts of 3.86 A are detected at -150°C with values of 3.98 A observed at room temperature. However, the anion is described as a double chain of cis-bridged MnF6 octahedra with 02 cations between the anion layers, and the resulting structure is not comparable with that of02[AsF6]. Moreover, susceptibility studies on this material would not be suitable for the purposes of this work as the magnetism would be complicated by the presence of Mn4+, a second paramagnetic center. With the sum of the van der Waals radius 3.02 A for two oxygen atoms, one would reasonably expect distances of 3.2-3.3 A as the limit for significant direct antiferromagnetic exchange. This estimation is corroborated by recent reports on the structures of the alkali metal ozonides K03 (40,41,42), Rb03 (40,4 1) and Cs03 (43), all of which exhibit significant anti ferromagnetism (34) than02[AsF6], and where 00 non-bonding contacts range between 3.01-3.15 A for K03 and 3.01-3.30 A for Rb03, respectively. For Cs03, where antiferro magnetic exchange is rather weak (36), only powder data are reported (43) and indexed, suggest ing Cs03 to be iso-structural with Rb03. The unit-cell dimensions indicate slightly longer 00 contacts for this compound (42). Contact distances of the order of those observed in the ozonides could be present in02[AsF61 and this could account for its observed magnetic properties. 129 5.3.5 I[SbF11] In a previous report from our group on12+[SbF1] it was shown that a TIP correction in excess of that assumed for the Br2+ compound was required to bring the experimental room- temperature moment into agreement with the value of 2.0 B predicted by theory (12). In addition, it was suggested in the earlier work that the observed decrease in the magnetic moment on decreasing the temperature to 80 K may arise from antiferromagnetic exchange between contiguous cations in the crystal lattice, where the shortest Fl non-bonding contact distance (5) is found to be 4.29 A (Figure 5.5). Magnetic measurements onI2[SbF1] down to 4.2 K reported here confirm the presence of antiferromagnetic coupling. The susceptibility rises to a maximum at —54 K, then decreases on further cooling (Figure 5.6). The rise again in suscep tibility on cooling below l0 K is probably due to trace amounts of paramagnetic structural impurity, as is often observed in antiferromagnetically coupled systems (13). In view of the clear evidence that this system is exchange-coupled, there is no justifica tion for arbitrarily reducing the room-temperature moment to 2.00 B with an appropriate TIP correction, as was done previously (12). Indeed, in view of the fact that the 23/2j 21t2g separation is significantly greater in the compound compared to Br2 (see Table 5.3), any TIP would be expected to be smaller in the former. TIP for (calculated as described above for Br2j is 68 x 10.6 cm3 mo11 and magnetic moments calculated employing this value are given in Table 5.2. The decrease in with decreasing temperature is consistent with anti ferromagnetic coupling; interestingly though, the absolute values in the high-temperature region are in excess of the theoretically predicted value of 2.0 B (11). An attempt is made here to analyze the magnetic susceptibility data of I2 (including the data from 295-130 K from ref. (12)) according to three theoretical models for antiferromagnetic exchange in one-dimensional systems. For a magnetically concentrated system with spins Si 130 Figure 5.5: Crystal Structure of Ij[Sb2F11] (redrawn using data from ref. 5) 2] r1+ T+4.29A £2 2 ISb2Fj Monoclinic I 13.283(5)A b L314(3)A C 5.57J(2)A P 103.75(2) V 597.5 A3 Dcsi 3.92gcm 2 2 X—X, bond enth 2.557(4) A X ... F, closest contact 2.89 A rSbF(termin.I) 1.85 A tSb_E(brIdgng) 2.00 A R brd./’ter 1.08 131 >- F— -j F— U- LU C U) D U) C-) F— LU z Figure 5.6: Magnetic Susceptibility vs. Temperature ofI2[SbF11]. (Circles are experimental data; Solid line is the best fit to the Ising S’ = 112 model) 0 rO C-) r 0 5 o0c1 0 0 0 100 200 300 TEMPERATURE (K) 132 and S, the nearest neighbor exchange coupling is given by the Hamiltonian Hex () N H = -23 Z aSjzSz + I3S1SJX + ySYSY [5.3] where S1Y and S1z are the components of the spin vector S1 of the ith atom, N the number of magnetic atoms and 3, the exchange coupling constant, which can either have a positive or negative value, i.e. ferromagnetic or antiferromagnetic interaction. The values of a, f3 nd y in equation (4.3) define the nature of the exchange coupling. When a = = = 1, the isotropic Heisenberg exchange Hamiltonian results. In the extreme situation where a = 1 and f = y = 0, the Ising model is described (anisotropic coupling). The isotropic Heisenberg Hamiltonian has been extensively examined but no exact solutions are presently known. However, results of many approximate methods exist (45). For the Ising model Hamiltonian, closed-form solutions are available for both the parallel and perpendicular magnetic susceptibilities of the S = 1/2 system (46). Considering the2113/2g ground state of I2 as having an effective spin S’ = 3/2 (47), our data were fitted to an isotropic Heisenberg model developed by Weng (48), employing the following polynomial expression and appropriate coefficients given by Hiller et al. (49) for the molar susceptibility: Ng23 XM = kT [A + Bx2][1 + Cx + Dx31’ [5.4] where x = kT/IJI. 133 In the above equation (and equations [5.6] and [5.7], see below) g is the Lande splitting factor, the Bohr Magneton, N the number of spins in the lattice, k Boltzmann constant, T the temperature and 3 the exchange coupling constant. The values of the coefficients for S=1/2 and S=3/2 are as follov s: S A B C D 1/2 0.2500 0.18297 1.5467 3.4443 3/2 1.2500 17.041 6.7360 238.47 Allowance was made for paramagnetic impurity by modelling it as a Curie paramagnet with a g value equal to that of the bulk sample, i.e. =Ng2f3S(S+1)/3kT. Expressing XM from equa tion [5.4] as hain’ the susceptibility expression including paramagnetic impurity is X [lP1x.jn +1Xpara The experimental data were fitted to this expression using g, J, and P as fitting parameters. The best fit was considered to be that set of fitting parameters which gave the minimum value of the function F (50): n F = [‘/ E (X1calcd - X’obs / X’b)2]1” [5.5] where n is the number of data points, and X’cacd and X’0b5 are the calculated and observed susceptibilities, respectively. For S’ = 3/2, the best fit of the data for I2 was with 3 = -7.2 cm1, the effective g = 1.23, P = 0.0035, and F = 0.0560. In view of the molecular anisotropy of the magnetic species, the anisotropic Ising model should be examined as well. Unfortunately, while an exact solution exists for the parallel susceptibility of the Ising S = 3/2 case (51), there is no solution for the perpendicular case and hence one cannot analyze the powder data according to 134 this model. An alternative approach to the data analysis is to consider the possibility that the four fold degeneracy of the ground 2fl3flg state has been lifted by second- or higher-order interactions with excited states, leaving a thermally isolated Kramer’s doublet as the only significantly thermally populated ground level. In this case one needs to consider an effective spin, S’ = 1/2, and here both Heisenberg and Ising models are available. Employing again the Weng model and equation [5.4], a best fit between theory and experiment was obtained with 3 = -28 cm1,effective g = 2.72, P = 0.0061, and F = 0.0487. As mentioned earlier, exact solutions for both the parallel and perpendicular suscep tibilities for the anisotropic Ising S’ = 1/2 case are available, and the data were analyzed using the equations of Fisher (46): (Ng22 \ /2IJ1 ‘\ Xii I lexp.i J [5.6] \ 4kT/ \kT/ / Ng232\ / IJI 131 / (IJI) = I 1[tanh j — J + Sech2 j JI [5.7] \81J1 / \kT/ kT \kT / Assuming Xwder = 1/3XII + 2/3, the best fit was generated with 3 = -38 cm1, effective g = 2.61, P = 0.0054, and F = 0.0368. The agreement between experiment and theory for this Isihg S’ = 1/2 case is illustrated in Figure 5.6 where the solid line is calculated from theory. The agreement between experiment and theory for the other two models is visually very similar to that shown in Figure 5.6, although in both cases the agreement is slightly poorer, as is indicated by the higher F values. In all three cases, as the temperature is lowered the experimental susceptibilities rise more steeply to a higher maximum value and then decrease more steeply 135 when compared to the calculated susceptibilities. The analysis of the magnetic susceptibility data clearly indicates that in the lattice of 12[SbF1] contiguous ions are relatively strongly antiferromagnetically coupled, although it does not provide a clear answer to the question of whether or not the thermally occupied ground state at low temperatures is a pure 23Rg state. Measurements of magnetic suscep tibilities at low temperatures on suitably oriented single crystals to obtain values (47) would be informative, but would require the synthesis of relatively large crystals of the material. It is important to note that the antiferromagnetic coupling observed here forI2[SbF11J is distinctly different from the dimerization process suggested for ‘2(1v) in HSO3F(2), just as the solid-state structure ofI2[SbF11] differs from the structures ofI42[AsF6] and I42[(SbF6) b3F14J(15) where square planar, diamagnetic 142+ cations are found. 5.4 Conclusion Magnetic susceptibility measurements to 4.2 K are reported for O2[AsF6j, Br2[Sb3F16], andI2[SbF11j. The data are interpreted utilizing previous results from photoelectron spectroscopy of the gaseous cations, known crystal structures, magnetic studies on the superoxide ion and the ozonide ion, and in the case of02+[AsF6],previous ESR studies. The magnetic properties of the three materials are quite different. Br2[Sb3F16]obeys Curie-Weiss law between 80 and 4 K: Cm = 0.49 ± 0.01 cm3 mo11 K and e = -0.74 ± 0.01 K (with TIP = 120 x 10-6 cm3 mol1). The magnetic moment decreases slightly from 2.04 B at room temperature to 1.93 B at 4 K. I2[SbF11] exhibits relatively strong antiferromagnetic coupling with a maximum in X observed at 54 K. the magnetic moment (corrected for a TIP contribution of 68 x 10 cm3 mo11) decreases from 1.92 B at 124 K to 0.41 B at 4 K. 136 Experimental susceptibilities for this compound over the range 300-4 K have been compared to values calculated using three different theoretical models for extended chains of antiferro magnetically coupled paramagnetic compounds. The major difference in magnetic behavior between Br2+ and 12+ is due to structural difference betweenBr2[Sb3F16]andI2[SbF1J. Magnetic exchange through contiguous ‘2 ions is suggested by the crystal structure ofI2[SbF11J. The shortest FI non-bonding contact is 4.29 A, comparable to the sum of the van der Waals radii of 3.96 A. In Br2[Sb3F16] the shortest Br”Br contact is 6.445 A, about 2.8 A longer than the van der Waals distance, and direct magnetic exchange becomes improbable as discussed above. O2[AsF6] exhibits Curie-Weiss behavior over the range 60-2 K (Cm = 0.34 ± 0.01 cm3 mo11 K, 8 = -1.90 ± 0.01 K). The magnetic moment, uncorrected for TIP, varies from 1.63 B at80Kto1•7B at2K. Finally, in theO2[AsF6],there appears to be weak antiferromagnetic coupling that may involve either super-exchange through intervening AsF6 anions (the smallest anion of the three encountered in this study) or even direct weak O0 interaction. References 1.a) N. Bartlett and D.H. Lohmann, Proc. Chem. Soc., 115 (1962). b) N. Bartlett and D.H. Lohmann, 3. Chem. Soc., 5253 (1962). 2. R.J. Gillespie and J.B. Milne, Inorg. Chem., , 1577 (1966). 3.a) R.J. Gillespie and M.J. Morton, Chem. Comm., 1565 (1968). b) R.J. Gillespie and M.J. Morton, Inor. Chem., II, 586 (1972). 137 4.a) A.J. Edwards, G.R. Jones, and R.J. Sills, Chem. Comm., 1527 (1968). b) A.J. Edwards and G.R. Jones, J. Chem. Soc. A, 2318 (1971). 5. C.G. Davies, R.J. Gillespie, P.R. Ireland, and J.M. Sowa, Can. J. Chem., 52, 2048 (1974). 6. R.J. Gillespie and 3. Passmore, Adv. Inorg. Radiochem., fl, 49 (1975) and references therein. 7. R.J. Gillespie and M.J. Morton, J. Mol. Spectrosc., 30, 178 (1969). 8. F.G. Herring and R.A.N. McLean, Inorg. Chem., 11, 1667 (1972). 9. D.W. Turner, C. Baker, A.D. Baker, and C.R. Brundle, “Molecular Photoelectron Spectroscopy”, Wiley, New York, 1970. 10. G. Herzberg, “Spectra of Diatomic Molecules”, van Nostrand, New York, 1950. 11. J.H. Van Vieck, “The Theory of Electric and Magnetic Susceptibilities, Oxford University Press, Oxford, 1932. 12. W.W. Wilson, R.C. Thompson, and F. Aubke, Inorg. Chem., j, 1489 (1980). 13. J.S. Haynes, K.W. Oliver, S.J. Rettig, R.C. Thompson, and J. Trotter, Can. J. Chem., 891 (1984). 14. A. Smaic, Inst. Josef Stefan, I.J.S. Report R. 612, 1 (1972); Chem. Abstr. 13032t (1973). 15. R.J. Gillespie, R. Kapoor, R. Faggiani, C.J.L. Lock, M.J. Murchie, and J. Passmore, Chem. Comm., 8 (1983). 16. J. Passmore, E.K. Richardson, and P. Taylor, J. Chem. Soc. Dalton Trans., 1006 (1976). 17. R.D.W. Kemmitt, M. Murray, V.M. McRay, R.D. Peacock, M.C.R. Symons, and T.A. O’Donnell, J. Chem. Soc. A, 862 (1968). 18. W.W. Wilson and F. Aubke, J. Fluorine Chem., j3, 431 (1979). 19. W.E. Falconer, F.J. DiSalvo, A.J. Edwards, J.E. Griffiths, W.A. Sunder, and M.J. Vasile, J. Inorg. Nuci. Chem. Suppl., 59 (1976). 138 20. A.J. Edwards, W.E. Falconer, J.E. Griffiths, W.A. Sunder, and M.J. Vasile, J. Chem. Soc. Dalton Trans., 1129 (1974). 21. N. Bartlett and S.P. Beaton, Chem. Comm., 167 (1966). 22. J.B. Baal, Jr., C. Pupp, and W.E. White, Inorg. Chem., , 828 (1969). 23. F.J. DiSalvo, W.E. Falconer, R.S. Hutton, A. Rodriguez, and J.V. Waszczack, J. Chem. Phys., 62, 2575 (1975). 24. I.V. Nikitin and V. Ya. Rosolovskii, Russ. Chem. Rev. (Engi. Trans.) 4, 889 (1971); Usp. Khim. 4, 1913 (1971) and references therein. 25. J. Shamir and J. Binenboym, Inorg. Chim. Acta, , 37 (1968); Inorg. Svnth., .14. 39 (1973). 26. J.N. Keith, I.J. Solomon, I. Sheft, and H.H. Hyman, Inorg. Chem., 2 230 (1968). 27. K.O. Christe, R.D. Wilson, and I.B. Goldberg, Inor. Chem., j, 1271 (1976). 28.a) V.1. Belova, Ya.K. Syrkin, D.V. Bantov, and V.F. Sukhoverkhov, Russ. J. Inorg. Chem. (Engi. Trans.) .j3, 765 (1968). b) V.1. Belova, V.Ya. Rosolovskii, and E.K. Nikitina, Russ. J. Inorg. Chem (Engi. Trans.)., 772 (1971). 29. D.R. Slim, Ph.D. Thesis, University of Birmingham, 1974. 30.a) P. Rigny and W.E. Falconer, 3. Chem. Phys., , 2581 (1975). b) C. Naulin and R. Bougon, 3. Chem. Phys., 4, 4155 (1976). 31. I.B. Goldberg, K.O. Christe, and R.D. Wilson, Inorg. Chem., 14, 152 (1975). 32. A. Grill, M. Schieber, and J. Shamir, Phys. Rev. Lett., 25, 747 (1970). 33. R.J. Myers, “Molecular Magnetism and Magnetic Resonance Spectroscopy”, Prentice- Hall Inc., Englewood Cliffs, New Jersey, 1973. 34. A. Bondi, 3. Phys. Chem., 68, 441 (1964). 35. A. Zumsteg, M. Ziegler, W. Kanzig, and M. Bosch, Phys. Cond. Matter, jJ, 267 (1974). 36. K. Lueken, M. Deussen, M. jansen, W. Hesse, and W. Schnick, Z. Anorg. Alig. Chem., 3, 179 (1987). 139 37. D.E. Cox, E.J. Samuelson, and K.H. Beckurts, Phys. Rev. 3102 (1973). 38. A.R. Young, II, T. Hirata, and S.I. Morrow, J. Am. Chem. Soc., 86, 20 (1964). 39. B.G. Mueller, J. Fluorine Chem., 12,409 (1981). 40. W. Schnick and M. Jansen, Angew. Chem. mt. Ed. (English’). 92. 54 (1985). 41. W. Schnick and M. Jansen, Z. Anorg. Aug. Chem., 32, 37 (1986). 42. W. Schnick and M. Jansen, Rev. Chem. Miner., 24, 446 (1987). 43. M. Jansen, and W. Hesse, Z. Anorg. Aug. Chem., Q, 47 (1988). 44. J.S. Miller, Ed. “Extended Linear Chain Compounds”, Vol. 3, Plenum Press, New York, 1983. 45. J.D. Johnson, J. Appl. Phys., , 1991 (1981). 46. M.E. Fisher, J. Math. Phys., 4, 124 (1963). 47. R.L. Carlin, “Magnetochemistry”, Springer-Verlag, Berlin, 1986. 48. C.H. Weng, Ph.D. Dissertation, Carnegie-Mellon University, Pittsburgh, PA, (1968). 49. W. Hiller, J. Strahle, A. Datz, M. Hanack, W.E. Hatfield, L.W. ter Haar, and P. Gutlich, J. Am. Chem. Soc., 106, 329 (1984). 50. W.V. Cicha, J.S. Haynes, K.W. Oliver, S.J. Rettig, R.C. Thompson, and J. Trotter, J. Chem., 63, 1055 (1985). 51. M. Suzuke, B. Tsujiyama, and S. Katsura, J. Math Phys., , 124 (1976). 52. 0. Edquist, E. Lindholm, L.E. Selin, and L. Asbrink, Phys. Scripta, 1, 25 (1970). 53. A.B. Comford, D.C. Frost, C.A. McDowell, J.L. Ragle, and l.A. Stenhouse, J. Chem. Phys., 4, 2651 (1971). 54. P. Venkateswarlu, Can. J. Phys., 47, 2525 (1969). 55. P. Venkateswarlu, Can. J. Phys., 4, 2525 (1969). 56. Landolt-Börnstein, Numerical data and functional relations in science and technology, Vol. 2, Magnetic properties of coordination and organometaflic transition metal com pounds, Springer-Verlag, Berlin, 1966. 140 CHAPTER 6 MAGNETIC EXCHANGE IN M(II) SULFONATES, M(ll) = Ni(I1), Pd(ll) AND Ag(ll) 6.1 Introduction The sulfonates described in this Chapter include the metal(II) fluorosulfates Ni(SO3F)2, Pd(SO3F)2,Ag(SO3F)2and “Pd(SO3F)”,more appropriately formulated as the mixed valency compound Pd(II)[Pd(IV)(SOF6],and the metal(ll) thfluoromethylsulfates Ni(SO3CF)2, Pd(SO3CF)2,and Ag(SO3CF)2.These compounds can also be considered as transition metal derivatives of the strong sulfonic acids I-{SO3Fand HSO3CF respectively. The anions SO3F and S03CF have the potential to coordinate to the metal ions as bidentate or tridentate bridging ligands (1-5), and consequently polymeric, layered materials with paramagnetic metal centers are formed. As pointed out in Chapter 3, only a limited number of paramagnetic binary fluoro compounds of divalent nickel, palladium and silver has been synthesized so far, and therefore it is of interest to study the magnetic properties of the above sulfonates which are rare binary fluoro derivatives of the respective divalent metals. Several transition metal fluorosulfate and trifluoromethylsulfate compounds have been studied previously for their magnetic properties in the higher temperature range, usually down to —80 K only (1-4,6-15). Interestingly, it appears that among the fluorosulfates, only Jr(SO3F)4 and its derivativeCs2[Ir(SO3F)](11) are magnetically concentrated, and in the trifluoromethyl sulfates magnetic exchange is detected only in Fe(SO3CF)(14) and Ag(SO3CF)2(4). The exchange coupling observed in all four of these derivatives is reported as antiferromagnetic. This is significant and not totally unexpected, since the majority of the magnetically concentrated transition metal fluoro compounds exhibit antiferromagnetism rather than ferro- or 141 ferrimagnetism (16). Ferromagnetic ordering is hence a rare phenomenon in ionic solids, and is largely confined to small groups of transition (d-block) and lanthanide metal -oxides, -chalcogenides and -halides with distinct structural features that permit one, two or three dimensional exchange via monoatomic anions (17, 18). However, in contrast the four metal(II) fluorosulfate compounds studied here contain the polyatomic SO3F anion and show significant ferromagnetic exchange at low temperatures. Previous magnetic measurements down to —80 K on the binary fluorosulfates Pd(SO3F)2(2), Ag(SO3F)(3) and “Pd(SO3F)”(2), showed that the compounds were relatively magnetically unconcentrated in that temperature range, and their susceptibilities followed the Curie-Weiss law closely with small positive Weiss constants. The three metal(lI) trifluoromethylsulfate compounds Ni(SO3CF)2,Pd(SO3CF)2and Ag(SO3CF)2investigated here for their magnetic behavior have all been synthesized previously (4, 5, 19) although no variable temperature susceptibility studies have been reported for the Ni(SO3CF)2and Pd(S03CF)2compounds until now. A previous high temperature study on the Ag(SO3CF)2compound indicated that the Ag(ll) ions in the sample were coupled antiferromagnetically (4). Therefore, the two analogous nickel and palladium derivatives were investigated in this work to determine whether these compounds are also magnetically concentrated with antiparallel spins. In addition, magnetic measurements were run on the Ag(SO3CF)2compound at lower temperatures to verify the previously reported higher tempera ture susceptibility data. It is significant to note here that the metal(ll) fluorosulfates and the irifluoromethyl sulfate analogs described in this Chapter are assumed to have the common CdC12-type structure, which is also seen in the metal(II) hexafluoroantimonates discussed in Chapter 3. Layered struc tures of the type seen in Sn(SO3F)2 (20a) and Ca(SO3H)2(20b) have been previously 142 proposed for a number of divalent metal sulfonates (6, 10, 13, 15), on the basis of magnetic properties and vibrational spectra. It was proposed that these compounds adopt a polymeric two-dimensional structure, where each metal site is surrounded by an octahedral arrangement (tetragonally distorted in the silver derivatives) of oxygen ligands, as illustrated for Pd(SO3F)2 in Figure 6.1. Figure 6.1: Proposed Structure of Pd(SO3F)2 0 0 Metal S 0 F 143 6.2 Experimental The synthetic procedures for the preparation of metal(ll) fluorosulfates Ni(SO3F)2(13), Pd(SO3F) (2), Ag(SO3F)2 (3), “Pd(SO3F)t’(2), and the metal(ll) trifluoromethylsulfates Ni(SO3CF)2(19), Pd(SO3CF2(5) and Ag(SO3CF)2(4) have been described in detail previ ously. These methods were followed in the present study as well. Attempts were made to synthesize “Pd(SO3CF)”,the corresponding trifluoromethylsulfate derivative of “Pd(SO3F)t’, but in all instances the solvolysis of “Pd(SO3F)”in excess HSO3CF (5) led to the binary Pd(SO3CF)2compound only. All the reactions were monitored by weight, and the purity of the products was determined by elemental analysis and IR spectroscopy. Magnetic data were corrected for Temperature Independent Paramagnetism (TIP) using the following reported lODq values [cm1]: Ni(SO3F)2,7340 (1); Pd(SO3F)2 11770 (1); Ag(SO3F)216600 (3) and Ni(SO3CF)2 7350 (19). 6.3 Results and Discussion The magnetic data obtained on the M(II) sulfonates with M=Ni, Pd and Ag, clearly indicate two distinct types of magnetic exchange present in the two groups of compounds. All the M(II) fluorosulfates studied here exhibit strong ferromagnetic exchange at lower tempera tures, whereas the M(II) trifluoromethylsulfates show antiferromagnetism with varying degrees of magnetic concentration over a wider temperature range. Therefore, the discussion of the magnetic results can be conveniently divided into two parts, ferromagnetism and antiferromag netism of the respective sulfonate derivatives. Ferromagnetism will be discussed first and in greater detail, since this phenomenon is rare and unusual in transition metal fluoro compounds. 144 6.3.1 Ferromagnetism of M(ll) fluorosulfates Ni(SO3F)2, Pd(SO3F)2, Pd(II)[Pd(IV)(SOF]and Ag(SO3F)2 Previous magnetic susceptibility measurements on the three palladium and silver fluorosulfates investigated here indicate that the compounds are relatively magnetically dilute down to —80 K (2,3). No detailed magnetic study exists in the case of Ni(SO3F)2. Therefore, measurements on Pd(SO3F)2,Ag(SO3F)2and “Pd(SO3F)”are extended down to —4 K (—2 K for Pd(SO3F)2)and Ni(SO3F)2is studied in the temperature range of —291 to 2 K. The pertinent magnetic data obtained with the vibrating sample magnetometer in the lower temperature range are presented in Tables 6.1, 6.2, 6.3 6.4, and the Gouy measurements of Ni(SO3F)2from —291 to 79 K are given in Table 6.5. A summary of the relevant magnetic parameters on all four fluoro sulfates is listed in Table 6.6 (see Appendices B-i to B-3 for additional magnetic data). For Pd(SO3F)2,“Pd(SO3F)”and Ag(SO3F)2,previous higher temperature measure ments were presented as plots of 1/XM vs. T, which were linear, and had positive Weiss constants (Table 6.6), indicative of Curie-Weiss behavior. The magnetic moments calculated for the three compounds (1-3) were in good agreement with expected values for approximately octahedrally coordinated Pd(II)(d8)and Ag(II)(d9)species. A similar situation is observed in the case of Ni(SO3F)2,where the inverse susceptibility vs. temperature graph in the higher temperature region also gives a straight line plot (Figure 6.2). However, the Weiss constant obtained here 0.41 K is relatively small and hence Ni(SO3F)2appears to be magnetically dilute down to —80 K. The 1eff value of about 3.27 B (Table 6.5) in this temperature range is not unexpected for Ni(SO3F)2,with the Ni(II) ions located in octahedral ligand sites (13,21; see also Chapter 3). 145 Table 6.1: Low Temperature Magnetic Data of Ni(SO3F)2 Temperature (K) XMCOff x 106 Peff (P.B)a (cm3mol1) 81.72 15520 3.16 77.72 16370 3.16 74.02 17250 3.17 69.72 18370 3.18 65.19 19740 3.18 59.86 21590 3.19 54.40 23880 3.20 47.60 27390 3.21 40.40 32800 3.24 31.25 43800 3.30 26.23 54310 3.37 20.65 71640 3.43 15.98 100200 3.57 11.17 170500 3.90 7.60 329600 4.47 6.30 502700 5.03 5.08 670900 5.22 4.24 766000 5.10 4.12 775700 5.05 3.37 839100 4.75 2.60 880500 4.28 2.30 900000 4.07 a = 2.828 [(xMcoff - TIP)T]112; TIP = 8N2/10Dq = 285 x 10 cm3 mo11 Magnetic field = 9225 G. 146 Table 6.2: Low Temperature Magnetic Data of Pd(SO3F)2 ., COlT X 106 Peff (IB)aTemperature (K) (cm3mol1) 82.06 19010 3.52 79.89 19560 3.52 78.33 20060 3.53 76.60 20610 3.54 74.47 21250 3.54 72.55 21950 3.55 70.28 22640 3.55 67.95 23520 3.56 65.77 24640 3.59 63.40 25740 3.60 60.73 27060 3.61 58.00 28620 3.63 55.00 30600 3.66 51.40 33020 3.67 47.90 36130 3.71 44.50 39870 3.76 40.60 45170 3.82 31.23 66750 4.08 21.35 127300 4.66 16.50 211800 5.29 11.25 574900 7.19 8.11 1015000 8.11 5.99 1162000 7.46 4.32 1215000 6.48 3.88 1236000 6.19 3.30 1248000 5.74 2.70 1258000 5.21 1.74 1271000 4.20 a = 2.828 .[(xMCOIT - TIP)T11/2;TIP = 8Nj32/lODq = 177 x 10 cm3 mo11 Magnetic field = 7501 G. 147 Table 6.3: Low Temperature Magnetic Data of Pd(H)[Pd(W)(SO3F6] Temperature (K) XMCO X 106 eff (IIB)a (cm3moP1) 82.28 21220 3.74 78.73 22130 3.73 74.64 23460 3.74 70.34 24860 3.74 65.88 26750 3.75 60.60 29130 3.76 55.00 32420 3.78 48.00 37180 3.78 44.20 40610 3.79 40.60 44880 3.82 31.65 60000 3.90 27.23 73720 4.01 21.75 98300 4.14 16.34 155000 4.50 11.17 355300 5.63 7.94 581500 6.08 5.98 663400 5.63 5.08 681600 5.26 a Peff = 2.828 [XMCOIT X TI1/2 Magnetic field = 7501 G. 148 Table 6.4: Low Temperature Magnetic Data of Ag(SO3F)2 Temperature (K) XM X 106 eff (JB)a (cm3mol1) 82.06 7950 2.27 78.28 8450 2.29 74.53 9070 2.32 70.34 9840 2.35 65.82 10840 2.38 60.50 12300 2.43 55.00 14270 2.50 47.95 17560 2.59 44.35 20080 2.66 40.55 23560 2.76 30.85 40400 3.15 26.58 58350 3.52 21.23 110400 4.33 16.34 277000 6.02 10.53 605800 7.14 7.40 688000 6.38 5.08 730300 5.45 4.54 735300 5.17 a Peff = 2.828 [(XMCOnI. - TIP)T1112; TIP = 4N132/lODq =63 x 10 cm3 mo11 Magnetic field = 7501 G. 149 Table 6.5: Magnetic Data of Ni(SO3F)2for the Temperature Range 291 to 79 K Temperature (K) XMCOff X 106 eff (B)a (cm3mo11) 291.3 4840 3.26 271.0 5240 3.28 253.1 5580 3.27 235.7 5970 3.27 219.0 6420 3.28 201.7 6940 3.28 177.6 7860 3.28 152.0 9120 3.28 127.5 10920 3.29 103.0 13250 3.27 87.0 15690 3.27 79.0 17100 3.26 a 1’eff = 2.828 [(XM’°’ - TIP)T]112; TIP = 8Nj32/lODq = 285 x 10 cm3 mo11 Magnetic field = 8000 G. 150 Table 6.6: Magnetic Parameters of Ni(SO3F)2,Pd(SO3F)2,Pd(ll)[Pd(IV)(SO3F6] and Ag(SO3F)2 Compound Temperature Weiss Const. ie j.Lff (max)b Tm (ff max)1’ Reference Range (K) 0 (K) (B) (RB) (K) Ni(SO3F)2 291—79 0.41±2 3.27 This work 81.7—2.3 5.27 4.78 Pd(SO3F)2 299—103 13±4 3.34 (1) 82.1—1.7 8.11 8.11 This work “Pd(SO3F)” 334—107 10±2 3.45 (1) 82.3—5.1 6.08 7.94 This work Ag(SO3F)2 301—80 20±2 1.92 (3) 82.1—4.5 7.14 10.53 This work a Values for 290 K. b Values for magnetic field = 7501 G. 151 Figure 6.2: -J 0 >- F— -J F a LU (-) U) (1) (-) F LU z 0 Inverse Susceptibility vs. Temperature of Ni(SO3F)2 r€) 250- () from Curie—Weiss Low (29V79K) K “ 200- 150- 100- 50- 0 Cm = 1 .34 cm3 moEt K e =0.41 . 0 50 I I 100 150 200 250 300 TEMPERATURE,K 152 The low temperature plots of 1/XM vs. T for Pd(SO3F)2,“Pd(SO3F)1’and Ag(SO3F)2are shown in Figure 6.3. Extrapolation of the higher temperature linear portion ((Pd(SO3F)2and “Pd(SO3F)”> 11 K, Ag(SO3F)2>14 K) of these plots produces intercepts on the temperature axis in excellent agreement with the values reported earlier (Table 6.6). In the lower temperature region all four compounds show a temperature independent 1/XM (and therefore XM) behavior as the temperature is lowered. This change in the temperature dependence of the magnetic susceptibility at low temperature is illustrated for Pd(SO3F)2and ‘Pd(SO3F)1’in Figures 6.4 and 6.5, where the field dependence of the susceptibility in this region is also shown for the two compounds. The approximate maximum susceptibilities expected for these compounds can be calculated with the following (spin only) expression given by Carlin (22): Msat = NgpS [6.1] where Msat = saturation magnetic moment (M/H = Xt) N = Avogadro’s number g Lande splitting factor = Bohr magneton S = Spin system, S=1 (Pd(II)) and S=1/2 (Ag(II)) The maximum or saturation magnetization situation corresponds to the alignment of all the magnetic spins parallel to the external field (H), where the magnetization becomes independ ent of field and temperature. The saturation susceptibility values calculated using equation [6.11 are compared with the observed maximum susceptibilities for Ni(SO3F)2, Pd(SO3F)2, “Pd(SO3F)”and Ag(SO3F)2in Table 6.7. 153 Figure 6.3: Inverse Susceptibility vs. Temperature of Pd(SO3F)2, Pd(ll)[Pd(IV)(SOF]and Ag(SO3F)2 140- I 120- o Pd(SO3F)2 0, • Pd(II)[Pd(V)(SO6] 100- — . Ag(SO3F)2 >- 80- 60- I (/) I D • LI) c . (_) 40- • 0 LJ z . 2O- t 0— 0 20 40 60 80 100 TEMPERATURE,K 154 Figure 6.4: 140- Magnetic Susceptibility vs. Temperature for Pd(SO3F)2at 7501 and 9625 G 120- S S100 80 60 c%J 0 x 0 C-) >- -J cL C.) D LI) C-) z C) 0 AT 7501G • AT 9625G 40- 20 0 0 • _ 0 oc• ooo 20 40 60 80 100 I I I I TEMPERATURE,K 155 C >< 0 0 >- I- -J cL i-J C) V.) V.) 0 I LJ zC, Figure 6.5: Magnetic Susceptibility vs. Temperature for Pd(II)[Pd(IV)(SO3F6]at 7501 and %25 G 799998- 399999- 200000- 0 o AT 7501G • AT 9625G 0 I I I 20 40 60 • TEMPERATURE,K 80 100 156 Table 6.7: Experimental and Calculated Saturation Magnetic Susceptibilities of Ni(SO3F)2,Pd(SO3F)2,Pd(II)[Pd(W)(SO3F6]and Ag(SO3F)2 % saturation = Compound XM sat. calculateda XM’ max. observeda Temp. (XM’ max) XM max. x 100 (cm3mo!4) (cm3 mo!) (K) XM sat. Ni(SO3F)2 1.488 1.041 2.69 70 Pd(SO3F)2 1.488 1.271 1.74 85 “Pd(SO3F)” 1.488 0.682 5.08 46 Ag(SO3F)2 0.744 0.735 4.54 99 a Susceptibility data for magnetic field = 7501 0. 157 It appears from the data in Table 6.7 that the magnetic behavior observed at lower temperature in Ni(SO3F)2,Pd(SO3F)2and Ag(SO3F)2may be largely a result of saturation magnetization. As a consequence, the magnetic moments of these compounds are temperature dependent and pacs through maxima at -5, —8, and —10.5 K respectively (Table 6.6). This is illustrated in the magnetic moment vs. temperature plot for Pd(SO3F)2and Ag(SO3F)2in Figure 6.6. Spin saturation in the mixed valent “Pd(SO3F)’ may be achieved at higher magnetic fields, since closer approach to saturation magnetization would be expected at stronger applied fields. Even at a field of 7501 G, the magnetic moments of this compound shows temperature dependent behavior and has a maximum at —8 K. In the Ni(SO3F)2,Pd(SO3F)2and “Pd(SO3F)”compounds, the paramagnetic Ni(ll) and Pd(II) ions have been shown to be in octahedral environments with 3A2g ground states (1,2,21). Although there is no orbital contribution to the susceptibility associated with this state, to first order, zero-field splitting via second order spin orbit coupling could lift the triplet spin degeneracy which can significantly affect the magnetic properties of such systems, particularly at low temperatures. The average susceptibility <X> of powder samples (X>=(X11+2X±)/3) in the presence of zero-field splitting for S=1 spin systems is obtained from the expression (22): = 2Ngt (2/x —2 exp(-x))/(x + exp(-x)) [6.2] 3kT 1 + 2 exp(-x) where x = DIkT, D = zero-field splitting parameter, k = Boltzmann’s constant, and N, g and B as defined for equation [6.1]. 158 Figure 6.6: Magnetic Moment vs. Temperature of Pd(SO3F)2and Ag(SO3F)2 I. z U a C) U z I I 40 60 TEMPERATURE,K o Pd(SO3F)2 • Ag(S03F)2 9- 8- 7- 6- 5- 4. 2 C 0 20 1 80 100 159 In general, the value of D is only a few wavenumbers, and with kT —205 cm1 at room temperature, D/kT <c 1. However, at lower temperatures D/kT becomes significant, thereby affecting the measured susceptibilities of these samples. The sign of D could be either positive or negative, and in the case of nickel, both signs have been observed (22). Attempts were made to fit the magnetic susceptibility data to equation [6.2]. However, only very poor fits were obtained, which indicates clearly that this effect cannot explain the magnetic behavior exhibited by the nickel and palladium complexes. Moreover, the magnetic behavior of Ag(SO3F)2cannot be rationalized by this effect, since in Ag(II) ions with 2Big ground states, the possibility of zero-field splitting does not exist. The magnetic properties observed in these four fluorosulfate compounds bear a strong resemblance to those reported for the binary chlorides FeCI2, CoCl2 and NiC12 by Starr et al. (23). Neutron-diffraction studies on the iron and cobalt compounds have revealed ferromagnetic coupling between the metal centers within each layer (intralayer exchange), and weak antiferro magnetic coupling between layers (interlayer exchange) (24). This magnetic behavior, termed metamagnetism, has been extensively reviewed for other compounds as well (25). Interestingly, the binary chlorides mentioned above have the typical CdCl2-type structure, with each layer of the metal atoms separated by two layers of chlorine atoms from the next metal atom layer. Each metal atom layer forms a two-dimensional hexagonal network in which every metal atom has six near neighbours (26). Furthermore, in a number of layer type fluorides containing Jahn-Teller ions, the intralayer ferromagnetic couplings are found to be much stronger than the interlayer anti ferromagnetic couplings, and as a result ferromagnetism is observed in these compounds (17). Similar interactions may be present in “the Jahn-Teller compound” Ag(SO3F)2studied in this work as well. 160 Similar crystal and magnetic structures are possible for the M(II) fluorosulfates examined in this work. Indeed, layer structures based on the CdC12 prototype have also been proposed for the compounds discussed here, as mentioned in the introductory section of this chapter (see also Figure 6.1). The apparent lack of solubility of these metal fluorosulfates in a suitable solvent such as fluorosulfuric acid, HSO3F, has so far precluded single crystal X-ray diffraction studies as well as magnetic measurements on oriented crystals leading to studies on paramagnetic anisotropy. It is significant to note that octahedral Pd(II) ions with a 3A2g ground state is found only in Pd(SbF6)2(Chapter 3), PdF2 (27), Pd(503F)2Pd(SO3CF)2,and in some of their cationic and anionic derivatives, just as AgF2 (28), Ag(SO3F)2and Ag(503CF)2have remained the only simple binary compounds of Ag(II) with a d9 configuration. However, the fluorides AgF2 (28), PdF2 (27), and NiF2 (30) are essentially antiferromagnetic compounds, and compare more appropriately with the trifluoromethylsulfate derivatives discussed in the next section. In PdF2, NiF2,and AgF2,weak ferromagnetism is observed due to a canting of the spins (27,28,30). The ternary Pd(II)[Pd(IV)F61is reported to have a significant ferromagnetic component at low temperatures, although the complex is weakly ferromagnetic over a wide temperature range (17, 18, 29). In contrast, the structurally similar Pd(II)[Pd(IV)(SO3F6]is found here as a strongly ferromagnetically coupled compound. The fluorosulfate ligand seems to bond strongly to the Pd(IV) center and weakly to Pd(ll), in an “anisobidentate” bonding mode (1,2). A large number of M(II) M’(IV)F6 type fluorides with both M and M’ transition metals such as Pd, Pt and Ni, have been studied for their ferromagnetic contribution (17,18) as already noted above in the case of the Pd2F6compound. Ferromagnetism in these compounds may be associated with cationic ordering, which is observed in Pd2F6 from neutron diffraction measurements (29). Magnetic ordering at low temperatures is explained in these ternary species by a mechanism where the spins of the eg2 electrons of the divalent metals (t2g6eg2) are ferromagnetically 161 coupled via a superexchange interaction involving the 2p fluorine orbitals and the empty eg orbitals of the tetravalent cations (17,18). Therefore, it has been shown that when cationic ordering in M(II) and empty eg orbitals on the transition metal M([V) are present, ferromagnetism can occur in these bimetallic compounds. This may be applicable to the Pd(II)[Pd(IV)(SO3F6]complex as well, although the much larger SO3F- anion may significantly affect the extent of ferromagnetic interaction. Interestingly, when the tetravalent cation is replaced by a non-transition metal ion such as Sn(IV) or Ge(IV), magnetic ordering is not observed even at 4.2 K (17). This observation appears to be valid for the ternary bimetallic fluorosulfates as well. A number of previously synthesized fluorosulfate derivatives with three different divalent metals where the tetravalent cation is Sn(IV) were studied for their low temperature magnetic properties in this work. The compounds chosen were Ni(ll)[Sn(IV)(SO3F6](31), Cu(II)[Sn(IV)(SOF6](31) and Ag(lJ)[Sn(IV)(SO3F61(3). The magnetic measurements obtained for these three samples are given in Appendices B-4, B-5, and B-6. It is clear from these data that the compounds are rela tively magnetically dilute to ‘—4 K. Furthermore, previous high temperature magnetic measurements on Pd(II)[Sn(IV)(SO3F6J(2) and Ag(II)[Sn(IV)(SO3F6j(3) indicated Curie-Weiss behavior for the compounds down to liquid nitrogen temperature. However, stronger magnetic exchange in teractions may be present at lower temperatures in Ag(II)[Pt(1V)(SO3F6j(3) and Pd(II)[Pt(IV)(SO3F & (2), where the tetravalent cation is a transition metal ion, i.e. Pt(IV), al though no magnetic exchange is observed in the two samples down to —80 K (3,2). Even though the above described superexchange mechanism may be valid for the Pd(II)[Pd(IV)(SO3F6]complex, for the binary Ni(SO3F)2,Pd(SO3F)2,and Ag(SO3F)2com pounds detailed magneto-structural relationships cannot be made, especially in the absence of 162 structural evidence from single crystal X-ray studies. This is also true in the case of the corresponding thfluoromethylsulfate derivatives discussed below. Crystal growth was, however, attempted for Pd(SO3F)2,“Pd(SO3F)’and Ag(SO3F)2, utilizing the method employed to obtain single crystals of Au(SO3F),the structure of which was reported recently by our group (32). Unfortunately, the highly polymeric compounds formed only microcrystalline materials, unsuitable species for single crystal measurements. In the following section, antiferromagnetic behavior in the divalent nickel, palladium and silver thfluoromethylsulfates will be discussed in some detail. 6.3.2 Antiferromagnetism of M(II) trifluoromethylsulfates Ni(SO3CF)2,Pd(SO3CF)2 and Ag(SO3CF)2 It was mentioned in the introduction that in previous magnetic susceptibility studies to —80 K, the only four sulfonates which had detectable magnetic exchange all showed antiferro magnetism, which is observed more frequently than ferromagnetism in magnetically concentrated transition metal compounds (16,33). In this study, the magnetic measurements on Ag(SO3CF)2were extended down to —4 K to complete the earlier Gouy work, and also to detect unusual magnetic properties, if any, at lower temperatures. The results of the low temperature study are given in Table 6.8, together with the previous Gouy data obtained for the compound. Similarly, the Pd(SO3CF)2and Ni(SO3CF)2complexes are investigated here for their possible antiferromagnetic behavior, and pertinent low temperature data for the two compounds are shown in Tables 6.9 and 6.10 respec tively. The nickel compound is also measured in the higher temperature range of —292 to 80 K by the Gouy method, and the results of this work are presented in Table 6.11. 163 Table 6.8: Magnetic Data of Ag(SO3CF)2for the Temperature Range 304 to 4 K ‘Temperature (K) XMCOff x 106 (ii (cm3mo11) 304 1100 1.64 278 1150 1.60 254 1200 1.56 226 1270 1.52 204 1320 1.47 179 1380 1.41 154 1420 1.32 128 1430 1.21 108 1390 1.10 82.06 650 0.66 78.28 630 0.63 70.34 620 0.59 65.82 610 0.57 60.38 570 0.53 47.80 550 0.46 40.40 540 0.42 31.85 540 0.37 26.35 520 0.33 21.40 520 0.30 16.55 510 0.26 11.20 510 0.21 8.04 510 0.18 5.34 510 0.15 4.39 510 0.13 a First nine data points from Ref. 4. b Not corrected for TIP; I1eff 2.828 [xMCOt X 11h12 164 Table 6.9: Low Temperature Magnetic Data of Pd(SO3CF)2 COlT x 106 eff (B)aTemperature (K) (cm3mo[1) 123.2 6950 2.62 118.0 7220 2.61 112.8 7460 2.59 108.2 7730 2.59 103.0 8030 2.57 97.97 8330 2.55 93.46 8690 2.55 88.31 9020 2.52 84.81 9350 2.52 81.95 9670 2.52 78.33 9970 2.50 74.47 10180 2.46 70.23 10540 2.43 68.46 10630 2.41 65.59 10990 2.40 60.40 11530 2.36 54.45 12190 2.30 51.40 12550 2.27 47.92 13030 2.24 44.20 13510 V 2.19 40.32 14140 2.14 31.60 15700 1.99 20.88 17970 1.73 16.00 19020 1.56 10.75 20220 1.32 7.52 21060 1.13 5.78 21330 0.99 5.62 21630 0.99 4.32 21720 0.87 3.70 21330 0.79 2.99 21270 0.71 2.40 21270 0.64 2.10 21240 0.60 a Not corrected for TIP; 11eff 2.828 [(XMCOff X T]1”2 165 Table 6.10: Low Temperature Magnetic Data of Ni(SO3CF)2 Temperature (K) XMCOff x 106 Ieff (B)a (cm3mol1) 81.78 16620 3.27 77.72 17360 3.26 74.19 18200 3.26 69.72 19240 3.25 65.19 20570 3.25 60.15 22160 3.24 54.00 24290 3.22 47.60 27270 3.21 40.10 31970 3.19 30.80 41390 3.18 25.93 48220 3.15 20.70 58650 3.11 15.90 74230 3.07 10.42 105700 2.96 7.16 140000 2.83 5.92 159800 2.75 4.85 181500 2.65 4.78 182100 2.64 3.96 200300 2.52 3.29 213900 2.37 3.11 220000 2.34 2.79 230400 2.27 2.60 234600 2.21 2.50 236200 2.17 a eff = 2.828 [(xMcoff - TIP)T11/2;TIP = 8N32/10Dq = 284 x 10 cm3 mo11 166 Table 6.11: Magnetic Data of Ni(SO3CF)2for the Temperature Range 292 to 80 K Temperature (K) XMCOff x 106 neff (B)a (cm3mo11) 291.8 5270 3.41 286.2 5390 3.42 269.3 5700 3.42 251.7 6050 3.41 234.5 6450 3.40 218.0 6890 3.39 200.3 7380 3.37 175.7 8310 3.36 150.0 9540 3.33 125.5 11320 3.33 103.0 13510 3.30 86.5 15740 3.27 79.6 16890 3.25 a 1eff = 2.828 [(xMcoff - TIP)T]112; TIP = 8N32/10Dq = 284 x 10 cm3 mo11 167 The previous magnetic report on Ag(SO3CF)2where antiferromagnetic behavior was seen with Xm at —138 K (4) is further confirmed by the present low temperature study. The magnetic moments decrease continuously with decreasing temperature, and the suscep tibility also folio ‘s a similar trend, falling rapidly in value below the Nel temperature (Table 6.8). The very small magnetic moments observed down to —4 K indicate that the compound has no ferromagnetic component present in the low temperature region. This behavior of Ag(SO3CF)2is in contrast to the magnetic behavior of AgF2, in which ferromagnetism is detected below 163 K (28). Therefore, it seems that structural differences in AgF2 and Ag(SO3CF)2play significant roles in determining the extent of antiferromagnetic coupling, although the exact exchange mechanism remains still unclear in the Ag(SO3CF)2compound. It has been shown that in octahedrally coordinated transition metal d9 fluorides, the Jahn-Teller effect leads to two crystallographic distortions, termed ferro and antiferrodistortive ordering, with the former favor ing antiferromagnetism and the latter ferromagnetism (34). In AgF2, the ferromagnetic units are coupled antiparallel, which give a bulk 3D antiferromagnetism to the compound, although spin canting produces a small ferromagnetic component below 163K (28). However, in layer type fluoride structures with Jahn-Teller ions, the intralayer ferromagnetic couplings are much stronger than the interlayer antiferromagnetic couplings, as postulated in the case of Ag(SO3F)2 discussed earlier, leading to ferromagnetism (17). But in contrast, the magnetic behavior observed in Ag(SO3CF)2may arise from strong interlayer antiparallel coupling of the spins with weak intralayer exchange, resulting in a bulk 3D antiferromagnetism for the compound. The magnetic data obtained for Pd(SO3CF)2also clearly indicate antiferromagnetic behavior, where the magnetic moments calculated in the temperature range —123 to 2K show typical temperature dependent low values (Table 6.9). The moments decrease with decreasing temperature, and a sharp decline becomes apparent below —40K, as seen in the magnetic 168 moment vs. temperature plot for the compound illustrated in Figure 6.7. From a previously unpublished study in our group, the room temperature magnetic moment for Pd(SO3CF)2was found as 2.90 B (35). This value seems to be reasonable when compared with the l.Leff of 2.62 B obtained at 123 K in this study (Table 6.9), and it also confirms the continuous decreasing trend of the magnetic moments with decreasing temperatures. The higher temperature eff value is not unexpected for an octahedrally coordinated Pd(II), where the moments are lowered by antiferromagnetic exchange. The magnetic susceptibility vs. temperature plot of Pd(SO3CF)2,shown in Figure 6.8, has a Xm at approximately 4K. This is conclusive evidence, as in Ag(SO3CF)2,that the spins in the palladium compound are coupled antiferromagnetically. The appearance of Xm at a very low temperature, in contrast to the Xm observed at 138 K in Ag(SO3CF)2(4), indicates that in the palladium compound the antiferromagnetic interaction is much weaker than in the corresponding silver species. It is not uncommon however, to observe stronger spin couplings in ad9-Jahn-Teller system than in ad8-octahedral system. Significant differences are also noted between the magnetic behavior of Pd(SO3CF)2 and its fluoride derivative PdF2. The difluoride, although antiferromagnetic with a Nel temperature of --217K, shows a ferromagnetic magnetization component (“weak ferro magnetism”) below this temperature (27). Consequently, the magnetic moment of 1.84 B found at room temperature has a much larger value at lower temperatures. This magnetic behavior of PdF2, as in the case of AgF2 (28), is accounted for by Moriya’s theory of single-ion magnetocrystalline anisotropy (30), where the preferred direction of magnetization is different for the nonequivalent magnetic ions, leading to a canting of the spins. This is in contrast to the magnetic properties of Pd(SO3CF)2,in which a continuous decreasing trend of the magnetic moments with decreasing temperatures is observed (Figure 169 I— z 0 Q w z 0 Figure 6.7: Magnetic Moment vs. Temperature of Pd(SO3CF)2 3 2.5 2 1.5 0.5 0 20 40 60 TEMPERATURE,K 80 120 140 170 (0 x 0 :2 N, 2 () I- -J a:i I— 0 1J C)(n C) I— U z C) :2 TEMPERATURE,K Figure 6.8: Magnetic Susceptibility vs. Temperature of Pd(SO3CF)2 25000 20000 15000 10000 5000 0 20 40 60 80 100 120 140 171 6.7), with the susceptibility also falling off below the Nel temperature of —4 K (Figure 6.8). It appears from these observations that in Pd(SO3CF)2the dominant magnetic interaction is antiferromagnetic and unlike PdF2, weak ferromagnetism is not seen at lower temperatures in the thfluoromethylsulfate compound. The magnetic results of Ni(SO3CF)2,illustrated in Tables 6.10 and 6.11, show tempera ture dependent magnetic moments that decrease with the gradual lowering of the temperature, indicative of antiferromagnetic interaction among the Ni(ll) centers. Excellent agreement is noted between the Gouy and the vibrating sample magnetometer data for the overlap region. The Gouy data (Table 6.11) in the temperature range 292 to 80 K are plotted as vs. T, and the following Weiss and Curie constants are obtained respectively: 8 = -12.56 K and Cm = 1.52 cm3 mo11 K. The negative Weiss constant is characteristic of antiferromagnetism, and this behavior in Ni(SO3CF)2is further confirmed by the low temperature magnetic data given in Table 6.10. The of 3.41 B calculated at room temperature falls within the expected range for an octahedrally coordinated Ni(II) species. However, curiously this value is slightly larger than the moment of 3.26 B found at the same temperature for the corresponding Ni(SO3F)2(Table 6.5), although both compounds appear to have similar electronic environments for the respective Ni(fl) ions with near identical lODq values (1,21,19). Furthermore, the antiferromagnetic Ni(SO3CF)2is expected to have a lower moment at room temperature than its ferromagnetic fluorosulfate derivative, analogous to the palladium and silver derivatives discussed previously. Interestingly, when the low temperature magnetic moments are plotted against tempera ture, the effects of magnetic coupling interactions for Ni(SO3CF)2and Ni(SO3F)2become observable nearly at the same temperature, but are of opposite nature, as illustrated in Figure 6.9. 172 Figure 6.9: I z 0 () bJ z Magnetic Moment vs. Temperature of Ni(SO3CF)2and Ni(SO3F)2 6- 5- 4- • Ni(SO3F)2 o Ni(SO3CF3)2 3- 2 20 40 60 TEMPERATURE,K 80 100 173 In contrast to Pd(SO3CF)2and Ag(SO3CF)2,the nickel derivative does not have a Xmax in its susceptibility data. Therefore, the decrease of the moment values with decreasing temperature is less pronounced in the case of Ni(SO3CF)2(Table 6.10). These observations indicate that the antiferromagnetic interaction in the nickel species is relatively weak in comparison to that in the palladium and silver derivatives. The magnetic behavior observed in Ni(SO3CF)2differs also from the reported antiferromagnetism of the corresponding binary fluoride NiF2, where a weak ferromagnetic moment is detected below the Ne1 temperature of 73.2 K, which is attributed, as in AgF2 and PdF2 discussed previously, to spin canting (30). The theoretical basis of this phenomenon has been developed extensively by Moriya in his study of the magnetic behavior of NiF2 (30). In concluding this discussion on the magnetic exchange interactions of the M(II) sulfonates, a few comments will be made here regarding the magnetic behavior of the corresponding copper derivatives. A previous magnetic study of Cu(SO3F)2and Cu(SO3F)2 (down to 100 and 127 K respectively) indicated that the salts are essentially magnetically dilute with moments normally observed for hexacoordinated copper(II) (15). For this work, the magnetic measurements on the Cu(SO3F)2is extended down to —4 K, and a summary of the data is given in Appendix B-7. The compound was synthesized according to the method described by Alleyne et al. (13). The magnetic moments calculated are independent of temperature and remain close to the expected value of —2.0 B down to —4 K. Good agreement is also noted in the overlap region between the previous high temperature neff values and the low temperature moments of this study. The reason for this magnetically dilute behavior of Cu(SO3F)2is not clear, and it is rather surprising to note that while Ag(SO3F)2shows strong magnetic exchange below —10 K, the corresponding Group 11 copper derivative is magnetically dilute down to very low temperatures. Interestingly, the copper(II) difluoride is antiferromagnetic with a Nel temperature of 69 K, and as seen previously in AgF2,PdF2 and NiP2, spin canting produces a 174 weak ferromagnetic moment below the Nel temperature in CuF2 as well (37). In summary, it may be envisioned that in the layer type metal(ll) sulfonates discussed in this Chapter, magnetic exchange may occur preferentially via one of two possible spin inter actions: ferromagnetism in the fluorosulfate compounds could arise from strong intralayer parallel spin coupling, while in the corresponding trifluoromethylsulfate derivatives predominant interlayer spin coupling could lead to an antiparallel arrangement of the spins in the lattice. Alternatively, the O-S-O bridging angle in the two types of sulfonates may favor ferromagnetism and antiferromagnetism for the fluorosulfates and trifluoromethylsulfates respectively. Although the exchange pathway cannot be stated clearly in these compounds due to a lack of X-ray crystal data, it is interesting to note that in the divalent sulfates FeSO4,NiSO4 and CuSO4 which also have oxygen bridging extended 3D lattice structures, antiferromagnetic ordering is postulated to occur through the O-S-O bridges (38,39). Furthermore, neutron diffrac tion data obtained on FeSO4 and NiSO4 indicate magnetically ordered sheet-type structures, and the structurally similar CrVSO4 appears to have ferromagnetically ordered sheets which stack antiferromagnetically (38). In other examples involving sulfate derivatives, the compounds are found as linear chains, and their magnetic properties have been analyzed utilizing either the Ising or Heisenberg exchange coupling models (40). In the case of the sulfonates discussed here, however, analyzing the magnetic data is made difficult by several factors. The choice of either the Ising or Heisenberg 2-D model is usually not appropriate for the metal(II) sulfonates. These models do not apply to a system where three-dimensional interactions are also present in the lattice structure. The one dimensional models of the type used in Chapter 5 cannot be utilized for the sulfonates for the same reason. The available 2-D Heisenberg model is not applicable in this instance, as this model only takes into account interactions between one paramagnetic center and only the four (but not six) nearest neighbors in a square array. In contrast, for the metal(II) sulfonates, the 175 proposed structure consists of each metal center being surrounded by six nearest neighbours (see Figure 6.1). It appears that the degree and sign of magnetic exchange in these compounds is a function of the O-S-O bridging angle, the M-O and S-O bond distances and the steric and electronic properties of the CF3 and F groups. Unfortunately, in the absence of any X-ray single crystal data it is rather difficult to make detailed magneto-structural correlations for these sul fonates in order to explain the observed magnetic interactions. 6.4 Conclusion The paramagnetic divalent fluorosulfates Ni(SO3F)2,Pd(SO3F)2,Pd(ll)[Pd(IV)(SO3F6], and Ag(SO3F)2 and their corresponding trifluoromethylsulfate derivatives Ni(SO3CF)2, Pd(SO3CF)2and Ag(SO3CF)2investigated for their magnetic properties show significant magnetic exchange, and except in Ag(SO3CF)2,the effects of magnetic exchange become observable at low temperatures. Two types of magnetic interactions are seen in the respective groups of compounds. The fluorosulfates exhibit ferromagnetism at all temperatures, whereas the trifluoromethylsulfates couple antiferromagnetically with the spin interactions noted over a wider temperature range. The fluorosulfates Pd(SO3F)2,“Pd(SO3F)”and Ag(SO3F)2were initially described as relatively magnetically dilute down to —80 K, and similarly, the Ni(SO3F)2compound studied here in the temperature range —291 to 2 K also follows the Curie-Weiss law between —291 and 79 K with Cm 1.34 ± 0.01 cm3 mo[1 K and 8 = 0.41 ± 2 K. For the ferromagnetic M(II) fluorosulfates, the following field dependent maximum magnetic moments are obtained in the temperature range —5 to 10.5 K: Ni(SO3F)2 5.22 B (5 K), PdSO3F)2 8.11 (8 K), “Pd(SO3F)”6.08 PB (8 K), and Ag(SO3F)27.14 B (10.5 K). Furthermore, the maximum magnetic susceptibility values of Ni(SO3F)2,Pd(SO3F)2and Ag(SO3F)2appear to indicate, for the magnetic fields used, saturation magnetization where all the magnetic spins align parallel to 176 the external magnetic field. Although the mixed valency fluorosulfate “Pd(SO3F)”shows significant ferromagnetism at low temperatures, the structurally similar bimetallic fluorosulfates Ni(ll)[Sn(IV)(SO3F6j,Cu(H)[Sn(IV)(SO3F6]and Ag(II)[Sn(IV)(SO3F6]are found to be magnetically dilute down to —4 K with calculated temperature independent magnetic moments of —3.3, —2.0, and —1.8 B respectively. In contrast to the fluorosulfates, the divalent trifluoromethylsulfate derivatives couple antiferromagnetically, and maxima in the susceptibility vs. temperature plots are noted for Pd(SO3CF)2and Ag(SO3CF)2at —4 and —138 K respectively. However, Ni(SO3CF)2does not show a Xm in its susceptibility plot, indicative of a weaker magnetic concentration in the compound. The magnetic moments of the three compounds decrease continuously with decreas ing temperatures, and hence no ferromagnetic contribution to the magnetic moments is detected at lower temperatures. Therefore, the antiferromagnetic behavior observed in Ni(SO3CF)2, Pd(SO3CF)2,and Ag(SO3CF)2seems to differ from that seen in the corresponding antiferro magnetic fluorides NiF2,PdF2, and AgF2,where a ferromagnetic magnetization component due to spin canting is detected at lower temperatures. Although the exchange pathways of the divalent sulfonates cannot be explained adequately due to a lack of X-ray single crystal data, the common layer type structure appears to indicate a possible intralayer vs. interlayer spin interaction, leading to predominantly parallel and antiparallel spin arrangements in the respective fluorosulfate and trifluoromethylsulfate lattices. However, the magnetic behavior observed in these sulfonates may be dependent on the O-S-O bridge angle, which may result in ferro- or antiferromagnetism for the respective groups of compounds investigated in this study. 177 References 1. K.C. Lee and F. Aubke, Can. J. Chem., 5, 2473 (1977). 2. K.C. Lee and F. Aubke, Can. J. Chem., 57, 2058 (1979). 3. P.C. Leung and F. Aubke, Inorg. Chem., fl, 1765 (1978). 4.a) P.C. Leung, K.C. Lee, and F. Aubke, Can. I. Chem., 57, 326 (1979). b) P.C. Leung, Ph.D. Thesis, University of British Columbia (1979). 5. S.P. Mallela, J.R. Sams, and F. Aubke, Can. J. Chem., j 3305 (1985). 6. J.R. Sams, R.C. Thompson, and T.B. Tsin, Can. 3. Chem., 55, 115 (1977). 7. J. Goubeau and J.B. Mime, Can. J. Chem., 45, 2321 (1967). 8. P.C. Leung and F. Aubke, Can. J. Chem., 2892 (1984). 9. P.C. Leung, G.B. Wong, and F. Aubke, I. Fluorine Chem., 35, 607 (1987). 10. J.M. Taylor and R.C. Thompson, Can. J. Chem., 4, 511 (1971). 11. K.C. Lee and F. Aubke, J. Fluorine Chem., 19, 501 (1982). 12. K.C. Lee and F. Aubke, Can. J. Chem., 59, 2835 (1981). 13. C.S. Alleyne, K.O. Mailer, and R.C. Thompson, Can. J. Chem., 5, 336 (1974). 14. J.S. Haynes, J.R. Sams, and R.C. Thompson, Can. 3. Chem., 52, 669 (1981). 15. A.L. Arduini, M. Garnett, R.C. Thompson, and T.C.T. Wong, Can. J. Chem., 5, 3812 (1975). 16.a) 0. Muller, Angew. Chem. mt. Ed. Engi., 2. 1081 (1987). b) L.N. Mulay in “Theory and Applications of Molecular Paramagnetism”, Eds. E.A. Boudreaux and L.N. Mulay, John Wiley and Sons, New York, 1976. 17. J.-M. Dance and A. Tressaud in “Inorganic Solid Fluorides”, Ed. P. Hagenmuller, Academic Press, New York, 1985. 18. A. Tressaud, J.-M. Dance, and P. Hagenmuller, Israel J. Chem., 17, 126 (1978). 19. M.T. Jansky and J.T. Yoke, 3. Inorg. Nuci. Chem., 41, 1707 (1979). 178 20.a) D.C. Adams, T. Birchall, R. Fa ggiani, R.J. Gillespie, and J.E. V ekris, Can. 3. Chem., , 2122 (1991). b) F. Charbonnier, R. Faure, and H. Loiseleur, Acta Crvst., 147 8 (1977). 21. D.A. Edwards, M.J. Stiff , and A.A. Woolf, Inorg. Nuci. Chem. Letters, ., 427 (1967). 22. R.L. Carlin, ‘Magnetoch emistry”, Springer-Verlag, New York, 1986. 23. C. Starr, F. Bitter, and A .R. Kaufmann, Phys. Rev., 5, 977 (19 40). 24. M.K. Wilkinson, J.W. C able, E.O. Wollan, and W.C. K oehier, Phys. Rev., IU. 497 (1959). 25. E. Stryjewski and G. Giordano, A dv. in Physics, , 487 (1977). 26. C. Starr, Phys. Rev., , 984 (1940). 27.a) N. Bartlett and P.R. Rao, Proc. Chem. Soc., 393 (1964). b) P.R. Rao, R.C. Sherwood, and N. Bartlett, 3. Chem. Phys., 49, 3728 (196 8). 28.a) E. Gruner and W. Klemm, Na turwissenschaften, 59, 25 (1937). b) P. Charpin, A.J. Dianoux, H. M arquet-Ellis, and C.R. Nguyen- Nghi, Acad. Sci. Fr., C264, 1108 (1967). c) P. Charpin, P. Plurien, and P. Me riel, Bull. Soc. Fr. Mineral Crist ollogr., 9, 7 (1970). d) P. Fischer, 0. Roult, and D. Schw arzenbach, 3. Phys. Chem. Soli ds, 3., 1641 (1971). 29. A. Tressaud, M. Winten berger, N. Bartlett, and P. Ha genmuller, C.R. Acad. ScL.Fr., C282, 1069 (1976). 30.a) T. Moriya, Phys. Rev., 112, 63 5 (1960). b) L.M. Matarrese and J.W. Stout, P hys. Rev., 24, 1792 (1954). V 31. S.P. Mallela, K. Lee, P .F. Gehrs, 3.1. Christensen, J.R . Sams, and F. Aubke, can. J. Chem., 5, 2649 (1987). 32. H. Wiliner, S.J. Rettig, 3. Trotter, and F. Aubke, Can. 3. C hem., , 391 (1991). 33. W.E. Hatfield, W.E. Es tes, W.E. Marsh, M.W. Pickens , L.W. ter Haar, and R.R. We ller in “Extended Linear Chain C ompounds’t,Ed. J.S. Miller, P lenum Press, New York, 1983. 179 34. D. Reinen and C. Friebel in “Structure and Bonding”, Eds. J.D. Dunitz, P. Hemmerich, C.K. Jorgensen, and D. Reinen, Vol. 37, Springer-Verlag, Berlin, 1979. 35. S.P. Mallela and F. Aubke, unpublished results. 36. F.A. Cotton and G. Wilkinson, “Advanced Inorganic Chemiostry”, 5th Edition, John Wiley and Sons, New York, 1989. 37. R.J. Joenk and R.M. Bozorth, J. Appl. Phys., 36, 1167 (1965). 38. B.C. Frazer and P.J. Brown, Phys. Rev., j, 1283 (1962). 39. I. Almodovar, B.C. Frazer, J.J. Hurst, D.E. Cox, and P.J. Brown, Phys. Rev., j., A153 (1965). 40. H.T. Witteveen and J. Reedijk, J. Solid State Chem., .IQ, 151 (1974). 180 preparation. The solvolysis product show temperature dependent low values, whereas the fluorination sample has unexpectedly high magnetic moments, which also decrease with decreasing temperature. As a preparative method, there appears to be a wide synthetic potential for this solvolysis route to the corresponding metal hexafluoro antimonates, since a large number of well charac terized transition metal fluorosulfates are available as precursors. The transition metal precursors of the hexafluoro antimonates, the divalent fluorosulfates Ni(SO3F)2,Pd(SO3F)2,Ag(SO3F)2,and the ternary “Pd(SO3F)1’all exhibit ferromagnetic exchange at lower temperatures. Additionally, the three binary fluorosulfate compounds indicate saturation magnetization at very low temperatures. In contrast, the corresponding trifluoromethylsulfates Ni(SO3CF)2,Pd(SO3CF)2 and Ag(SO3CF)2 couple antiferro magnetically, and for the last two compounds Xm are also found in their susceptibility vs. tem perature plots. The antiferromagnetism observed in these compounds differ from that seen in the corresponding binary fluorides, in that no ferromagnetic magnetization component due to spin canting is detected even at very low temperatures. Interestingly, both the divalent hexafluoro antimonates and the sulfonates studied in this work have a common layered structure which is based on the CdCl2prototype. Furthermore, the two post-transition metal layered compounds Sn(SO3F)2 and Sn(SbF6)2,which are structurally similar to their transition metal derivatives mentioned above, form it-arene adducts with mesitylene(mes) to give the weakly bound complexes Sn(SO3F)2mes and Sn(SbF6)2mes in high yield. The reduction of the lattice energies in the layer structures of the parent tin compounds by the wealdy nucleophilic anions SbF6 and SO3F appear to facilitate adduct formation, with the weaker nucleophile SbF6 been the more effective of the two ions, leading to the 2:1 complex with the arene. Mössbauer data of the adducts indicate 182 partial back-donation of the 5s electrons of tin to the antibonding it’’ orbitals of mesitylene to further stabilize the tin-arene bond, giving rise to synergic bond characteristics. The remaining group of non-transition metal fluoro complexes, the molecular species O2[AsF6],Br2[Sb316] andI2[SbF1i]’ which were investigated for their low temperature magnetic behavior, exhibit magnetic properties that are quite different for the three derivatives. This is reflected in the magnetic data of the respective complexes, measured down to -4K. Of the halogen compounds, Br2[Sb3F16J is magnetically dilute to low temperatures, whereas I2[SbF1] show relatively strong antiferromagnetic coupling with a Xm at 54K. As in Br2[Sb3F16J, the O2[AsF6] compound also exhibits Curie-Weiss behavior down to low temperatures, but weak antiferromagnetic exchange seems to be present in the very low end of the temperature region. InBr2[Sb3F16],the shortest non-bonding Br”Br distance is too large to invoke direct orbital overlap, but in12+[SbF1]magnetic exchange can occur via contiguous 12+ ions, where the non-bonding II distance is comparable to the sum of the van der Waals radii. The low values observed forO2[AsF6],which are below the spin only magnetic moment value, result from crystal field interactions in the solid lattice that partially quench the orbital contribution to the magnetic susceptibility. 183 APPENDICES 184 APPENDIX A A-i: Standard Reduction Potentials of Selected (M/M9 Couples* Electrode Potential (V) Ni2 + 2e —> Ni -0.250 Pd2 + 2e —> Pd +0.987 Cu + e — Cu +0.521 Cu2 + 2e — Cu +0.337 Cu2 + e — Cu -i-0.153 Ag + e — Ag +0.7991 Ag2 + e — Ag +1.980 Au + e —> Au +1.691 Au3 + 3e —> Au +1.498 Sn2 + 2e — Sn -0.136 Sn + 2e — Sn2 +0.15 Sb205+ 6H + 4e —> 2SbO + 3H20 +0.58 1 Sb205+ 2W + 2e —> 2SbO4+ H20 +0.479 + 4e — 2H0(1) +1.229 02 + 4W + 4e — 2H0(g) +1.185 * From J.E. Huheey, “Inorganic Chemistry”, 2nd Ed., Harper and Row, New York, 1978. 185 A -2 : Fr eq ue nc y R an ge o fV ib ra tio na lF un da m en ta ls fo rF lu or os ul fa te G ro up 0 0 - Bo nd ing Fr eq ue nc y R an ge (cm - 1 ) o r E Co or di na tio n — _ _ _ _ _ _ _ _ M od e St re tch in g Ba nd D ef or m ati on Bo nd CO VA LE NT i’ (S — O ) v ,, ,. (S — O) v (S — F) I F i TR IDE NT AT E > . (s o 3) - P (- ) PU RE LY v ., ., ,(S — O) i.( S— F) 61_ , . , , (S 0 3) 6 (S O 3) U c i w ’z z i c ri c i ION IC = is( S— O) v (S — F) 7 ,, i. (S O F ) ION IC V /Z Z Z J v ,, ,, (S O 2) L i 9 fi 7 (5 .J ) D O PE RT UR BE D rz z ’ r (5 02 F) (S Oi ) . (s— a) - . ‘ CO VA LE NT p0i, ,, (S O ) ( S O 7) Q (5 0 ) (S O 2 F) 0 I - ‘ CE ] O D D MO NO DE NT AT E 8 .( S O i) ? ., (s -F ) C i v (S — F) 7 (S O 2F ) v .. ym (S 0 2) CO VA LE NT w (S— O) is( S— F) 6 (S 0 7 ) i, (S -F ) r (S O 2F ) BID EN TA TE [f // A c ii i r j D 0 , , , , (S O 2) (S O 2) 7 (so2 F) 14 00 10 00 80 0 I I 12 00 60 0 40 0 A-3: X-ray Powder Data for Ni(SbF6)2 [Ni(SO3F)2+ SbF5la [Ni + F2 + SbF5] [NiF2 + SbF5]C d-space (A), Intensity d-space (A), Intensity d-space (A), Intensity 4.60 rn-s 4.61 rn-s 4.56 rn 4.17 s 4.24 s 4.26 m 4.03 w 4.17 rn 3.71 s 3.75 s 3.68 s 2.702 rn 2.744 rn-s 3.58 rn 2.534 m-w 2.576 m-w 2.70 rn 2.344 m-w 2.363 m 2.51 w 2.273 rn-s 2.252 s 2.33 w 2.224 rn 2.207 rn-w 2.22 m 2.167 rn 2.125 rn-w 2.10 w 1.859 rn 1.881 rn 1.86 rn 1.841 m 1.819 rn 1.768 w 1.746 w 1.714 rn-s 1.728 s 1.704 rn 1.681 m 1.70 m 1.648 rn 1.646 m 1.625 w 1.61 m 1.553 vw 1.550 vw 1.534 rn-w 1.521 rn 1.50 w 1.493 m 1.491 m 1.47 w 1.465 rn 1.447 m-w 1.44 w 1.426 w 1.420 rn-w 1.398 w 1.377 w 1.352 w 1.332 m a This work b Christe et at, J. Fluorine Chem., 4, 287 (1987) C Gantar et at, 3. Chem. Soc. Dalton Trans., 2379 (1987) 187 A-4: The Assignments and Energies of Octahedral Ni(IT) and Pd(II) (d8) Ligand Field Spectra, According to Lever,* are as Follows: 3A2g _> 3T2g 1)1 = lODq 3A2g _* 3Tig(F) = 7.5B + l5Dq - 1/2(225B + lOODq2 - 18ODqB)’a 3A2g 3Tig(P) = 7.5B + l5Dq + 1/2(225B + lOODq2- 18ODqB)12 and,3+u2-3i1=15B * A.B.P. Lever, J. Chem. Educ., 45,711(1968). [see also: Yu-Sheng Dou, J. Chem. Educ., 67, 134 (1990)]. 188 A-5: Low Temperature Magnetic Data of Ni(SbF6)2Made from Ni, F2 and SbF5 Temperature [K] x 106 [cm3mol1] Peff B1a 81.83 17140 3.35 78.06 17930 3.35 74.19 18740 3.33 69.72 19830 3.33 65.19 21030 3.31 60.15 22660 3.30 55.00 24620 3.29 47.90 27610 3.25 40.60 32010 3.22 30.80 41040 3.18 26.00 47840 3.15 21.00 57360 3.10 15.90 73240 3.05 10.30 105600 2.95 7.16 140500 2.84 6.30 153800 2.78 5.18 174200 2.69 4.40 191900 2.60 3.96 201700 2.53 3.46 207900 2.40 3.11 215800 2.32 2.89 219900 2.25 2.79 221800 2.22 a Uncorrected for TIP 189 A-6: Low Temperature Magnetic Data of Ni(SbF6)2Made from NiF2 and SbF5 in HF Temperature [K] XMCOff x 106 [cm3mol4] Peff B1a 81.83 12000 2.80 78.06 12490 2.79 74.19 13170 2.79 69.72 14000 2.79 65.19 14900 2.79 60.15 16130 2.79 55.00 17750 2.79 47.90 20070 2.77 40.60 23450 2.76 31.23 30160 2.74 26.40 35040 2.72 21.00 43070 2.69 15.90 55980 2.67 10.75 78900 2.60 7.16 114600 2.56 6.15 132600 2.55 5.16 151700 2.50 4.40 173500 2.47 4.04 190700 2.48 3.70 199400 2.43 3.29 211000 2.36 3.11 218100 2.33 2.89 225600 2.28 a Uncorrected for TIP 190 A-7: Low Temperature Magnetic Data of Au(SO3F)(SbF6 Temperature [K] XMCOff x 106 [cm3mol4]a ff [j.t] 81.78 410 0.52 74.24 460 0.52 69.72 560 0.56 65.25 630 0.58 60.21 690 0.58 54.20 820 0.60 47.40 980 0.61 40.00 1250 0.63 30.85 1820 0.67 25.60 2340 0.69 20.60 3070 0.71 15.85 4240 0.73 10.30 7010 0.76 6.84 11500 0.79 6.30 14800 0.86 4.32 20570 0.84 4.20 21090 0.84 3.37 22170 0.77 2.69 24300 0.72 a Experimental MW of compound = 734.88 g mol1 191 A-8: Crystal Structures of Some Arene-Sn(U) Complexes (a) C6HSnC1 (Aid4) (b) p-(CH3)2C64SnCJ (AICJ4) Cr3) CI(3 (c) [(C6H2SnCI (AICI4)12 (d) C6HSn(AICI4)2C6H C(2) (1) 192 A-9: Qualitative MO Diagram of [(arene)Ga] Complex (C6 Symmetry) E a Ga c6v HOMO e a LUMO a 193 APPENDIX B B-i: Low Temperature Magnetic Data of Ni(SO3F)2at Magnetic Field = 7501 G Temperature [K] XMCOff x 106 [cm3mol1] eff [IB]a 81.61 15470 3.18 77.72 16310 3.18 74.19 17180 3.19 69.49 18380 3.20 65.19 19700 3.21 59.86 21590 3.22 54.00 23950 3.22 47.60 27460 3.23 40.02 33160 3.26 30.35 45210 3.31 25.75 55730 3.39 20.48 73180 3.46 15.75 104700 3.63 10.42 200000 4.08 7.04 417400 4.85 6.30 534400 5.19 5.18 726200 5.49 4.78 813200 5.58 4.40 873100 5.54 4.32 879100 5.51 3.37 963100 5.10 2.89 1017000 4.85 2.69 1041000 4.73 a Uncoffected for TIP 194 B-2: Low Temperature Magnetic Data of Pd(SO3F)2at Magnetic Field = 9625 G Temperature [K] XMCOff x 106 [cm3mol1J 14ff 1JLB]a 81.50 18900 3.51 77.89 19840 3.52 70.17 22590 3.56 60.50 27090 3.62 54.40 30610 3.65 47.90 36090 3.72 40.30 45460 3.83 31.50 64760 4.04 7.42 875200 7.21 5.64 955100 6.56 4.32 984400 5.83 3.54 998000 5.32 2.60 1006000 4.57 2.20 1010000 4.22 1.84 1013000 3.86 a Uncorrected for TIP 195 B-3: Low Temperature Magnetic Data of Pd(JJ)[Pd(IV)(SO3F6]at Magnetic Field = %25 G Temperature [KJ ZMCOff x 106 [cm3moP1] eff B]a 82.33 21070 3.73 74.98 23360 3.74 61.00 30070 3.83 48.60 38330 3.86 32.10 59150 3.90 22.30 96310 4.14 12.15 285700 5.27 6.84 537300 5.42 5.08 604400 4.96 a Uncorrected for TIP 196 B-4: Low Temperature Magnetic Data of Ni(II)[Sn(IV)(SO3F6] Temperature [KJ XMCOff x 106 [cm3mol1] I.Lff [1.LBIa 81.56 16610 3.29 77.83 17460 3.30 73.91 18470 3.30 69.49 19560 3.30 65.02 20830 3.29 59.98 22600 3.29 53.85 24950 3.28 47.45 28330 3.28 40.30 33300 3.28 30.70 43750 3.28 25.85 51580 3.27 21.05 63720 3.28 15.50 85710 3.26 10.06 134900 3.30 6.44 223400 3.39 4.46 321200 3.39 4.24 340600 3.40 a Uncorrected for TIP Magnetic field = 7501 G 197 B-5: Low Temperature Magnetic Data of Cu(II)[Sn(IV)(SO3F6] Temperature [K] XMCOff x 106 [cm3mol1] ff []a 81.50 6540 2.06 77.83 6830 2.06 74.08 7100 2.05 69.83 7550 2.05 65.31 8010 2.05 60.09 8600 2.03 54.10 9520 2.03 47.55 10700 2.02 40.00 12730 2.02 30.80 16280 2.00 26.10 19360 2.01 21.35 23630 2.01 16.15 30650 1.99 10.48 46260 1.97 7.19 66730 1.96 5.84 82150 1.96 5.00 95270 1.95 4.78 98560 1.94 4.32 109700 1.95 3.02 140500 1.84 2.40 168800 1.80 a Uncorrected for TIP Magnetic field = 9225 G 198 B-6: Low Temperature Magnetic Data of Ag(II)[Sn(W)(SO3F6] Temperature [K] x 106 [cm3mol1] ff [& 81.61 5010 1.81 77.55 5230 1.80 73.68 5490 1.80 69.26 5970 1.82 64.90 6330 1.81 59.68 6890 1.81 53.12 7710 1.81 46.94 8730 1.81 40.20 10300 1.82 30.36 13640 1.82 25.30 16640 1.83 20.18 20070 1.80 15.05 26320 1.78 9.76 39680 1.76 6.44 57840 1.73 6.15 61370 1.74 5.34 68870 1.72 4.62 77760 1.70 3.96 91000 1.70 3.37 103600 1.67 2.69 112400 1.56 2.50 116200 1.52 a Uncorrected for TIP Magnetic field = 9225 G 199 B-7: Magnetic Data of Cu(SO3F)2for the Temperature Range 312 to 4 K Temperature [K] XMCOff x 106 [cm3mol1] Peff 312 1830 2.08 300 1900 2.08 287 1970 2.07 272 2090 2.08 240 2350 2.08 211 2660 2.08 181 3060 2.07 152 3620 2.07 123 4460 2.07 100 5460 2.07 81.78 6500 2.05 78.06 6850 2.05 74.36 7290 2.07 69.94 7780 2.07 65.77 8320 2.08 59.98 8950 2.06 55.00 9880 2.07 48.10 11180 2.06 41.60 13100 2.08 31.80 17280 2.09 26.65 20600 2.09 21.83 25170 2.09 16.90 33010 2.11 11.13 52110 2.15 8.70 71180 2.22 5.91 77690 1.91 4.20 85130 1.69 a First ten data points from Ref. 15, Chapter 6. b Corrected for TIP (TIP = 100 x 10*6 cm3 mo11,Ref 15, Chapter 6) Magnetic Field = 9225 G 200


Citation Scheme:


Usage Statistics

Country Views Downloads
China 2 1
City Views Downloads
Beijing 2 1

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}


Share to:


Related Items