UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Fluorosulfates of the noble metals, and their use in novel superacid systems Lee, Keith Cheuk-Lap 1980

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

Item Metadata


831-UBC_1980_A1 L35_3.pdf [ 12.57MB ]
JSON: 831-1.0060736.json
JSON-LD: 831-1.0060736-ld.json
RDF/XML (Pretty): 831-1.0060736-rdf.xml
RDF/JSON: 831-1.0060736-rdf.json
Turtle: 831-1.0060736-turtle.txt
N-Triples: 831-1.0060736-rdf-ntriples.txt
Original Record: 831-1.0060736-source.json
Full Text

Full Text

FLUOROSULFATES OF THE NOBLE METALS, AND THEIR USE IN NOVEL SUPERACID SYSTEMS by KEITH CHEUK-LAP LEE B . S c , Honours, The University of British Columbia, 1976 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY - . . i in THE FACULTY OF GRADUATE STUDIES Department of Chemistry We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September 1980 0 Keith Cheuk-Lap Lee, 1980 i i Authorization 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 i t 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. 3Ad.d..m.Q. Date Keith Cheuk-Lap Lee i i i Abstract This study was initiated by the search for novel superacid systems based on HSO^F, the strongest simple protonic acid. Therefore, the synthesis of binary metal fluorosulfates, M(S0 3F) n, was investigated, with special emphasis on compounds where M is in a medium or high oxidation state and formally coordinatively un-saturated, i.e., n = 3,4 or 5. Preferably, the resulting fluoro-sulfate would also display a sufficiently high thermal s t a b i l i t y with respect to dissociations giving rise to SOg or $20,^2, a n c' reasonable solubility in HSOgF. In order to obtain suitable binary fluorosulfates, a new, generally applicable synthetic method was developed, involving the oxidation of a metal with ^2^s^2 l n ^ e P r e s e n c e °^ HSOgF: HSO-jF M + n/2 S 20 gF 2 ^—> M(S0 3F) n , where M is primarily a 4d or 5d transition metal, in particular, palladium, platinum, gold and iridium. This fluorosulfonation mixture combines the strong oxidizing a b i l i t y of $2®6^2 w^t'1 t ' i e good solvating property of HSO^F towards ionic solutes. The a b i l i t y of these binary fluorosulfates to act as fluorosulfate ion acceptors was established by the synthesis of anionic complexes of the type CM(S0 3F) n + r T i] m", where m = 1,2. Both Au(S0 3F) 3 and Pt(S0 3F) 4 were found to act as ansolvo-acids in HS03F by the removal of fluorosulfate ions from the solvent ionization equilibrium: iv 2HS03F x H 2S0 3F + + S G y " , _ 2-to form [AiKSOgF^] or [Pt(S0 3F)g] . Results from conductivity measurements indicate the resulting superacids have acid strengths comparable to that of the HSO^F-SbFg-SO^ system and definitely superior to the HSO^F-SbFg system. In both cases, the variety of ternary fluorosulfates synthesized included unusual cations such as B r 3 + and Br^ +. In the case of platinum, evidence for poly-nuclear oligomeric anions of the type of [Pt(S0 3F) j - ] n n ~ were obtained; similar observations were also made in the tin system. The oxidation of palladium yielded PdtSO^F^ and a mixed valency compound, Pd(II)[Pd(IV)(S03F)g],- both containing Pd(II) 3 in a rather uncommon octahedral coordination sphere and a Ag q electronic ground state. These magnetically dilute compounds allowed meaningful magnetic results and ligand f i e l d parameters to be interpreted. Both Ir(S0 3F) 3 and Ir(S0 3F)^ could be synthesized, along with some tenary fluorosulfate complexes. Some preliminary results were also obtained on the fluorosulfates of germanium, molybdenum, niobium and tantalum. V TABLE OF CONTENTS page Abstract i i i Table of Contents v List of Tables v i i i List of Figures x Acknowledgement xi i Chapter 1 Introduction l.A General Introduction 1 l.B Properties of HS03F 6 l.C Superacids Systems based on HSO3F 8 l.D Applications of Superacids 14 l.E Preparation of Metal Fluorosulfates 15 l.F Transition Metal Fluorosulfates 21 1. G Vibrational Modes of the S0 3F Group 22 Chapter 2 Experimental 2. A Introduction 27 2.B General Equipment 29 2. C Chemicals 39 Chapter 3 Palladium-Fluorosulfate 3. A Introduction 42 3.B Experimental 46 3.C Discussion 56 3.C.1 Synthesis and General Discussion 56 3.C.2 Vibrational Spectra 65 3.C.3 Electronic Spectra 75 3.C.4 Magnetic Susceptibility 82 3.C.5 Solution Studies in HS03F 92 3.D Conclusion 95 Chapter 4 Gold-Fluorosulfate 4.A Introduction 96 vi 4.B Experimental 100 4.C Discussion 109 4.C.1 Synthesis and General Discussion 109 4.C.2 Vibrational' Spectra 118 4.C.3 Magnetic Susceptibility 131 4.C.4 Solution Studies in HS03F 133 4. C.5 Raman Studies of Au(S0 3F) 3 in Liquid S0 2 151 4. D Conclusion 155 Chapter 5 Platinum-Fluorosulfate 5. A Introduction 157 5.B Experimental 160 5.C Discussion 165 5. C.1 Synthesis and General Discussion 165 5.C.2 Vibrational Spectra 169 5.C.3 Magnetic Susceptibilities 178 5. C.4 Solution Studies in HS03F 179 5. D Conclusion 197 Chapter 6 Iridium-Fluorosulfate 6. A Introduction 198 6.B Experimental 200 6.C Discussion 204 6. C.1 Synthesis and General Discussion 204 6.C.2 Vibrational Spectra 206 6.C.3 Magnetic Susceptibility 210 6. C.4 Solution Studies in HS03F 219 6. D Conclusion 223 Chapter 7 Molydenum (VI)-Fluorosulfate 7. A Introduction 224 7.B Experimental 226 7.C Discussion 228 7. C.1 Synthesis and General Discussion 228 7.C.2 Vibrational Spectra 229 v i i 7. C.3 19F-n.m.r. Spectra of MoCKSGVK 234 7. D Conclusion 236 Chapter 8 Germanium- and Tin-Fluorosulfates 8. A Introduction 237 8.B Experimental 239 8.C Discussion 243 8. C.1 Synthesis and General Discussion 243 8.C.2 Vibrational Spectra 245 8.C.3 Conductometric Studies in HS03F 254 8.C.4 n.m.r. Studies in HS03F 263 8.D Conclusion 268 Chapter 9 Conclusion 270 References 273 Appendix A Conductivity Calculations 284 Appendix B Gold-Trifluoromethylsulfate 293 Appendix C List of Abbreviations 296 LIST OF TABLES Table page 1.1 Properties of H S O 3 F , H 2S0 4 and H20 7 2.1 Diamagnetic Susceptibilities 38 2.2 Chemicals used with Purification 38 3.1 Vibrational Frequencies of Pd(S0 3F) 2 66 3.2 Vibrational Spectra of Cs 2[Pd(S0 3F) 6], (NO),[Pd(S0 3F) 6], (C10 2) 2[Pd(S0 3F) 6], Ba[Pd(S0 3F) 6], Pd[Pd(S0 3F) 6], Pd[Pt(S0 3F) 6] and Pd[Sn(S0 3F) 6] 70 3.3 i . r . Frequencies for (N0) 2-, Cs 2- and Ba-[Pd(S0 3 FK] ... 74 3.4 Electronic Transitions and Ligand Field Parameters for some Palladium (II) Compounds and Ni(S0 3F) 2 77 3.5 Electronic Spectra and Ligand Field Parameters for [Pd(S0 3 F K ] 2 " 80 3.6 Magnetic Properties of Pd(S0 3F) 2 85 3.7 Magnetic Properties of Pd 2(S0 3F) 6 85 3.8 Magnetic Properties of Pd[Pt(S0 3F) 5] 86 3.9 Magnetic Properties of Pd[Sn(S0 3F) 6] 86 3.10 Magnetic Properties of some Pd(II) Compounds at Room Temperature 88 3.11 Magnetic Properteis of some Ni(II) Compounds 90 3.12 Magnetic Susceptibilities for Diamagnetic Palladium Fluorosulfates 91 3.13 Conductivity of Cs 2[Pd(S0 3F) 6] in HS03F ,.. 94 4.1 Vibrational Frequencies of M[Au(S03F)[+] 121 4.2 Vibrational Frequencies of Au(S0 3F) 3 125 4.3 Vibrational Frequencies of [X(S03F)2] • [Au(S03F)i»] 128 4.4 Vibrational Frequencies of Br p[Au(S0 3F) l t] 131 4.5 Magnetic Susceptibilities of Au(S0 3F) 3 and Br n[Au(S0 3 F k ] 133 4.6 Conductometric Titration of Au(S0 3F) 3 with KS03F in HS03F 140 4.7 Raman Frequencies for Cs[Au(S0 3F) 4] ,Br3[Au(S03FK] and Au(S0 3F) 3 in HS03F 148 ix 4.8 Raman Frequencies for Au(S0 3F) 3 in Solid State and in £-S0 2 Solution at 298 K 152 5.1 Vibrational Frequencies of Pt(S0 3F) 4 and Au(S0 3F) 3 170 5.2 Raman Frequencies of [Pt(S0 3F) 6] 2" Complexes 173 5.3 Vibrational Frequencies of CsPt(S0 3F) 5 and Related Compounds 177 5.4 Diamagnetic Susceptibilities for some Pt(IV) Fluorosulfate Compounds at Room Temperature 179 5.5 Conductivity of (C10 2) 2[Pt(S0 3F) 6] and CsPt(S0 3F) 5 in HS03F 182 5.6 Conductivity of Pt(S0 3F) l f in HS03F 184 5.7 Conductometric Titration of PtCSOsF)^ with KS03F in HS03F 189 6.1 i . r . Frequencies of Ir(S0 3F) 3 and Ir(S0 3F) 4 207 6.2 Vibrational Frequencies of [ I r ( S 0 3 F ) 6 ] 2 " 209 6.3 Magnetic Properties of (C10 2) 2[Ir(S0 3F) 6] 211 6.4 Magnetic Properties of Cs 2[Ir(S0 3F) 6] 211 6.5 Magnetic Properties of Ir(S0 3F ) i+ 212 6.6 Magnetic Properties of (C10 2) 2[Ir(S0 3F) 6] diluted in (C10 2) 2[Pt(S0 3F) 6] 216 6.7 Conductivity of (C10 2) 2[Ir(S0 3F) 6] in HS03F 221 7.1 Vibrational Frequencies of Mo0 2(S0 3F) 2 231 7.2 Vibrational Frequencies of MoCKSCVK 233 8.1 Raman Frequencies of CsSn(S0 3F) 5 246 8.2 Vibrational Frequencies of GeF 2(S0 3F) 2 248 8.3 Stretching Force Constants of some Group IV Element-Fluorine Bonds 251 8.4 Vibrational Frequencies of (C10 2) 2[Ge(S0 3F) 6] 253 8.5 Conductometric Titration of KSn(S0 3F) 5 with KS03F in HS03F 256 8.6 Conductivity of (C10 2) 2[Ge(S0 3F) 6] in HS03F 262 X LIST OF FIGURES Figure page 1.1 -HG Functions of some Acids and Bases in HS03F 11 1.2 Relationship between I/E and -H0 13 1.3 Vibrational Band Pattern for the S0 3F Group 24 2.1 Examples of Glass Apparatus 28 2.2 Conductivity Cell 31 2.3 Weight Dropper 31 2.4 Graduated Buret 33 2.5 Low Temperature Raman Cell 36 3.1 Proposed Structure of Pd(S0 3F) 2 . . v 66 3.2 Raman Spectrum of Pd 2(S0 3F) 6 69 3.3 Raman Spectrum of Pd[Pt(S0 3 F ) 6 ] 69 3.4 Anisobidentate Bridging Mode for S03F 72 3.5 Energy Level Diagram for a d 8 Ion in an 0 n Field 77 3.6 d-Orbital Energy Level Diagram for a Square Planar Complex 80 3.7 Curie-Weiss Plot for Pd(II)-Fluorosulfate Derivatives .. 87 3.8 Conductivity of Cs 2[Pd(S0 3F) 6] in HS03F 93 4.1 Raman Spectrum of K[Au(S0 3 F)iJ 119 4.2 Raman Spectrum of C10 2[Au(S0 3F)J 120 4.3 Conductivity of [Au(S03F)iJ-Complexes in HS03F 136 4.4 Conductivity of Au(S0 3F) 3 in HS03F 137 4.5 Conductometric Titration of Au(S0 3F) 3 with KS03F in HS03F at 25°C 139 4.6 19F-n.m.r. Spectra of Au(S0 3F) 3 and Cs[Au(S03F)t+] in HS03F 145 5.1 Infrared Spectrum of (C10 2) 2 [Pt(S0 3F) 6] 174 5.2 Electrical Conductivity of Cs 2[Pt 2(S0 3F) 1 0] and (C10 2) 2[Pt(S0 3F) 6] in HS03F at 25°C 181 5.3 Conductivity of Pt(S0 3FK in HS03F 183 5.4 Conductometric Titration of PtCSOsF)^ with KS03F in H S O 3 F 187 xi 5.5 1 9F-n.m.r. of Cs 2[Pt(S0 3F) 6] in HS03F .. 195 6.1 Magnetic Moments of (C10 2) 2[Ir(S0 3F) 6] and Cs 2[Ir(S0 3F) 6] 213 6.2 Magnetic Moment of (C10 2) 2[Ir(S0 3F) 6] diluted in (C10 2) 2[Pt(S0 3F) 6] 217 6.3 Conductivity of (C10 2) 2fIr(S0 3F) 6] 220 7.1 1 9F-n.m.r. Spectra of MoO(S03F)4 234 8.1 Conductometric Titration of KSn(S0 3F) 5 with KS03F in HS03F -255 8.2 Electrical Conductivity of KSn(S0 3F) 5 in HS03F at 25°C 259 8.3 Conductivity of (C10 2) 2[Ge(S0 3F) 6] in HS03F 261 8.4 119Sn-n.m.r. Spectra in HS03F 264 8.5 19F-n.m.r. of [Sn(S0 3F) x] n" in HS03F 266 xii ACKNOWLEDGEMENT I would like to express my sincere appreciation to Professor F. Aubke, my research supervisor, for his inspirations and guid-ance during my years of graduate study. Thanks are also extended to other members of the faculty and staff of this department for their assistance in various aspects of this research, in particular, Professor R.C. Thompson and Dr. P.C. Leung for stimulating discussions, Dr. S.O. Chan for taking some of the n.m.r. spectra, and the glass-blowing and mechanical shops for the construction of most of the apparatus used. Pro-fessors J . Trotter and N.L. Paddock are thanked for reading parts of the manuscripts. Mrs. B. Krizsan is thanked for some of the illustrations. I would also like to thank my typists for doing such a fine job: Mr. Derek Lee and Miss Roselyn Lee, who did most of the typing, and Ms. Dorathea Baker and Mrs. Bev Gray. The receipts of a National Science and Engineering Research Council Post Graduate Scholarship and a Dr. F.J. Nicholson Fellowship are gratefully acknowledged. T O 1 CHAPTER 1 INTRODUCTION l.A GENERAL INTRODUCTION Fluorosulfuric acid, HS0 3F, was f i r s t prepared by Thorpe and Kirman by the reaction between sulfur trioxide and anhydrous hydrogen fluoride 1: HF + S0 3 • HSO3F . (1.1) Although completely ionic salts of the acid, such as KS0 3F, can be recrystal1ized from their neutral aqueous solutions 2, HS0 3F i t s e l f is nevertheless unstable in water, in which i t is hydrolyzed almost spontaneously according to: HSO3F + H20 • HF + H2S0n . (1.2) The instability of the S-F bond towards hydrolysis is also illustrated by the fact that most fluorosulfates are extremely sensitive to moisture. While there have been questionable reports on the existence of pure aqueous solutions 3> h and hydrates 5 of the acid, HS0 3F is incompatible with the aqueous system for a l l practical purposes 6" 9. On the other hand, the acid has been studied extensively as an ionizing and protonic non-aqueous solvent. This can be reflected in a number of reviews on HSO3F and its derivatives, in particular those by Lange 1 0 , Gillespie et al n > 1 2 , Thompson13, Jache et al 1 1 + » 1 5 and Olah et al 1 6 . The subject has also been described in related articles by Cady 1 7 , Williamson 1 8 , and, De Marco and Shreeve 1 9 . In addition, the versatility of HS0 3F can be illustrated by i t s 2 use as a fluorinating agent in reactions with oxides and oxyacid salts, summarized in an article by Englbrecht 2 0 . H S O 3 F is the strongest simple protonic acid, more so than H S O 3 C F 3 , H 2S 20 7, H2SGV, H S O 3 C I and H C l C v Its high dielectric constant makes i t a very suitable medium for the formation and generation of ionic solutes 2 1 . Since the neat solvent displays such a high acidity, i t is not surprising that HSO3F is widely used as a basis for superacid systems (sometimes called 'magic acids' by some workers in the f i e l d of carbonium ion chemistry 2 2 > 2 3 ) . The definition of a superacid is by no means universal; some refer to i t as any protonic acid with an acidity greater than that of 100% r^SO^ 1 2 . According to this classification, HS03F i t s e l f would be a superacid. Another definition, preferred and subsequently used in this thesis, views a superacid as a multicomponent system with an acidity unattainable by simple protonic acids. In terms of solvo-system terminology, an acid in HSO3F increases the acidium ion concentration, [H 2S0 3F +]. It can accomplish this either by the protonation of the solvent according to: HA + HSO3F *• H2S03F+ + A" , (1.3) or by the removal of the fluorosulfate ion, SOsF", from the self ionization equilibrium by: A + 2HS03F y H 2S0 3F + + A-S03F" . (1.4) Due to the extremely high acidity of HSO3F, the direct protonation of HSO3F by a simple protonic acid is not possible. 3 While most solutes are bases in H S O 3 F , only a few strong Lewis acids have been found to act as ansolvo-acids according to equation ( 1 . 4 ) . The most thoroughly studied superacid system in HS03F contains the fluoride or fluoride-fluorosulfates of antimony (V) as the ansolvo-acids. The fluoride-fluorosulfates are usually formed in situ in these mixtures by the addition of S0 3 in a reaction such as 2 4 : SbF 5 + S0 3 • SbFi+(S03F) . (1 .5) These compounds have been found to be stronger acids than the pentafluorides in HSO3F; SbF 2(S0 3F) 3, identified only in HS0 3F solutions, gives rise to the strongest acid of the group. It is apparent that the successive substitution of fluorine atoms by fluorosulfate groups increases the acidity of the superacid system 1 1 . This leads to the expectation that the pentakis-(fluorosulfate) should yield a solution of even greater acidity than the fluoride-fluorosulfates. Unfortunately, a higher degree of substitution is not possible in this system, and only SbFi+(S03F) 2 5 and SbF 3(S0 3F) 2 2 6 have hitherto been isolated; attempts to prepare Sb(S0 3F) 5 have failed 2 7 . Although there is no implied correlation between the acidity and basicity of a solvent, the existence of the highly electrophilic acidium ion, H2S03F+, in a superacid system indicates that other unusual cations may be stabilized in the medium. The complexation of the strongest base in the fluorosulfuric acid system — S03F~, to give larger, less nucleophilic anions such as [ S b F 2 ( S 0 3 F ) i t ] i s expected to decrease the basicity of the resulting solution. Therefore, in order for an ansolvo-acid to give rise to a strongly acidic and weakly basic solution in H S O 3 F , i t must possess strong Lewis acidity towards S03F~ with the resulting complex anion, A'S03F~, showing minimal reverse dissociation according to: A-S03F~ > A + SQaF" . (1.6) In the antimony (V) fluoride-fluorosulfate superacid system, additional modes of dissociation of the compounds in solution by 1 2> 2 h: [SbF 2(S0 3F)J" - y [SbF 3(S0 3F) 3]" + S0 3 , (1.7) and the decomposition of the ansolvo-acids themselves by 2 6 : SbF 3(S0 3F) 2 *• SbFjS0 3F) + S0 3 , (1.8) can lead to the formation of S0 3, a reasonably powerful oxidizing agent. The presence of S0 3 may give rise to oxidative side reactions in these media, in particular with species such as carbonium ions 2 2 and polyhalogen cations 2 8 - 3 0 > Hence, i t becomes necessary to look for other fluorosulfates that may be better ansolvo-acids in the HS03F solvent system than those presently known. A good fluorosulfate acceptor should have the following properties: a) a reasonable thermal s t a b i l i t y , especially towards S0 3 elimination and ligand redistribution reactions, b) an incompletely f i l l e d coordination sphere in order to f a c i l i t a t e the addition of S03F group(s), c) a high solubility in HS03F to give reasonable 0 5 concentrations, and d) a low oxidative power, to reduce side reactions and to allow the study of a wide range of species. A review of the previously reported fluorosulfates reveals that only a few form anionic complexes, and of these, none are really suitable as ansolvo-acids. For example, although salts containing the anion - [Sn(S0 3F) 6] 2~ can be prepared 3 1> 3 2, SnfSOgF)^ is virtually insoluble in HSO3F 3 3 . Both Br(S0 3F) 3 and I(SO3F)3 form anionic complexes of the type of [Hal (SC^F^] - 3 i > 3 4 , 3 5 S th ey c a n a ] s o e x i s t in cationic species of the type - [Hal(S0 3F) 2] + 2 8> 3 1. This indicates that the Hal(S0 3F) 3 can act as both an acceptor and a donor towards a S03F group. Not surprisingly, I(S0 3F) 3 has been found to be amphoteric in HS0 3F, capable of both acidic and basic reactions in the solvent 3 6 . Furthermore, both these compounds are oxidizing 3 7> 3 8, and thus are unsuitable as ansolvo-acids in the HS0 3F acid system. It appeared that other fluorosulfates would have to be synthesized, and possibly by new methods. Fluorosulfates of the 4d and 5d elements have not been studied extensively, but are like l y suitable for the following reasons: a) The halides of these metals are capable of the formation of halocomplexes, indicating Lewis acidity in the halides system which most lik e l y should extend to the fluorosulfate system. b) The Lewis acidity of this type of fluorosulfates is 6 expected to increase with the metal's oxidation number, and high oxidation states of these metals are relatively easily attainable. In order to establish that the metal fluorosulfate will accept fluorosulfate ions, the successful synthesis of anionic fluorosulfato-complexes will provide the f i r s t test. The characterization of these ternary fluorosulfates by various spectroscopic means, including Raman, i . r . , n.m.r. and u.v.-visible, should f a c i l i t a t e their subsequent detection in superacid systems. Their acidic behavior in H S O 3 F can then be assessed by the use of electrical conductivity measurements. In H S O 3 F , as in H2S0i+ and H 2 0 , the ions produced by autoprotolysis display much higher mobilities than any other ions in the solvent. This has been explained by the concept.that they conduct by the proton-transfer mechanism in these strongly hydrogen bonded solvents 3 9 . Consequently, strong acids and strong bases give rise to highly conductive solutions in H S O 3 F , and titrations with K S O 3 F (a strong base in H S O 3 F ) can then be used to distinguish between acidic and basic conductivities. In the following sections in this chapter, a few important topics will be reviewed in more detail. I.B PROPERTIES OF H S O 3 F The properties of HS03F are compared with those of the more familiar protonic solvents, H2 S0i + and H 2 0 , in Table 1.1. As can be seen, HS03F has a long liquid range; and a low TABLE 1.1 Properties of HS03F, H7SOk and H?0 (at 25°C) HS03F 1 3 H ?S0 4 39 H90 k 0 -88.98 10.371 0.00 162.7 290 100.00 1.7264 1.8269 0.9971 1.56 24.54 0.8904 120 100 78.33 l.ossxio- 1* 1.044x 10-2 4.3x10" 185 250 350 135 166 200 3.8xl0 - 8 2.7x10 -4 1.0x10" 15.07 11.93 -7.00 Freezing point, °C Boiling point, °C Density (g-cc - 1) Viscosity (cp) Dielectric constanl < (n^cm - 1) A+ A" K a p (mol 2kg- 2) -H0 viscosity. These two factors have made H S O 3 F a very useful medium for n.m.r. studies at low temperatures 22,24,41-45^ In such conditions, individual components of an exchange reaction can be identified. In contrast to h^SO^, in which a series of dissociative reactions, represented by equations (1.9) to (1.12), have been observed 3 9 , 2H?S0i4 >• H 3S0u + + HSOu" K=2.7xl0",+ (1.9) 2H2S0k - > H 30 + + HS 20 7" K=5.1xl0" 5 (1.10) H 2S 20 7 + H2S0k- >• H 3S0 u + + HS 20 7" K=1.4xl0~2 (1.11) H20 + H2SO4 >• H 30 + + HS0h" K=l , (1.12) HS03F undergoes only two minor autodissociative reactions in the pure state k 6 : 2HS03F y H7S0gF+ + SOgF" K=3.8xl0 - 8 (1.13) HS03F >• HF + SOa K<3xl0" 7 . (1.14) For a l l practical purposes, these dissociations do not interfere with the interpretation of conductivity and cryoscopic data in 8 HS03F 1 3 . A decrease in intermolecular association in HS03F, due to the presence of one less hydrogen per molecule than in H2SG\, can be reflected in i t s lower viscosity and the lower molar conductivities for H 2S0 3F + and SC^F". Nevertheless, the ion mobilities of the autoprotolysis ions are s t i l l much higher than those of the other ions in HS03F. For example, K+ has a A + value of only 17 in HS03F (the data for H 2S0 3F + and S0 3F" are listed in Table 1.1). The high acidity of HS03F as compared to other strong acids in terms of H0 of the pure acids is illustrated below 12,^7-49. HF(-ll)<H2S0tt(-11.93)<HS03Cl(-13.80)<HS03CF3(-14.00)<HS03F(-15.07). l.C SUPERACID SYSTEMS BASED ON HSO3F As mentioned previously, the SbF5-3S03 superacid has been found to be the strongest acid in HS03F. Solutions of SbF 5 in HS03F, although slightly less acidic, have been used extensively in organic chemistry for the stabilization of carbonium ions, because they contain less S0 3. The SbF5-HS03F system is very complex, and can be represented by the following equilibria 2 HS03F + SbF 5 • H[SbF5(S03F)] (1.15) H[SbF 5(S0 3F)] + HSO3F y H 2S0 3F + + [SbF 5(S0 3F)]" (1.16) 2H[SbF5(S03F)] • H 2S0 3F + + [Sb 2F 1 0(S0 3F)]' (1.17) HSO3F ^ HF + S03 (1.18) 2HF + 3SbF5- • HSbF6 + HSb 2F n (1.19) 3S03 + 2HS03F y HS 20 6F + HS 30 9F (1.20) 9 This interactive series of reactions gives rise to different dominant species in solutions, depending on the concentration of SbF5. This can lead to side reactions and may present d i f f i c u l t i e s when attempts are made to isolate a particular product from the mixture. Also, because the mixture is so complex, the use of spectroscopic techniques such as F-19 n.m.r. and Raman are d i f f i c u l t . Since SbF 5 is an ansolvo-acid of the HS03F system, other pentafluorides of Group V elements, such as AsF 5 and TaF 5, should also be expected to behave as acids in the solvent. While they do give rise to conducting solutions in HS03F, these solutions are only very weakly acidic (when either PF5 or NbFs is dissolved in HSO3F, the resulting solutions are hardly conducting). Titrations of these solutions with KS03F also indicate only weakly acidic behaviors. Broad and ill-defined conductivity minima at the base/acid ratios of 0.06, 0.01, and 0.31 are found for solutions of BiF 5, AsF 5 and AsF 5«S0 3, respectively 5 0 . In the HS03F-AsF5 system, S0 3 presumably behaves as i t does in the HS03F-SbF5 system. While the conductometric titration with KSO3F can provide a * semi-quantitative assessment of a solute's acidity in HSO3F, a ranking of it s acidity can be made by two other methods discussed below: a) The Bronsted acidity of a solution governs the degree of protonation of bases in i t . By using a series of weakly basic nitro-aromatic compounds, the Hammett Acidity Function,-H 0, has been determined for a number of strong acids and superacids 4 8, i + 9. The results parallel those obtained from conductometry, namely, that the effectiveness of a solute in increasing the acidity of H S O 3 F increases in the order of S03<AsF5<SbF5<SbF5-3S03. The behavior of S O 3 as an extremely weak acid in H S O 3 F , forming polyfluorosulfuric acids of the type H(S03)nF, has also been demonstrated by the spectroscopic methods of Raman and F-19 n.m.r. 1+1»'+3. A 7% molar solution of SbF 5«3S0 3 in H S O 3 F gave an H0 value of -19.35, the highest one measured thus far. The lack of suitable bases prevented the extension of the work to more concentrated, and presumably more acidic solutions. By an extra-polation of the results obtained at lower concentrations in the HS03F-SbF5 system, the 1:1 molar mixture - sometimes called 'magic acid'. - has an H0 value of -19.5, making that solution an extremely acidic medium. Figure 1.1 shows the H0 function plot for a number of solutes in HS03F 1 + 9. While the original H 0 measurements, as defined by Hammett, were obtained by the spectrophotometry monitoring of the concen-trations of the base and i t s protonated species, an alternative method has been proposed recently by Sommer 5 1> 5 2. This method is based on determining the increase in the barrier of rotation around the phenyl-carbonyl bond, in compounds such as p-methoxy-benzaldhyde, upon protonation as measured by n.m.r. shifts. A value equivalent to H0 of -21.4 was obtained for a 25 mol% of SbF 5 in H S O 3 , suggesting an extremely high proton donating a b i l i t y in the medium 5 1 . However, the increase in the rotation barrier for the C-C bond in the indicator could conceivably be dependent FIG 1.1 on the increased viscosity of these strongly hydrogen bonded solutions containing high concentrations of SbF 5. Thus, this may not be a true measure of the protonating a b i l i t y of such highly concentrated superacids. This could explain the conclusion reached in the study that HF-SbF5 (which has a higher viscosity than HS03F-SbF5) is the much more acidic medium of the two 5 2 . In any event, i t should again be pointed out that the usefulness of a superacid system is reflected in the protonating a b i l i t y , but also on the lack of basicity in the system. The greater charge delocalization in the H S O 3 F based superacid is illustrated in the following basicity measurements. b) The nucleophilicity, or basicity of a solution,determines the s t a b i l i t y of a cationic species in i t . An alkyl cation, for example, can rearrange with or without deprotonation during i t s lifetime in the acid, depending on the nucleophilicity of the acidic species. Using tritium-labelled alkanes, Krammer studied the extent of the isomerization and exchange reactions of alkanes in a series of acids containing different Lewis acids 5 3> 5 t t. A selectivity parameter, I/E, defined as the ratio of: k(isomerization)/k(exchange) , was derived for each of the mixtures studied. The results, represented in Figure 1.2, show that I/E parallels H0 within a given acid, but the correlation is solvent dependent. Alterna-tively, a strong acid is usually also a weak base in a given solvent system. FIG 1.2 RELATIONSHIP BETWEEN I/E AND -H0 l.D APPLICATIONS OF SUPERACIDS The low bascity of superacid systems has also been used to advantage in the study of highly electrophilic, non metallic polyatomic cations, in particular, polyhalogen cations 2 5 » 5 5 . The 'hardness' and electrophilicity of a particular cation are expected to increase with a decrease in its size. Thus, while I 7 + , I 5 + and I 3 + are stable in 100% H2SO4 5 6 , the smaller I 2 + can only be obtained either in oleum 5 7 or in HS03F 5 6 . In the bromine system, the oxidation of Br 2 in the superacid medium of HS03F-SbF5-S03 according to 2 8 » 5 8 : 3Br 2 + S 20 6F 2 • 2Br 3 + + 2S03F" , (1.21) gives rise to B r 3 + , but, in the more basic HS03F, the cation disproportionates: B r 3 + + S0 3F" >• Br 2 + BrS03F . (1.22) Similarly, the smaller B r 2 + ion is found to be unstable even in the extremely weakly basic medium of HS03F-SbF5-S03 2 8 : 2Br 2 + + 2HS03F y B r 3 + + BrS03F + H 2S0 3F + . (1.23) Contrary to earlier reports 5 9 , there is no evidence for the existence of either C l 2 + or C l 3 + in superacid solutions 2 9 . However, Cl 3 +-containing salts have been identified in the reaction of a mixture of C l 2 and C1F with either AsF 5 or HF-SbF5 at -78°C 6 0 . The fact that strong protonic acids can catalyze rearrange-ment and nucleophilic substitution reactions involving large organic molecules has been known for a long time. Very recently, 15 Olah found that superacid systems, at high temperatures and pressures, can protonate small alkanes such as methane to form species like CH 5 + 1 6 . The subsequent condensation of such cations to form larger, longer chained alkanes is a process of obvious industrial importance. l.E PREPARATION OF METAL FLUOROSULFATES Fluorosulfate chemistry displays a similarity to halogen chemistry in that the S0 3F group can be viewed as a pseudohalogen. Consequently, many of the synthetic methods used commonly in the synthesis of halides can be adapted to the fluorosulfate system with only minor changes. l . E . l SOLVOLYSIS OF ACID SALTS IN H S O 3 F As is common for the preparation of salts of a strong acid, fluorosulfates can be made by simple displacement reactions such a s 2 , 2 1 , 6 1 - 6 3 . H S O 3 F + KC1 c HC1 + KSO3F . (1.24) The oxidation state of the metal involved remains unchanged and since this is an ionic reaction, i t is especially applicable to the synthesis of alkali or alkali earth fluorosulfates. If the acid salt is polymeric, an incomplete reaction often occurs. For transition metal fluorosulfates that show limited solubilities in H S O 3 F , such as Cu(S03F)2, a complete reaction does not take place without the addition of KS0 3F to increase the basicity of 16 the medium 6 2 . As expected, the formation of a volatile acid such as HC1 or HBr, as the by-products from the reaction greatly simplifies the preparation. Conversely, for carboxylates or other salts that give rise to non-volatile acids, the products, which need to be insoluble in HS03F in these cases, have to be separated by f i l t r a t i o n . Furthermore, the solvolysis of compounds such as carbonates or nitrates will give H20 as one of the products, and therefore cannot be used. I.E.2 OXIDATION OF METALS BY HSO^F Hydrohalic acids are frequently used to react with metals to give metal halides directly. Just as some metals can reduce r\2S0^ to S0 2, HSO3F has been found to be reduced to S0 2 and even elemen-tal sulfur. Besides, most metals are inert even in the boiling acid, either because HS03F cannot oxidize the metal or because an insoluble surface layer is formed . I.E.3 INSERTION OF SO3 INTO M-F BONDS The formation of fluoride-fluorosulfates of As(V) and Sb(V) in H S O 3 F by the insertion of S0 3 into the metal-fluorine bonds has been discussed 2 1 +, 5 0: [SbF 5(S0 3F)]" + nS03 »• [SbF(, 5. n)(S0 3F) n + 1]" . (1.25) Species with n of up to 3 have been identified in solution. When no acid is present, the substitution is incomplete, forming only one product when an excess of SO3 is used 6 5 . 17 SbF 5 + S0 3 • SbF1+(S03F) . (1.26) When this method is applied to other metal fluorides, complete reaction is found only for alkali and alkaline earth metal fluorides 2 » 6 6 > 6 7 . For example, the difluorides of nickel and copper, when subjected to S O 3 , yielded only about 20% conversions 6 8 . NiF 2 + 2S03 • Ni(S0 3F) 2 . (1.27) The reaction of CrF 5 with S0 3 did not give a fluorosulfate of Cr(V), but did provide a novel synthesis of S 20 6F 2 6 9 . CrF 5 + 5S03 v S 20 6F 2 + Cr(S0 3F) 3 . (1.28) A final obstacle in these reactions is the tendency for S0 3 to polymerize, giving rise to asbestos-1 ike materials of rather low v o l a t i l i t y , making product separation d i f f i c u l t and awkward. To summarize, the above three synthetic routes are of limited use in the synthesis of metal fluorosulfates. Furthermore, they do not provide a means to obtain metals in high oxidation states. I.E.4 OXIDATIVE FLUOROSULFONATION WITH X S O 3 F ; X=Ha1 OR S O 3 F This general scheme combines fluorosulfate addition together with oxidation, analogous to the direct reaction of a metal or its salts with a halogen. I.E.4.1 BIS(FLUOROSULFURYL) PEROXIDE S 20 6F 2, discovered by Cady and Dudley, is most eff i c i e n t l y prepared by the flow reaction of S0 3 with F 2 in the presence of 18 AgF2 as a catalyst 7 0 » 7 1 . 2S03 + F 2 • S 20 6F 2 . (1.29) It has a low dissociation energy of 22 kcal/mol, yielding S O 3 F radicals, and can be regarded as a pseudohalogen in this respect 7 2 5 7 3 : F0?S0-0S0?F >• 2-0S0?F . (1.30) Oxidative reactions with S 20 5F 2 can give a variety of fluoro-sulfate-containing compounds, depending on the choice of starting materials. A few il l u s t r a t i v e examples are given below: a) Oxidative addition of S03F MF3 + S 20 6F 2 • MF 3(S0 3F) 2 M=As 7 \ Sb 2 6 (1.31) CeF3 + JsS 20 6F2 CeF 3(S0 3F) ^ (1.32) b) Halogen displacement MClx + 2S 20 6F 2 • M(S0 3F) x + |C12 (1.33) M=Ga(III) 7 6 , Sn(IV) 3 3 . c) Metal oxidation I Mo + 3S 20 6F 2 y Mo0 2(S0 3F) 2 + 2S 20 5F 2 7 5_ (1.34) S 20 6F 2, being a nonionizing liquid, has very poor solvating properties towards ionic solids. Many metals are unreactive towards this strong oxidizing agent because of the rapid passiva-tion of the metal surfaces. On the other hand, S 20 6F 2 is easily transferred and handled in glass vacuum lines and i t can be obtained readily and in good yields by reaction (1.29). The solubility problem mentioned previously may be solved by the availability of a suitable solvent to increase the reactivity 19 of S 20 6F 2. I.E.4.2 HALOGEN (I) MONOFLUOROSULFATES Halogen (I) mono-fluorosulfates, formed by the direct reaction between the halogen and S 20 6F 2, Hal 2 + S 20 6F 2 >- 2HalS03F Hal=F, Cl, Br, I , (1.35) are themselves also very strong oxidizing agents and should therefore be very useful in the preparation of metal fluorosul-fates. Their relative effectiveness as the active reagents in such reactions is outlined below: a) Fluorine (I) monofluorosulfate FS03F is sometimes formed as a by-product during the synthe-sis of S 20 6F 2 by reaction (1.29) when fluorine is present in excess. It is an extremely powerful oxidizing agent, and reactions with i t usually lead to the simultaneous addition of fluorine and fluorosulfate, as in 2 6 : SbF 3 + FS03F • SbFjSOc-F) . (1.36) The major obstacle to the investigation of FS03F chemistry is it s unpredictable tendency to detonate. b) Chlorine (I) monofluorosulfate Like F S O 3 F , CISO3F is the only binary fluorosulfate of the halogen. It can be made only by a high temperature (125°C) reaction which takes a week to complete. It must also be stored and handled in metal apparatus since i t reacts with trace amounts of moisture in glass to give a red viscous liquid 7 7 . It has not 20 been used extensively in synthetic fluorosulfate chemistry also because of the ready availability of BrS03F which can substitute for ClS0 3F in most cases but does not have any of it s disadvan-tages. c) Bromine (I) monofluorosulfate BrS03F, although extremely hygroscopic like the other halogen fluorosul fates, has been found to be an extremely versatile oxidative fluorosulfonating reagent. This easily synthesized red liquid appears to be associated and supports the solvation of ionic species 3 8> 7 8. i t can be safely stored and handled in glass and can be d i s t i l l e d in vacuo at room temperature. BrS03F reacts with chlorides or bromides to give the corresponding fluorosulfates 7 9 according to: MCl n + nBrS03F • M(S0 3F) n + sBr 2 + §C1 2 • (1-37) When alkali metal chlorides are used, complexation takes place and salts of the type M(I)Br(S0 3F) 2 are formed 8 0> 8 1. in the few reported reactions involving noble metals, BrS03F was found to oxidize and dissolve the metals directly: M + nBrS03F y M(S0 3F) n + §Br 2 (1.38) M=Au, Pt 8 2 , Ag 8 3 , 8 \ n=3, 4, 1.33 . d) Iodine (I) monofluorosulfate ISO3F is expected to be only a mildly oxidizing and fluoro-sul fonating agent. Since i t is a solid which is d i f f i c u l t to prepare, i t has mainly been used for the study of poly- and inter-halogen cations 3 7>85,86 -21 In conclusion, for the purpose of this study, a l l the pre-viously described synthetic reagents have their advantages and disadvantages (either in terms of their physical or their chemical properties). It would be desirable to devise a general synthetic method which can be applied to specific compounds with only minor modifications needed. l.F TRANSITION METAL FLUOROSULFATES The chemistry of the 3d metals i s , in general, better explored and understood than that of the heavier transition elements. It is therefore not surprising that most of the fluoro-sulfates of the 3d elements are known, with the metals in the normal oxidation numbers of +2 and +3 5> 8 7. These are easily prepared by the solvolysis of the corresponding chloride or other salts in HS03F followed by f i l t r a t i o n . These fluorosulfates are insoluble in the acid, indicating high degrees of polymerization in the solid state. This is supported by results from vibration-al spectroscopy 6 2> 8 7. In addition, the oxyfluorosulfates V0(S03F) 3 8 8 and Cr0 2(S0 3F) 2 3 1 + have also been reported. These fluorosulfates are unsuitable as ansolvo-acids in HS03F. For the heavier elements, the binary fluorosulfates of Cd 8 7 and Hg 8 9 , prepared by: Cd(C 6H 5C00) 2 + 2HS03F • Cd(S0 3F) 2 + 2C6H5C00H (1.39) Hg + S 20 6F 2 • Hg(S0 3F) 2 (1.40) are also polymeric, with tridentate S0 3F groups 8 7 . Gold and platinum dissolve in BrS03F and adducts have been isolated from 22 the reaction mixtures 8 2 . Oxyfluorosulfates with the metals usually in the highest attainable oxidation states are known; examples are: NbO(S03F)3, TaO(S0 3F) 3, Re0 3(S0 3F), Re0 2(S0 3F) 3 8 8 , Mo0 2(S0 3F) 2 9 0 and W0(S03F)4 2 7 . With the high positive charge on the metals in these compounds, they should be expected to possess Lewis acid properties. Since these compounds have not been f u l l y characterized, a more detailed investigation of them and of related systems is also the intention of this study. l.G VIBRATIONAL MODES OF THE S03F GROUP The analysis of the vibrational spectrum of the S0 3F group can provide information regarding the type of coordination and the extent of interactions with the cations 9 1 . This is a distinct advantage a S03F group has over monatomic ligands such as F or Cl". A free SO^F" ion has C^v symmetry, but this can be reduced to C s or even Cj with a corresponding increase in the number of vibrational modes from 6, (3Aj + 3E),to 9, (6A' + 3A"), when one or two oxygen atoms are coordinated. Coordination of a l l three oxygen atoms or even including the fluorine atom in a tridentate or tetradentate manner will restore symmetry to the ligand and reduce the number of bands back to six, (3A^ + 3E). All of the vibrational modes of SO^F are both i . r . and Raman active. Hence, the vibrational spectra of SO^F groups can be divided into two main groups: 23 a) 6-band spectra produced from ionic, tridentate and tetra-dentate SO^F groups with C 3 v symmetry. b) 9-band spectra produced from mono- and bi-dentate SO^F groups which have C g or symmetry. Since there are three degenerate E modes when the fluoro-sulfate group has C 3 v symmetry, these are susceptible to splittings by site symmetry or by aspherical cations producing polarization effects. The splittings can be seen in the SO^ stretching region at -1200 to 1300 cm - 1. Within the two groups of 6- and 9-band spectra, a further distinction can be made based on the band positions. F i r s t l y , coordination of oxygen or fluorine will weaken and thus lower the stretching frequencies of the S-0 and S-F bonds. Secondly, electron withdrawal via the coordination of oxygen will strengthen the remaining S-0 and S-F bonds. This, according to Cruickshank and Webster 9 2 , is due to the increased back donation from the f i l l e d pir orbitals of oxygen or fluorine to the empty dir orbitals on sulfur now that the sulfur atom has become more positive. Hence the S-F stretching frequency can be used as a diagnosis to determine i f coordination of the oxygen atom is present. Invariably, the S-F stretch shifts to a higher frequency for covalently bound SO^F groups. Examples of the band patterns for the various types of coordination of the SO^F group are shown in Fig 1.3 and discussed subsequently: a) ionic KSO3F, being an ionic crystalline salt 9 3 , Fig 1.3 VIBRATIONAL BAND PATTERN FOR THE S03F GROUP 'SO 'SF °SO,F Ionic, C 3v Ionic (Perturbed), C 2v Monodentate, Cs Bidentate, C s Tridentate, C 3 V Tetradentate, C 3 V KSOoF 9 3 NOSO.F 9 4 KBr(S03F)1+ 3 5 (CH 3) 2Sn(S0 3F) 2 9 7 N i(S0 3 F ) 2 8 7 T i 3 C l 1 0 ( S 0 3 F ) 2 1 0 0 1500 1000 500 I 200 cm - l ro 25 gives a vibrational spectrum representative of ionic fluorosulfate with bands at 1285, 1079, 745, (594, 586), 570 and 407 cm"1. b) ionic perturbed With strongly polarizing and sometimes asymmetrical cations such as N0+ and N0 2 +, splittings of the E modes are observed. For example, v^, (S0 3 stretch), becomes s p l i t by about 35 cm - 1 in N0(S03F) 9 l +. c) covalent monodentate In Hal-S03F, the symmetry of the molecule is reduced to C"s, and the vibrational spectra show nine distinct bands. For example, in C1S03F, v(S-F) is increased to 830 cm-1 and v(S-O) are found at 1478, 1225 and 856 cm-1 9 5 . d) covalent bidentate As a bidentate ligand, the S0 3F group is usually found in the bridging mode between two metal ions, as in (CH 3) 2Sn(S0 3F) 2 9 6 ' 9 7 . For this type of coordination, v(S-O) are found at -1400, ~1150 and ~1070 cm - 1. While the example given here can be classified as containing bridging S03F groups spanning two identical metal ions, another type, which may be termed 'aniso-bidentate bridging' seem to occur in cases when the two metal ions have different charges. The spectra for such species, as in M(II) [M(IV)(S0 3F) 6] 8 1 +> 9 8 contain bands typical for both monodentate and bidentate S03F groups. e) covalent tridentate The 3d transition metal bis(fluorosulfates) are polymeric, each S03F ligand being coordinated to three different metal centres 8 7> 9 9. The symmetry of the anion is restored to C^. This type of coordination can be distinguished from the.ionic C q u situation by the high v(S-F) 26 of about 850 cm - 1 for these compounds. f) covalent tetradentate In the unique case involving Ti 3C1io(S0 3F) 2, the vibrational spectra for the compound indicate a C^v symmetry for the S03F group. v(S-F), however, is s i g n i f i -cantly lowered to 660 cm - 1, presenting evidence for the coordination of the fluorine atom as well 1 0 ° . Ionic fluorosulfates are not the only type that can show cation-anion interactions; in fact, most spectra of fluorosulfates have additional complexities caused by solid state splittings. Generally, vibrational spectra do provide useful information concerning the environment of the fluorosulfate group and the type of coordination. A number of diagnostic band positions together with their intensities normally allow a clear dis-tinction between monodentate and bidentate coordinations, the two most commonly encountered types. / 27 CHAPTER 2 EXPERIMENTAL 2.A INTRODUCTION This chapter will deal only with general experimental techniques and the sources of starting chemicals used in this study; specific syntheses will be described in the appropriate chapters. Since most of the compounds involved are extremely moisture sensitive, an environment excluding a i r , (or more specifically,water), had to be employed. Standard vacuum line techniques were used for the handling of volatile materials; a l l stop-cocks were of the greaseless teflon stem valve type from Kontes. Reactions were performed in one-part Pyrex reaction vials of about 30mL capacities, fitted with the teflon stop-cocks. Some typical glass apparatus used in this study is shown in Fig 2.1. When high gas pressures were anticipated in a reaction, 3mm thick-wall reactors were used in place of the standard lmm-wall reactors. In order to reduce the sources of contaminations, teflon coated stirring bars were used only when necessary. For the preparation of solutions for conductometric studies that required sti r r i n g , a glass-encapsulated stirring bar was used in a f l a t end reactor. In the course of the study, some specific situations required special modifications to the standard apparatus; a few are described in the appropriate chapters or shown in Fig 2.1. A standard reactor vacuum f i l t e r FIG 2.1 EXAMPLES OF GLASS APPARATUS 29 All reactions were monitored by weight differences in the reaction vials. The removal of volatile materials after a reaction had completed usually proceeded by evacuating f i r s t at room temperature or below, followed by a further evacuation at ~40°C. The latter step was necessary especially in cases involving HSO3F or BrSOsF, as both have extremely low vapor pressures at room temperature. Occasionally, the calculated yields from some reactions appeared to be low, with attack of the reaction vessel and the subsequent weight loss as S i t h e primary reason. In particular, high temperature reactions involving any fluorosulfate-containing reactants and extended reaction times were vulnerable to such an attack. Nevertheless, the reduction in the weights of the reactors were usually below lOmg and did not justify the use of less convenient vessels such as those made of fluoro-plastics (Kel-F) or metal. For reactions involving BrF 5 or fluorine at high pressure, Monel reactors were used. 2.B GENERAL EQUIPMENT  2.B.1 DRYBOX The handling of hygroscopic materials that had limited v o l a t i l i t i e s at room temperature necessitated the use of a dry box — a Vacuum Atmosphere Corporation Dry-Lab, Model HE-43-2 fitted with a Dri-Train Model HE-93-B recirculating unit. The inert atmosphere of dry nitrogen gas was constantly drawn through 30 a molecular sieve chamber to maintain its purity. 2.B.2 ELECTRICAL CONDUCTIVITY The conductivity c e l l , shown in Fig 2.2, is similar to the design of Barr et al 2 1 . This cell allowed the measurement on solutions with between ~7 to ~100mL of volumes. The platinum-black coating of the electrodes was renewed from time to time by electroplating from a solution of H 2PtCl 5 1 0 1 . The cell constants were determined using aqueous KC1 solutions of ~0.01M 1 0 2 and had values from (5.367 ± 0.007) to (5.353 ± 0.017) cm"1. The quantitative measurements of conductivity were carried out using a Wayne-Kerr Universal Bridge, Model B221A. The cell temperature of (25.00 ± 0.01)°C was maintained by immersion into an o i l bath of ~35L capacity, regulated by a Sargent Thermonitor, Model ST. It was noted that after about 5 minutes, most of the solutions studied did not show any further variations in their conductivities. The equilibration period for a l l solutions was about 10 minutes in the o i l bath after each addition of new materials to the c e l l . In the beginning of this study, the addition of solutes or solvent to the cell was by the use of weight droppers, shown in Fig 2.3. It had previously been noted that the introduction of small amounts of moisture into the cell during the insertion and removal of the dropper led to a small but consistent rise in the conductivity of pure HS03F of the order of 10 - 5 f r ^ m - 1 1 0 3 . 32 Because of the small quantities of solutes used in this study (~1 mmol), and the low molecular weight of water (18 g-mol - 1), i t was decided that this interference was not acceptable. Furthermore, the formation of HF and H 2 S0i t from the hydrolysis of HSO3F would introduce two basic solutes in HSO3F. Obviously, conductometric studies of acids in HS03F, the main objective of this study, would be adversely affected by this side reaction. The volumetric addition of solutions, although limited in accuracy by the buret (±0.02 mil), provided an alternative that excluded the exposure to moisture. Such a buret is illustrated in Fig 2.4. Designed to be f i l l e d in the Dry box, which limited i t s maximum capacity to ~20mL, the buret's accuracy was maximized by using large volumes of dilute solutions. The actual measurement of the density of the solution to be added did not have to be performed as the buret was used as a fractionating device only. As long as the i n i t i a l and final volumes and weights were known the intermediate quantities were calculated as a fraction of the total quantity. This also minimized errors due to the increase in the internal bore of the buret from etching, which would make any density measurements derived from other sources meaningless. This volumetric addition produced results that were better reproducible and less error prone than the method which monitors successive additions by weight. FIG 2.4 VOLUMETRIC DROPPER graduated buret 34 2.B.3 I.R. SPECTROMETRY Routine i . r . spectra at room temperature were recorded on either a Perkin Elmer, Model 457,or a Pye Unicam Model SP1100, Grating Spectrophotometer, providing low frequency limits of 250 and 400 cm - 1, respectively. Spectra were taken on thin solid films pressed between AgBr, AgCl or BaF2 windows with a trans-mission range down to ~250 cm - 1, ~400 cm - 1 or ~800 cm - 1, respectively. The high reactivity of the fluorosulfates studied precluded the use of mulling agents or other window materials. For solution spectra in HSO^F, thin teflon films were used to protect the windows, (only the 0-H region was of interest). For low temperature spectra at ~80K, a Perkin-Elmer Model 225 Grating Spectrophotometer with a low frequency limit of 200 cm - 1 was used. The cell was of the Hornig-Wagner design consisting of an evacuable Pyrex Dewar with a pair of Csl windows. The Dewar cooled a brass sample window holder directly. Volatile samples could be condensed onto a cooled Csl window and solid samples could be directly deposited. All spectra were calibrated with a polystyrene reference. 2.B.4 RAMAN SPECTROMETRY Raman spectra were usually obtained with a Spex Ramalog 5 spectrophotometer equipped with a Spectra Physics Model 164 argon ion laser and using the 514.5 nm exciting line. The samples were contained in melting point capillaries. For compounds that were 35 brown or red, the green laser was found to cause local decompo-sition. Therefore, for these compounds, Raman spectra of reduced J resolution were sometimes obtained using a Cary 81 spectrophoto-meter equipped with a Spectra Physics Model 125 He-Ne laser with a 632.8 nm exciting line. The samples were contained in 5 mm f l a t end Pyrex tubes. A low temperature cell for the Ramalog 5, shown in Fig 2.5, was developed to enable the recordingof spectra of compounds that were unstable to the laser at room temperature. The sample capillary was held in the brass sample holder (cooled to -80K) with grease. Generally, compared to spectra obtained at room temperature, better resolution was obtained. Shifts of bands to lower wavenumber, of the order of 3 to 5 cm - 1, were also noted. 2.B.5 ELECTRONIC SPECTROMETRY Diffuse reflectance spectra of powdered solids from 350 to 740 nm were obtained on a Bausch and Lomb Spectronic 600 Spectrophotometer. For absorption spectra on mulls (in C8Fi 7S0 2F or fluorolube o i l ) , or solutions in HS03F, a Cary 14 Spectrophoto-meter with a range from the near i . r . to 210 nm, was used. Standard s i l i c a solution cells used were of 1 or 10 mm path lengths. Solid mulls were pressed between quartz plates of 2.5 cm diameters and held inside a teflon sample holder. FIG 2.5 LOW TEMPERATURE RAMAN CELL inner part of Dewar brass sample holder quartz windows 37 2.B.6 MAGNETOCHEMISTRY -Magnetic susceptibilities were measured using a Gouy appara-tus described by Clark and O'Brien 1 0 5 . The calibration was performed using Hg[CoCSCN)^] 1 0 6 . The temperature of the sample was controlled by the rate of evaporation of liquid nitrogen around the chamber. Diamagnetic corrections were taken from Landolt-Bb'rnstein 1 0 7 and listed in Table 2.1. 2.B.7 NUCLEAR MAGNETIC RESONANCE XH — Using a Varian T-60 or a Varian EM-360, both operated at 60 MHz, spectra were recorded of solutions contained in 5 mm n.m.r. tubes and referenced to external TMS. 1 9 F — C.W. spectra were recorded on a Varian EM-360 operated at 56.45 MHz using 5 mm n.m.r. tubes. The lock signal was provided by either external CFC13 or internal H S O 3 F . F.T. spectra were recorded on a Varian XL-100 operated at 94.1 MHz locked on external D6-acetone. All 1 9F-spectra are referenced to CFC13 with shifts to lower f i e l d considered negative. 1 1 9Sn, 1 9 5 P t — F.T. spectra were obtained on a Briiker Spectro-spin 80 operated at 29.88 MHz for 1 1 9 S n and 21.72 MHz for 1 9 5 P t , 10 mm n.m.r. tubes were used, and the spectrometer was locked to external D6-acetone. No specific references for the chemical shifts were used. Table 2.1 DIAMAGNETIC SUSCEPTIBILITIES 1 0 7 (in cgs units) Au 3 + — 27xlO" 6 Cs+ 3 1 x l 0 - 6 I r 3 + 3 4 x l 0 " 6 2 8 x l 0 " 6 Pd 2 + 2 5 x l 0 - 6 Pd 3 + 2 2 x l 0 - 6 Pd 4 + I 8 x l 0 " 6 Pt^ + 2 8 x l 0 - 6 S n ^ 1 6 x l 0 " 6 SO3F- — 4 0 x l 0 " 6 (by analogy with S ( V " ) Pascal constants : Br 3 0 . 6 x l 0 - 6 , Cl 2 0 . 1 x l 0 " 6 , 0 — ~ - 4 x l 0 - 6 Table 2.2 CHEMICALS USED WITHOUT PURIFICATION Chemical Source Purity (%) Au, -20 mesh Alfa 99.995 Au 20 3 Alfa reagent grade Ge, -325 mesh Alfa 99.999 Ge02 MCB 99.999 Ir, -60 mesh Alfa 99.9 Mo, -200 mesh Alfa >99.9 Pd, -60 mesh Alfa 99.9 Pt, -60 mesh Alfa 99.9 PtCl 2 Strem 99.9 Sn, 20 mesh BDH 99.97 W, -100 mesh Alfa 99.98 39 2.C CHEMICALS 2.C.1 USED WITHOUT PURIFICATION These are mostly metal powder of high purity for which no purification was necessary or practical. They are listed in Table 2.2 along with their purities and their sources. 2.C.2 PURIFICATION REQUIRED a) The alkali metal chlorides were dried in a 120°C oven KC1 - 99.9% from MCB CsCl - 99.5% from BDH b) BaCI? was obtained by the dehydration of BaCl 2«2H 20 (99.0%, Analar, BDH) in a 120°C oven. c) Br? from Milinkroft (AR grade) was purified by storage in a Pyrex vessel containing P 20 5 to remove moisture and KBr to remove C l 2 . d) Cl ? from Matheson was passed through KMnOi+ to remove HC1 and bubbled through concentrated H2SO4 to remove moisture. e) SO? from Matheson, was liquidfied in a Pyrex vessel containing P 20 5 as the desiccant. f) N0C1 from Matheson, was liquidfied and stored in a Pyrex vessel with CaS04 as the desiccant. g) H S O 3 F , from Allied Chemical, was purified by double di s t i l l a t i o n s in a Pyrex apparatus at 1 atm. of dry nitrogen 2 1 . The constant boiling fraction at 163°C was either collected into a conductivity cell or into a large Pyrex storage vessel for 40 synthetic uses. The freshly d i s t i l l e d acid had a specific conductance of l . l - 1 . 3 x l 0 _ L f sT^nf h) H S O 3 C F 3 from Minnesota Mining and Manufacturing Company was d i s t i l l e d from concentrated H2S0t+ at about 15 torr of dry nitrogen. The fraction with a boiling point of 120°C was collected. 2.C.3 SYNTHESIS REQUIRED a) S o O p F ? was prepared by the reaction between S 0 3 and F 2 in a N 2 flow system catalyzed by AgF2 7 1 . The crude product was condensed as a liquid in Pyrex traps cooled by dry ice. Most of the F S O 3 F was allowed to escape by warming the product to room temperature. Any excess SO3 was removed by washing the crude product with concentrated H2S0i+ followed by the subsequent separa-tion of the two immiscible layers using a separation funnel. The product was then exacuated at dry ice temperature to remove the last traces of F S 0 3 F and then vacuum d i s t i l l e d into a large one-part Pyrex storage vessels stoppered with Kontes teflon stopcocks. b) BrS03_F was prepared by reacting Br 2 with a slight excess of S 20 6F 2 7 8 in a long stem Pyrex reactor. The excess of S 20 6F 2 was required to prevent the presence of unreacted Br 2. The crude product was pumped briefly to remove any oxygen formed and then stored. c) C I S O 3 F was prepared by the reaction of a mixture of C l 2 with S 20 6F 2 in a Monel reactor at 125°C for 5 days 7 7 . Excess 41 reactants were removed by evacuation at dry ice temperature. The product was stored in the monel reactor. d) CIO? was prepared from the reaction of NaC103 and H2SO4 in the presence of oxalic acid according to Brauer 1 1 8 . The product was used immediately for the preparation of C102S03F. e) C10?S03F was synthesized by d i s t i l l i n g a large excess of S 20 5F 2 into a thick wall reactor containing C10.2 at -198°C 1 0 9 . The reactor was allowed to warm up slowly to room temperature. The resulting 2-layer liquid was pumped in vacuo to remove the excess S 20 6F 2. 2.C.4 ELEMENTAL ANALYSES I These were performed by Analytische Laboratorien (formerly A. Bernhardt) Gummersbach, West Germany. 42 CHAPTER 3 PALLADIUM-FLUOROSULFATE 3.A INTRODUCTION In the periodic group of Ni-Pd-Pt, the chemistry of palla-dium has received the least amount of attention in the past. As is commonly found in transition metal chemistry, a gradual in-crease in the s t a b i l i t y of higher oxidation states of the metal is noted as one goes down the group n o . Therefore, nickel is usually found in the +2 oxidation state, Ni(IV) being the highest one observed; platinum, on the other hand, exhibits both the +2 and +4 oxidation states, Pt(VI) being the highest one known. Both the +2 and +4 oxidation states of palladium are also quite common, but i t is unlike platinum in that the divalent state is found to be more prevalent, examples for Pd(IV) being somewhat rare. The highest obtainable oxidation state of palladium is +5, and as is usually the case, i t is found in a fluoride, the recent-ly reported 0 2 +[PdF 6]", and possibly 0 2 + [ P d 2 F 1 1 ] " i n . While many organometal1ic compounds of palladium, like those of nickel and platinum, display oxidation states of 0 and +1, i t is inte-resting to note that Pd(III), and also Pt(III), are exceedingly rare, and only a few examples such as the dicarbollyl of Eti+N[Pd-( T T - 2 , 3 B 9 C 2 H I I ) 2 ] 1 1 2 have been reported. Generally, these com-pounds with palladium in the formal oxidation state of +3 are hard to make (electrochemical methods are often used), and require chelating or IT bonding ligands to delocalize the metal's charge. 43 In the coordination chemistry of Ni(II), the d 8 ion usually exists in either one of three configurations: octahedral, e.g. [Ni ( H 2 0 ) 6 ] 2 + ; square planar, e.g. [Ni (CN)i+]2"; or, tetrahedral, e.g. [NiCliJ 2-. Magnetochemistry can provide an excellent means of differentiating between the three geometries:- square planar complexes are diamagnetic while both octahedral and tetrahedral ones are paramagnetic, with magnetic moments that should corre-spond to two unpaired electrons. Because the tetrahedral config-uration has a 3 T : ground state, a large orbital contribution to its magnetic moment is expected, resulting in y e f f values being much higher than the spin only value of 2.83 yg. This can also lead to a complex temperature dependence of the magnetic suscep-t i b i l i t y , rather than the Curie-Weiss law behavior observed for octahedral compounds. For Pd(II) and Pt(II), diamagnetic square planar complexes are almost exclusively found. The increase in ligand f i e l d splitting for these two heavier elements of the group causes the dx2_y2 orbital to be extremely high in energy, i.e., antibonding. Thus, a square planar configuration, in which the strongly antibonding d x2_ v2 can remain vacant, is favored. Also due to the large A 0 , Pd(IV) and Pt(IV) are invariably octahedral, with a xA 2g electronic ground state. The palladium-fluorine system represents some rather inter-esting anomalies and will be briefly reviewed because similarities between the corresponding fluoro- and fluorosulfato- compounds are to be expected. 44 PdF2, formed by the reduction of the trifluoride by either SFi+ or SeF^ 1 1 3 , is the only paramagnetic binary Pd(II) compound. A magnetic moment of 1.84 U g , suggesting the presence of a magnetically concentrated system, is reported. The rutile structure of PdF2 1 0 \ with an octahedral coordination for Pd(II):, and almost linear Pd-F-Pd bonds, allows antiferromagnetic coupling via the superexchange mechanism 1 1 5 . Therefore, i t is not surprising to find that in mixed crystals of PdF2-ZnF2, the effective magnetic moment of Pd(II) increases with the amount of ZnF2 diluent present, and at infinite dilution, an extrapolated y e f f of ~3.2ug for PdF2 is obtained 1 1 6 . Paramagnetism is also found for some derivatives of PdF2, such as CsPdF3 1 1 7 , ZnPdFi+ 1 1 6 and Pd(II)M(IV)F 6, with M=Ge, Sn, Pd, Pt 1 0 8 ; but BaPdF4 is reported to be diamagnetic 1 1 6 . Although PdO is diamagnetic 1 1 9 , a compound which may be formulated as K 8Pd(II)Pd(IV)30n is paramagnetic, with a magnetic moment of 2.76yg for Pd(II) 1 2 ° . In contrast to the difluoride, the other palladium dihalides and their complexes are diamagnetic. For example, B-PdCl2, isomor-phous with the corresponding platinum chloride, consists of P d 6 C l i 2 clusters with square planar [PdCliJ units linked by chlorine-bridges 1 2 1 . While the other halogens form only the dihalides, fluorine, because of it s high oxidizing power, is capable of supporting two additional binary fluorides: Pd 2F 6 and PdFi+. PdF6, expected to be an even stronger oxidizing agent than PtF 6, has not been 45 successfully synthesized 1 2 2 . Palladium dibromide dissolves in BrF3 to form a compound of the composition of PdF 3«BrF 3 which, upon heating, decomposes to form the trifluoride 1 2 3 . The trifluoride is paramagnetic, and it s formulation as PdF3 gives a magnetic moment of 2 . 0 4 u g , suggesting a low spin d 7 electronic configuration for Pd(III) 1 2 4 . The absence of Jahn-Teller distortion in the solid state led Bartlett to invoke a mixed valency composition - Pd(II)Pd(IV)F 6, for the compound 1 0 8 . The recalculated magnetic moment of 2 . 8 8 y g for Pd 2F 6 is consistent with the presence of high spin d 8 Pd(II) and low spin d 6 Pd(IV), both in octahedral ligand fields. A recent neutron diffraction study shows Pd 2F 6 to be isomorphous with Pd(II)M(IV)F 6, with M=Ge, Sn, Pt 1 2 5 , and M(II)Pd(IV)F 6, with M=Mg, Ca, Zn, Cd 1 2 6 . Although the two types of palladium species in the solid state cannot be distinguished in terms of their ionic r a d i i , they can be found in two regular octahedral sites of different dimensions. The tetrafluoride can be obtained by the direct fluorination of the trifluoride at 300°C and 7 atm. 1 2 7 . PdF^ is diamagnetic 1 0 and a neutron diffraction study indicates octahedral coordination for Pd(IV) 1 2 8 . In view of the high oxidation state for palladium in PdFit, the low spin d 6 electronic configuration found is not unexpected. Diamagnetic complexes of the type of M 2[PdF 6], with M=K, Rb, Cs, can be readily obtained by the fluoridation of the complex chloride with either fluorine 1 2 9 or B r F 3 1 3 0 . 46 Both these Pd(IV)-containing fluorides are strong oxidizing agents, for example, they are capable of converting XeF2 to XeF^ 1 3 1 . This is consistent with the generally observed trend that Pd(IV) is usually more strongly oxidizing and more d i f f i c u l t to obtain than the corresponding Pt(IV) compounds. To summarize, the fluorides of palladium show some marked differences when compared to the other halides: a) higher oxidation states can be obtained in the fluoride system; and b) the difluoride and some of i t s derivatives deviate from the square planar geometry in favor of octahedral structures. Futhermore, the existence of [PdHalJ 2- and [PdHal 6] 2 _, with Hal=F, Cl, indicates halide ion acceptor properties of the palladium halides. The formation of [P d C l 6 ] 2 - complexes is interesting in that the parent compound -PdCl 4, has not been prepared. The investigation into the hitherto unknown palladium fluoro-sulfate system is expected to contribute further insights into both the halide and fluorosulfate systems 9 8> 1 3 2. 3.B EXPERIMENTAL 3.B.1 SYNTHESIS OF PALLADIUM BIS(FLUOROSULFATE)  3.B.1.1 OXIDATION OF PALLADIUM METAL WITH BrS03F Pd + 2BrS03F • Pd(S0 3F) 2 + Br 2 (3.1) In a typical reaction, palladium metal (508 mg, 4.77 mmol), was allowed to react with an excess of BrS03F(~5 mL) at ~110°C for 2 weeks. This slow reaction led to the formation of a purple powder in the mixture. The removal of a l l volatile materials 47 yielded Pd(S03F)2,(1.437 g, 4.72 mmol). Pd(S0 3F) 2 is a light purple, hygroscopic, polycrystalline solid. It appears to lack any solubility in either BrS03F or HS03F. When heated to above ~250°C, i t decomposes, with conden-sations of S0 3-like crystals forming on the cooler part of the melting point capillary, suggesting a decomposition scheme of: Pd(S0 3F) 2 • PdF2 + 2S03 (3.2) Analysis Pd S F Calculated 0/ 35.10 21.15 12.49 Found % 34.94 21.06 12.47 3.B.1.2 THERMOLYSIS OF Pd 2(S0 3F) 6 Pd 2(S0 3F) 6 2Pd(S0 3F) 2 + S 20 6F 2 (3.3) The decomposition of Pd 2(S0 3F) 6, (323 mg, 0.343 mmol), at ~160°C in vacuo yielded Pd(S0 3F) 2, (277 mg, calc'd 244 mg), detected as the sole solid product by its Raman spectrum. S 20 6F 2 was identified as the only volatile product (plus traces of SiFi+) by both i . r . and Raman spectroscopy. The reaction cannot be taken to completion because of the presence of small amounts of v undecomposed Pd 2(S0 3F) 6 on the cooler part of the reaction vial inaccessible by heating. 3.B.1.3 REDUCTION OF Pd 2(S0 3F) 6 WITH Br 2 Pd 2(S0 3F) 6 + Br 2 • 2Pd(S0 3F) 2 + 2BrS03F (3.4) Br 2, (-3 mL) was added to a sample of Pd 2(S0 3F) 6, (443 mg, 0.549 mmol), and the mixture was heated at 100°C for h hour. 48 The removal of a l l volatile material yielded Pd(S0 3F) 2, (339 mg, 1.113 mmol), identified by the i . r . spectrum of the purple solid obtained. 3.B.2 SYNTHESIS OF PALLADIUM TRIS(FLUOROSULFATE) 2Pd(S0 3F) 2 + S 20 6F 2 • Pd 2(S0 3F) 6 (3.5) An excess of S 20 6F 2(~5 mL) was added to Pd(S0 3F) 2, (398 mg, 1.31 mmol). An immediate reaction was evident from the gradual darkening of the solid. The mixture was then heated at ~80°C for 12 hours to insure a complete conversion. After the removal of a l l volatile materials, Pd 2(S0 3F) 6 was obtained, (5.18 mg, 0.642 mmol). Pd 2(S0 3F) 6 is a dark brown, hygroscopic powder. It displays no observable solubility in either S 20 6F 2 or HS03F. It is ther-mally stable up to ~180°C at atmospheric pressure, although decom-position occurs at the lower temperature of ~120°C in vacuo according to equation (3.3). Analysis Pd S F Calculated % 26.36 23.83 14.12 Found % 26.54 23.68 14.01 3.B.2.2 OXIDATION OF PALLADIUM METAL WITH S 20 6F 2/HS0 3F 2Pd + 3S 20 6F 2 MaL> Pd 2(S0 3F) 6 (3.6) A mixture of S 20 eF 2/HS0 3F, (~5 mL), was d i s t i l l e d into a reactor containing palladium metal (215 mg, 2.02 mmol). After heating the mixture at ~100°C for 3 days, a reddish-brown solution 49 containing dark brown solids was formed, and a l l metal had reacted. The removal of al l volatile materials yielded Pd 2(S0 3F) 6, (794 mg, 0.984 mmol). 3.B.3 SYNTHESIS OF HEXAKIS(FLUOROSULFATO)PALLADATE (IV) COMPLEXES  3.B.3.1 PREPARATION OF Cs 2 [Pd(S0 3F) 6] CsCl + HS03F • CsS03F + HC1 (3.7) 2CsS03F + Pd + S 20 6F 2 H S 03 F> Cs 2[Pd(S0 3F) 6] (3.8) HS03F (~3 mL) was added to a mixture of palladium metal, (100 mg, 0.940 mmol), and CsCl, (316 mg, 1.877 mmol). After the removal in vacuo of a l l of the HC1 formed from the i n i t i a l solvo-lysis of CsCl, a mixture of S 20 6F 2/HS0 3F, (~5 mL), was d i s t i l l e d into the reactor. An immediate reaction occurred at room temper-ature, with the color of the solution taking on a deep red color. Further heating of the reactants at ~100°C for 3 days led to the dissolving of a l l the metal powder. The removal of al l volatile materials yielded a solid which analyzed as Cs 2[Pd(S0 3F) 5], (902 mg, 0.933 mmol). The compound can also be made by the reaction of a 2:1 ratio of CsCl and palladium metal with ~5 mL of BrS03F. A typical reac-tion requires about 14 days at ~100°C. Cs 2[Pd(S0 3F) 6] is a dark red, hygroscopic, crystalline solid which is soluble in both BrS03F and HS03F. It decomposes at ~200°C. 50 Analysis Calculated % Found % Cs Pd F 27.50 11.01 11.79 27.25 10.85 11.67 3.B.3.2 PREPARATION OF Ba[Pd(S0 3F) 6] BaC1.2 + 2HS03F * Ba(S0 3F) 2 (3.9) Ba(S0 3F) 2 + Pd + 2S 20 6F 2 H S 0 3 F > Ba [Pd(S0 3F) 6] (3.10) In a manner analogous to the preparation of Cs 2[Pd(S0 3F) 6], a mixture of Ba(S0 3F) 2, formed from the solvolysis of BaCl 2, (310 mg, 1.490 mmol),in HS03F, and palladium metal, (158 mg, 1.490 mmol), was reacted with S 20 6F 2/HS0 3F, (~5 mL). An immediate reaction also occurred at room temperature, with the formation of a bright red precipitate. The reaction mixture was heated at ~90°C for 7 days as the reaction was quite slow due to the insol-u b i l i t y of the product. The removal of a l l volatile materials yielded Ba[Pd(S0 3F) 6], (1.233 g, 1.471 mmol). An attempt to react a stoichiometric mixture of BaCl 2 and palladium metal with BrS03F was unsuccessful as Ba(S0 3F) 2 was found to be insoluble in BrS03F. Ba[Pd(S0 3F) 6] is an orange-red, hygroscopic powder. It is slightly soluble in HS03F and is thermally stable up to ~200°C. Analysis Ba Pd S F Calculated % 16.39 12.70 22.95 13.60 Found % 16.24 12.64 22.89 13.82 51 3.B.3.3 PREPARATION OF (C1Q 2) 2[Pd(S0 3F) 6] 2C102S03F + Pd + 2S 20 6F 2 • (C10 2) 2[Pd(S0 3F) 6] (3.11) Palladium metal, (129 mg, 1.212 mmol), was reacted with a mixture of C102S03F (~0.5 mL) and S 20 6F 2 (~5 mL). After maintain-ing a reaction temperature of ~70°C for 3 days, dark red crystals were formed together with the disappearance of a l l the metallic reactant. The removal of a l l volatile materials at ~60°C yielded (C10 2) 2[Pd(S0 3F) 6], (1.013 g, 1.212 mmol). (C10 2) 2[Pd(S0 3F) 6],(749 mg, 0.896 mmol), was also prepared by the reaction of Pd 2(S0 3F) 6, (322 mg, 0.446 mmol), with a mixture of C102S03F, (-0.5 mL), and S 20 6F 2, (-2 mL), according to: 2C102S03F + y>d 2(S0 3F) 6 + %S 20 6F 2 • (C10 2) 2[Pd(S0 3F) 6] (3.12) The product is a very dark red, hygroscopic, crystalline solid. It is soluble in both HS03F and C102S03F, and melts with decomposition at ~200°C Analysis Cl Pd F Calculated % 8.48 12.73 13.64 Found % 8.60 12.82 13.68 3.B.3.4 PREPARATION OF (N0) 2[Pd(S0 3F) 6] 4N0C1 + Pd y (N0) 2[PdCl 4] + 2N0 (3.13) (NOhfPdClJ + 3S 20 6F 2 H S Q 3 F > (N0) 2[Pd(S0 3F) 6] + 2C12 (3.14) By the reaction of palladium metal, (110 mg, 1.034 mmol), with a large excess of N0C1, (~3 mL), at room temperature for 52 12 hours, (N0)2[PdCli»], (316 mg, 1.025 mmol), was obtained. This intermediate, insoluble in N0C1, was identified by i t s i . r . spec-trum which has two bands at 2150, (vN-0), and 332 cm - 1, (vPdCl^). This solid was then reacted with a mixture of S 20 6F 2/HS0 3F, (~6 mL), at 80°C for 1 hour. The removal of a l l volatile materials at ~70°C yielded (N0) 2[Pd(S0 3F) 6], (780 mg, 1.025 mmol). It can also be obtained by the oxidation of (N0) 2[PdCliJ by BrS03F, according to: (N0)2[PdCl,J + 6BrS03F • (N0) 2[Pd(S0 3F) 6] + 2C12 + 3Br 2 , (3.15) although a large excess of BrS03F has to be used for the forma-tion of a pure product. (N0) 2[Pd(S0 3F) 6] is a bright red, hygroscopic, crystalline solid. It is soluble in HS03F and less so in BrS03F. It melts with decomposition at ~200°C. Analysis N Pd S F Calculated % 3.68 13.99 25.29 14.99 Found % 3.44 13.81 25.11 15.17 3.B.4 SYNTHESIS OF PALLADIUM(II) HEXAKIS(FLUOROSULFATO) METALLATES(IV)  3.B.4.1 PREPARATION OF Pd Pt(S0 3F) 6 Pd(S0 3F) 2 + Pt(S03FK H S ° 3 F > Pd[Pt(S0 3F) 6] (3.16) Platinum metal, (224 mg, 1.148 mmol), was converted into Pt(S03F)i+ by oxidation with a mixture of S 20 6F 2/HS0 3F. Pd(S0 3F) 2, 53 formed by the reaction of palladium metal, (121 mg, 1.128 mmol), with a mixture of S206F2/HS03F followed by the subsequent reduc-tion of Pd 2(S0 3F) 6 with Br 2, was added to a solution of the Pt(S0 3F ) i + in HS03F. No reaction was apparent at room temperature. After heating the mixture at ~50°C for 3 days, a green homogeneous precipitate formed. The solid obtained after a f i l t r a t i o n of the suspension was washed with HS03F and dried at ~80°C in vacuo. Pd[Pt(S0 3F) 6] is a light green, hygroscopic powder. It appears to be insoluble in HS03F and is thermally stable up to ~200°C. Analysis Pd S F Calculated % 11.88 21.80 12.72 Found % 11.62 21.74 12.88 3.B.4.2 PREPARATION OF Pd Sn(S0 3F) 6 Pd(S0 3F) 2 + Sn(S03FK ^M^. Pd[Sn(S0 3F) 6] (3.17) Pd[Sn(S0 3F) 6] was prepared in a reaction similar to that for Pd[Pt(S0 3F) 6]. Pd(S0 3F) 2 (539 mg, 1.770 mmol) was mixed with a stoichiometric amount of SnfSOsF)^ formed from the reaction of tin metal (210 mg, 1.770 mmol) with S 20 5F 2/HS0 3F. No reaction occurred when HS03F was added to the mixture; when the reactor was heated at ~70°C, a slow reaction, evident from the gradual clearing of the i n i t i a l l y cloudy solution, took place. After 7 days in this condition, a l l the purple Pd(S0 3F) 2 solids had changed into a blue powder, and the supernatant became clear. 54 The solids were f i l t e r e d , washed with HS03F and dried at ~70°C in vacuo. Pd[Sn(S0 3F) 6] is a light blue, hygroscopic powder. It is apparently insoluble in HS03F and decomposes at ~250°C to a light purple solid. Analysis Pd Sn S F Calculated % 12.98 14.48 23.47 13.41 Found % 13.06 14.61 23.31 13.72 J 3.B.5 SYNTHESIS OF TETRAKIS(FLU0R0SULFAT0)PALLADATE(II) COMPLEXES  3.B.5.1 REDUCTION OF Cs 2 [Pd(S0 3F) 6] BY Br 2/S0 2 Cs 2[Pd(S0 3F) 6] + 2S02 B r 2 > Cs 2[Pd(S0 3F) t }] + 2S 30 8F 2 (3.18) Cs 2[Pd(S0 3F) 6], was prepared by the reaction of palladium metal, (100 mg, 0.940 mmol), and CsCl, (316 mg, 1.877 mmol), with S 20 6F 2/HS0 3F. After a l l volatile materials were removed in vacuo, S02(~4 mL) and B r 2 ( ~ l mL) were added to the solid. The reaction mixture was stirred at room temperature for 1 day. The removal of al l volatile materials gave a light brown material which analyzed as Cs 2[Pd(S0 3 F K ], (721 mg, 938 mmol). Cs 2[Pd(S0 3F) 4] is a light brown hygroscopic powder which melts with decomposition at ~195°C. It is slightly soluble in HS03F, but deposits Pd(S0 3F) 2 immediately; the fil t e r e d solid shows only the i . r . spectrum of Pd(S0 3F) 2. Analysis Cs Pd S F Calculated % 34.59 13.85 16.69 9.89 Found % 34.68 13.66 16.54 10.01 55 3.B.5.2 REDUCTION OF Ba [Pd(S0 3F) 6] BY Br 2/SQ 2 Ba[Pd(S0 3F) 6] + 2S02 B r ? > BaIPd(S0 3FK] + 2S 30 8F 2 (3.19) In a manner analogous to the preparation of Cs 2[Pd(S0 3F) t |], Ba [Pd(S0 3F) 6] (196 mg, 0.941 mmol), prepared from the reaction of BaCl 2 and palladium metal (100 mg, 0.940 mmol) with S 20 6F 2/HS0 3F } was reduced with a mixture of Br 2/S0 2 to give a compound suggesting the formulation of Ba[Pd(S0 3F) l t], (602 mg, 0.941 mmol). Ba [Pd(S03F)iJ is a very light brown hygroscopic powder. It decomposes at ~200°C to a purple solid. It is insoluble in HS03F. (N0) 2[Pd(S0 3F) 4] can also be prepared likewise. Analysis : Ba Pd S F Calculated % 21.46 16.63 20.04 11.88 Found % 21.22 16.82 20.03 11.92 56 3.C DISCUSSION 3.C.1 SYNTHESIS AND GENERAL DISCUSSION  3.C..1.1 BINARY FLUOROSULFATES Pd(S0 3F)2 can be prepared by one of three methods: a) the oxidation of palladium metal with BrS03F, b) the reduction of Pd 2(S0 3F) 6 with Br 2, and c) the controlled pyrolysis of Pd 2(S0 3F) 6. Reaction a) has two precedents: the interaction of gold and platinum with BrS03F is reported to yield the corresponding Au(III) and Pt(IV) fluorosulfates 8 2 . However, palladium metal is found to be much less reactive towards BrS0 3F, and long reaction times at high reaction temperatures of ~160°C are required. Further differences can be found in the solu b i l i t i e s of the three products in the reaction media: while both Au(S0 3F) 3 and PtlSOsF)^ dissolve readily to form rather stable intermediates, Pd(S0 3F) 2 precipi-tates from the reaction mixture and no evidence of complex forma-tion i s evident. The high thermal s t a b i l i t y , and i t s seeming insolubility in both BrS0 3F and HS03F even at temperatures up to ~160°C, are consistent with Pd(S0 3F) 2 having a polymeric structure. Although reaction a) provided the only direct synthesis of Pd(S0 3F) 2, the reduction of Pd 2(S0 3F) 6 by Br 2 was subsequently found to be the preferred method because of the ready availability of the trisfluorosulfate. Also, the product from reaction a) was found to contain traces of a brown substance, most lik e l y 57 Pd 2(S0 3F) 6, which had to be decomposed by evacuation at ~120°C. The observation seems to imply that BrS03F is capable of the oxida-tion of palladium metal to the tetravalent state, and this is indeed found in the preparation of soluble [Pd(S0 3F) 6] 2 _ complexes. The successful reduction of Pd 2(S0 3F) 6 to Pd(S0 3F) 2 by Br 2, commonly used as an oxidizing agent, indicates the strong oxidiz-ing a b i l i t y of Pd(IV), in keeping with general observations regard-ing the chemistry of palladium 1 1 9 , i 3 1 . The purple color of Pd(S0 3F) 2 provides an interesting com-parison with other binary Pd(II) compounds: PdF2 is also purple in color, but the other dihalides are brown. This may suggest an octahedral coordination for Pd(II) in the bis(fluorosulfate), as in the difluoride. The facile preparation of Pd 2(S0 3F) 6 by the oxidation of the metal with a solution of S 20 6F 2 in HS03F illustrates the impor-tance of the presence of a suitable solvating medium in these reactions, even though the product may only be slightly soluble in HS03F. Without the addition of HS03F, the oxidation of the metal hardly proceed at a l l , even at ~160°C at a high pressure of S 20 6F 2 in a Monel reactor. An alternate formulation of the tris(fluorosulfate) as Pd(S0 3F) 3 would imply the presence of Pd(III), a very uncommon oxidation state and in contrast to the trif l u o r i d e , which is the mixed valency compound of Pd(II) Pd(IV)F 6 1 0 8 . It is very likely that such an analogy exists in the fluorosulfate system, and 58 experimental evidence consistent with the formulation will be discussed subsequently. 3.C.1.2 FLU0R0SULFATO COMPLEXES OF PALLADIUM Both the binary fluorosulfates of palladium are insoluble in HSO3F, thus precluding their uses as superacids in H S 0 3F. However, the formulation of the tris(fluorosulfate) as a mixed valency compound suggests the existence of a [Pd(IV)(S0 3F) 6] 2~ species and a Pd(II) [M(IV)(S0 3F) 6] species, much like [PdF 6] 2 _ and Pd[M(IV)F6] in the fluoride system. Therefore, complexation reactions were attempted. A general preparative route to complexes containing [Pd(S0 3F) 6] 2 - is the oxidation of the metal by S 20 6F 2 / H S 0 3F in the presence of a basic fluorosulfate. This method has also been used successfully in the preparation of complexes containing other metals in high oxidation state. The reaction involving C102S03F did not require HSO3F as a solvent because the former is a super-cooled liquid at room temperature and can act as an ionizing solvent 1 3 4 . The synthesis of (N0) 2[Pd(S0 3F) 6] by the reaction of (N0) 2[PdCl 4] with S 20 sF 2 illustrates the oxidation of Pd(II) to Pd(IV) with the replacement of Cl" by S03F". The moderate reaction temperature of ~80°C minimized the formation of CIO2SO3F (by the oxidation of C l 2 with S 20 6F 2 in the presence of glass) although a small amount of a C10 2S0 3F-like v red liquid was 59 detected as a by-product. The complete removal of this contami-nant in vacuo at ~70°C is indicated by the agreement in the ana-lyt i c a l result for the N0+-complex. Since both the C10 2 +- and the NO+-complexes of [Pd(S0 3F) 6] z~ have nearly identical thermal s t a b i l i t i e s , the preferential formation of the latter may be due to: a) the greater v o l a t i l i t y of C102S03F, and/or b) the higher lattice energy expected for (N0) 2[Pd(S0 3F) 6] due to the smaller size of the cation. Although only four examples of complexes containing [Pd(S0 3F) 6] 2~ are reported here;, other complexes containing M(I) and M(11) cations are expected to be obtainable using the general methods described here. The synthesis of Pd(II) [M(IV)(S0 3F) 6] involved the addition of a slight excess of M(S0 3F) 4, where M=Pt, Sn, to Pd(S0 3F) 2 followed by digestion in HS03F and subsequent f i l t r a t i o n of the insoluble products. The excess is needed to ensure a stoichio-metric ratio in the end products because of d i f f i c u l t i e s encoun-tered in weighing the fluorosulfates accurately in the dry box. The rather low solubility of Sn^C^F)^ in HS03F 3 3 dictated a rather long reaction time. However, in the reaction involving the very soluble PtCSOsF)^, only a very slight increase in the reaction rate was noted; this provided more evidence for the extremely high degree of polymerization and insolubility for Pd(S0 3F) 2. The colors of these compounds (light blue for 60 Pd [Sn(S0 3F) 6] and light green for Pd[Pt(S0 3F) 6] ), also departed from the dark colors usually associated with square planar Pd(II) compounds. Thus the Pd(II) ion in these two fluorosulfato-complexes may be octahedrally coordinated. In the preceeding reaction, Pd(S0 3F) 2 acts formally as a S0 3F" donor, i.e., a base. Attempts were also made to involve the bis(fluorosulfate) in anionic complex formation; i n i t i a l l y , these attempts were unsuccessful: a) CsS03F was found not to undergo any direct complexation reaction with Pd(S0 3F) 2, even after heating the mixture in HS03F at ~160°C. The solid obtained after f i l t r a t i o n gave only an i . r . spectrum of unreacted Pd(S0 3F) 2. b) The attempted solvolysis of (N0) 2 [PdClJ in HS03F at ~160°C produced an inhomogeneous mixture of white and brown solids (the starting material is brown). The i . r . spectrum taken of this mixture showed only bands due to N0S03F. The weight of the final product suggests that the solvolysis reaction must have occurred according to: (N0) 2[PdCl 1 +] + 2HS03F • 2N0S03F + PdCl 2 + 2HC1 . (3.20) Subsequently, a sample of PdCl 2 was indeed found not to react with HS03F at ~160°C, most likely a result of its high degree of poly-merization. c) The solvolysis of (N0) 2[PdCliJ in BrS0 3F also led to the oxidation of Pd(II) to Pd(IV). This implies that the formation of Pd(S0 3F) 2 from the reaction of BrS03F with the metal may be 61 due to the insolubility of Pd(S0 3F) 2 in the reaction medium rather than as a consequence of the limited oxidizing a b i l i t y of BrS03F. d) Because of the a b i l i t y of Br 2 to reduce Pd 2(S0 3F) 6 to Pd(S0 3F) 2 at elevated temperature, the reduction of Cs 2[Pd(S0 3F) 6], Ba [Pd(S0 3F) 6] and (N0)2[Pd(S03F)6] with Br 2 were attempted. But even at a reaction temperature of ~120°C, only (N0) 2[Pd(S0 3F) 6] showed any signs of a reaction taking place, and the product is a mixture of unreacted (N0) 2[Pd(S0 3F) 6], and some white/purple solids which could be N0S03F and Pd(S0 3F) 2. e) As the lack of a suitable solvent could have led to the demise of the above reduction attempt, a mixture of Br2/HS03F was heated with Ba[Pd(S0 3F) 6] at ~70°C for 1 day. The removal of volatile materials yielded a light brown product with a weight corresponding to the composition of BaPd(S03F)i+. An i . r . spectrum of the product, however, showed the presence of some bands due to Pd(S0 3F) 2, and therefore, although a reduction had taken place, some decomposition of the product had also occurred. Since the reduction of a fluorosulfate with Br 2 usually leads to the formation of BrS03F, i f i t can be removed from an equ i l i -brium, the reaction may be encouraged to proceed more effectively. S0 2 was used as such a BrS03F scrubber. The successful reduction of [Pd(S0 3F) 5] 2 _ complexes with a Br 2/S0 2 mixture at room tempera-ture probably proceeds via the following scheme: [Pd(S0 3F) 6]2- + Br 2 ^ > [Pd(S0 3FK]2- + 2BrS03F (3.21) 62 BrS03F + S0 2 >• Br 2 + S 30 8F 2 . (3.22) S0 2 alone, like Br 2 by i t s e l f , did not react with Cs 2[Pd(S0 3F) 6] at temperatures up to ~100°C. Equation (3.21) appears to be shifted to the l e f t , since the soluble [Pd(S0 3F) 6] 2 _ complexes can be synthesized in a Br 2/BrS0 3F mixture, By the careful removal of a l l the S0 2 and most of the Br 2, the slightly less volatile S 30 8F 2 was identified by i t s gas i . r . spectrum. This procedure employing Br 2/S0 2 as the reduction couple could be applied to the preparation of Pd(S0 3F) 2 from Pd 2(S0 3F) 6 and conceivably other systems in which the presence of BrS03F may be unacceptable. The faster reaction time at a lower reaction temperature and the presence of relatively inert and volatile by-products are the major advantages of this method. The difference in reactivity of the ionic [Pd(S0 3F) 6] 2" complexes as compared to Pd 2(S0 3F) 6 may be attributed to their different thermal s t a b i l i t i e s . While Pd 2(S0 3F) 6 decomposes at ~120°C in vacuo to give S 20 6F 2, the ionic complexes are stable up to ~200°C, and their decomposition products contain S0 2F 2, iden-t i f i e d by its i . r . spectrum. S0 2F 2 is also found in the gaseous thermolysis products of Au(S0 3F) 3 and Pt(S0 3F) J +, and is a result of complete breakdown of the compound rather than via step-wise reduction. 3.C.1.3 ATTEMPTED SYNTHESIS OF Pd(S0 3F) t t The synthesis of [Pd(S0 3F) 6] 2~ strongly suggests that the 63 binary compound, PdCSOsF)^, may be obtainable as well. Although solid PdCSOsF)^ could not be isolated, some evidence for i t s v existence in strongly oxidizing solution has been obtained: a) Pd 2(S0 3F) 6 is virtually insoluble in HS03F, but the clear solution containing the solid immediately took on a dark red color when S 20 6F 2 was added. The u.v. spectrum of the solution is identical to that produced by (N0) 2[Pd(S0 3F) 6] in HS03F, with a Xmax at 320 nm. As expected, the solubility of PdCSOsF)^ in such a mixture increased with increasing proportion of HSO3F (visually monitored), but even at a S 20 6F 2/HS0 3F ratio of less than 1/20, the solubility of Pd 2(S0 3F) 6 is less than a few mg per -20 mL. b) Using a Soxhlettype extraction device made of sintered-glass, the only product obtained after 21 days of extraction at ~70°C was Pd 2(S0 3F) 6, formed slowly in the filtered liquid. c) The oxidation of palladium metal by S 20 5F 2/HS0 3F from ~25°C to ~100°C did not lead to the formation of PdCSOsF)^. In order to minimize the decomposition of Pd(S0 3Fk at elevated temperature in vacuo, the removal of S 20 6F 2/HS0 3F was performed at room temp-erature (a very long process). Nevertheless the solid isolated afterwards was invariably Pd 2(S0 3F) 6. d) The attempted high pressure oxidation of Pd 2(S0 3F) 6 at 200°C for 7 days with an excess of S 20 6F 2 in a Monel reactor did not lead to a further oxidation of palladium. To rationalize a l l the above observations, some speculations concerning the palladium-fluorosulfate system can be made: 64 a) There are many precedents where high oxidation states may be obtained in anionic complexes when the parent compounds are not known. Some pertinent examples are the [PdCl6] 2" complexes and 0 2 +[PdF 6]~, with neither PdCl^ nor PdF5 obtainable so far. b) The decrease in decomposition temperatures from Pd(S03F)2 (250°C) to Pd 2(S0 3F) 6 (120°C) indicates a marked drop in the thermal stability in the series. Pd(S0 3F)i t, i f obtainable, may not be very stable at room temperature, c) The equilibrium: Pd2(S03F)6+ S 2 0 6 F 2 2Pd(S03FK , (3.23) is expected to be strongly dependent on the relative concentra-tions of HS03F and S 2 0 6 F 2 . In order to obtain a high concentra-tion of the presumably soluble Pd(S03F)i+ in HS03F, a low S 2 0 6 F 2 / HS03F ratio is required because S 2 0 6 F 2 , being a covalent l iquid, decreases the ionizing ability of the solvent. This in turn would almost certainly drive the equilibrium back to the left as the oxidizing strength of the solvent mixture is decreased. As an added disadvantage, the subsequent removal of both S 2 0 6 F 2 and HS03F at the end of the reaction favors the formation of the insoluble and more thermally stable Pd 2(S0 3F) 6. d) The observed formation of S 2 0 6 F 2 , presumably via S03F« radicals, in the pyrolysis of Pd 2(S0 3F) 6 is very unusual in terms of the thermal decomposition paths of metal fluorosulfates 6 2 . Only a few other examples can be found: Xe(S03F)2, XeF(S03F) 1 3 5 » 1 3 6 and Ag(S03F)2 8 k ; and all of these are extremely strong oxidizing 65 agents. Furthermore, although the [Pd(S0 3F)g] z ~ complexes are very stable entities, they are extremely corrosive materials capable of oxidizing Br 2 to BrS03F and producing the most serious attacks on AgCl i . r . windows of a l l the solid compounds examined in this study. All these observations point to an extremely strong oxidizing a b i l i t y of Pd(IV), and.the energy required for i t s form-ation must be recovered from sources such as increased lattice energy for the ionic complexes,and polymerization for Pd 2(S0 3F) 6. In summary, the inability to obtain pure Pd(S0 3F) i + may be attributed to i t s limited thermal s t a b i l i t y , the lack of solubil-ity of Pd 2(S0 3F) 6 in HS03F, and the high oxidizing power of Pd(IV). 3.C.2 VIBRATIONAL SPECTRA  3.C.2.1 Pd(S0 3F) 2 The i . r . and Raman data for Pd(S0 3F) 2 are listed in Table 3.1 and compared to the reported i . r . frequencies for Ni(S0 3F) 2 8 7 . The spectra are assigned in terms of a C^v symmetry for the S0 3F group except for the two low frequency bands, For Ni(S0 3F) 2, the 290 cm"1 band has been assigned to lattice modes 8 7 ; for Pd(S0 3F) 2, the very weak band at 354 cm - 1 could be due to impurities as i t was not present in the i . r . The position of the S-F stretch at ~860 cm - 1 implies a tridentate coordination mode for the anions rather than a true S03F" ion. Thus, Pd(S0 3F) 2 is isostructural TABLE 3.1: VIBRATIONAL FREQUENCIES OF Pd(S03F)2 Pd(S03F) 2 Ni(S0 3F) 2 8 7 Assignment R IR IR 1231 m 1240 s,b 1262 vs v*(E) un ms 1090 ms 1120 s vi(Ai) 869 m 860 s 859 s «2(Ai) 612 w 610 s 619 s v 5(E) 552 w 565 m 568 m v3(Aij 430 m 421 m 422 m v 6(E) 354 vw impurity Latti ce 290 m mode FIG.3.1 PROPOSED STRUCTURE OF Pd(S0 3F) 2 1 3 7 Metal O © O F 67 with a general class of bis(fluorosulfate)s which includes those of Ni, Mg, Ca, Zn, Cd, Hg, Fe and Co 8 7>". This highly poly-meric structure is also consistent with the observed physical properties of Pd(S03F)2, discussed in the previous section. The absence of splittings for the three degenerate E modes suggests a regular octahedral environment for Pd(II), in contrast to the extensive splittings observed in the spectra for Cu(S03F)2 and Mn(S0 3F) 2 6 2 , 8 7 » " , which implies a distortion of the octahedral coordination spheres for the M(II) cations in these compounds. The proposed structure of Pd(S0 3F) 2, consistent also with i t s magnetic and electronic spectral information, is shown in Fig. 3.1 1 3 7 . 3.C.2.2 [Pd(IV)($0 3F)6l 2~ AND Pd(II)[M(IV)(S0 3F) 6] Both i . r . and Raman spectra of the fluorosulfato complexes containing Pd(IV) were not easily obtainable for a number of reasons. The oxidizing a b i l i t y of these compounds limited the use of i . r . window materials to BaF2, with a transmission range down to only about 800 cm"1. The dark colours of these compounds also interfered with the recording of Raman spectra at room temp-erature when the 514.4 nm line of the argon ion laser was used. Thus, sample cooling down to ~80K had to be employed in order to reduce laser induced decomposition. Alternatively, Raman spectra of rather low resolution were obtained on another spectrometer equipped with a He-Ne laser (623.8 nm). 68 For the two Pd[M(IV)(S03F)6] complexes, with M = Pt, Sn, i .r . spectra down to 450 cm - 1 were obtained (AgCl plates). For Pd[Sn(S03F)6], because of the presence of strong fluoresence, a Raman spectrum could not be obtained. The Raman spectra for Pd 2(S0 3F) 6 and Pd[Pt(S03F)6] are shown in Fig. 3.2 and Fig.3.3, respectively. The observed vibrational frequencies for all the [Pd(S03F)6]2~ complexes and the two Pd[M(IV)(S03F)6] compounds are listed in Table 3.2 together with the literature values for K2[Sn(S03F)6] 3 2 . With the exception of Pd[Sn(S03F)6], the Raman frequencies are reported, since they are generally of better resolution and extend to lower frequen-cies. The corresponding i.r. spectra show no major discrepancies in most of the band positions. Except for a few additional features which will be discussed later, the spectra of these hexakis(fluorosulfato). complexes correlate well with each other and the main bands can be attributed to monodentate fluorosulfate groups in an anionic environment. A common, tentative assignment for these octahedral complexes is also attempted in Table 3.2. As noted previously for the [Sn(S03F)G]2" cases 3 2 , band proliferation, especially in the S03F stretching region (-800 cm"1) is found. This is due, presumably, to solid state splittings and, probably more importantly, to extensive vibrational couplings suggesting, in turn, strong covalent bondings to the central M(IV) atoms. This is particularly true for the [Pd(S03F)6]2" 70 I—! (/I — c ' E CC r~>JO 1—I •X3 < 'r—i — l ^ 5 c o " '—' 1/1 •— XJ e a. u •—- < O o — — E TJ U a. •—< i — i > CM < O T3 ^ stretch stretch ! 1 s~ 1 s „ m is; •g * g 55= X J E > X E s g B Is i ! 1293 1240 1215 > 5 g is 1290 1235 1209 , i 15 CO O .C •o CU E .O qj .O > CO o o t i * SS ii % iJ ^- E E 41 i/l ^ * E C " £-S O CM fwo O t o o * ' Q o VI £ E Em » t/i Pi ori «i E > *0 "» E Lf) CO ° J^J ^ CO CO ifl ifl w I E X X E ^ * 1 E E E " >(=5 E ? > > J E ? J > E _ O r^ . O n O " - 0 1 ' " T M O ,— o c o tn f i <— cri tn OJ — i— n o o w oo co l O i / i u i ^ r ' r ^ ' t v i c v i JZ x s x E E > 1 E > > E > e ^ 5 ?>x E > X E S ^ E o =5 at a. E X O •a O 01 * J o "S GLJ cn u "O a» c a. * 71 derivatives, in which strong Raman bands at "1240 cm - 1 (vS0 2 asym) and ~1020 cm - 1 (vS-OPd) have weaker side bands at ~1210 and ~980 cm"1, respectively, with the intensities reversed in the i . r . spectra. The spectra for the three Pd(II) containing compounds have bands at ~1170 to 1190 and ~1040 to 1050 cm"1 that are not present in the spectra of the other complexes, including Ba [Pd(S0 3F)6l• Bands in these regions, also reported for Ag[Pt(S0 3F) 6] and Ag[Sn(S0 3F) 6] 8 \ are usually associated with bidentate bridging fluorosulfate groups 3 3 . Their occurrence in the spectra of these particular compounds may be due to the strong polarizing a b i l i t i e s of the divalent transition metal cations. Ba 2 + , although being a divalent cation, is not expected to have as great a polarizing effect on the fluorosulfate groups because i t lacks the necessary d-orbitals. Further evidence in support of the coordination of Pd(II) will be presented in subsequent sections. This type of coordination, shown in Fig. 3.4, in which the fluorosulfate ligands are bonded primarily to M(IV) and coordinated weakly to M(II), can be con-sidered as bridging anisobidentate. Hence, the ionic formulation of these compounds as consisting of discrete M(II) and [M( IV)(S0 3F) 6] 2" units may be somewhat misleading. This highly polymerized structure is also inferred from the noted lack of solubility of these complexes, (Ba [Pd(S0 3F) 6] is slightly soluble). Strong Raman bands at ~620 to 640, ~400 to 460 and ~270 to 72 FIG.3.4 ANISOBIDENTATE BRIDGING MODE FOR S03F 280 cm - 1, observed for a l l the anionic complexes are not assign-able to absorptions due to the fluorosulfate groups alone. Their positions appear to show a slight dependence on the nature of both the M(IV) species and the counter-cations. These bands, occasionally s p l i t into doublets, are most likely due to M(IV)-0 vibrations, possibly coupled with ligand deformation modes. The' evidence for strong M(IV)-0 bonds leading also to the coupling of S-0 stretching modes has been discussed earlier. For the complexes with polyatomic heterocations, additional bands due to the cations are observed and they compare favorably with previously reported data. For the N0+ complex, the N-0 stretch is found at 2330 cm"1 in the i . r . ; in (N0) 2[Sn(S0 3F) 6], the same vibration is found at 2325 cm"1 3 2 . Bands due to C10 2 + are found at 1293 cm"1 (split into 1300 and 1290 cm"1 for the 73 stronger i . r . band) and 1050 cm"1, consistent with literature values 3 2> 1 3 8. The bending mode for C10 2 +, expected to be only weakly Raman active, and usually found at ~520 cm"1, is not observable in this case. To summarize, the apparent spectral similarities indicate strong structural similarities in these compounds. The formula-tion of the tris(fluorosulfate) as the ternary compound, Pd(II) [Pd(IV)(S03F)e] is supported by its vibrational spectra. The strong vibrational coupling between the fluorosulfate groups and the [M(IV)06] skeleton confirms the covalent character of the M(IV)-0 bond. 3.C.2.3 [Pd(S0 3FK] 2-Because of the light brown color of the samples containing [Pd(S0 3F) i + ] 2", no Raman spectrum was obtainable. The i . r . fre-quencies are listed in Table 3.3. A good correlation can be made between the spectra of [Pd(S0 3F)i t ] 2 _ and those from the presum-ably square planar, [Au(S0 3F)i+]", suggesting structural similari-ties. The difference in the positions of a few diagnostic bands can be interpreted as being caused by the replacement of Au(III) by Pd(II). It has been hypothesized that the weakening of an S-0 bond due to the coordination of the oxygen leads to the streng-thening of the remaining S-F and S-0 bonds 9 2 . The lowering of the S-F stretching frequency by about 20 cm - 1 in the Pd(II) compound is significant, and must be attributed to a decrease in 74 TABLE 3.3: I.R. FREQUENCIES FOR (NO)2-. Cs 2- AND Ba- [Pd(S0 3F) l t] [Pd(S0 3F) u] (N0) 2[Pd(S0 3FK] -2330 vw Cs 2[Pd(S0 3FK] Cs[Au(S0 3FK] Assignment v N-0 1340 vs 1350 vs 1340 vs 1440 vs V 7 asymS03 1200 vs 1200 vs 1200 vs 1210 vs V i symS03 1030 vs 1020 vs 1000 vs 930 vs vi+ asym SO 3 800 s 800 s 795 s -820 s v 2 S-F 640 s 660 s 655 s,sh 645 s 675 s v 3 S03F rock + v MO sym 580 s 580 s 575 s 585 s v 5 S03F def 550 s 560 s 555 s 550 s v 8 S03F def 450 m -420 m -420 w 460 m v 9 S03F rock + v MO asym 75 covalent nature of the Pd-0 bond. This can also be illustrated by the decrease in the splitting in the asymmetric SO3 stretch (vi4 and v 7 ) for [Pd (S0 3F) 4] 2". (For [Pd ( S 0 3F) k] 2 _ , v7-v t t=~340 cm"1; for [Au^OgF)^]"", v7-vi+=~470 cm"1). This seems to indicate that the three S-0 bonds are more similar in the [Pd(S03F)iJ 2" species than the ones in [Au ( S 0 3F ) i t ]also not unexpected. The colors of these Pd(II) complexes, together with evidence from vibrational spectroscopy, seem to indicate a square planar environment for the Pd(II). Magnetic susceptibility measurements, to be discussed later, support such a conclusion. 3.C.3 ELECTRONIC SPECTRA The Pd(II) containing compounds of Pd(S0 3F) 2, Pd[Sn(S0 3F) 6] and Pd[Pt(S0 3F) 6] are purple, blue and green, respectively, quite in contrast to the usually brown colors of Pd(II) compounds with Pd(II) in a square planar environment. Similar blue or purple colors are also found for PdF2, CsPdF3 1 0 7 , ZnPdF^ 1 1 6 , and Pd[M(IV)F6] 1 0 3 , with M = Ge, Sn, Pt, a l l with Pd(II) in an octa-hedral environment. However, a detailed analysis of their elec-tronic spectra is not reported. It is therefore not surprising that the magnetic results and the vibrational spectra also supported a regular octahedral environment for Pd(II) in Pd(S0 3F) 2 and Pd[M(IV)(S0 3F) 6]. 76 Pd(S0 3F) 2 shows a three-band electronic spectrum that can be assigned as due to d-d transitions. Although no extinction coefficients to support the assignment were obtainable because of the insolubility of a l l the compounds in H S O 3 F , i t appears that a l l the bands are of about the same intensities and the and B values obtained from such an assignment are very reasonable. The electronic ground state for a d 8 ion in an octahedral f i e l d is 3A2g> and three d-d transitions are expected. The energy level diagram for a d 8 ion in an octahedral f i e l d is shown in Fig 3.5. The band positions of the observed electronic spectra and the calculated ligand f i e l d parameters of Pd(S0 3F) 2, Pd[Sn(S0 3F) 6], Pd[Pt(S0 3F) 6] and Ni(S0 3F) 2 6 1 are listed in Table 3.4. There is generally a good correspondence between spectra obtained from mulls and those obtained by diffuse reflec-tance. In the platinum compound, v 3 is not observed because the strong charge transfer band due to Pt(IV), with X m a x at 245 nm and e of ~1.5xl0 4 M^cm-1, extends well into the visible region. This further substantiates.the premise that the remaining bands are due to the d-d transitions with lower intensities. The good agreement in band positions for the spectra for the three compounds suggests a similar environment for Pd(II) and an 3A 2g ground state in a l l cases. Although no splitting of any of the bands is observable, i t does not confirm or deny the existence of distortion in the coordination sphere as the bands are quite broad. In any case, no distortion is evident from the vibrational 77 8 FIG. 3.5: ENERGY LEVEL DIAGRAM FOR A d ION IN AN 0. FIELD 1 i i v3 v 2 1 1 lV F ) vj = lODq v 2 = %(15B + 30Dq) - J*[(15B - lODq)2 + 128-lODq]^  v 3 = Js(15B + 30Dq) + %[(15B - lODq) 2 + ^B-lODq] 3 5 TABLE 3.4: ELECTRONIC TRANSITIONS AND LIGAND FIELD PARAMETERS  FOR SOME PALLADIUM (II) COMPOUNDS AND Ni(S07Fh ("in c m - 1 ) Pd[Sn(S0 3F) 6] Pd[Pt(S0 3F) 6] Pd(S0 3F) 2 Ni(SC 3F) 11,050 11,400 11,770 v 2 16,950 16,950 17,400 26,850 26,480* 27,030 Dq 1,105 1,140 1,177 734 B 717 642 633 905 B/B° 0.863 0.774 0.763 0.838 C a l c u l a t e d value _ l Diffuse r e f l e c t a n c e values f o r v 2 are 16,890 cm f o r Pd[Sn(S0 3F) 6J and 16,810 cm" 1 f o r Pd[Pt(S0 3F) 6] 77 8 F I G . 3 . 5 : E N E R G Y L E V E L D I A G R A M F O R A d I O N I N AN 0. F I E L D < i 1 i ^3 v 2 vj = lODq v 2 = % ( 1 5 B + 30Dq) - ^ [ ( 1 5 B - lODq) 2 + IZB-lODq]*5 v 3 = % ( 1 5 B + 30Dq) + Js[(15B - lODq) 2 + IZB-IODqf2 T A B L E 3.4: E L E C T R O N I C T R A N S I T I O N S A N D L I G A N D F I E L D P A R A M E T E R S  F O R S O M E P A L L A D I U M ( I I ) COMPOUNDS A N D N H S O T F ) ? ("in cm - 1) Pd[Sn(S0 3F) 6] Pd[Pt(S0 3F) 6] Pd(S0 3F) 2 Ni(SC 3F) ^1 11,050 11,400 11,770 v 2 16,950 16,950 17,400 V3 26,850 26,480* 27,030 Dq 1,105 1,140 1,177 734 B 717 642 633 905 B / B ° 0.863 0.774 0.763 0.838 * calculated value , Diffused reflectance values for v 2 are 16,890 cm for Pd[Sn(S0 3F) 6] and 16,810 cm"1 for Pd[Pt(S0 3F) 6] 78 analysis of the compounds. By using Tanabe-Sugano diagrams as suggested by Lever 1 3 9 , the positions of a l l the bands are confirmed, and v 3 for Pd[Pt(S0 3F) 6] is calculated using this method. The octahedral splitting 10 Dq, and the interelec-tronic repulsion term B, are also obtained. The increase in Dq and the decrease in 3, (defined as the ratio of B over B°, the free ion value), in going from Ni(II), (3d 8), to Pd(II), (4d 8), are not unexpected in view of the higher nuclear charge and more spacially diffused 4d orbitals for palladium 1 1 0 . It appears that within the series of Pd(S0 3F) 2, Pd[Pt(S0 3F) 6] and Pd[Sn(S0 3F) 6], 10 gradually decreases while 3 increases, with B approaching the B° value of-830 cm"1 l k l . The presence of the M(IV) species, with their higher cationic charge, is expected to compete favorably for the coordination of the S0 3F ligands. A lessening of Pd-S03F interaction is expected to decrease 10 D and also increase the interelectronic repulsion as the orbitals are more confined to the metal. For the two complexes, the one that contains Sn(IV) has the lowest 10 D and the highest 3 value. This cannot be due to structural dissimilarities, because this was not noted in the vibrational spectra, and must be a t t r i -buted to a greater a b i l i t y of Sn(IV) to bond to the S0 3F groups. This has also been observed in the following two complexes Ag(II)[M(IV)(S0 3F) 6], with M = Pt, Sn 8 h . (N0) 2[Pd(S0 3F) 4], Cs 2[Pd(S0 3F) 1 +] and Ba [Pd(S03F)1+] are light brown solids, and the bands of the diffuse reflectance spectra 79 of the latter two compounds are compared with those of K 2[PdCli t] in Table 3.5. The energy level diagram for a d 8 ion in a square planar ligand f i e l d is shown in Fig-3.6. As can be seen, a good comparison can be made between the spectra for the two types of compounds, suggesting the presence of species with similar elec-tronic configuration. Strong absorptions are found for K o J P d C l i J at 278 and 223 nm l k 2 , and are assigned to L —*• M charge transfer bands. For [PdfSOsF)^] 2", such absorptions were not observed as they are beyond the range of the instrument, but an extremely strong background was observed at the high energy end of the spectrum, giving rise to quite poor resolution for the two bands at ~450 nm. An assignment of the obtained spectra by analogy with K2[PdCli+] gives Ai=d5,400 cm"1, A2=4600 cm" , and A3^2200 cm"1 A i is equivalent to 10 i f i t can be assumed, (from a crystal f i e l d point of view at least), that a square planar structure is derived from the gradual elongation of an octahedron, and that the energy difference between d(x 2-y 2) and d(xy) remains 10 D 1 4 3 . M A comparison to a A x of ~16,700 cm - 1 for K 2 [ P d C l i J , is reasonable on the basis that the fluorosulfate group is generally known to produce a weaker ligand f i e l d than a chloro group 6 1 . Some generalizations concerning the relative a b i l i t y of a S O 3 F group to coordinate to a metal ion depending on how the other two oxygen atoms are bonded can be made. For the Pd(II)-fluorosulfate system, three types of coordination of a S03F group are evident from results obtained in this study: 80 FIG. 3.6: d-ORBITAL ENERGY LEVEL FOR A SQUARE PLANAR COMPLEX 1 1 + 2 i i i A 2 < i A 3 \ > 2 2 d(x -y ) d(xy) d(z 2) d(xz), d(yz) v 2 = A i + A 2 v 3 = A x + A 2 + A 3 TABLE 3.5: ELECTRONIC SPECTRA AND LIGAND FIELD PARAMETERS FOR [Pd(S0 3F)J 2- (in cm"1) Cs 2[Pd(S0 3F) l t] Ba[Pd(S0 3F) i +] [PdCiiJ -15,400 w 15,400 w 16,700 2 - l<t2 1 1 v i . Ai - A 2 l 3 l 3 v 2 . A, - B, I s 1 3 -3. A l g + E l g Charge Transfer Charge Transfer •20,000 s.sh 22,200 s -20,000 s.sh -22,700 s 21,500 23,300 36,000 44,900 A l A 2 A 3 15,400 4,600 2,200 15*400 4,600 2,500 16,700 4,800 1,800 81 a) monodentate, as in [ P d ( S 0 3 F ) i J 2 _ , b) aniso-bidentate bridging between Pd(II) and a M(IV) cation, as in Pd[M(IV)(S0 3F) 6],and c) bridging tridentate to three equivalent Pd(II) in Pd(S0 3F) 2. Most li k e l y a result of less competition for ligand electrons from another metal centre, a monodentate S0 3F produces the strongest ligand f i e l d . The withdrawal of electron density from a S03F group due to i t s bonding to a highly charged M(IV) cation can drastically reduce the ligands' coordinating a b i l i t y to Pd(II). In fact, the reduction can make the S03F ligand in the former co-ordination mode bind more weakly to Pd(II) than does a tridentate S03F group. By extrapolation, a bidentate S03F group bridging two Pd(II) should produce a ligand f i e l d splitting somewhat be-tween the ones present in [Pd(S0 3F ) i t ] 2 _ and Pd(S0 3F) 2. Similar observations have been made for a number of transition metal t r i -fluorides, MF3, with M = T i , V, Cr, Mn, Fe and Co, and their mixed ternary salts of the general formula of M'2M"[MF6]. The fluorine atoms in the binary fluorides are shared by two M(III) cations while in the [MF 5] 3 _ complexes, the fluoride ligands are essentially coordinated to M(III) with weak interactions towards M", an alkali metal ion. An increase in A of ~1% is noted in going from MF3 to [MF 6] 3" l k h . For the [Pd(S0 3F) 6] 2~ containing compounds, less informative electronic spectra were obtained. Pd 2(S0 3F )6 is a dark brown solid, and i t s u.v.-vis spectrum consists of a very intense, 82 featureless band which extends over most of the spectrum. This must be due to charge transfer because a) the high cationic charge of Pd(IV) enhances L —> M transfer, and b) the mixed valency nature of the compound which can give rise to M —• M1 transfer. As mentioned, the solution of (N0) 2[Pd(S0 3F) 6] in HS0 3F shows a single absorption peak at 320 nm with an extinction co-efficient of 1.3x10^ M_1cm"1; charge transfer orgin of the L —*- M type is also suggested here. 3.C.4 MAGNETIC SUSCEPTIBILITY As mentioned in the introduction, most palladium compounds, regardless of the oxidation state of palladium, are diamagnetic. This is because a) both Pd(II) and Pd(IV) are even-electron species, b) Pd(IV) is invariably octahedral low spin with high 10 Dq values, and c) Pd(II) is usually in a square planar,and thus diamagnetic, environment. Octahedral Pd(II), d 8, is expected to be paramagnetic, with two unpaired electrons corresponding to a 3A 2g ground state. Therefore, a direct orbital contribution to the magnetic moment 83 is not expected, but through spin-orbital coupling, which is expected to be quite large for 2nd and 3rd series transition metals, the observed magnetic moment should be in excess of the spin-only value of 2.83 y^. 3.C.4.1 PARAMAGNETIC Pd(II) FLUOROSULFATES Paramagnetism is found for Pd(S0 3F) 2, Pd 2(S0 3F) 6, Pd[Pt(S03-F) 6] and Pd[Sn(S0 3F) 6], a l l with magnetic moments indicative of two unpaired electrons for Pd(II). As mentioned, the electronic spectra for these compounds have shown that a 3A 2g ground state can best be assigned to Pd(II) in these compounds. The magnetic susceptibility of such a species can be represented by the f o l -lowing equation, according to Figgis K0 , 8kfN6 = s.o. f l _ 8k£x 0 ) + 8 k 2NB 2 ( 3 24) XA XA U 10 Dn j 10 Dr K i'^> where x A S'°" = spin only susceptibility, k = electron delocalization factor, X Q = spin-orbit coupling constant = -1,12 for d 8, 8 = Bohr Magneton, N = Avogardro's Number 10 Dq = Ligand Field Splitting. The f i r s t correction term involving x 0, is due to the spin-orbital coupling, and the second (and much smaller) term arises from 2nd order Zeeman Effect and is usually called Temperature Independent Paramagnetism (T.I.P.). K, the spin-orbital coupling 84 parameter, has been estimated as 1460 cm - 1 for Pd(II) 1 4 5 . Therefore, with x o neqative, the observed magnetic susceptibility should be higher than the spin only value. The magnetic susceptibilities of the four paramagnetic Pd(II) compounds are listed in Tables 3.6 to 3.9 and their Curie-Weiss Law dependence is represented diagrammatically in Fig 3.7. A comparison of these results with the corresponding fluoride system is made in Table 3.10. Having obtained the ligand f i e l d s p litting parameter, D, from the electronic spectra of some of H these compounds, i t is possible to estimate the electron derea-lization factor, k, and also to verify the value of the spin orbital coupling constant X . As can be seen, the four paramagnetic compounds a l l obey the Curie-Weiss Law in the temperature range of the investigation with very small Weiss constants as extrapolated from ~80 K. This suggests the presence of magnetically dilute systems in these cases, 1n contrast to the antiferromagnetic couplings observed in the fluoride system 104,108,116. n - j s noi expected that such couplings via the superexchange mechanism 1 1 5 , l t t 6 can occur ef-fectively through the polyatomic fluorosulfate ligands. The experimentally obtained magnetic susceptibilities and moments are very close to the ones calculated from the Curie-Weiss Law rela-tionship. Small temperature dependence of the susceptibilities as a result of positive Weiss constants, can be decreased by applying T.I.P. to the two complexes containing Sn(IV) and Pt(IV). 85 TABLE 3.6 MAGNETIC PROPERTIES OF Pd(S0 3F) 2 T c xm "Xm l / x m (calc) V f u e f f (calc) V f ' 103 1. 462x10" 2 68.40 67.43 3.47 3.50 3.45 125 1. 163x10" 2 85.18 83.32 3.43 3.46 3.41 150 9. 678x10" 3 102.2 101.4 3.43 3.44 3.40 183 8. 119x10" 3 121.7 125.2 3.47 3.42 3.44 207 7. 083x10" 3 139.1 142.5 3.45 3.41 3.42 225 6. 313x10" 3 155.8 155.5 3.40 3.40 3.36 250 5. 655x10" 3 173.6 173.6 3.39 3.39 3.35 274 5. 067x10" 3 193.3 190.9 3.37 3.39 3.32 299 . 4. 668x10" 3 209.5 209.0 3.38 3.38 3.33 K cgs cgs cgs yB yB yB l / x m C (calc) = (T - 10)/1.38 TIP = 140xl0"6 cgs units k - 0.89 TABLE 3.7 MAGNETIC PROPERTIES OF Pd 2(S0 3F) 6 T . c xm l / x mC (calc) y e f f v e f f (calc) V f T 334 4.419x10" 3 226.3 224.3 3.44 3.45 3.38 302 4.926x10" 3 203.0 202.2 3.45 3.46 3.40 279 5.410x10" 3 184.8 186.3 3.48 3.46 3.43 253 5.987x10" 3 167.0 168.3 3.48 3.47 3.44 229 6.616x10" 3 151.1 151.7 3.48 3.48 3.44 204 7.472x10" 3 133.8 134.4 3.49 3.48 3.46 179 8.621x10" 3 116.0 117.1 3.51 3.50 3.48 153 9.985x10" 3 100.2 99.12 3.50 3.51 3.47 131 1.185x10" 2 84.37 83.91 3.52 3.53 3.50 107 1.468x10" 2 68.13 67.30 3.54 3.57 3.53 K cgs cgs cgs yB VB u B u l / x m C (calc) = (T - 10)/1.45 TIP - 150xl0"6 cgs units k - 0.92 to 0.95 86 TABLE 3.8 MAGNETIC PROPERTIES OF Pd[Pt(S0 3F) 6] T xm l / x mC (calc) veff P e f f (calc) y e f f ' 89 1. 4.47x10" 2 69.09 69.82 3.21 3.19 3.20 119 1. 106x10" 2 90.41 90.95 3.24 3.24 3.23 142 9. 350x10" 3 107.0 107.1 3.26 3.26 3.24 167 7. 975x10" 3 125.4 124.8 3.26 3.27 3.24 193 6. 895x10" 3 145.0 143.1 3.26 3.29 3.23 222 6. 122x10" 3 163.3 163.5 3.30 3.30 3.27 246 5. 544x10" 3 180.4 180.4 3.30 3.30 3.27 278 4. 934x10" 3 202.7 202.9 3.31 3.31 3.27 302 4. 565x10" 3 219.1 219.8 3.32 3.32 3.28 K cgs cgs cgs yB yB yB T l / x m C (calc) = (T + 10)/1.42 TIP = 121x10"6cgs units k ~ 0.81 TABLE 3.9 MAGNETIC PROPERTIES OF Pd[Sn(S0 3F) 6] T c xm l / x mC (calc) p e f f w e f f (calc) y e f f T 80 1. 829x10" 2 54.69 55.41 3.42 3.40 3.40 107 1. 407x10" 2 71.09 72.09 3.47 3.45 3.45 129 1. 171x10" 2 85.38 85.68 3.48 3.47 3.45 154 9. 833x10" 3 101.7 101.1 3.48 3.49 3.45 178 8. 522x10" 3 117.3 116.0 3.48 3.50 3.45 203 7. 564x10" 3 132.2 131.4 3.50 3.52 3.46 229 6. 784x10" 3 147.4 147.5 3.53 3.52 3.48 252 6. 121x10" 3 163.4 161.7 3.51 3.53 3.46 280 5. ,609x10" 3 178.3 179.0 3.54 3.54 3.49 306 5. .172x10" 3 193.3 195.0 3.56 3.54 3.49 K cgs cgs cgs yB yB y B u 1/x m C (calc) = (T + 10)/1.62 TIP = 185x10"6 cgs unit k - 0.99 87 FIG 3.7 CURIE-WEISS PLOT FOR Pd(II)-FLURQUSULFATE DERIVATIVES 0 100 200 300 T(K) 88 TABLE 3.10: MAGNETIC PROPERTIES OF SOME Pd(II) COMPOUNDS AT ROOM TEMPERATURE 3 c ™ x m y eff temperature dependence Ref. PdF 2 1.28 1.75 antiferromagnetic coupling 108,151 Pd[PdF6] 3.31 2.81 e = -28K Pd[PtF 6] 3.10 2.72 e = -IK II Pd[GeF6] 3.34 2.82 e = -31K Pd[SnF 5] 3.37 2.83 e = -28K n Pd(S0 3F) 2 4.80 3.38 X m = 1.38/(T-10) Pd[Pd(S0 3F) 6] 5.01 3.46 x c = i.45/(T-10) Pd[Pt(S0 3F) 6] 4.61 3.32 = 1.42/(T+10) Pd[Sn(S0 3F) 6] 5.26 3.54 X^ = 1.62/(7+10) 89 For the two compounds with negative Weiss constants, no improve-ments could be expected from the T.I.P. adjustments, but s t i l l , the deviations are very small. No great significance should be attached to these small Weiss constants because they are extrapo-lated over a temperature range of close to 100 K. Magnetic orderings, should they occur, cannot be inferred from these measurements; ferromagnetic orderings have been found for Pd 2F 6 and Pd[PtF 6] below 13 K and 21 K, respectively l l f 7. However, the similarity in the magnetic moments for a l l the four fluorosulfate-containing compounds indicates a similar environment for Pd(II) in a l l cases. It also shows, conclusively, that the tris(fluorosulfate) should be formulated as containing high spin, paramagnetic Pd(II), and low spin, diamagnetic, Pd(IV), in a mixed-valency configuration. At f i r s t , the magnetic moments of these fluorosulfates appear to be much higher than the spin only value for Pd(II); however, an investigation into the isoelectronic nickel (II) system shows that even Ni(II), with a spin orbit coupling para-meter of only ~630 cm - 1 l t + 5, can have magnetic moments as high as 3.5 yg. Some representative y e f f values at 298 K are listed in Table 3.11. As mentioned, although the magnetic moment of PdF2 is lowered by antiferromagnetic exchange, at infinite dilution of the fluoride by ZnF2, an extrapolated moment of 3.2 yg is ob-tained 1 1 6 , quite close to the values for the fluorosulfate-90 Table 3.11 MAGNETIC PROPERTIES OF SOME Ni(II) COMPOUNDS Ni(S0 3F) 2 yeff 3.49 Temperature Dependence Curie-Weiss Reference 148 (NH^)2NiFlt«2H20 3.24 Curie-Weiss, 6 = -13 K 149 NiF 2-4H 20 3.16 Curie-Weiss, e = -19 K 149 LiNiF 4 3.19 Curie-Weiss, e ~ 0 K 149 Ni F 2 2.85 Antiferromagnetic 11+9,150 NiF2/MgF2, 5.48 moUNi 3.04 Curie-Weiss, e ~ -17 K 150 NiCl 2 2.91 Ferromagnetic 150 NiCl 2/MgCl 2, 3.88 moUNi 3.20 Curie-Weiss, e ~ 6 K 150 containing compounds reported here. The increase -in the y e f f at room temperature in the order of Pd(S0 3F) 2 - Pd[Pt(S0 3F) 6] < Pd 2(S0 3F) 6 < Pd Sn(S0 3F) 6 parallels a similar trend observed in the fluoride system 108,151. Refer-ring to equation (3.24), the only factors expected to have a dependence on individual Pd(II) compounds are 10 D the l i g -and f i e l d s p l i t t i n g , and k the electron delocalization factor which leads to an effective reduction of A from the free ion value, x 0. Although no ligand f i e l d parameters can be obtained from the electronic spectrum of Pd 2(S0 3F) 6, the similarity of it s magnetic moment to those of the other compounds ju s t i f i e s using the known 10 D^'s for the three compounds to arrive at an e s t i -mate for k for Pd(II) in Pd 2(S0 3F) 6. The values of k obtained appear to increase in a similar trend as does y e f f , approaching the free ion value of 1 for the Sn(IV) compound. However, these differences which are obtained from small deviations in the magnetic measurements from Curie-Weiss Law 91 should not be overemphasized quantitatively since: a) they are derived from values that test the limit of accuracy of the Gouy method, and b) a purely Russell-Saunders coupling scheme, although assumed to apply in the calculations, may not accurately describe the behavior in heavy metal ions 1 1 + 1. Nevertheless, the magnetic result from these Pd(II) compounds confirms the conclusions arrived at from electronic and vibra-tional spectroscopy. The Sn(IV) complex has the smallest 10 D^ , and both p and k are almost equal to 1, suggesting that of the four compounds, i t most closely approaches a free-ion environment for Pd(II). An octahedral coordination for Pd(II), evident from vibrational spectroscopy, is also substantiated by these measure-ments. 3.C.4.2 DIAMAGNETIC FLUOROSULFATO COMPLEXES OF PALLADIUM As expected, Pd(IV) was found to be diamagnetic, suggesting a xAig ground state. For [Pd(S0 3F ) i J 2", diamagnetism was also observed, indicating a square planar coordination sphere the type usually encountered for P'd(II). The magnetic susceptibility results are listed in Table 3.12. Table 3.12 MAGNETIC SUSCEPTIBILITIES FOR DIAMAGNETIC PALLADIUM  FLUOROSULFATES (in c.g.s. units) xg xm E x d i a Ba[Pd(S0 3F) 6] -3.15±0.27 x IO"7 -264±23 x IO"6 -202 x IO - 6 Cs 2[Pd(S0 3Fk] -2.78±0.24 x K T 7 -214±18 x IO - 6 -247 x IO - 6 92 3.C.5 SOLUTION STUDIES IN HS03F Due to the observed lack of any appreciable solubility of most of the fluorosulfate-containing palladium compounds in H S O 3 F , solution studies in these systems were rather limited. Although Cs 2[Pd(S0 3F ) iJ and (N0) 2[Pd(S0 3F)i t] are soluble in HS03F, the resulting solutions are very unstable, with a precipi-tation of Pd(S0 3F) 2 forming almost immediately upon the addition of the compounds. Pd(S0 3F) 2 was identified by i . r . spectroscopy as the only solid l e f t after a f i l t r a t i o n of the mixture. Pre-sumably, the high lattice energy of the extensively polymerized Pd(S0 3F) 2 favored i t s formation and subsequent removal from the equilibrium of: [Pd(S0 3Fk] 2- Pd(S0 3F) 2 + 2S03F~ . (3.25) This is also consistent with the observation that no reversal of equation (3.25) takes place when alkali metal fluorosulfates are mixed with Pd(S0 3F) 2. In order to provide further evidence for the existence of [Pd(S0 3F) 6] 2 _ as an ionic entity, electrical conductivity mea-surements for solutions of Cs 2[Pd(S0 3F) 6] in HS03F were underta-ken. The results, shown in Fig 3.8, agree quite well with the literature values for an analogous compound, K2 [Sn(S0 3F) 6] 32. The basic dissociation for the anions [M(S0 3F) 6] 2 _, where M = Pd, Pt, Ir, Sn, will be discussed in detail in Chapters 5 and 9. An ionization scheme, at least in a f i r s t approximation, can be FIG.3.8 CONDUCTIVITY OF Cs 2[Pd(S0 3F) 6] IN H S O 3 F 94 represented by: Cs 2[Pd(S0 3F) 6] !«2tf 2 C s | s o , v ) + [ P d ( S 0 , F ) 6 ] 2 f s o l y ) (3.26) TABLE 3.13 CONDUCTIVITY OF Cs7[Pd(SO F) 1 IN HSO F_ (interpolated) 3 6 3 m K 0 .0000 1.220 0 .0020 7.275 0 .0040 11.45 0 .0060 15.02 0 .0080 17.91 0 .0100 20.48 0 .0120 23.00 0 .0140 25.14 mol.kg" l O - ^ c m " 1 units 95 3.D CONCLUSION The study into the fluorosulfates of palladium yielded a number of new compounds with the metal in the +2 and the +4 oxi-dation states. Two binary fluorosulfates, Pd(S03F) 2, and the mixed valency compound, Pd(II)Pd(IV)(S03F) 6 , were obtained and characterized. Pd (S03F ) 2 has the metal in the rather uncommon octahedral and thus paramagnetic environment. Dilute magnetic systems are also encountered for Pd(II)[M(IV)(S0 3 F) 6 ], where NI = Pd, Pt, Sn. Ligand f i e l d parameters have also been determined for a l l the above compounds, except for P d 2 ( S 0 3 F ) 6 , by electronic spectro-scopy. Diamagnetic complexes containing the presumably square planar [ P d ( S 0 3 F ) t J 2 _ can also be obtained. While a number of [Pd(S03F) 6] 2~ complexes can be synthesized and characterized, the parent compound Pd(S03Fk has thus far eluded a l l attempts. Indirect evidence seems to indicate its existence in S 2 0 6 F 2 / H S 0 3 F and that i t has a strong fluorosulfate-accepting a b i l i t y . 96 CHAPTER 4 GOLD-FLUOROSULFATE 4.A INTRODUCTION Although gold is a member of the copper triad copper, silver and gold , unlike in most other groups, there is l i t t l e resemblance between the chemistry of the three metals. While both the noble metals can exist in the monovalent state, Au(III) is the only commonly found high oxidation state. Both Ag(II) and Ag(III) are extremely strongly oxidizing and cannot be obtained easily. Compounds have frequently appeared in the literature in which gold has formally the oxidation number of +2. Most of them, however, are actually mixed valency compounds containing Au(I) and Au(III) because they are diamagnetic. A true Au(II) compound should be paramagnetic because of the presence of a d 9 electronic configuration for the metal. The few truly Au(II) compounds con-tain dithiolate-type ligands which are very effective in electron delocalization. These compounds are paramagnetic, as studied by bulk magnetic susceptibility measurements and e.s.r. 1 3 3 » 1 5 2 . The f i r s t example of a +5 oxidation state for gold is the [AuF 6] _ complexes, formed by the fluorination/complexation reac-tion of AuF3 according to 1 5 3> 1 5 t +: 2XeF2 + AuF3 + 5F 2 • [Xe 2F n] + [AuF6] " (4.1) The pentafluoride, AuF5, can be prepared by the pyrolysis of 0 2[AuF6l" 1 5 5 . These fluorine coordinated compounds of Au(V) represent a frequent observation in transition metal chemistry 97 that the highest possible oxidation state of a metal is usually found in such systems. The general chemistry of gold has been reviewed in a number of recent publications 1 5 6 - 1 5 9 . Again, some brief comments on the halide system of gold can provide some pertinent background information to the corresponding fluorosulfate system. The monohalides, Au-Hal, with Hal=Cl, Br, I, the trihalides, Au-Hal3, with Hal=F, Cl, Br, and the previously mentioned AuF5, are known. All of these binary halides, because of their unsatu-rated coordination spheres, can exhibit Lewis acidity towards halide ions or other suitable donor ligands to form complexes. For examples, neutral adducts of the types Au(I)-Hal«L, and Au(III)-Hal3«L, where L=neutral ligands, and anionic complexes of the types of [Au(I)-Hal 2]", [ A u d l l J - H a l J ' and [AuF 6] _, can readily be prepared. The chemistry of the monohalides, by v i r -tue of i t s similarities to those of Ag(I) and Cu(I), will not be discussed here. The three trihalides are polymeric to varying degrees: Au2Cle and Au 2Br 6 are the species present in vapor, solid state and solutions in covalent solvents. AuF3, on the other hand, is extensively polymerized, and consists of chains of fluorine-brid-ged [AuFJ units packed in a helical manner 1 6 ° . The tetrahalo-complexes, similar to the binary trihalides, are square planar and therefore diamagnetic 1 6 1 . Actually, a l l gold compounds, except for the very few that contains Au(II), are diamagnetic; 98 the rationale generally invoked to explain the diamagnetism of most Pd(II) compounds is even more applicable to Au(III) because of the l a t t e r 1 s larger ligand f i e l d splitting. Au(S0 3F) 3, f i r s t reported by Cady in 1972 8 2 , appears to be the only well documented binary oxyacid salt of Au(III). While the existence of the nitrate is questionable, i t s anionic complex, K[Au(N03)[J , can be prepared and crystallographic results indi-cate the compound has monodentate nitrate group in a square planar configuration around Au(III) 1 6 2 . Au(S0 3F) 3 may behave as an ansolvo acid in the HS03F solvent system for the following reasons: a) The reaction of gold metal with BrS03F, as reported by Cady et a l 8 2 , resulted in the formation of a solution. Since BrS03F is an ionizing solvent 3 8 , the solvation of the product indicates some type of dissociation process of the solvent taking place with Au(S0 3F) 3. b) The product isolated from the reaction mixture described above is a solid of the approximate composition of Au(S03-F)3«2BrS03F. This could be a donor-acceptor adduct, and since BrS03F is not known to be a strong S03F acceptor 1 6 3 , Au(S0 3F) 3 could be the acceptor, with the formation of a species which may be formulated as [Br 2S0 3F] + [Au(S0 3F ) i + ] - . c) Solutions of AuF3 in HS03F have been reported by Woolf to show acid behavior, although the actual nature of the species present in solution was not identified 1 6 1 +. Therefore, a f u l l investigation into the chemistry and solu-tion behavior of Au(S0 3F) 3 in HS03F was undertaken, in the hope of establishing its existence as a novel superacid system 1 6 5 , 1 6 The gold-trifluoromethylsulfate system was also studied for pos-sible superacid behavior in HS03CF3; the results are described in Appendix B. 100 4.B EXPERIMENTAL 4.B.l SYNTHESIS OF Au(S0 3F) 3 4.B.1.1 OXIDATION OF GOLD METAL BY S 20 6F 2/HS0 3F 2Au + 3S 20 6F 2 - ^ J ; 2Au(S0 3F) 3 (4.2) In a typical reaction, S 20 6F 2 (~2 mL) and HS03F (~3 mL) were d i s t i l l e d into a reactor containing gold powder (579 mg, 2.940 mmol). The reaction proceeded slowly at room temperature, with the metal dissolving over a period of about 6 hours to give a yellow to orange solution. The removal of a l l volatile materials yielded Au(S0 3F) 3, (1.460 g, 2.955 mmol). Au(S0 3F) 3 is a yellow to orange, hygroscopic, crystalline solid. It is sublimable at ~120°C in vacuo and melts with decom-position at ~140°C to a dark red liquid. It is very soluble in HS03F, with concentrations of up to ~7 m having been obtained. The elemental analysis of Au(S0 3F) 3 was reported by Cady et al 8 2 . 4.B.1.2 S0LV0LYSIS OF AuCl 3 IN BrS03F 2AuCl 3 + 6BrS03F • 2Au(S0 3F) 3 + 3Br 2 + 3C12 (4.3) AuCl 3, (168 mg, 0.554 mmol), dissolved in BrS03F, (~6 mL), with gas evolution. The vapor had the appearance of C l 2 , which could be in equilibrium with BrCl. To ensure a complete reaction, the solution was heated at 60°C for 4 hours. After the removal of a l l volatile materials, Au(S0 3F) 3 was obtained, (276 mg, 0.599 mmol). 101 Au(S03F)3 can also be obtained from, a) the reaction of Au 20 3 with S206F2/HS03F: Au 20 3 + 3S 20 6F 2 2Au(S0 3F) 3 +. 30 2 , (4.4) or b) the reaction of gold metal with BrS03F followed by the thermal decomposition of the intermediate as reported by Cady et al 8 2 . 4.B.2 SYNTHESIS OF IONIC [Au(S0 3 F K r COMPLEXES CsCl + HS03F • CsS03F + HC1 (4.5) 2CsS03F + 2Au + 3S 20 6F 2 2Cs [Au(S0 3F)J (4.6) HS03F (~3 mL) was added to a mixture of CsCl (142 mg, 0.843 mmol) and gold powder (166 mg, 0.843 mmol). The HC1 formed was removed by evacuating a l l the solvent. S 20eF 2, (~2 mL) and HS03F, (~4 mL), were then added to the solid mixture and the reactor was subsequently heated at ~60°C for 12 hours. After which time a l l the metal had reacted. The removal of a l l volatile materials yielded Cs [Au(S0 3F) l t], (622 mg, 0.844 mmol). An alternative synthetic route involves solvolysis/oxidation in BrS03F according to: CsCl + Au + 4BrS03F • Cs[Au(S0 3F)J + 2Br 2 + %C1 2 . (4.7) The equimolar mixture of CsCl and gold metal was heated in an excess amount of BrS03F at ~60°C for 2 days, followed by the removal 102 of a l l volatile materials. Cs[Au(S0 3F)iJ is a yellow, crystalline, hygroscopic solid. It is soluble in both HS03F and BrS03F, and melts at 105°C into a deep orange liquid which decomposes at ~300°C. Analysis Cs Au F Calculated % 18.30 27.13 10.47 Found % 18.14 26.98 10.30 4.B.2.2 PREPARATION OF K[Au(S03F)iJ KC1 + HS03F • KS03F + HC1 (4.8) 2KS03F + 2Au + 3S 20 6F 2 2Cs [Au(S0 3F) 4] (4.9) In an analogous manner to the preparation of the correspon-ding Cs + salt, KC1 (78 mg, 1.05 mmol) and gold powder (206 mg, 1.05 mmol) were reacted with a mixture of S20 6F 2 and HS03F to yield K[Au(S0 3F) l t], (692 mg, 1.09 mmol). Similarly, K[Au(S0 3F)i t] can also be obtained from the co-sol volysis of KC1 and AuCl 3 in an excess of BrS03F, followed by the removal of a l l volatile materials. K[Au(S03F)iJ is a yellow, crystalline, hygroscopic solid. It is soluble in both HS03F and BrS03F; i t melts with decomposi-tion at ~230°C. Analysis K Au F Calculated % 6.18 31.15 12.02 Found % 6.22 30.80 12.28 103 4.B.2.3 PREPARATION OF Li [Au(S03FK] Li 2C 0 3 + S 20 6F 2 • 2LiS0 3F + C0 2 + h02 (4.10) 2LiS0 3F + 2Au + 3S 20 6F 2 2Li [Au(S0 3F) 4] (4.11) An excess of S 20eF 2 (~1 mL) was added to a mixture of Li 2C 0 3 (25 mg, 0.68 mmol) and gold powder (132 mg, 0.67 mmol). After the gas evolution from the f i r s t reaction had ceased, a l l volatile materials were removed and a mixture of S 20 6F 2 (~2 mL) and HS03F (~2 mL) was d i s t i l l e d into the reactor. A slow reac-tion occurred, and the reactor had to be kept at 100°C for 2 days order for a l l the metal to react. A precipitate formed, and the removal of a l l volatile materials yielded Li [Au(S0 3F) i t], (406 mg, 0.68 mmol). Li[Au(S0 3F) 4] is a yellow, crystalline, hygroscopic solid which does not melt below 280°C. It is only slightly soluble in HS03F at room temperature. Analysis Li Au F Calculated % 1.16 32.87 12.66 Found % .1.21 32.59 12.51 4.B.2.4 PREPARATION OF (C102) [Au(S03FK] C10 2S0 3F + Au(S0 3F) 3 • (C102) [Au(S0 3F) 4] (4.12) Au(S0 3F) 3, (224 mg, 0.453 rnrnol), was dissolved in an excess of C102S03F (~2 mL). The subsequent removal of a l l excess C102-S03F yielded (C10 2)[Au(S0 3F) 4], (288 mg, 0.436 mmol). 104 The above compound can also be obtained from a reaction of gold metal with a C102S03F/S206F2 mixture, according to: 2C102S03F + 2Au + 3S 20 6F 2 • 2(C102) [Au(S03F)lt] . (4.13) (C102) [Au(S03F)tt] is a yellow, crystalline, hygroscopic solid. It is soluble in C102S03F at 25°C to give a light orange solu-tion. It melts at 101 - 103°C to an orange liquid which decom-poses at ~170°C. Analysis Cl Au S Calculated % 5.23 29.81 19.41 Found % 5.23 30.00 19.24 4.B.2.5 PREPARATION OF (NO)[AU(S0 3F)LJ 2AU + 3Br 2 • 2AuBr3 (4.14) 4N0C1 + AuBr3 • (NO) [AuClk] + 3N0Br (4.15) (N0)[AuClJ + 4BrS03F • (NO) [Au(S0 3F)J + 2Br 2+2Cl 2 (4.16) Gold powder (112 mg, 0.569 mmol) was reacted with Br 2 at 80°C for 12 hours to give AuBr3,.(247 mg, 0.565 mmol), which was reacted with an excess of N0C1 (~5 mL) at 80°C for h day. The insoluble, light brown, powder was identified as (NO) [AuCl i + ] , (209 mg, 0.567 mmol), by its Raman spectrum which had bands at 2205 and 2183 cm"1, (vN-0), and also at 345 and 321 cm"1, (v[Au-Clitl). (NO) [Au(S0 3F) i t ], (348 mg, 0.558 mmol) was prepared by reacting the (NO) [AuCli+] with an excess of BrS03F, (~2 mL), and heating the reaction mixture at 80°C for h hour. The product was 105 isolated after the removal of a l l volatile materials. (NO) [Au(S03F)(+] is a yellow, crystalline, hygroscopic solid. It is soluble in both BrS03F and HS03F. It melts at 192 - 195°C to an orange liquid which decomposes at ~260°C. Analysis N Au F Calculated % 2.25 31.61 12.19 Found % 2.10 31.85 12.32 4.B.3 SYNTHESIS OF [AuCSOgFK] COMPLEXES CONTAINING HALOGENO- FLUOROSULFATO CATIONS 4.B.3.1 PREPARATION OF [Br(S0 3F) 2] [Au(S0 3FkJ Br 2 + 3S 20 5F 2 • 2Br(S0 3F) 3 (4.17) 2Br(S0 3F) 3 + 2Au + 3S 20 6F 2 > 2[Br(S0 3F) 2] [Au(S03F)tf] (4.18) Br 2, (37 mg, 0.231 mmol) and gold powder (90 mg, 0.457 mmol) were reacted with S 20eF 2 (~3 mL) at 70°C for % day. The oxidation of Br 2 proceeded quickly, as evident by the formation of a yellow solution. Soon afterwards, the gold metal dissolved to form two immiscible, light yellow layers, with the top layer presumably containing mostly S 20 6F 2. The bottom layer crystallized when the mixture was cooled at 0°C for 3 days. The removal of a l l volatile materials at room temperature yielded [Br(S0 3F) 2] [Au(S0 3F) [ +], (408 mg, 0.468 mmol). [Br(S0 3F) 2] [Au(S03F)!+] is a light yellow, hygroscopic, cry-106 stalline solid that has a tendency towards supercooling at room temperature. It melts at 52 - 55°C. Analysis Br Au S F Calculated % 9.17 22.61 22.08 13.08 Found % 9.26 22.45 21.84 13.13 4.B.3.2 PREPARATION OF [I(S0 3F) 2] [Au(S0 3F)J I 2 + 3S 20 6F 2 • I(S0 3F) 3 (4.19) 2I(S0 3F) 3 + 2Au + 3S 20 6F 2 • 2 [I (S0 3F) 2] [Au(S0 3F)J (4.20) In a manner similar to the last preparation, I 2 (92 mg, 0.363 mmol) and gold powder (143 mg, 0.726 mmol) were allowed to react with S 206F 2 (~3 mL) at ~80°C for 12 hours. After a l l the metal had reacted, crystalline [I(S0 3F) 2] [Au(S0 3F) 4] was formed upon cooling the mixture to room temperature (650 mg, 0.708 mmol). [I(S0 3F) 2] fAu(S0 3F)i 1 ] is a light orange to yellow, hygrosco-pic crystalline solid. It has a melting range of 70 - 73°C. Analysis I Au F Calculated % 13.82 21.45 12.41 Found % 14.04 21.61 12.67 107 4.B.4 SYNTHESIS OF [Au(S03FKr COMPLEXES CONTAINING POLYBROMINE  CATIONS 4.B.4.1 PREPARATION OF (Br3)[Au(S03F)„] Au + 4BrS03F • (Br 3) [Au(S0 3Fk] + JgBr2 (4.21) An excess of BrS03F (~3 mL) was d i s t i l l e d onto gold powder, (293 mg, 1.488 mmol), contained in a reaction v i a l . A vigorous, exothermic reaction occurred at room temperature. To ensure a complete reaction, the mixture was heated at 60°C for 6 hours. The removal of a l l volatile materials at room temperature yielded a compound that analyzed as (Br 3) [Au(S0 3F) 1 +], (1.239 g, 1.488 mmol). (Br 3)[Au(S0 3F) iJ is a dark brown, hygroscopic, polycrystal-line solid. It decomposes at 105°C to form Au(S0 3F) 3. It is soluble in BrS03F and HS03F. Analysis Br Au F Calculated % 28.78 23.65 9.12 Found % 28.64 23.89 9.15 4.B.4.2 PREPARATION OF (Br5)[Au(S03F)»J (Br 3)[Au(S0 3FM + Br 2 * (Br 5) [Au(S03F)tt] (4.22) Gold powder (175 mg, 0.888 mmol) was converted into (Br 3)-[Au(S03F)iJ as described above. Br 2 (~5 mL) was d i s t i l l e d onto the intermediate and the mixture was heated at 70°C for 6 hours. The removal of a l l volatile materials at room temperature yielded (Br 5)[Au(S0 3F) 4], (875 mg, 0.882 mmol). 108 The same compound can also be obtained by a) the addition of excess Br 2 to a 2:1 mixture of Au(S0 3F) 3 and S 20 6F 2, according to: 5Br 2 + 2Au(S0 3F) 3 + S 20 6F 2 • 2(Br 5)[Au(S0 3F),J , (4.23) or b) the addition of excess Br 2 to [Br(S0 3F) 2][Au(S0 3F) 4]: [Br(S0 3F) 2] [Au(S03F)il]+3Br2 • (Br 5)[Au(S0 3F) 4] + 2BrS03F . (4.24) (Br 5) [Au(S0 3F) i t ] is a dark brown, polycrystalline, hygrosco-pic solid. It is insoluble in Br 2, and melts with decomposition at ~65°C. The compound shows a slight decomposition pressure at room temperature in vacuo but is stable i f kept in 1 atm nitrogen. Analysis Br Au F Calculated % 40.25 19.84 7.66 Found 40.20 20.07 7.79 109 4.C DISCUSSION 4.C.1 SYNTHESIS AND GENERAL DISCUSSION  4.C.1.1 Au(S0 3F) 3 For the intended solution study on Au(S0 3F) 3 in HS03F, a convenient and fast synthetic route to very pure samples of the compound was needed. The original synthesis, described by Cady et al 8 2 , involved the oxidation of the metal with BrS03F followed by the pyrolysis of the intermediate with the reported composition of Au(S0 3F) 3«2BrS0 3F, led to slightly impure products. Samples of Au(S0 3F) 3 obtained by this method were found to be slightly para-magnetic, with a magnetic moment of ~0.5 and displaying Curie-Weiss law behavior from ~300 to ~100K. Interestingly, the original synthesis of AuF3 by an almost parallel method, repre-sented below: Au + 2BrF 3 >- AuF 3«BrF 3 + %Br 2 (4.25) AuF 3-BrF 3 — A u F 3 + BrF 3 , (4.26) also produced weakly paramagnetic materials 1 6 7 , while pure AuF3 obtained from the direct fluoridation of AuCl 3 is diamagnetic 1 6 0 , 1 6 Although the presence of Au(II) or paramagnetic bromine-containing species such as B r 2 + may be the cause of the observed paramagnetism, a detailed investigation was considered beyond the intent of this study. The reaction of gold metal with S 20eF 2 resulted in only a slow surface attack and required a reaction temperature of ~160°C for ~14 days. The addition of HS03F as the solvent allows the 110 synthesis of pure and diamagnetic Au(S03F)3 in a fast and e f f i c i -ent manner. As an added advantage, this method allows the direct preparation of solutions of S 20 6F 2 in HS03F because the excess S 20 6F 2 can be removed readily. As shown by Raman spectroscopy of the resulting solutions, for an i n i t i a l 1:1 S 20 6F 2/HS0 3F mixture, no trace of S 20 6F 2 (detected by vO-O at 900 cm - 1) 9 5 was l e f t after about 70% of the solution was removed by vacuum. The alternative synthetic route to Au(S0 3F) 3 by the solvolysis of AuCl 3 in BrS0 3F offers no distinct advantage over the S 20 6F 2/ HS03F method. The attempted solvolysis of AuCl 3 in HS03F did not proceed even at ~110°C. 4.C.1.2 IONIC [AutSOgF^r COMPLEXES The preparation of these complexes utilizes the various synthetic methods described in Chapter 3 for the preparation of Pd(IV) complexes. Due to the lack of reactivity of gold metal towards HS03F 6 h , stoichiometric mixtures of gold and an alkali metal chloride can be solvolysed to give the corresponding M(I)(S0 3F). The addition of S 20 6F 2/HS0 3F results in the oxidation of gold and the formation of the alkali metal tetrakis(fluorosul-fato) aurate(III) salts. Because of the deliquescent nature of Li Cl, the carbonate was used instead; but because the solvolysis of C0 3 2~ in HS03F would lead to water being one of the products, according to: C0 3 2" + 2HS03F • H20 + C0 2 + 2S03F" , (4.27) I l l S 20 6F 2 was used as the fluorosulfonating agent. These pseudo-one-step reactions are very reliable synthetic routes which exclude the handling of hygroscopic intermediates. NO[AuCli+] is usually prepared by the reaction between N0C1 and AuCl3 in the presence of a solvent such as diethyl ether or THF 1 6 9 . To avoid having to dry and use such hygroscopic solvents, an alternative synthetic method was chosen, involving the solvoly-sis and subsequent complexation of AuBr3 in N0C1. A very interesting trend can be noted in the melting points of a few of these complexes: C10 2 +, (~102°C), ~Cs +, (~105°C), < N0+, (~193°C), K+, (~230°C), L i + , (~280°C). This may be due to lattice energy effects which depend on the size of the cation, and anion-polarization by the cation resulting in anion-cation interactions. The effect of N0+ is very similar to that of K+; this is consistent with crystal!ogra-phic results from other studies which have shown that the two cations have almost identical effective ionic radii in the solid state 9 4 . 4.C.1.3 COMPLEXES CONTAINING PQLYBROMINE CATIONS In order to characterize the intermediate of Au(S0 3F) 3* 2BrS03F from the Au + BrS03F reaction, thought to contain a [Br 2(S0 3F)] + ion, the published formation reaction was reinvesti-gated 8 2 . While interhalogen cations such as B r 2 C l + , I 2C1 + and I 2 B r + 8 5 are known, such an analogy does not exist in the form of 112 [Br 2(S0 3F) ] + . Although the auto-ionization of BrS03F has been postulated as 3 8 : 3BrS0aF *• [Bro(S0 3F)] + + [Br(S0 3F) 2]" , (4.28) only thermally unstable salts containing [Br(S0 3F) 2]" are known 80,81,163. The oxidation of gold with an excess of BrS03F did yield a dark brown polycrystalline material: however, i t s chemical analy-sis revealed i t s composition as Br 3Au(S0 3F) t f, suggesting a formu-lation as Br 3 +[Au(S0 3F) 1 +]". The B r 3 + cation has been identified in solid compounds of the type of Br 3 +EF n", where EF n" = AsF 6" 1 7 ° , BF./ and SbF 6" 1 7 1 , and in solutions of strong protonic acids 25,28,55. jhe addition of Br2 to BrS03F giving B r 3 + was also shown by low temperature i . r . spectroscopy 1 7 2 . The formation of B r 3 + in the reaction of gold with BrS03F may proceed via the following reaction scheme: Au + 3BrS03F • Au(S0 3F) 3 + 3 / 2 B r 2 (4.29) Br 2 + BrS03F >- B r 3 + + S0 3F" (4.30) B r 3 + + S0 3F' + Au(S0 3F) 3 B r S 0 3 F > Br 3 +[Au(S0 3F) 4]" , (4.31) with the precipitation of the final product a necessary driving force to i t s formation. Evidence for reaction (4.30) has been found in superacid solutions 2 8 and low temperature i . r . study 1 7 2 Unlike the previously reported Br 3 +-containing compounds, which a l l show appreciable decomposition pressures at 25°C, Br 3 +[Au-(S0 3F ) iJ~ is found to be thermally stable at room temperature, 113 with no decomposition evident in a dynamic vacuum at 25°C. The thermal decomposition of B r 3 + [Au(S0 3F) iJ~at ~100°C in vacuo to give slightly paramagnetic Au(S0 3F) 3 may proceed via a Br 2 +-containing species. The 2 n 3 / 2 g ground state would give rise to the observed paramagnetism. Therefore, attempts were made to synthesize Br^tAuCSOsF)^], by reacting Br 2 with stoichiometric amounts of BrS03F or S 20 6F 2 and Au(S0 3F) 3, according to: hBr2 + BrS03F + Au(S0 3F) 3 *• Br 2 +[Au(S0 3F) 1 +]" (4.32) Br 2 + J 2 S 20 6F 2 + Au(S0 3F) 3 • Br 2 +[Au(S0 3F) 1 +]" . (4.33) However, both these reactions led to the formation of rather inhomogeneous products, even when C 8F 1 7S0 2F was used as the inert reaction medium for Br 2. The observation of a strong resonance Raman spectrum of B r 2 + , with the principle band at 360 cm - 1, indicates the cation was present 2 5 » 2 8 , but an analytically pure product of the desired composition was not isol able. The further addition of bromine to Br 3[Au(S0 3F ) i J resulted in the formation of Br 5[Au(S0 3F ) i 4 ], a compound of relatively low thermal st a b i l i t y (although the decomposition at room temperature is rather slow, thus allowing i t s isolation). A formulation invoking the hitherto unknown B r 5 + ion is possible. In the iodine system, I 5 + has been reported in solid I 5A1Clu 1 7 3 as well as in HS03F solution together with I 7 + 5 6 . The limited thermal st a b i l i t y of B r 5 + is not unexpected, and the compound can only be formed in a large excess of Br 2, whether the starting material is Br 3[Au(S0 3F ) tJ or [Br(S0 3F) 2][Au(S0 3F ) i J. Being an easily oxidiz-114 able species, with a formal oxidation number of +1/5 for bromine, B r 5 + is unstable in the presence of any significant quantities of BrSOsF; for example, when gold is reacted with a mixture of Br 2/ BrS03F of approximately 1:1 by volume, the only product isolated is the tribromine species. 4.C.1.4 COMPLEXES CONTAINING HALOGENO-FLUOROSULFATO CATIONS There is only one precedent reported for complexes contain-ing [Hal(S0 3F) 2] + cations, with Hal = Br and I, — t h o s e of [Sn(S0 3F) 6] 2" 3 1 . The synthesis of the corresponding [AuCSOsF)^]" containing complexes was undertaken to gain more insights into this group of compounds and to compare the S03F~ acceptor a b i l i -ties of Au(S0 3F) 3 and Hal(S0 3F) 3. The successful synthesis of [Hal (S0 3F) 2] + [Au(S0 3F) i t] _ would establish Au(S0 3F) 3 as the strong-er ansolvo acid and confirm the amphoteric behavior of the halo-gen trifluorosulfates. The reaction of stoichiometric amounts of bromine or iodine and gold with an excess of S 20 6F 2 resulted in the formation of the corresponding [Hal(S0 3F) 2] [AuCSOsF)^], with the formulation supported by a characterization of the compounds. S 20 6F 2 acts as both the oxidizing agent and the solvent in these reactions by dissolving the halogen tris(fluorosulfate) formed in the i n i t i a l reactions. The reaction can be viewed as between a Lewis acid and a Lewis base, according to: Hal(S0 3F) 3 + Au(S0 3F) 3 • [Hal(S0 3F) 2][Au(S0 3F) 4] . (4.34) 115 The alternative reaction scheme involving Au(S0 3F) 3 as the base is not supported by vibrational spectroscopy. As the compound originally thought to be a 2:1 adduct be-tween BrS03F and Au(S0 3F) 3 turned out to contain the B r 3 + cation instead, attempts were made to synthesize the adduct by alternate methods. Although Au(S0 3F) 3 was found to be soluble in BrS03F, the only product isolated upon the removal of a l l volatile mate-ri a l s was Au(S0 3F) 3. It therefore appears that the [Br 2(S0 3F)] + cation has very limited s t a b i l i t y with respect to dissociation in vacuo at room temperature. 4.C.1.5 REACTION OF Au(S0 3F) 3 WITH NEUTRAL DONOR MOLECULES Au(S0 3F) 3 + :L — Au(S0 3F) 3-L (4.35) Although many attempts were made to synthesize adducts of the type of Au(S0 3F) 3«L according to equation (4.35), no stable products could be isolated. Au(S0 3F) 3 was found to be slightly soluble in S0 2, and the resulting solution was investigated by Raman spectroscopy, but the removal of a l l S0 2 at 0°C yielded only unreacted Au(S0 3F) 3. The use of other Lewis bases such as pyridine, dimethyl sulfoxide, and acetonitrile a l l led to the reduction of Au(S0 3F) 3 to gold metal. It therefore appears that: a) With the formation of SbF 5*S0 2 1 7 h , SbF 5 is a stronger Lewis acid towards S0 2 than Au(S0 3F) 3 i s , although this could be 116 due to the inability of S0 2 to compete favorably against the bridging S O 3 F groups in the Au(S03F)3 polymer. In any case, S0 2 is a very weak Lewis base; SbF 5«S0 2 has only limited thermal sta-b i l i t y 1 7 4 and S0 2 has been used effectively as a diluent in superacid systems k 5 . b) The observed reduction of Au(S0 3F) 3 by the ligands with stronger Lewis basicity may proceed by the displacement of the coordinating S O 3 F group followed by a reductive decomposition of the Au(III) species to Au(0). The substitution of S O 3 F groups in transition metal bis(fluorosulfate)s by pyridine and bipyridine to form ionic fluorosulfates have been reported 8 3 , 8 4 , 1 7 5 . 4.C.1.6 ATTEMPTS TO STABILIZE OTHER POLYHALOGEN AND  INTERHALOGEN CATIONS The only reported example for a polychlorine cation is C l 3 + , found in thermally unstable complexes with [AsF 6]~ or [SbFel as the counter-anion 6 0 . The reasonable thermal st a b i l i t y of Br 3[Au(S0 3F) iJ suggests that the chlorine analogue may exist as well. Also, interhalogen cations formed between bromine and chlorine are uncommon, with B r C l 2 + and B r 2 C l + identifiable only in SbF 5 solutions 1 7 6 ; the interaction of B r 3 + and B r 5 + with C l 2 may be a possible synthetic route to such compounds. Unfortunately, both Br 3[AuCSOaFK] and Br5.[Au(S03F) J , when reacted with chlorine gas at room temperature, or with liquid chlorine at ~-78°C, resulted in the formation of a product with 117 the characteristics of AuCl3 as followed by weight and i . r . spec-troscopy. No compounds containing mixed chloro-bromo cations were isolated from these reactions. AuCl 3 reacted with C1S03F to give a mixture identified by Raman spectroscopy as consisting of C102[AuCSOsFk] and Au(S0 3F) 3, with the former arising as a result of C l S O 3 F 1 s reaction with traces of moisture in the quartz reactor 1 1 . The attempted reductive replacement of (C10 2) + in C102[AufSOsF)^] with C l 3 + by reaction with liquid chlorine was unsuccessful; an analogous reaction using Br 2 yielded Br 3[Au(S0 3F ) iJ . No reaction was evident either when AuCl3 was mixed with a ~1:3 mixture of C1S03F/C12 at ~-23°C. Therefore, the basicity of the [AuCSOsF)^]~ anion may not be low enough for the stabilization of C l 3 + , B r C l 2 + or B r 2 C l + ; a l l of them are expected to be more elec-trophilic than Br3 +. The borderline stability of B r 2 + in the Au(III)-S03F system may be due to a similar explanation. 4.C.1.7 ATTEMPTED FURTHER OXIDATION OF GOLD Although the most common oxidation state of gold is Au(III), Au(V) can be found in the fluoride system 1 5 3 _ 1 5 5 , Considering the strong similarity between fluorides and fluorosulfates, the possibility of such an analogy in the fluorosulfate system was explored. Neither Au(S0 3F) 3 nor Cs[Au(S0 3F)iJ, when exposed to u n f i l -tered u.v. radiation from a Hg-discharge lamp in the presence of S 20 6F 2 at temperatures up to ~100°C, was found to react. The 118 addition of 0 2 to the above reactor containing Au(S03F)3 and S 20 6F 2 did not yield any 0 2 + containing products. F S O 3 F , a very powerful oxidant, was added to Cs[Au(S03F)i4] and warmed up to room temperature; no fluoride-fluorosulfate of Au(V) was obtained from the reaction. 4.C.2 VIBRATIONAL SPECTRA 4.C.2.1 IONIC [AuCSLVK]-COMPLEXES AND Au(S0 3F) 3 The Raman spectra of K[Au(S0 3F)i4] and (C102) '[AufSOaFK] are shown in Fig. 4.1 and 4.2, respectively. As can be illustrated in the figures, these complexes are excellent Raman scatterers, and spectra of extremely high signal to noise ratio and good resolution are generally obtained even with laser powers of "15 mW. The Raman frequencies of M[Au(S03F)iJ , with M = Cs, K, L i , NO, and C102, are listed in Table 4.1 and are compared with those of K[I(S03FK] ' 3 5 . The i . r . frequencies of Cs[Au(S0 3F)iJ are also included in the table to illustrate the intensity difference between the two types of spectra. Except for occasional small splittings, identical spectra are obtained for a l l the complexes discussed here. Both band positions and relative intensities agree well with the previously reported spectra for [USOaF)^]" and [Br(S0 3F)i+]~, suggesting some structural similarities and similar vibrational mode assign-FIG.4.1 Raman Spectrum of K[Au(S0 3 F) 4 ] 1300 9 0 0 Wavelength co 00 r - in in mm 01 m Oco 500 3 0 0 100 FIG.4.2 Wavenumber cm-TABLE 4.1 K[I(S0 3F)H] 3 5 M = Ci0 2 R R 1423 m 1409 m 1416 m 1375 w 1250 s 1222 w,sh 1002 m 837 m 620 vs 582 m 554 m 442 s 1300 vw 1228 vs 1208 w 1050 m 1025 s 985 m.sh 974 m 922 vw 835 m 818 tn 643 vs 587 w 576 w 555 w 546 w 518 w 448 s 407 w 417 w 397 w 260 s 277 vs 239 vs 162 w 125 w VIBRATIONAL FREQUENCIES OF M[Au(S0 3F)„] M = NO R 2331 w 1417 in 1395 in 1231 vs 1204 w M = Li R M = K R 1434 w 1430 w 1409 w.sh 1415 m 1400 w 1398 m 1390 w,sh 1242 vs 1227 s 1210 w,sh M = Cs R 1404 m 1380 w.sh 1018 s 1029 s 999 s 1011 s 978 w 985 w 985 w 960 vw 933 vw 950 vw 945 vw 846 m 835 w 845 vw 830 w 818 m 815 m 828 w 814 w 651 s 642 vs 660 vs 649 vs 643 s 580 w 582 w 583 w 580 w 573 w 576 w 546 w 555 w 551 w 550 w 455 in 453 s 459 m 452 s 455 m 402 vw 415 w 407 w 405 w 391 w 384 vw 391 w 278 vs 280 vs 282 vs 282 s 279 vs 150 w,sh 151 w 179 w 150 m 133 in 146 s 127 m 116 m M = Cs IR 1400 s,b 1239 vs 1240 s 1205 w.sh 1208 vs 1010 w.sh 970 s.sh 930 vs,b 830 s 810 s 678 s 585 s 550 s 456 s Assignment vN0+ v a s s o 3 v a s C£0* v s S0 3 v s Q0* v a sso 3 vSF vMO + def 6S02bend 6S02 6 C nO * vMO + def P S0 3 vMO + def Lattice Modes 122 merits 3 5 . The position of the SO3 - stretching modes at ~1400, ~1220, and ~1000 cm"1 indicate monodentate S O 3 F groups in an over-al l anionic environment 3 2> 3 5. The reversal of the intensities of these bands in the i . r . and Raman spectra, best illustrated by the two groups at ~(1240, 1200 cm-1) and "(1010, 950 cm"1), are presumably due to vibrational couplings between identical S0 3F groups through the [AuOiJ skeleton to produce in-phase and out-of-phase vibrations, as discussed in Chapter 3. An extreme case of intensity reversal in the two types of spectra can be found in the 640-680 cm-1 region, where internal fluorosulfate modes are commonly not expected for monodentate S O 3 F groups. Strong contributions from [AuOiJ stretching vibra-tions coupled to ligand deformation modes may explain the mutual exclusion of i . r . and Raman bands in this region. In addition, strong Raman bands are found at ~450 and ~280 cm"1 for a l l the complexes; they may have their origins in the same sort of vibra-tional couplings discussed for the ~650 cm"1 band, not entirely unexpected for such a complex molecule with a cation of +3 charge. A Raman spectrum was taken of the melt of Cs[Au(S03F)[t] at ~150°C; except for the presence of line broadening, the spectrum obtained was identical to that at room temperature. Depolarization ratio measurements revealed a l l the stronger bands, for which p , the depolarization ratio, could be obtained, have a large contri-bution from asymmetric vibrations, with a l l p values of the order of 0.1 (for a totally symmetric vibration p should be close to 123 zero). This is expected for a molecule with such a low symmetry (C s) and supports the argument for the presence of strong vibra-tional couplings. An interesting comparison of the [AuOtJ vibrations can be made with the vibrational spectrum of KtAuCNOs)^], which has an extremely intense and depolarized Raman band at 360 cm - 1 assign-able to [vg-AuOi+l 1 7 7 . Furthermore, the splittings in the N03-stretching modes can best be explained as due to solid state effects since the intensities after splitting are very similar. On the contrary, as previously mentioned, in the fluorosulfate system, some of the bands are s p l i t into strong/weak pairs, evi-dent of strong vibrational couplings. The crystal structure of K[Au(N0 3)iJ 1 6 2 indicates a significant lengthening of the Au0-N02 bond to 1.37 A from 1.25 A in Ni(N0 3) 2-4H 20 1 7 8 , in which N0 3 _ is uncoordinated. The other two N-0 bond lengths are (1.20±0.03) A, shorter than that found in free nitrates. As a comparison, in (CH 3) 2Sn(S0 3F) 2, in which the S03F groups are bidentate bridging, al l the SnO-S bond lengths are equivalent within experimental error 9 7 . Therefore, as far as vibrational analysis is concerned, [AU (N0 3)LJ~ may be viewed as consisting of [AuOJ---[N0 2], where-as in [Au(S0 3F)tJ", a much closer association between the two units is evidently present, giving rise to the observed extensive vibrational couplings. For the complexes containing the alkali metal ions, the add-itional splittings observed in the S0 3-stretching region are most likely due to anion-cation interaction, which can be a form of solid state effects; the degree of splitting appears to increase with decreasing cation sizes, suggesting an increase in the pola-rizing power of the cation. The manifestation of these effects in producing splittings in the SO3 stretching modes rather than in producing any significant frequency shifts provides further evi-dence for a very stable [Au(S03F)4] unit. The presence of ionic species in a l l the complexes is sup-ported by the almost invariable band position of the anion and the agreement of the vibrational frequencies for the two hetero-cations, C102+ and N0+, with published values. The fundamentals for the chloronium cation are found at 1300, 1050 and 518 cm"1 1 7 9 similarly, the N-0 stretching frequency is found at 2331 cm - 1, compared to ~2300 cm"1 for N0S03F 9 1 +. Splittings in the vibra-tional spectra due to the aspherical nature of these two multi-atomic cations are not unexpected. Finally, the similarities in the vibrational spectra extend also to Au(S0 3F) 3 and the halogen tris(fluorosulfate)s, Br(S03F) 3 and I ( S O 3 F ) 3, illustrated in Table 4.2. The spectra contain contributions from both monodentate and bridging bidentate fluoro-sulfate groups. The diagnostic band for the bidentate bridging S O 3 F ligand is found at 1135 cm - 1 1 8 0 , and also in the S0 3-stretch ing, a strong doublet is present at 1442 and 1425 cm - 1 in the i . r . spectrum, indicating two types of S03F coordinations. The s p l i t -tings in the [AuOiJ related vibrational modes at (682, 670 cm - 1), Table 4.2 VIBRATIONAL FREQUENCIES OF Au(S0 3F) 3 Au(S0 3F) 3 Au(S0 3F) 3 I(S0 3F) 3 3 5 R IR R 1449 m 1442 vs 1469 m 1423 m,sh 1425 s,sh 1381 w 1240 s 1227 vs 1220 s,sh 1233 vs 1182 w,sh 1102 w 1135 ms 1076 w,sh 1053 m 1055 s 1050 m 955 m,sh 960 s,b 963 s 936 s 920 s,sh 899 s 895 s,b 873 m,sh 869 s 822 w 820 s 826 w 682 s 670 s,sh 650 vs 642 vs 610 w 610 w 619 w 590 s 583 w 582 s 580 m,sh 546 m 550 s 540 m,sh 465 m 460 m 457 s 444 w 430 w 412 m 386 m 348 m 327 w,sh 292 vs 290 vs 285 s 270 s,sh 189 181 m 170 149 148 m 126 (465, 444 cm"1) and (292, 285 cm - 1) are most lik e l y the result of the presence of non-identical S03F groups around Au(III) and the coupling of kuQh vibrations of adjacent [AuG\] units through bridging S0 3F ligands. While splittings previously assigned to solid state effects are observed in the Raman spectra of some [Au(S03F)tt] complexes in the same regions, they are of much small-er magnitudes and are not present in their i . r . spectra. The proliferation of low frequency bands for Au(S0 3F) 3, especially in the i . r . spectrum, must be a result of structural complexities leading to a reduced local symmetry for the [AuG\] unit. A band of medium intensity at 348 cm - 1 in the Raman spectrum of Au(S0 3F) 3 may be assigned as due to vAu-0, which, because of the distortion in the AuGV skeleton from square planar geometry, becomes a sepa-rate vibration in addition to those that are coupled to ligand deformations. Thus, i t can be concluded that Au(S0 3F) 3 contains both bridg-ing- and terminal-fluorosulfate groups, in possibly a polymerized structure with a distorted square planar environment for Au(III). 4.C.2.2 [X(S0 3F) 2] [Au(S0 3FK], X=Br, I Both the t i t l e compounds are excellent Raman scatterers when sample cooling to ~80 K is employed; the spectra thus obtained are of extremely high resolution. They were found to either melt or decompose at room temperature upon their exposure to laser radiation. Their vibrational frequencies are compared with those 127 of [Br(S0 3F) 2] [Sn(S0 3F) 6] 3 1 and Cs [AU(S0 3F)LJ in Table 4.3. As can be seen, both new compounds give rather complex vibrational spectra, and a complete band assignment is therefore not possible. It is evident that the strong characteristic Raman bands for [Au(S0 3F)iJ" are present in the spectra of the two compounds; on the other hand, referring to the spectrum of K[I(S0 3F) 4] listed in Table 4.1, bands due to [HSOsF)^]- could not be found in the iodine-containing complex discussed here. For example, the strong [I(S0 3F) 4]" bands at 1250, 1002, 620, 442, 260 and 239 cm"1 are absent in the gold-containing complex. This argues against the alternative formulation for these compounds as [Au(S0 3F) 2]-[X(S0 3F ) iJ, with Au(S0 3F) 3 formally acting as the S03F-donor. This is also consistent with the observation that while both halogen fluorosulfates I(S0 3F) 3 3 6 and Br(S0 3F) 3, are weak electrolytes in HS03F, Au(S0 3F) 3 forms highly conducting solutions in HS03F and as will be discussed in the following section, has been found to be a strong S03F acceptor. Further supporting evidence for the suggested formulation can be found in the bromine-containing complex. By comparison to the reported spectra for Br(S0 3F) 3 3 5 and [Br(S0 3F) 2][Sn(S0 3F) 6] 3 characteristic bands due to [Br(S0 3F) 2] + are readily recognized in the [Au(S0 3F)iJ" complex. In particular, the pair of bands at 1050 and 1491 cm - 1, assigned to an asymmetric S0 3-stretch, are found in other cases at such a high frequency only for volatile S0 3F-derivatives such as S 20 6F 2 9 5 , S 20 5F 2 and S 30 8F 2 1 8 1 . It 128 TABLE 4.3: VIBRATIONAL FREQUENCIES OF [X(S0 3F) 2]-[Au(S0 3F) L |] [Br(S0 3F) 2]- * ~ in(S0 3F) 6] 3 i X • = Br X • = I X = I CstAufSOaF), R R, -80K R, -80K IR R 1500 ms 1505 ms 1476 w 1458 w.sh 1487 mw.sh 1491 w.sh 1479 w 1460 w 1440 w.sh 1420 mw 1425 ms 1413 ms 1410 s.sh 1404 ms 1416 m 1400 m 1400 mw 1388 vw 1388 s 1387 ms 1389 mw 1380 w 1382 vs 1380 vw 1252 m.sh 1248 s 1235 s 1237 m.sh 1239 vs 1197 s 1220 ms 1227 ms 1225 s.sh 1205 w 1195 ms 1217 vs 1198 vs 1145 s 1174 m 1174 ms 1175 s.sh 1130 m.sh 1135 m 1130 m.sh 1092 s ' 1039 ms 1020 ms 1022 s 1022 ms '1025 w i o n s 1008 m.sh 990 vw 985 m 963 mw,b 964 ms 960 s,b 960 vw,b 921 vw 930 vw 925 vs 887 vw 865 w 860 m 865 vs 864 m 855 w 850 ms 852 m 842 w 840 m,sh 830 mw 830 w 817 w,b 830 vw 745 s 745 ms 813 ms 815 vs 814 w 773 vw 640 s 648 vs 651 vs 670 m.sh 646 vs 649 vs 631 s ,sh 635 s 596 mw 581 w 589 w 579 s 580 w 571 w 569 w 558 m 549 w 553 mw 555 m.sh 550 w 530 m 530 vw 548 w.sh 542 s 464 s 452 s 453 s 452 s 452 s 430 vw 414 ms 410 vw 399 vw 405 w 386 m 390 vw 390 vw 391 w 309 s 308 vs 298 s.sh 296 vw 290 s.sh 264 ms 281 vs 278 vs 282 vs 213 w 210 vw 205 w 181 w 151 m 159 w 150 m 108 w 135 w 127 m 100 w.sh 109 m 129 also appears that the increase in positive charge on the bromine atom increases the frequency of this vibrational mode, as can be illustrated below for the three compounds: [Br(S0 3F) 2] + [Au(S0 3F)J, Br(S0 3F) 3 3 5 , K[Br(S0 3FK] 3 5 v = 1505 , 1490 , 1424 A speculation as to the cause of this trend could be the streng-thening of the Br-0 bond upon the removal of a S03F group as the positive charge becomes more concentrated on the bromine atom. Nevertheless, an ionic formulation of these compounds consisting of [X(S0 3F) 2] + and [Au(S0 3F)iJ~ moieties may be an oversimplification. Prominent bands in the region of ~1150 cm - 1, characteristic for bidentate bridging fluorosulfate groups 1 8 0 , indicate the presence of covalent cation-anion interaction in both the [Au(S0 3FK] -- and [Sn(S0 3F) 6] 2"-derivatives. Such an interaction is not unexpected between these large anions and cations; in the molecular structure of the related interhalogen cation complexes of [BrF 2][SbF 6] 1 8 2 and [IC1 2][SbCl 6] 8 6 , short interionic distances suggest the presence of strong interactions, but not to the extent that the compounds could be classified as covalent solids. 4.C.2.3 Br n[Au(S0 3F) t f], n=3, 5 Because of their dark colors, both polybromine-containing complexes gave rather poorly resolved Raman spectra, even when the samples were cooled to ~80 K. Their i . r . spectra at room 130 temperature show a band pattern typical of [Au(S03F)4]", with no cation bands observable down to ~450 cm - 1. Low temperature i . r . spectra using Csl windows allowed the extension of the transmis-sion range down to ~200 cm - 1; but because the sample could only be deposited as a powder, the spectra have peaks with broad con-tours due to the Christiansen effect. The vibrational frequencies of the two polybromine complexes are compared with the Raman fre-quencies of Cs [Au(S03FK] in Table 4.4. As mentioned previously, i t becomes evident from the vibra-tional spectra that the only fluorosulfate-containing species is [Au(S0 3F) iJ". The band positions of the spectra from the two compounds are very similar, consistent with the almost invariant spectra observed for the ionic [AufSOsF)^]"-complexes. Unfortu-nately, a clear identification and assignment of bands due to the two polybromine cations are rather d i f f i c u l t . The low frequency range expected for v Br-Br (<300 cm - 1) presented an instrumenta-tion problem combined with the possible confusion of bands obser-ved at such low frequencies with lattice modes. Furthermore, the very intense Raman band at ~282 cm"1 observed for the [Au(S0 3F) 4]" ion is in the same region where the stretching fundamentals for B r 3 + are expected. (In superacid solutions, only one peak at ~290 cm - 1 was reported for B r 3 + , attributed to a coincidence of the two stretching modes) 5 5 » 1 8 3 . A relatively strong Raman band at 220 cm - 1 observed for Br 3 [Au(S03F)iJ may be due to the bending mode of B r 3 + . Two other weak bands are found at 188 and 178 cm"1 Table 4.4 VIBRATIONAL FREQUENCIES OF Br n [Au(SQ 3 FKl Br 3[Au(S0 3 F K ] R, ~80 K IR, ~80 K Br 5[Au(S0 3 F K ] R, ~80 K IR, ~80 K Cs[Au(S0 3 F) 4 ] R? ~298 K 1408 ms 1 1386 m 1410 vs,sh 1392 vs 1409 m 1392 m 1380 s,b 1404 m 1380 w,sh 1218 s 1192 m 1220 s 1190 s,sh 1224 s 1201 mw 1200 s,b 1239 vs 1205 w 1016 m 960 vw 1020 w 955 vs,sh 930 vs 1020 m 1014 m 960 s,sh 920 s,b 1011 s 960 vw 829 mw 814 w 830 s,b 820 vw 800 s 830 mw 814 w 645 s 678 s 650 s 645 s,sh 670 s 649 vs 590 vw 585 s 575 s 580 w 545 w 548 s 540 m 550 w 520 w 520 vw 454 s 462m 452 ms 450 m 452 s 405 w 400 vw 405 w 319 w 286 vs 300 ms 285 w 304 m 294 m 282 m 305 m 295 m 275 s,sh 267 m 279 vs 260 w 282 vs 220 s 188 w 178 w,sh 167 w 150 w 205 ms,b 150 m 150 m 132 in the Raman spectrum, with a band at 300 cm - 1 observable also in the i . r . spectrum. The assignment of these weak bands is most like l y lattice modes. Br 5 + , because of its low symmetry (C^ or even Cj), is ex-pected to have nine vibrational modes, a l l i . r . and Raman active. (VSEPR predicts a linear Br3 unit with bent terminal bromine at-tachments at each end free to rotate). For Br 5[Au(S0 3F) iJ , bands that may be due to B r 5 + are found at 304, 295, 267 and 205 cm - 1 in the Raman spectrum, and at 305, 295 and 260 cm - 1 in the i . r . spectrum. It is noteworthy that the Raman spectrum does not provide any evidence for the presence of free Br 2, with v(Br-Br) expected at 320 cm"1 2 8 . Therefore, the presence of B r 5 + can best be arrived at by default of a l l the other possibilities that the five bromine atoms can exist in, especially since there is no evidence for a breakdown of the [Au(S0 3F ) i J" anion and an involvement of bromine in the formation of Au-Br bonds. Unfortunately, both the poly-bromine-containing complexes are polycrystalline and a l l attempts at a crystal structure determination have been unsuccessful. 4.C.3 MAGNETIC SUSCEPTIBILITY Except for a paramagnetic contaminant present in samples of Au(S0 3F) 3 prepared by the pyrolysis of Br 3[Au(S0 3F ) i J , the 133 tris(fluorosulfate) is diamagnetic. Diamagnetism is also observed for the two polybromine-containing complexes, further confirming the presence of a square planar low-spin electronic configuration for Au(III). The magnetic susceptibility results are listed below in Table 4.5. Table 4.5 MAGNETIC SUSCEPTIBILITIES OF A u ( S 0 3 F ) 3 AND B r n [ A U ( S 0 3 F ) L J x g(cgs units) Au(S0 3F) 3 -(2.94±0.03)xl0-7 Br 3tAu(S0 3F) 4] -(2.88±0.03)xl0-7 Br 5[Au(S0 3F) i + ] -(3.34±0.15)xl0-7 x m(cgs units) z x d i a -(147±2)xl0 - 6 - 1 5 2 x l 0 - 6 -(240±3)xl0-6 - 2 8 4 x l 0 - 6 -(331±15)xl0-6 -345xl0" 6 4.C.4 SOLUTION STUDIES IN HS03F With the exception of Li [Au(S03F)tJ , a l l complexes contain-ing [Au(S03F)i t]" are very soluble in HS03F at room temperature. As expected for a polymeric material, Au(S0 3F) 3 dissolves only very slowly; but when the i n i t i a l l y produced suspension is a l -lowed to stand for several hours at room temperature, or i f heating is applied at ~100°C for ~1 h, clear solutions can be formed. In addition, when Au(S0 3F) 3 is prepared from S 20 6F 2/ HS03F, the removal of a l l S 20 6F 2 results in the formation of very concentrated solutions of Au(S0 3F) 3 in HS03F (~7 m) and these can be easily diluted for solution studies. 134 4.C.4.1 ELECTRONIC SPECTRA With the exception of the polybromine complexes, a l l solu-tions containing the gold-fluorosulfato species are yellow to deep orange in color, depending on the concentration. Both Au-( S 0 3 F ) 3 and Cs [AU(S0 3F)LJ give rise to identical absorption spec-tra in HS03F, with A m 3 V at 278 nm and e of 1.5X101* M^cm"1. This 3 max rather broad band, best interpreted as due to charge transfer, extends well into the visible region and obscures any band which may be due to d-d transitions (the half-width of the band is about 120 nm). In [Pd(S0 3F)iJ 2~, the charge transfer bands are higher in energy, presumably due to the lower effective nuclear charge on the metal, and therefore bands arising from d-d transitions are observed. Solutions of the polybromine-cation-containing complexes, although reddish brown in color when they are concentrated, gave identical spectra with * m a x =278 nm. Bands due to bromine con-taining species, ( B r 3 + at 375 nm, Br 2 at 410 nm, and B r 2 + at 510 nm) 2 8 » 5 8 , a l l with e of the order of ~1600 M^cm - 1, could not be detected. 4.C.4.2 CONDUCTIVITY MEASUREMENTS The results of electrical conductivity measurements on d i -lute solutions of K[Au(S0 3F) 4] and Au(S0 3F) 3 in HS03F are shown 135 in Fig. 4.3 and 4.4 respectively. While both solutes behave as strong electrolytes in H S O 3 F , as indicated by the almost linear concentration dependence of the specific conductance, their ef-fects on the dissociation equilibrium of the solvent are quite different. It appears that the ionization of K[Au(S0 3F)tJ in H S O 3 F can best be described according to: K[Au(S0 3F)J K + ( s Q l v ) + t A u ( S 0 3 F ) J - ( s o l v ) (4.36) Although the basic dissociation of [AutSOsF)^]- according to: [Au(S0 3 F K r ( s o l v ) A u ( S 0 3 F ) 3 ( s o l v ) + S 0 3 r ( s o l v ) (4.37) is possible, the conductometric titration of Au(S0 3F) 3 with KS03F, (the reverse of the above reaction) giving a minimum con-ductance at a mole ratio of 1:1, does not provide any supporting evidence for such a dissociation occurring to any great extent; furthermore, the equilibrium constant for equation (4.37) thus calculated is of the order of 10"6, On the other hand, Au(S0 3F) 3 exhibits a much higher conduc-t i v i t y in HS03F, and as can be seen from Fig 4.4, i t approaches the conductivity observed for solutions of SbF2(S0 3F) 3 2 k . Since this can only be due to the presence of large amounts of either H 2S0 3F + or S03F", as they possess very high ionic mobilities in HS03F 1 1 , Au(S0 3F) 3 is therefore either a strong acid or a strong base. With preliminary information concerning the Lewis acidity of Au(S0 3F) 3 obtained from the synthesis of complexes containing FIG. 4.3 CONDUCTIVITY OF [Au(S03F)4]-COMPLEXES IN HS03F FIG.4.4 CONDUCTIVITY OF Au(S0 3F) 3 IN HS03F m (mol/kg) 138 [Au(S03F)4], i t is not surprising therefore to find that Au(S0 3F) 3 behaves as a strong acid by i t s conductometric titration with KS03F solutions, shown in Fig.4.5 and Table 4.6. As expected for the titration of a strong acid (the only other example of such a moiety in HS03F being SbF 2(S0 3F) 3) 2 h , with a strong base, the conductance decreases gradually upon the addition of KS03F, until the endpoint at the KS0 3F/Au(S0 3F) 3 ratio of 1 is reached. The coincidence of the endpoint with the point of lowest conductance is typical for a strong, monoprotic acid in a protonic solvent. In comparison, SbF 5, a weaker acid, reaches its minimum conduc-tance in a titration at a KS03F/SbF5 ratio of ~0.4, and the re-gion of low conductance is broad and ill-defined 2 4 . The titration can be formulated as the neutralization of a strong acid H[Au(S0 3F) i +] or its equivalent, with a strong base KS03F, according to: KS03F + H[Au(S0 3Fk] • K[Au(S03F)lt] + HS03F . (4.38) Whether H[Au(S03F)tt] or solvated Au(S0 3F) 3 is actually the acid present in solution is immaterial to this discussion, since they are both nonconducting species. Since KS03F solutions were used in the titration instead of solid KS03F, the total concentration of the system decreased as the titration proceeded. This is the major factor contributing to the curvature of the conductivity plot; a correction, although not entirely theoretically j u s t i f i e d , can be made and is also shown in Fig. 4.5. It can be seen that this second curve, in FIG.4.5 0 0.5 1.0 K/Au TABLE 4.6 CONDUCTOMETRIC TITRATION OF AufSQgF), WITH KSO,F IN HSO.F R Kobs < c a l c l Kobs' Kcalcl I (calc) 0 99.92 99.92 0 0.031 0.1 86.82 86.72 0.10 0.030 0.2 75.21 75.04 0.16 0.030 0.3 64.84 64.64 0.20 0.030 0.4 55.52 55.30 0.22 0.030 0.5 47.09 46.88 0.21 0.030 0.6 39.43 39.23 0.20 0.030 0.7 32.42 32.26 0.16 0.029 0.8 25.99 25.88 0.11 0.029 0.9 20.07 20.01 0.06 0.029 1.0 15.60 15.60 0 0.029 lO^n-^m"1 lO^n^cm - 1 lO-'+fWcnr1 mol.kg - 1 141 which the conductance has been corrected to correspond to the i n i t i a l concentration of Au(S03F)3, is almost a straight line, especially at the i n i t i a l points. The fact that the concentration corrected curve is not a straight line can be caused by: a) the polymerization of Au(S0 3F) 3, b) the dependence of molal conductivity, X , with respect to concentration, and/or c) the incomplete ionization of the superacid in solution. The polymerization of the ansolvo acid, found in SbFs and SbF2-(S0 3F) 3 2 U , does not seem to occur in the HS0 3F-Au(S0 3F) 3 system, at least not in the dilute solution (~0.04 m) range that the t i -tration was performed in. For an ionic electrolyte in dilute solutions, the molal conductivity is expected to decrease as the ionic strength, I, which can be approximated by molality for a 1:1 electrolyte (I=%im-jZj2, where z is the ionic charge), increases. This rela-tionship, which is really only applicable at ionic strengths of less than 10~3 m, can be represented below 1 8 4 : x* = xj - kl% where x* = x* as I —+ 0 ic and k = func(x o, e , T, n, z). At higher ionic strengths, X* decreases much faster than can be described by the above equation. In this particular t i t r a t i o n , by coincidence, the ionic strength remains relatively constant 142 throughout the ti t r a t i o n , and therefore the variation of X* due to changes in I is not very serious. Assuming K+ and [Au(S0 3F)iJ _ are the only conducting species in solution at the endpoint, (this is not entirely correct and a solvent conductivity of l x l O - 4 fT^m - 1 has been subtracted from the interpolated K of 1 . 5 6 x l 0 - 3 o^cm - 1 at the endpoint), some estimates concerning the X* values and the equilibrium constant for the acid dissociation of H [Au(S0 3F)iJ according to: H [ A u ( S 0 3 F K ] + HSO3F H 2S0 3F + + [Au(S0 3F) 4] ", (4.40) can be made. Using the literature value of 320 for x*(H2S03F+) and a value of 30 for X*(K+) 2 1 , and applying the X* correction procedure described in Appendix A, a value of 24 is obtained for X*( [Au(S0 3F) l +] -), together with Ka of 0.051 mol/kg. The X* value for the anion is in a reasonable range for a singly charged ion in HS0 3F; by comparison, X*( [SbF 2(S0 3F) 1 +]") is ~13 at 0.086 m'2tf. This latter low value may be attributed to i t having been obtain-ed from a solution of much higher ionic strength and perhaps, more importantly, the larger degree of solvent-solute interaction in the antimony system, as illustrated by i t s tendency towards dimerization. The correspondence between the calculated conductance, using the above parameters, and the experimentally obtained data is quite good, considering the crude adjustments made to the X* values. The remaining deviations, which are also listed in Table 4.6, are quite small and can be rationalized in terms of non-143 idealities in the solution behavior of the species. An earlier conductometric titration of AuF3 in HSO3F with KSO3F gave a conductivity minimum at a KS03F/AuF3 ratio of ~0.4 1 6 h . It is lik e l y that the titration is not between KS03F and AuF3, but between KS03F and a series of mixed fluoride-fluorosulfate of of Au(III) formed by the solvolysis of AuF3 in HSO3F: AuF3 + HSO3F * AuF ?(S0 3F) + HF , etc. (4.41) The HF formed, being a base in HSO3F 2 1 , would increase the basi-city of the solution according to: HF + HSO3F - »• H 2F + + SO3F- . (4.42) The conductivity result on the Br n +[Au(S0 3F) t t] species is also shown in Fig. 4.3. As can be seen, they have an almost iden-tic a l conductance/concentration dependence as K[Au(S0 3F ) t i ] in HSO3F. This observation is not totally unexpected for either of the two complexes, i f they ionize as 1:1 electrolytes; for Br 3-[Au(S03F)it], the following dissociation is most lik e l y : Br 3[Au(S0 3F) 1 +] ( s ) B r 3 + ( s o l v ) + [ A u ( S 0 3 F ) i + ] ~ { s q ] y ) .(4.43) However, for Br 5[Au(S0 3F) iJ , although a similar dissociation is consistent with the observed conductivity, i t is not expected to occur for a number of reasons. F i r s t l y , B r 5 + has never been observed in solution, and excess Br 2 can be reportedly removed from superacid solutions containing B r 3 + by evacuation 2 8 . Sec-ondly, when Br 5 [Au(S03F)iJ is dissolved in HSO3F, the evolution of Br 2 is evident from the color of the gas phase above the solu-tion. This suggests a dissociation of Brs + into B r 3 + and Br 2, 144 the latter being a non-electrolyte, would give rise to a similar conductivity as Br 3[Au(S0 3F)iJ solutions: Br 5[Au(S0 3F),] ( s ) B r 2 ( s o l v ) r B r 3 + ( s o l v ) + r A u ( S ° 3 F K r ( s o l v ) . (4-44) 4.C.4.3 N.M.R. SPECTRA The 19F-n.m.r. spectra of solutions of Au(S0 3F) 3 and Cs[Au-(S0 3F)iJ are shown in Fig. 4.6. All spectra show the presence of two resonances: the strong one at ~-40 ppm due to HS03F and a weaker one slightly downfield assigned to Au(III)-fluorosulfate-containing species. The spectrum for the Cs[Au(S0 3F)iJ solution is very straight forward, with the weaker peak due to [Au(S0 3Fk] _ arising as a result of a dissociation similar to the type of equation (4.36). The presence of two very sharp signals indicates negligible ex-change in the system. This is most likely a consequence of the formation of a coordinatively saturated species with l i t t l e or no dissociation giving rise to other S0 3F-containing moieties. This hypothesis is also supported by the observation that the solvent peak at -40.8 ppm is not shifted from that of pure HS03F (-40.6 ppm) within experimental errors. For Au(S0 3F) 3, the n.m.r. spectra of i t s solutions in HS03F also show two resonances, but they seem to indicate a much more complex solution behavior. The spectrum for the dilute solution already shows 1ine broadening for both resonances. When the con-145 FIG 4.6 19F-N.M.R. SPECTRA OF Au(S0 3F) 3 AND Cs[Au(SQ3F),J IN HSQ3F CsAu(S0 3 F) 4 0 . l 4 no K CFCl . A u ( S 0 3 n 3 0.3 m 290 K so 40 A u ( S 0 3 F ) 3 0.6 m 310 K 30 20 10 0 ppm 146 centration is increased the position of the solute resonance becomes increasingly uncertain due to line broadening, and ~2 m, the peaks are found at -41.2 and ~-46.5 ppm. For a more concentrated solution (~7 m), the two resonances collapse into a single, extremely broad, (half width of ~5 ppm) and asymmetric peak. Although this concentration dependence is also partly due to viscosity effects, the presence of two sepa-rate resonances for dilute solutions and the asymmetric single resonance in the concentrated solution argues against a rapid ex-change rate for equation (4.40). Since i t is not expected that the exchange of a proton would drastically alter the magnetic environment for fluorine in a fluorosulfate group, its effect on line broadening is not as great as that obtained i f a SO3F group actually dissociates from [Au(S03F)i+]". The latter process pre-sumably takes place rapidly in the HSO3F-I(S0 3F)3 system, as a single resonance was observed in its 1 9 F - n.m.r. spectrum even at -90°C 3 6 . The appearance of a single 1 9F-resonance for the solute in the HS03F-Au(S03F)3 is consistent with the presence of a mono-meric species (at least in dilute solutions), rather than SO3F bridged species. Again, due to line broadening, combined with the expectation that the chemical environment for the fluorine atoms in both configurations should not be very different, such a conclusion may not be warranted from these results alone. In addition, a downfield shift of the solvent resonance to 147 10.30 ppm was observed in the ^-n.m.r. spectrum of the concen-trated solutions, while for dilute solutions, the resonance remains unchanged at 9.93 ppm. Since the resonance for the pure solvent is already broad due to extensive hydrogen bonding, no further broadening was discernable. The downfield shift of the resonance, however, has a precedence in the H2S03F-H[B(S03H)iJ superacid system 1 8 5 . 4.C.4.4 VIBRATIONAL SPECTRA Raman spectra were obtained on solutions of Au(S03F)3 and Cs [Au(S03F)i t] in HS03F. The observed Raman shifts are listed in Table 4.7. As expected for solution spectra, the bands are rather broad, and since the fluorosulfate group is common to both the solvent and solute, some overlap is encountered. However, the AU-SO3F species in solution is found to be an excellent Raman scatterer, allowing the observation of bands at quite low concentrations. For more concentrated solutions, suppression of the solvent bands occurs, leading to the observed trend in relative intensities. The spectra of Au(S0 3F) 3 in solution are very similar to that obtained for Cs [Au(S03F)iJ in both the solid state and HS03F solution. This suggests a breakdown of the binary fluorosulfate's polymeric structure once i t becomes solvated. Any polymeric fluorosulfate bridged species, observed in the solid state vib-rational spectra of Au(S0 3F) 3, are not evident in even very TABLE 4.7: RAMAN FREQUENCIES FOR Cs[Au(SQ 3F)J, B r g l X S O ^ K ] and Au(SQ 3F) 3 IN HS03F HS03F Cs[Au(S0 3FK] Br 3[Au(S0 3F) 4]* Au(S0 3F) 3 Au(S0 3F) 3 Au(S0 3F) 3 -1 m ~1 m -0.15 m -2.3 m -7 m 1440 ms 1440 m,b 1420 m 1417 w 1430 ms 1420 m,sh 1412 m,b -1360 m,b 1420 m,b -1360 vw 1230 ms.b 1178 m 1232 s 1190 m,sh 1230 vw 1175 s 1060 w 1228 vs 1185 s.sh 1227 vs -1190 vw 1226 vs -1090 vw 3v,Br2+ 960 s 958 m 910 vw 1020 w 968 m 960 vw 910 vw 960 w,b 910 w,b 997 m,sh 960 m 912 m 850 vs 849 s -820 m.sh 850 ms 718 w 848 s 846 m 830 m.sh 840 m 820 m.sh 2v,Br2+ 649 vs 652 ms 647 vs 647 vs 647 vs 560 vs -570 s.sh -570 s,sh 580 -570 m,sh 555 sh 553 s 553 m 550 s 547 m 546 s 453 s 456 m 455 457 s 457 s 405 s 406 s 400 w -395 ms,b -400 w,b -400 w,b 393 s,sh 392 s 391 w 362 m -350 vw,b v,Br2+ 278 vs 284 vs 206 w 276 vs 276 vs 276 vs 6,Br3+ 150 m 156 m 150 mw 151 mw 148 mw *Frozen solution at ~80K 149 highly concentrated solutions of Au(S0 3F) 3 in HS03F. Very weak new bands at ~1360, ~1090 and 350 cm - 1 for the almost saturated (~7 m) solution may be due to combination bands, the presence of small amounts of bridged species or laser induced decomposition products. It should also be mentioned that the most prominent bands at ~645, ~450 and ~275 cm - 1, previously assigned to modes containing Au-0 vibrations, are a l l present in the solution spec-tra as single bands with no noticeable splittings. Overall, the solution spectra of HS0 3F/Au(S0 3F) 3 can best be regarded as the composite of HS03F and [Au(S0 3F) iJ". The concentration range covered by the Raman experiment overlaps well with those used to obtain n.m.r. spectra. However, more dilute solutions were used to obtain conductivity measure-ments. Nevertheless, i t seems rather unlikely that polymerization of Au(S0 3F) 3 should exist at such low concentrations, evidence for such a process can hardly be detected in very concentrated solutions. The results presented here suggest that the species in solution can best be described as H[Au(S03F)tJ in equilibrium with [Au(S0 3F)J~. Finally, as previously reported for pure HS03F 8, bands due to the 0-H stretching vibration could not be detected in the sol-ution Raman spectra obtained here. I.R. spectra, obtained by using Teflon films between BaF2 plates, show a shift of the sol-vent band from 3125 cm"1 1 8 6 to ~2950 cm"1 for a 2.3 m solution. However, the extreme broadness of the band (half width ~700 cm - 1) 150 makes i t d i f f i c u l t to attach much significance to the frequency shift. For the two polybromine-containing species, only Br3 +[Au-(S03F)tJ- was found to be stable in HS03F solutions, and even for this solution, the sample had to be cooled to ~80 K in order to obtain a Raman spectrum. The Raman frequencies are also listed in Table 4.7. The peaks are generally quite broad, like in a typical solution spectrum. All the bands due to [Au(S03F)tJ" and HSO3F can be distinguished, along with a series of resonance Ra-man bands originating at 362 cm - 1, and assignable to B r 2 + . The concentration of B r 2 + must be very low, as bands due to this species are extremely intense due to the resonance Raman effect. The source of B r 2 + may be due to the dissociation of B r 3 + accor-ding to: B r w 1 ^ > Br?"1"/ i \ + %Br 9/ , N , (4.45) 3 (solv) 4 2 (solv) 2 2(solv) ' v ' with the equilibrium lying mostly to the l e f t . The strong band at 284 cm - 1, usually found at 276 cm - 1 for other [Au(S03F)i+] complexes, may contain some contribution from Br3 + vibrations, expected at ~290 cm - 1 2 8 . This upward shift in band position is also noted in its solid state spectrum. Finally, the weak band at 206 cm - 1 may be due to 6 Br3 +, which, because of the solvation of the cation by HSO3F, becomes shifted to a lower frequency from 220 cm - 1 in the solid. 151 4.C.5 RAMAN STUDIES OF Au(S0 3F) 3 IN LIQUID SQ2 Sulfur dioxide, owing partly to its low Lewis acidity and basicity, is commonly used as a solvent in superacid systems, since i t is a relatively poor ionizing solvent which can dissolve organic compounds. It was originally thought that because of the observed high Lewis acidity of Au(S0 3F) 3 towards S03F", adduct-formation according to: Au(S0 3F) 3 + S0 2 • [Au(S0 3F) 3-S0 2] , (4.46) may be possible. Although no such adduct could be isolated, i t was hoped that Raman spectroscopy may provide some evidence for its existence in £-S0 2 solutions. The spectra were recorded in high pressure capillaries containing saturated solutions of Au(S0 3F) 3 in i~S02 at room temperature. The solubility of Au(S0 3F) 3 in £-S0 2 is quite low, but solute peaks with intensi-ties comparable to the two weaker solvent peaks at 1337 and 521 cm - 1 can be obtained. The Raman frequencies for such a solution are listed in Table 4.8. Most compounds containing S0 2 as a ligand are very unstable with respect to S0 2 elimination 1 8 7 . The Lewis acid-Lewis base adduct of SbF 5«S0 2, with an oxygen-bonded S0 2 ligand, is only thermally stable up to ~50°C 1 8 8 . The i . r . spectrum of this complex has bands at 1625(w), 1323(s), 1145(w, sh), 1100(m), 696(s), and 480 cm-1(w) 1 7 \ It has also been detected in the HS03F-SbF5-S02 superacid system by the presence of a band at 1106 cm - 1 in the Raman spectrum 2 2 . It seems that in the presence Table 4.8 RAMAN FREQUENCIES FOR Au(S0 3F) 3 IN SOLID STATE AND IN Jl-S02 SOLUTION AT 298 K solid state z-S02 solution Assignment 1449 m 1423 m,sh 1427 m 1385 m 1337 s v a s £-S0 2 1227 vs 1276 vw 1224 vs 1145 vvs 1122 m v s £-S0 2 v s £-S0 1 60 1 8 1102 w 1053 m 955 m,sh 936 s 899 s 1071 w 1031 w 970 m 915 m 822 w 820 w 650 vs 657 m,sh 610 w 583 546 m 543 w,sh 521 s 6 a-S02 465 m 444 w 458 m 348 m 327 w,sh 365 w 292 vs 285 s 286 vs 272 vs 153 of liquid S0 2 as the solvent, the only diagnostic Raman band for 0-bonded S0 2 is a band at ~1100 cm-1. Most transition metal -S02 complexes are S-bonded, for [Ru(NH3)1+(S02)Cl ]C1, the S0 2 vibrations are found at 1301(s), 1278(s), 1110(s), and 555 cm-^m), in the i . r . 1 8 9 . A comparison of the Raman frequencies of solid Au(S0 3F) 3 and i t s solution in z-S02 reveals that except for some small shifts in band positions, the spectra are almost identical. The band at 1122 cm"1 is due to the isotope effect of S 1 60 1 80 and is confirmed by its presence in the Raman spectrum of pure £-S02. This band, although expected to be ~10"3 times as intense as the band due to the more abundant S 1 60 2, arises because this symmetric mode is ~103 times more intense than the bands in the rest of the spectrum. Some increased splittings in the [AuOiJ related vibrations are observed and may be interpreted in two opposite ways: a) In the formation of smaller oligomers such as Au 2(S0 3F) 6, the coupling between the [AUOLJ vibrations in adjacent [AU(S0 3F)LJ units via S0 3F bridging groups may be stronger than in the solid state where increased cross-linkage serves to diffuse the coupling. b) If Au(S0 3F) 3'S0 2 is formed, the replacement of a bidentate S0 3F group by S0 2 will lead to a greater distortion in the AuO^ unit from Dt+h symmetry, leading to an increased splitting in the bands. The band at 365 cm - 1 can conceivably be due to vAu-OSO or 154 vAu-S02, but the position of vAu-(S03F bridging) at 348 cm - 1 in Au(S0 3F) 3 makes such an assignment rather ambiguous. In the S0 3-stretching region for bridging S03F group at ~1000 cm - 1, the relative intensity of the three bands at 1031,970, and 915 cm - 1 is very similar to that found for solid Au(S03F)3-; quite unlike that observed for monodentate S03F groups such as [Au(S03F)iJ Therefore, i t does not appear that the relative proportion of monodentate/bidentate S03F groups changes upon the dissolution of Au(S0 3F) 3 in 2-S02. With direct evidence supporting the formation of an adduct absent, i t must be concluded that i f i t does exist, i t must be present in quite a low concentration, The species most lik e l y to be present in ^-S02 are smaller oligomers of Au(S0 3F) 3, such as Au 2(S0 3F) 6. 155 4.D CONCLUSION Au(S0 3F) 3 has been shown to act as a strong ansolvo-acid in HS03F, giving rise to a new superacid system. From conductivity measurements and the conductometric titration of H[Au(S03F)iJ with KS03F, the acidity of the new superacid is found to be comparable to that of H[SbF 2(S0 3F)i f]. However, unlike the anti-mony system, no extensive polymerization of the ansolvo-acid is evident in solution, although Au(S0 3F) 3 i t s e l f is a polymer, with bridging S03F groups. A series of ionic salts containing the [Au(S0 3F)iJ" anion + + can be synthesized with alkali metal ions, NO and C102 . These have been characterized using vibrational spectroscopy. The Lewis acidity of Au(S0 3F) 3 with respect to S03F abstrac-tion can be demonstrated in i t s reaction with the halogen t r i s -(fluorosulfate)s of Br(S0 3F) 3 and I(S0 3F) 3,to give the correspon-ding [Hal(S0 3F) 2] + cations; in this system, extensive anion-cation interaction is evident from the proliferation of bands in the vibrational spectra, suggesting the presence of bridging S03F groups. The low oxidizing power of Au(S0 3F) 3 is illustrated by the successful synthesis of polybromine cations-Br 3 + and B r 5 + , stab i l -ized by [Au(S0 3F)iJ" The stability of B r 3 + in the gold complex with respect to dissociation is higher than a l l previously repor-ted examples of B r 3 + complexes, and the B r 5 + cation is the only such example so far. The st a b i l i t y of solutions of Br 3[Au(S0 3F) iJ in HSO3F also points to the low basicity of the [Au(S03F)iJ anion in solution, a property essential to any further attempts in the stabilization of other novel cations. Au(S03F)3, [AuCSOsF)^]", and the two polybromine cations are a l l diamagnetic, suggesting a square planar coordination of S03F for Au(III) giving rise to a !A Xg electronic ground state. 157 CHAPTER 5 PLATINUM-FLUOROSULFATE 5.A INTRODUCTION The chemistry of platinum has been studied extensively in the past and is one of the earliest systems investigated in coor-dination chemistry. The principle oxidation states of platinum are +2 and +4; the higher states of +5 and +6 occur only in fluorine-containing compounds. The investigation into the chemis-try of PtF 6, one of the most powerful oxidizing and fluorination agents, led to the discovery of the f i r s t noble gas compound. Platinum forms a wide variety of complexes with platinum in both very high and very low oxidation states. Simple anionic complexes containing [Pt-Halt+]2~ and [Pt-Hal 6] 2~, together with [PtF 6] - and [ P t 2 F 1 1 ] ~ are known and can be easily obtained. Their existence suggests halide ion accepting properties of the binary platinum halides. In contrast to palladium, the relative ease with which the +4 oxidation state can be accessed is illustrated by the exist-ence of PtBrti and P t l ^ , formed by direct halogenation of the metal. The reaction takes place even when neither Br 2 nor I 2 are usually considered to be strong oxidizing agents 1 9 0 . Some structural aspects and some principle features of the halides of platinum will be reviewed briefly. As mentioned before, most of the compounds in the series of Pt-Hal 2, Pt-Hal 3 and Pt-Hal 1+ are known but in keeping with the 158 trend observed for palladium, the trivalent oxidation state of platinum is very uncommon and the trihalides are best regarded as mixed valency compounds of P t ( I I ) P t ( I V ) - H a l N o t unexpected, the d i - and tetra-halides are a l l diamagnetic, implying a square planar d 8 electronic configuration for Pt(II) and an octahedral low-spin d 5 configuration for Pt(IV). In general, Pt(IV) com-pounds are diamagnetic because of the high ligand f i e l d splitting possible with the 5d metal in such a high oxidation state. The 'trihalides', with the exception of P t 2 F 6 , only recently synthe-sized 1 9 1 , are also diamagnetic, providing one the most conclu-sive evidence for the mixed valency nature of the compounds (Pt(III), d 7, with an odd number of electron, must be paramagne-t i c ) . P t 2F 6, a black solid whose existence was suggested at a very early time in fluorine chemistry by Moissan, is the only example for Pt(II) in an octahedral environment with a 3A 2g elec-tronic ground state and a magnetic moment of 3.05 yg. PtF 2 has not been successfully synthesized so far, and i t has been sug-gested that this fluoride may not be stable with respect to a disproportionate into platinum metal and PtFi* 1 9 2 . The two higher-valent fluorides, [PtFsl^ and PtF 6, are ob-tained by the fluorination of platinum, or lower-valent platinum halides, at different conditions 1 9 2 » 1 9 3 . Both have found use as fluorinating agents in the synthesis and stabilization of easily reduced and novel cations such as 0 2 +, XeF+ and C1F 6 + 1 2 2 , 1 9 ^ Fluoridation reactions involving PtF 6 usually lead to the forma-159 tion of [PtF 6]~ or [P t 2 F n ] " , both are anions of low basicity. The solution behavior of PtF^ in HS03F, like that of AuF3, was studied by Woolf; however, due to the extremely low s o l u b i l i -ty of PtFit in HS03F, no quantitative results were obtained, even though acidic behavior was suggested 1 6 4 . The only known fluoro-sulfate of platinum, PtCSOsF)^, was synthesized in a method ana-logous to the preparation of Au(S0 3F) 3 from a reaction betwe-en the metal and BrS03F 8 2 . The i n i t i a l product from the above reaction with BrS0 3F is likewise reported to give an adduct of the approximate composition of PtCSOsF)^-4BrS03F. These rather preliminary results raise the possibility that PtCSOsF)^ may act as a S03F~ acceptor, leading to the formation of [Pt(S0 3F) 6] 2~ and therefore a superacid system in HS03F. 160 5.B EXPERIMENTAL 5.B.1 PREPARATION OF Pt(S0 3FK Pt + S 20 6F 2 ^ 4 Pt(S0 3FK (5.1) In a typical reaction, platinum powder (233 mg, 1.19 mmol) was reacted with a mixture of S 20 6F 2/HS0 3F at 120°C for 2 days. The orange colored solution gave, after the removal of a l l volatile materials, dark yellow crystals of Pt(S0 3F) t f, (692 mg, 1.07 mmol). Pt(S0 3F ) i t is a deep yellow to light orange crystalline, hy-groscopic solid. It melts with decomposition at 220°C and is very soluble in HS03F. The micro-analysis of the compound had been reported previously 8 2 . Pt(S0 3F)t t can also be prepared by the reaction of platinum metal with BrS03F followed by the pyrolysis of the intermediate as reported earlier by Cady et al 8 2 . 5.B.2 SYNTHESIS OF COMPLEXES CONTAINING [Pt(S0 3F) 6] 2~ 5.B.2.1 PREPARATION OF Cs 2 [Pt(S0 3F) 6] CsCl + HS03F > CsS03F + HC1 (5.2) 2CsS03F + Pt + 2S 20 6F 2 ^ £ Cs 2 [Pt(S0 3F) 6] (5.3) CsS03F, formed by the solvolysis of CsCl (302 mg, 1.80 mmol) in HS03F, was mixed with a stoichiometric amount of platinum metal powder (175 mg, 0.897 mmol) and reacted with S 20 6F 2 (~3 mL) 161 and HSO3F (~5 mL). The reaction required heating at 80°C for 3 days, after which time a light yellow solution had formed. The evacua-tion of a l l volatile materials yielded crystalline Cs 2[Pt(S0 3F) 6]. Cs 2[Pt(S0 3F) 6] is a light yellow, crystalline, hygroscopic solid. It is soluble in HS03F and melts with decomposition at ~260°C. Analysis Cs Pt S F Calculated % 25.19 18.49 18.23 10.80 Found % 25.03 18.39 18.32 10.72 5.B.2.2 PREPARATION OF (C10 2) 2[Pt(S0 3F) 6] 2C102S03F + Pt(S0 3FK • (C10 2) 2[Pt(S0 3F) 6] (5.4) C102S03F (~1 mL) was added to a sample of Pt(S0 3FK (828 mg, 1.400 mmol). A portion of Pt(S0 3F) i t was found to dissolve in the liquid which was heated at ~60°C for 1 day. The removal of a l l volatile materials at ~70°C gave (C10 2) 2 [Pt(S0 3F) 5]. The compound can also be prepared by the reaction of p l a t i -num metal with a mixture of S 20gF 2 and C102S03F according to: Pt + 2C102S03F + 2S 20 6F 2 — — • (C10 2) 2 [Pt(S0 3F) 6] (5.5) at 160°C for 2 days followed by the removal of a l l volatile mater-ials at 70°C. (C10 2) 2[Pt(S0 3F) 6] is a light yellow, crystalline, hygrosco-pic solid. It is soluble in HS03F and melts with decomposition at ~195-200°C. 162 Analysis Cl Pt F Calculated % 7.67 21.11 12:33 Found % 7.60 21.24 12.42 5.B.2.3 PREPARATION OF Ba[Pt(S0 3F) R] BaCl 2 + 2HS03F • Ba(S0 3F) 2 (5.6) Ba(S0 3F) 2 + Pt + 2S 20 6F 2 • Ba[Pt(S0 3F) 6] (5.7) Anhydrous BaCl 2 (88 mg, 0.423 mmol) was converted into the fluorosulfate by the reaction with HS03F (~5 mL). After the removal of the HC1 evolved, S 20 6F 2 (~2 mL) was d i s t i l l e d onto the equimolar mixture of Ba(S0 3F) 2 and platinum metal powder (82 mg, 0.420 mmol) in HS03F. After heating at 140°C for 3 days, a l l the metal had dissolved and a light yellow precipitate had formed. The removal of a l l volatile materials yielded Ba[Pt(S0 3F) 6]. Ba[Pt(S0 3F) 6] is a light yellow, hygroscopic powder. It is only sparingly soluble in HS03F and melts with decomposition at ~190°C. Analysis Ba Pt F Calculated % 14.82 21.05 12.03 Found % 14.83 21.03 12.27 5.B.2.4 PREPARATION OF ( B r 3 ) 2 [Pt(S0 3F) 6] Pt + 6BrS03F (Br 3) 2[Pt(S0 3F) 6] (5.8) Platinum metal powder (157 mg, 0.805 mmol) was reacted with a large excess of BrS03F (~10 mL) at 140°C for 7 days. About 1 mL 163 of Br 2 was added into the mixture which was heated at ~70°C for about 5 minutes. The subsequent removal of a l l volatile materials at room temperature yielded a compound analyzed as (Br 3) 2[Pt(S0 3F) G] (998 mg, 0.787 mmol). (Br 3) 2[Pt(S0 3F) 6] is a dark brown, polycrystalline, and hygroscopic solid. Prolonged standing at room temperature in vacuo causes the slow evolution of Br 2, but the compound is rea-sonably stable in an atmosphere of nitrogen where i t decomposes at ~90°C. Analysis Br Pt F Calculated % 37.78 15.37 8.98 Found % 38.05 15.36 8.84 164 S.B.3 PREPARATION OF CsPt(S0 3F) 5 CsCl + HSO3F • CsS03F +.HC1 (5.9) CsS03F + Pt + 2S 20 6F 2 CsPt(S0 3F) 5 (5.10) Stoichiometric amounts of CsCl (102 mg, 0.606 mmol) and platinum metal powder (118 mg, 0.605 mmol) are mixed in a reactor. HSO3F (~2 mL) is added to solvolyze the CsCl, and the resultant solution evacuated to remove a l l HS03F. A mixture of S 20 6F 2/HS0 3F (~4 mL) is then added and the reaction vial held at 80°C for 3 days, after which time a l l the metal had dissolved and a light orange solution had formed. The removal of a l l volatile materials yielded a compound which analyzed as CsPt(S0 3F) 5 (503 mg, 0.608 mmol). CsPt(S0 3F)5 is a light yellow crystalline solid, soluble in HS03F. It melts with decomposition at ~154°C to an orange liquid. Analysis Cs Pt F Calculated % 16.11 23.35 11.53 Found % 15.93 23.41 11.63 165 5.C DISCUSSION 5.C.1 SYNTHESIS AND GENERAL DISCUSSION  5.C.1.1 PLATINUM TETRAKIS(FLUOROSULFATE) PtfSOsFK can be prepared, in a manner analogous to the synthesis of Au(S03F)3, by the reaction of platinum metal with a) S 20 6F 2 / H S O 3 F , or b) BrS03F followed by the subsequent decom-position of the bromine-containing intermediate. Method a) is the preferred synthetic route as Pt(S0 3F) 4 is formed directly. Samples of the compound obtained via b) were always darker in color (Cady reported the color of Pt(S0 3F)i , . as being dark brown) and contained paramagnetic impurities. It seems that the high temperature (~120°C) involved in the pyrolysis of the intermedi-ate (although the intermediate show signs of decomposition at ~90°C, a higher temperature was needed in order to shorten the time for the reaction), may also lead to the thermal decomposi-tion of the desired product. This intermediate, again similar to the Au(S0 3F) 3-BrS0 3F system, is best regarded as (Br 3) 2[Pt(S0 3F) 6] instead of an adduct between Pt(S0 3F ) i t and BrS03F. There are, however, a few major differences between the synthesis of Pt(S0 3 FK and Au(S0 3F) 3. While both metals can be synthesized by the same methods, platinum displays a much lower degree of reactivity towards either BrS03F or S 20 5F 2/HS0 3F. Furthermore, the removal of HS03F from solutions containing Pt(S0 3F ) i + proceeds only with d i f f i c u l t i e s at room temperature, suggesting possibly a greater degree of solvent-solute interac-166 tion in the platinum system. 5.C.1.2 COMPLEXES CONTAINING [Pt(S0 3F) 6] 2~ 2 + The synthesis of the three ionic complexes containing Ba , C10 2 + and Cs + follows a general scheme to the preparation of transition metal fluorosulfato complexes. Their synthesis pro-vides the f i r s t evidence for the fluorosulfate-accepting property of Pt(S0 3F) 4. Complexes of a less ionic nature, exemplified by Pd[Pt(S0 3F) 6], discussed in Chapter 3, and Ag[Pt(S0 3F) 6] 8 4 , can also be readily prepared. The reaction of BrS03F with platinum metal leads to the formation of a complex containing the B r 3 + cation. Although this reaction supposedly proceeds via an analogous mechanism as the corresponding reaction in the gold system, some differences exist. The formation of (Br 3) 2[Pt(S0 3F) 5] according to equation (5.8) requires that a l l the bromine formed from the reduction of BrS03F be used up, while in the reaction involving gold metal, and excess of half a mole of Br 2 is l e f t per mole of gold reacted. Since B r 3 + is not expected to be very stable in the presence of BrS0 3F, a strong oxidizing agent, the complete conversion of a l l Br 2 into the complex may not occur. Experimentally, yields corresponding to as low as a 1:1 mixture of Pt(S0 3F)i t and (Br 3) 2[Pt(S0 3F) 6] have been obtained for reactions involving large excess of BrS03F. The situation is further aggrevated by the observed low thermal sta b i l i t y of the complex in vacuo. The synthesis of analytically 167 pure (Br 3) 2[Pt(S0 3F) 6] therefore requires the introduction of a small quantity of Br 2 (~1 mL) into the reaction mixture after a l l the metal had dissolved. The preparation can also be accomplished by the in situ synthesis of BrS03F in the reaction vial using S 20 6F 2 and a slight excess of Br 2. This latter method eliminates the i n i t i a l vacuum d i s t i l l a t i o n of BrS03F, which has a low vapor pressure at room temperature. The thermal stability of (Br 3) 2[Pt(S0 3F) 6] is comparable to that of Br 5 AuCSOsF)^ , both showing signs of decomposition, giving off Br 2, at room temperature in vacuo. It is therefore not surprising that further addition of bromine to B r 3 + , forming longer chained polybromine cations, does not occur in this system. It is possible that the large number of bromine atoms, already six in ( B r 3 ) 2 [Pt(S0 3F) 6], will drastically reduce the lattice energy in a complex such as (Br 5) 2[Pt(S0 3F) 6]. If this postulation is correct, then the lower thermal st a b i l i t y of (Br 3) 2[Pt(S0 3F) 6] as compared to Br 3[Au(S0 3F) 4] is not necessarily a consequence of the lower Lewis acidity of PtCSOsF)^. 5.C.1.3 SYNTHESIS OF CsPt(S0 3F) 5 The synthesis of CsPt(S0 3F) 5 was sought as supporting evi-dence for the polymerization of PtCSOsF)^ as inferred from the conductometric titration of the fluorosulfate with KS03F solu-tions. Being a S03F~ acceptor, forming [Pt(S0 3F) 6] 2~ in the presence of an excess of S03F~, CsPt(S0 3F) 5 is made in a reaction 168 essentially between CsS03F and PtCSOaF)^, a l l the HC1 formed in the i n i t i a l solvolysis of CsCl in HS03F having been removed by evacuating the CsS03F solution to dryness (the formation of a mixture of complexes containing Cs + and C102+ is a li k e l y occur-rence i f any chlorine-containing substance was present). This complex appears to be a dimeric species from an analysis of i t s vibrational spectra. 5.C.1.4 ATTEMPTED SYNTHESIS OF FLUOROSULFATES CONTAINING  PLATINUM IN OTHER OXIDATION STATES The ease with which Pt(IV) could be obtained suggests that Pt(V) fluorosulfate or fluorosulfato complexes might be obtain-able. This was found not to be the case, and no reaction was observed when Pt(S0 3F) 4 was heated at 100°C with S 20 6F 2. It seems logical to assume that, with the reactivity of S 20 6F 2 great-ly enhanced when i t is dissolved in HS03F, the oxidation of a transition metal in S 20 6F 2/HS0 3F should lead to the highest oxida-tion state attainable in the metal-fluorosulfate system. The attempted synthesis of Pt(S0 3F) 2 by the solvolysis of Pt C l 2 in H S O 3 F , consistent with attempted reactions involving other noble metal chlorides examined in this study, with or with-out the presence of KS03F, did not give rise to any reaction pro-duct. The heating of a sample of K 2PtCl i ( in HS03F at ~80°C for 2 days led to the decomposition of the sample , although an ionic fluorosulfate, probably KS03F, identified by i t s i . r . spectrum, 169 was the only fluorosulfate-containing species present. The resi-due gave a positive chloride ion test. Furthermore, the extremely polymeric nature of PtCl2 can be illustrated by an unsuccessful chloride abstraction reaction with Sn(S0 3F) 4, which shows no signs of any reaction after 3 days at ~90°C. While the reduction of Pd 2(S0 3F) 6 with Br 2 gave rise to Pd(S0 3F) 2, such an analogous reaction involving Pt(S0 3F) L f or any of i t s complexes may not be too lik e l y . The formation of ( B r 3 ) 2 -[Pt(S0 3F) 6], containing bromine in the +1/3 formal oxidation state and in the presence of an excess of bromine, seems to indi-cate that the fluorosulfate of Pt(IV), likewise Au(III), is com-patible with Br 2, and therefore no reduction should be expected to take place. 5.C.2 VIBRATIONAL SPECTRA 5.C.2.2 PLATINUM TETRAKIS(FLUOROSULFATE) Due to the low Raman scattering efficiency of PtCSOsF)^ coupled with its limited s t a b i l i t y in the laser beam, only poorly resolved Raman spectra could be obtained. The vibrational fre-quencies of PtCSOsF)^ are compared to the i . r . frequencies of Au(S0 3F) 3 in Table 5.1. As can be seen, a good correspondence between the major bands in the spectra of the two compounds can be made, indicating 170 Table 5.1 VIBRATIONAL FREQUENCIES OF Pt(S0 3F) t t and Au(S0 3F) 3 Pt(S0 3FK Pt(S0 3Fk Au(S0 3F) 3 Sn(S0 3Fk 3 3 R. I.R. I.R. R. Assignment 1445 m ~1400 vs,b 1442 vs 1425 s,sh 1431 s 1420 s,sh ^asS03 1230 vs 1225 vs 1240 s 1220 s,sh 1233 s vsS02 1060 w 900 s,b -1150 s,vb "1070 vs.vb ~1000 vs,vb -900 vs,vb 1135 ms 1055 s 960 s,b 920 s,sh 895 s,b 1124 s 1075 s 986 w 911 m v a sS0 3 -822 vw ~820 s,vb 820 s,b 845 m 827 m VS-F 640 s ~615 w,sh -670 s,b ~640 s,b 682 s 670 s,sh -650 w,sh 610 w 640 w,sh 632 s v S M _ 0 + MO' 580 s 590 s 582 s 589 s S03F def 545 w 540 m 550 s 552 m S03F rock 454 m 460 m 427 s vasM-0 + MO-S def. 293 s 320 w M _ Q + s o F r o c k 268 s s o 171 the presence of similar bonding modes for SO3F groups in them. In the i . r . spectrum of Pt(S03F)Li, a very strong and broad peak extends from ~800 to ~1150 cm-1; some minor features are discern-able only when very thflf-film samples are used. This is in con-trast to Au(S03F)3, for which relatively well resolved bands can be observed in this same region. This complexity in the S-0 and S-F stretching modes of PtCSOaF)^ may not be totally unexpected. If the coordination of Pt(IV) is octahedral, for each Pt(IV), there will be two monodentate and two bridging bidentate S 0 3 F groups. By comparison, Au ( S 0 s F ) 3 would have only two monodentate and one bridging bidentate S 0 3 F groups for each Au(III). This increased proportion of bridging SO3F groups would give rise to a proliferation of vibrational modes, in particular in the S 0 -stretching region, unless their local symmetry is identical, as in SnfSOsF)^. In contrast to Pt(S0 3F) l t, SnfSOsF)^ has relatively well-resolved bands in the ~800 to ~1100 cm - 1 region in its i . r . spec-trum 3 3 , while both compounds are expected to give the same mono-dentate to bidentate S03F ligand ratio. The reason may l i e in the proposed structure for SnfSOaFK, which has trans-monodentate SO3F groups and a square planar arrangement for the bridging S0 3F groups. This would mean that a l l the bridging S03F groups are well ordered and in relatively similar structural arrangements. Conversely, in Pt(S03F) 4, the bridging may not lead to the forma-tion of such a well ordered polymeric structure. This may be 172 illustrated by a difference in the physical properties of two tetrakis(fluorosulfates). While Sn(S0 3F) i t is virtually insoluble in H S O 3 F , Pt(S0 3F) l t is extremely soluble, although the break down of the polymer in the latter case is very slow. 5.C.2.2 COMPLEXES CONTAINING [Pt(S0 3F) 6] 2~ Unlike the parent compound, the three ionic complexes con-taining the [Pt(S0 3F) 6] 2" anion were found to be excellent Raman scatterers, giving rise to well resolved spectra. (Br 3) 2[Pt(S0 3F) 6], however, presented a similar problem as the analogous Au(III) complex; as a result of its dark color and low Raman-scattering efficiency, only poorly resolved Raman spectra could be obtained, even when the sample was cooled to ~80 K. The Raman frequencies of the four [Pt(S0 3F) 6] 2 _ complexes are listed in comparison with those of Cs 2[Pd(S0 3F) 6] in Table 5.2. The i . r . spectrum of (C10 2) 2IPt(S0 3F) 6] is shown in Fig. 5.1. The vibrational spectra of Pd[Pt(S0 3F) 6] have already been discussed in Chapter 3, and are consistent with the presence of anisobidentate bridging S0 3F groups spanning Pd(II) and Pt(IV), both in octahedral coordina-tion spheres. A comparison of the vibrational spectra of [Pt(S0 3F) 6] 2" and [Pd(S0 3F) 6] 2" reveals that the compounds both contain monoden-tate S0 3F groups, and that they appear to be isostructural. Slight shifts at ~630, ~450 and ~280 cm - 1 are presumably due to the different M(IV) cations, supporting the assignment of these TABLE 5.2: RAMAN FREQUENCIES OF [ P t ( S 0 3 F ) 6 ] 2 - COMPLEXES X = ( C * 0 2 ) 2 1406 m 1375 m 1300 vw 1250 vs 1220 vw.sh 1208 m 1045 s 1022 w 1003 w 852 w 826 vw 807 vw 631 s 580 vw 546 vw 517 w 445 m ~410 vw,b 283 s 273 s X = C s 2 1416 w.sh 1410 m 1250 vs 1219 m 1043 s 1010 m -800 w,b 634 vs 579 vw 550 vw 444 m 280 vs X = Ba 1397 w 1386 m 1258 vs 1218 m 1033 m 1012 w 857 w 838 vw 629 vs 583 vw 549 vw 460 m 422 vw 411 w 283 vs X = ( B r 3 ) 2 -1400 vw -1377 w 1229 m 1199 w 1045 w -1005 vw 634 w 450 w 295 m 272 m C s 2 [ P d ( S 0 3 F ) 8 ] 1405 w 1395 vw.sh 1236 vs 1212 m 1020 vs 995 ms 835 vw 805 vw 785 w 618 ms -585 vw.sh 542 w 441 ms -400 vw 270 s Assignment v a s S 0 2 v a s C * 0 2 + v s S0 2 v s C * 0 2 + v a s s o 2 vS-F v M-0 + S0 3F def s 3 def S O 3 F 6Cn0 2 + vacM-0 + S0 3F def aS S O 3 F def Br 3 + v$M-0 + S O 3 F FIG 5.1 The infrared spectrum of (Cl02)2[Pt(SO„ F ) J between 1500 and 400 cm-' 1400 1200 1000 800 600 400 175 vibrational modes as having substantial contributions from the [M06] skeletal vibrations. The i . r . spectrum of (C10 2)2[Pt(S0 3F) 6] shows almost a mutual exclusion of the S O 3 F stretching modes and the MO vibrations with the Raman frequencies l i s t e d , especially in the "1050 to ~900 cm-1 region. Extensive couplings within the same vibrational modes between different S O 3 F groups must be pre-sent to account for the intensity differences. For the C10 2 +-containing complex, v a s C10 2 + are found at 1298 and 1285 cm - 1 in the i . r . spectrum, s p l i t by the two naturally occurring isotopes of 3 7C1 and 3 5C1. v s - and 6- C10 2 + are found at 1045 and 517 cm - 1 respectively, supporting an ionic formulation for the compound. The i . r . spectrum of (Br 3) 2[Pt(S0 3F) 6], down to the low fre-quency limit of ~400 cm"1 for the AgCl windows used, indicates the presence of bands due only to the [Pt(S03F) 6] 2~ anion. In the low temperature Raman spectrum, a l l the major bands due to the anion are also present, together with new features at 295, 355 and ~700 cm-1. The two higher frequency bands show an inten-sity difference that may be due to the Resonance Raman effect of B r 2 + , and the peak at 295 cm"1 compares well with the value of 290 cm"1 in superacid solutions 2 8 and the band at 280 cm - 1 in Br 3[Au(S0 3F) iJ , both assigned as due to B r 3 + . The presence of B r 2 + seems to be suggested in a l l the Raman spectra of B r 3 + -containing complexes studied here; i t may be a decomposition product or in equilibrium with B r 3 + in these systems. 176 5.C.2.3 CsPt(S0 3F) 5 Both the Raman and i . r . spectra of this compound have fea-tures that can be ascribed to the presence of both monodentate and bidentate S0 3F groups. The vibrational frequencies of CsPt-(S0 3F) 5 are listed in Table 5.3 together with those of CsSn(S0 3F) 5 and CsRu(S0 3F) 5. It is evident from the Table that the characteristic strong Raman band at 900 cm - 1 for PtCSOsF)^ is missing from the Raman spectrum of CsPt(S0 3F) 5, and a general simplification of the S-0 stretching region in the i . r . spectrum of the latter compound is also noticed. This argues against the possibility that the com-pound is simply a mixture of PtCSOsF)^ and Cs 2[Pt(S0 3F) 6]. Not surprisingly, the spectra are s t i l l quite complex, as even for a dimeric [ P t 2 ( S 0 3 F ) 1 0 ] 2 " species, there exists three different types of S03F groups equatorial bridging, equatorial monoden-tate and terminal monodentate. A good agreement between the vibrational frequencies of the compounds listed in Table 5.3 suggests the presence of similar S03F groups in a l l three cases. The i . r . active asymmetric S0-stretching vibration for CsSn(S0 3F) 5 is at ~1000 cm"1, almost 80 cm"1 higher than that found for the other complexes. An ana-logous observation is also found for [Sn(S0 3F) 6] 2~ 3 2 ; whereas in [M(S0 3F) 6] 2", with M=Pd, Pt, Ir, and [Au(S0 3F) 4]", the band is ^shifted to ~930 cm"1. A slightly different mode of electronic interaction may take place in the complexes involving transition TABLE 5.3: VIBRATIONAL FREQUENCIES OF C s P t ( S 0 3 F ) 5 AND RELATED COMPOUNDS C s P t ( S 0 3 F ) 5 CsRu(S0 3 F) 5 2 0 9 C s S n ( S 0 3 F ) 5 R IR IR R IR 1442 vw.sh 1430 vw 1412 m 1400 vs 1405 vs ,b 1418 w -1400 vs ,b v S0 3 1390 vw 1250 vw.sh rn 1240 s 1265 m 1260 m V " 2 1225 w.sh 1210 vs 1215 vs -1220 w,b 1220 vs 1190 s ,b 1130 w 1125 s ,b 1112 w 1080 s ,b 1120 s 1140 m  v S 0 2 bridge  1080 vw 1058 w 1062 w 1050 w,sh 1040 m 1045 m 1005 m -1002 vw v a s S ° 3 967 w,b 950 vs ,b -1000 vs ,b 930 w 920 vs .b 925 vs ,b 880 w 835 w 850 w -830 w.sh 817 w 800 vs ,b 810 s ,b 820 w 800 vs ,b 658 w 670 w,sh 635 s 650 m,b 660 m 630 m 625 m,b 605 w 605 vw 590 vw 595 w 580 vw 580 m 580 ms 585 w 587 m 547 w 540 m 550 ms 560 w 560 m 510 vw 515 vw.sh 447 m -400 w,sh 440 mw -430 w,sh -450 vw.sh vSF v s M-0 + S0 3 F def S0 3 F def v M-0 + S0 3 F def 410 vw as 275 s v sM-0 + S 0 3 bend metals. The vibrational spectra of CsSn(S0 3F) 5 will be discussed in more detail in Chapter 9. Therefore, i t appears that the existence of monomeric [Pt-(S0 3F) 5]~ with monodenfate S0 3F groups is not supported by the vibrational spectra of the compound. A polymeric structure, con-taining [Pt(S0 3F)5] n~ n, must therefore be present, as there is no evidence that the compound is a mixture of Pt(S0 3F ) i + and [Pt-(S0 3F) G] 2~. A dimer would have two [Pt(S0 3F) 6] octahedra joined by a common edge, a trimer or high oligomers would be a cyclic structure joined via cis-corners of the octahedra. The exact structure cannot be elucidated by vibrational spectroscopy alone and an X-ray diffraction study is required. 5.C.3 MAGNETIC SUSCEPTIBILITY All the Pt(IV)-S0 3F compounds investigated in this study are diamagnetic, suggesting a low-spin octahedral ligand f i e l d for the d 6 Pt(IV) and a lk1g electronic ground state. The experi-mentally obtained susceptibilities at room temperature are listed in Table 5.4. As i t was required for a dilution experiment in Chapter 6, the temperature dependence of the magnetic susceptibi-l i t y of (C10 2) 2[Pt(S0 3F) 6] was investigated from 300 to 81 K. The diamagnetic susceptibility of -(250 ± 1 5 ) x l 0 - 6 cgs units was found to have no discernable temperature dependence. The sum 179 of the Pascal constant for C 1 0 2 + and the diamagnetic correction for [ P t ( S 0 3 F ) 6 ] 2 " 1 0 7 is - 3 2 0 x l 0 - 6 cgs units. Therefore, (70 ± 15)*l0~ 6cgs units of paramagnetic susceptibility must be due to Temperature Independent Paramagnetism, expected to be only about 5 0 x l 0 - 6 cgs units for Pt(IV) l k l . The presence of T.I.P. is also evident for the other compounds listed in Table 5.4. Table 5.4 DIAMAGNETIC SUSCEPTIBILITY FOR SOME Pt(IV)-FLU0R0SUL- FATE COMPOUNDS (IN cgs UNITS) AT ROOM TEMPERATURE Xg x m E x d i a P t ( S 0 3 F k -(1.52±31)xl0-7 -(81±17)xl0-6 -188xl0~ 6 Ba[Pt(S0 3F) 6] -(2.95±7)xl0"7 -(271±6)xl0"6 -292xl0" 6 (C10 2) 2[Pt(S0 3F) 6] * -(2.70±16)xl0~7 -(250±15)xl0"6 - 3 2 0 x l 0 - 6 ( B r 3 ) 2 [ P t ( S 0 3 F ) 6 ] -(3.47±16)xl0~7 -(440±20)xl0"6 -454xl0" 6 * 81 < T < 300 K 5.C.4 SOLUTION STUDIES IN HS0 3F  5.C.4.1 ELECTRONIC SPECTRA Solutions of Pt(S0 3F) l t and Cs 2[Pt(S0 3F) 6] have identical absorption spectra when dissolved in HS0 3F, supporting the pre-sence of similar Pt(IV)-fluorosulfate species. The spectrum con-sists of a strong broad peak at 245 nm (half width of ~70 nm), £-1.5x10^ M_1cm_1. This band must be of charge transfer origin; no other absorptions due to d-d transition could be observed. 180 5.C.4.2 ELECTRICAL CONDUCTIVITIES Both (C102)2[Pt(S03F)6] and Pt(S03F)4 dissolve in HS03F to give conducting solutions, and, as in the HS03F-Au(S03F)3 system, the binary fluorosulfate gives rise to strongly conducting solu-tions. CsPt(S03F)5 also dissolves in HS03F to give solutions with conductivities slightly higher than those of (C102)2[Pt(S03-F) 6 ] , when the conductivities are calculated for equivalents of cations, i.e. Cs + and C10 2 +. The conductivities of solutions of (C102)2[Pt(S03F)6] and CsPt(S03F)5 are shown in Fig.5.2 and l i s t -ed in Table 5.5. Those of PtCSOsF)^ are shown in Fig.5.3 and Table 5.6. The conductivities of solutions of (C102)2[M(S03F)6], where M=Pt, Sn 3 2 , Ir, Pd, are all very similar, indicating a similar dissociative process in HS03F. A qualitative titration using small amounts of KS03F indicates no acidic property for (C10 2) 2-[Pt(S03F)6]. If (C102)2[Pt(S03F)6] undergoes a simple ionic dissociation in solution, like K[Au(S03F)iJ , according to: (C102)2[Pt(S03F)6] + HS03F > 2 C 1 ° 2 + ( s o l v ) + E p t ( S 0 3 F ) 6 ] 2 - ( S o l v ) > ( 5 - n > the conductivity of the resultant solution should be about 1.5 times that of a solution of K[Au(S0 3F)iJ at the same concentra-tion. This is expected because a) A*(K+)=30, while X*([Au(S03F)4]")=24, b) x0(K ) seems to be slightly higher than Ao(C102 ) 1 0 9 , and, 181 FIG 5.2 K , xlO 4 1 30-1 The electrical conductivity of Cs 2 Pt2(S03F) 1 0 and (Cl0 2) 2Pt(S0 3F)6 in H S 0 3 F at 2 5 ° C (C10 2) 2Pt(S0 3F-) 6 20 H Cs 2Pt 2(S0 0 .01 .02 m .03 182 Table 5.5 CONDUCTIVITY OF (C10 2) 2[Pt(S0 3F) 6] AND CsPt(S0 3F) 5  IN HS03F m* (C10 2) 2[Pt(S0 3F) 6] [Cs(Pt(S0 3F) 5] 0.000 1.720 1.148 0.002 5.185 5.925 0.004 7.986 10.62 0.006 10.59 14.64 0.008 13.54 0.010 15.54 0.012 17.82 0.014 20.11 0.016 22.29 0.018 24.59 0.020 27.46 0.022 30.43 0.024 32.86 0.026 34.10 mol-kg"1 lO'V^cm - 1 l O ^ s ^ c m - 1 units * For comparison purposes, the molality of CsPt(S0 3F) 5 is calculated for two cations, i.e., [CsPt(S0 3F) 5] 2 FIG 5.3 CONDUCTIVITY OF Pt(S0 3FK IN HS03F .01 , 02 .03 .04 m (mol/kg) Table 5.6 CONDUCTIVITY OF Pt(S0 3FK IN HS03F m 0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.014 0.016 0.018 0.020 0.022 0.024 0.026 0.028 0.030 0.032 0.034 0.036 0.038 0.040 mol'kg" ic(obs) I. 134 II. 83 12.43 16.05 22.05 28.05 32.85 37.08 41.73 46.78 51.58 56.03 60.25 64.41 68.50 72.44 76.31 80.14 83.96 87.77 91.57 l O - ^ c r r r 1 ic(calc) 1.100 6.75 13.09 19.08 24.77 30.20 35.40 40.38 45.18 49.81 54.29 58.62 62.82 66.90 70.87 74.73 78.50 82.17 85.76 89.27 92.70 10 _ i +si~ 1 cm - 1 units 185 c) A*[Pt(S0 3F) 6] 2- should be less than X*( [Au(S0 3F) l +]") due to its double negative charge which may result in more extensive solution. This, however,is not found for (C10 2)2[Pt(S0 3F) 6] or any of the [M(S0 3F) 6] 2" complexes; on the contrary, the conductiv-ity of these solutions is about twice that of KtAuCSOsF)^], at the same concentration. Furthermore, whereas the molal conductiv-it y of K[Au(S03F)iJ remains relatively constant with respect to concentration, the same is not true for the [M(S0 3F) 6] 2" complexes; these seem to decrease much faster as the concentration increases. The high conductivity, the large concentration dependence of the molal conductivity and the lack of acidic behavior for the solutions seem to suggest the existence of an equilibrium giving rise to S0 3F" in the process, i.e., a basic dissociation for [M(S0 3F) 6] 2". Further evidence for this will be presented in the titration of PtCSOsF)^ with KS03F. As can be seen from Fig 5.3, the conductivity of PtCSOsF)^ in HS03F is comparable to that of Au(S0 3F) 3, which has been shown to be a superacid in HS03F; but being a solute that can conceiv-ably give rise to two acidium ions per mole of PtCSOsF)^ dissolv-ed, the conductivity appears quite low, or about a half as expect-ed for a complete dissociation. As with any diprotic acid, H 2[Pt(S0 3F) 6], the hypothetical solvated acid in HS03F, should undergo the following two-step deprotonation reaction: HS03F + H 2[Pt(S0 3F) 6] H 2S0 3F + + H[Pt(S0 3F) 6]" (5.12) 186 HS03F + H[Pt(S0 3F) 6]~ 7 ^ H 2S0 3F + + [Pt('s03F)6] 2 _ . (5.13) Ki should also be expected to be much larger than K2, by analogy with multiprotic acids in the aqueous system, such as H 2S0 4 and H3PCV With the high conductivity of solutions of HSOaF-PtfSOaFK, H 2[Pt(S0 3F) 6] must be almost completely dissociated via equation (5.12), and the almost identical conductivity of HSOsF-PtfSOsF)^ with that of HS0 3F-Au(S0 3F) 3 suggests an almost monoprotic system, or Ki >> K2. The conductometric titration of H 2[Pt(S0 3F) 6] with KS0 3F, a typical plot of which is shown in Fig 5.4, supports both the above suggestions. In the beginning of the t i t r a t i o n , and up to a K/Pt ratio of about 1, the conductance decreases almost linear-ly with the addition of KS0 3F, indicating that a strong acid is present. But further into the ti t r a t i o n , a gradual increase in the slope of the titration curve occurs, and a broad conductivity minimum is found at a K/Pt ratio of ~1.5. The titration curve crosses the equivalent point of 2 without any changes in the slope that would suggest that i t is reached. This indicates a very weakly acidic solution, and K 2[Pt(S0 3F) 6], the empirical formula of the species present at the endpoint, is a base in HS03F. The titration was repeated 6 times, using both weight-and volume-buret addition of KS0 3F, and the result is reprodu-cible. Using the procedure described in Appendix A, a theoretical titration curve was calculated; this is shown in Fig 5.4 and FIG. 5.4 ^ calc 'd ' 1 1 - l 0 1 2 3 K/Pt 188 listed in Table 5.7. The following parameters were used: K a p = [H 2S0 3F +]- [S0 3F _] = 3 . 8 x l 0 - 8 mol /kg. This would give a basic dissociation, K_2, for [ P t ( S 0 3 F ) 6 ] 2 " in HS03F, according to: [Pt(S0 3F) 6] 2" + HS03F -^=±* H[Pt(S0 3F) 6r + S03F" , (5.14) of 1.2 10~3 mol/kg, i.e., [Pt(S0 3F) 6] 2~ is a reasonably strong base when dissolved in HS03F. This is consistent with the high conductivity of solutions of [Pt(S0 3F) 6] 2~ and the basic behavior of the solution near the endpoint of the Pt(S0 3F)i +/KS0 3F t i t r a -tion. By analogy, the other [M(S0 3F) 5] 2", where M=Sn, Pd, Ir, because of the similarity in the conductivity of their solutions, should have very similar K_2 values and thus are strong bases in HSO3F. • The f i t between the calculated and observed titration points is acceptable at the ratios K/Pt at ~0 and at ~2, the deviation increases quite rapidly as K/Pt approximates 1, a higher acidity than can be accounted for by equations (5.12) and (5.13) was ob-served. The x * values are in the range-expected for the various species, although the apparently high mobility of [Pt(S0 3F) 6] 2" xJ(H 2S0 3F +) X*(S03F~) OK+) X*(H[Pt(S0 3F) 6]") x J([Pt(S0 3F) 6] 2") Ki K2 320, 227, 29, 29, 32, 6.3xl0~ 2 mol/kg, 5.0*10~ 5 mol/kg, and Table 5.7 CONDUCTOMETRIC TITRATION OF Pt(S0 3F)^ WITH K S O 3 F IN HSQ3F R 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 ic(obs) 104.8 79.43 60.09 45.51 35.40 28.71 24.75 22.73 21.69 21.62 24.61 fi-1cm" <(calc) 104.8 79.78 59.48 42.77 28.93 19.03 17.33 17.81 18.93 20.95 24.61 I 0.032 0.031 0.031 0.030 0.030 0.031 0.038 0.045 0.051 0.056 0.060 mol'kg - 1 units * calculated using the following parameters: X*(H 2S0 3F +) = 320 X*(S03F") = 227 Aj(K +) = 29 X*(H[Pt(S0 3F) 6]") = 29 X*([Pt(S0 3F) 6] 2-) = 32 [H 2S0 3F +]- [ S O 3 F - ] = 3.8xl0-8 mol 2.kg - 2 Ki = 0.063 mol-kg - 1 K2 = 5.0xl0-5 mol-kg"1 190 should require some explanation. Although one would normally expect that a doubly charged ion would have a less mobility than its singly charged counterpart because of the increased size of the former's solvation -"Sphere, such a phenomenon may not be very important when the ion i t s e l f is already very large, like the [M(S0 3F) 6] 2~ studied here, and the charge can be delocalized more effi c i e n t l y than a point charge such as K+. K+ may be extensive-ly solvated in HSO3F since X*(Ba 2 +) is about 35 in Ba(S03F)2 sol-utions (x 0(K ) = 29) 2 1 . For similarly solvated ions of different charges, the higher the charge on the ion, the higher i t s mobility should be, since i t would experience an increased force from the electric f i e l d . At the intermediate K/Pt ratios during the ti t r a t i o n , the excess acidium ion may be indirectly due to a dimerization of [Pt(S03F) 6]-containing species according to: H 2[Pt(S0 3F) 6] ^ ± ^ H 2 [ P t 2 ( S 0 3 F ) 1 0 ] + HSO3F (5.15 ) H[Pt(S0 3F) 6]" 7 ^ % H 2 [ P t 2 ( S 0 3 F ) 1 0 ] + S O 3 F " (5.16) [Pt(S0 3F) 6] 2- % [ P t 2 ( S 0 3 F ) 1 0 ] 2 - + S O 3 F " . (5.17) All these dimerization reactions have the result of decrea-sing the acidity of the system; two also have a strong dependen-ce on [S03F~] and conversely [H 2S0 3F +]. At the i n i t i a l conditions of high acidity, i f the hypothesis is correct, dimerization or even higher oligomer formation is favored. But as the titration proceeds (by the addition of K S O 3 F solution which also dilutes the system), the i n i t i a l l y formed oligomers will gradually depoly-191 merize, and increase the acidity of the solution by the reverse reaction which can be represented by: %H ?[Pt 9 (S0.F) i n] + 3HSO3F T ^ U [ P t ( S 0 3 F ) 6 ] 2 " + 2 H 2 S 0 3 F + . (5.18) The net effect of this extra quantity of H 2 S0 3F + would con-tribute to an increase in conductivity of the system. Further-more, the actual Ki would be higher than the one calculated here and the actual K2 would be lower. This is because of the calcu-lation procedure used; also because of this, X 0 ( [ P t ( S 0 3 F ) 6 ] 2 " ) would be lowered, but since the calculated value is so high to start with, this should not be a problem. This extra set of equilibria serves to buffer the system when the acidity is not very high, and the observed broad conductivity minimum seems to be consistent with this. It is possible to include equations (5.15) to (5.17) into the calculation, but the procedure outlined in Appendix A already involves four variables and the addition of three more is certainly beyond the intent of this study. Quali-tatively, the addition of equations (5.15) to (5.17) appears to be able to shift the calculated curve closer to the observed one. It is now necessary to present some evidence for polymeriza-tion in the platinum(IV)-fluorosulfate system. P t ( S 0 3 F ) i + is a polymer in the solid state, much like Au (S0 3F) 3 i s , and both dis-solve in HS0 3F only with d i f f i c u l t i e s , although very concentrated solutions can be formed. Pt (S0 3F )L, . shows a much greater reluc-tance to dissolve, and this has been discussed in the vibrational 192 spectra section. Solutions of PtCSOsF)^ obtained in this manner all show lower conductivity than those obtained without f i r s t isolating PtCSOaF)^ as a solid (i.e., formed from platinum metal and S 20 6F 2/HS03F). The complete removal of a l l S 20 6F 2 in the latter case is shown by Raman spectroscopy for one such sample. It therefore appears that oligomers, possibly of a much higher degree of polymerization than 2, may be present when solid Pt-(SOsF)^ is dissolved in HS03F. Furthermore, as mentioned before, the apparent molal conductivity of (C10 2) 2[Pt(S03F) 6], and also those of CsPt(S0 3F) 5, CsSn(S0 3F) 5 and Pt(S0 3F) t t a l l show a large concentration dependence. This is especially evident for CsSn-i (S03F) 5, for which a larger concentration range was investigated. A detailed discussion on the conductivity of CsM(S03F)5 and [M-( S 0 3 F ) 6 ] 2 _ , M=Pt, Sn, will be presented in Chapter 9. All these observations suggest the occurrence of polymerization in solutions containing these species. Supporting evidence can also be found in the vibrational spectra of CsPt(S0 3F) 5 (also CsSn(S0 3F) 5), which indicate the presence of both monodentate and bidentate S03F groups. It can be visualized that CsPt(S0 3F) 5 is the compound isolated at K/Pt=l during the ti t r a t i o n , and that i t is formed either by the desol-vation-dimerization of H[Pt(S03F) 5]" according to: 2H[Pt(S0 3F) 6]" v [ P t 2 ( S 0 3 F ) 1 0 ] 2 " + 2HS03F , (5.19) or, i f [ P t 2 ( S 0 3 F ) i 0 ] 2 " is the predominant species present at that condition, by a desolvation process. The n.m.r. and vibrational 193 spectra evidence to be presented later seem to indicate quite a low extent of polymerization in solution, and favors equation (5.19) (this does not negate i t s contribution to the conductivity though, because the deviation between the two curves in Fig 5.4 is quite small). CsPt(S03F) 5 dissolves in HS03F to give a solution which shows acidic conductivity. Its higher conductivity than solutions of (C10 2) 2[Pt(S0 3F) 6] (Fig 5.2) is consistent with H 2S0 3F + having a greater ionic mobility than S03F". 5.C.4.3 SOLUTION VIBRATIONAL SPECTRA It was hoped that Raman spectra of solutions of Pt(S0 3F)it would show signs of polymerization. This, however, was not ob-served up to a concentration of about 0.12 mol/kg. Actually, the only bands that are definitely due to Pt(S0 3F) 4 are at 638, 455 and 276 cm - 1, as a l l other bands are overlapped by broad solvent bands. This suggests that solutions of Pt(S03F)i+ are not as ef-ficient Raman scatterers as those of Au(S0 3F) 3. In addition, because of the broadness of the peaks, splittings observed in the solid state spectra of CsPt(S0 3F) 5 are not evident here. 5.C.4.4 SOLUTION N.M.R. All three species, Pt(S0 3F) [ +, CsPt(S0 3F) 5 and Cs 2[Pt(S0 3F) 6] gave identical 19F-n.m.r. spectra in HS03F solutions. This is 194 in sharp contrast to what was observed in the Au(S0 3F) 3-HS0 3F system. A typical 19F-n.m.r. spectrum of Pt(S03F)i+-HS03F is shown in Fig. 5.5. One of the natural occurring isotope of platinum, 1 9 5 P t , has a nuclear spin of and i t s coupling to 1 9 F supposedly gives rise to the splitting of the fluorine resonance with J(Pt-F) = 31 Hz. Since 1 9 5 P t is present in only 33.8% natural abundance, about 2 / 3 of the fluorine signal should not be affected, and this was shown by an integration of the peaks. Therefore,the coupling occurs via a sulfur and an oxygen atom. [ P t F 6 ] 2 _ on the other hand, presumably due to the proximity of the two atoms, has a large J(Pt-F) coupling constant of ~2000 Hz 1 9 5 . For the species [ P t 2 ( S 0 3 F ) 1 0 ] 2 - , the coupling scheme for that ion alone will be a 1:8:2:8:1 pattern, each separated by ~15 Hz. This is expected since the chemical environment for a fluorosulfate group should not be very different whether the S0 3F ligand is monodentate or bidentate. However, no satellites on either side of the t r i p l e t was observed, at concentrations up to ~0.12 mol/kg. This indicates that even i f the dimer is present, i t must be in quite a low concentration. The sharpness of the resonances does not necessarily mean a lack of exchange between the platinum-fluorosulfate-containing solute (at -47.8 ppm) and the solvent (at -40.8 ppm), as these resonances are 7 ppm apart, and proton exchange evidently does occur in this system. FIG. 5.5 F-19 nmr of Cs 2 CPt (SO3F) 6J in HSO3F (conc. = 0.11 m) -47.75 ppm J = 31 Hz -4077 ppm H S O / ,1. 196 These 19F-n.m.r. results show that identical species con-taining platinum and fluorosulfate groups are present in solution, whether the starting solid is the binary fluorosulfate or either of i t s two types of complexes. Since 1 9 5 P t is a nucleus accessible by n.m.r. spectroscopy, attempts were made to obtain 1 9 5Pt-n.m.r. spectra of these solu-tions. For the spectrum of [Pt(S0 3F) 6] 2~, a septet of the rela-tive intensity of 1:6:15:20:15:6:1 and separated by J(Pt-F)=31 Hz is expected, while for [ P t 2 ( S 0 3 F ) 1 0 ] 2 " , two groups of septets separated by J(Pt-Pt) should be present i f Pt-Pt coupling occurs. The observation of the latter group of peaks should provide the most conclusive evidence for the existence of the dimer in solu-tion. This, unfortunately, did not occur, and no 1 9 5Pt-n.m.r. spectra was obtainable for any one of the fluorosulfates, even when solutions of up to ~1 mol/kg concentration were used, a l -though a resonance from a [ P t C l 6 ] 2 " test solution was obtained. The wide range of chemical shifts possible in 1 9 5Pt-n.m.r. spectra (~7000 ppm from [ P t F 6 ] 2 " to [ P t C l 6 ] 2 - ) combined with i t s low sen-s i t i v i t y (only ~0.3 % that of proton when natural abundance is included) 1 + 0 probably contributed to this experiment's demise. 197 5.D CONCLUSION HSOsF-PtCSOsFK has been found to be a superacid in the HS03F solvent system. In a per-mole basis, its acidity is comparable to that of HS0 3F-Au(S0 3F) 3 and slightly less than that of HS03F-SbF2-(S0 3F) 3. The decrease in acidity in the system, similar to other diprotic acids, is due to the low equilibrium constant for the dissocia-tion of the second proton. In addition, the oligomer-formation of the [Pt(S0 3F) 6]-containing species is also a contributing fac-tor. These have been shown by conductometric titrations of the superacid with KS03F. As expected, complexes containing [ P t ( S 0 3 F ) 6 ] 2 _ can readily be formed, including one which contains the tribromine cation — -(Br 3) 2[Pt(S0 3F) 6]. The anion dissociates in HS03F to give basic conducting solutions, and appears to be isostructural with a general class of [M(IV}(S0 3F) 6] 2" anions. A complex with the empirical formula of CsPt(S0 3F) 5 can also be synthesized this is a weak acid in HS03F and is not monomeric in the solid state. The Pt(IV)-fluorosulfate system investigated here is diamag-netic, thus indicating a 1Aig, low-spin electronic ground state for the d 6 metal ion in an octahedral ligand f i e l d . 198 CHAPTER 6 IRIDIUM-FLUOROSULFATE 6.A INTRODUCTION Iridium is one of the rarest members ofthe platinum metals, but because of its possible role in homogeneous catalysis, espe-c i a l l y with Ir in the lower oxidation states, the chemistry of iridium has been a f i e l d of active research 1 9 8 > 1 9 6. Iridium can exist in a variety of oxidation states, from -1 to +6, the possible exception being Ir(II), which is isoelectronic with Pd(III) and Pt(III). As far as the halogen chemistry of iridium is concerned, i t is much like that of platinum, with IrF 6, [IrFslLi and Ir-Hal 3, Hal=F, Cl, Br, I, a l l having been synthesized. Al-though the binary tetrahalides of iridium are unknown, i t forms many very stable [ I r - H a l 6 ] 2 " complexes 1 9 0 . The instability of Ir-Haltt, at least in the case of IrF 4, cannot be attributed to the lack of oxidative a b i l i t y of the halogen, and may be due to the more favorable formation of either IrF 3 or [IrFsli,. The nor-mal product of the fluoridation of iridium metal, depending on the reaction condition, is IrF 6 or t l r F s l ^ ; IrF 3 can only be ob-tained by the reduction of the higher fluorides. While the gold and the platinum systems discussed in previous chapters provided some interesting solution chemistry, and novel superacid systems in HS03F, they are lacking in having any para-magnetic compounds. One of the most easily accessible oxidation states of iridium, at least in anionic complexes, is Ir(IV), which 199 is d 5. So regardless of whether the ion is high SDin or low spin, this should provide a paramagnetic species. The high ligand f i e l d s plitting expected for Ir(IV) would almost certainly force the electronic state to be 2T 2g, low-spin. Although this would imply quite complicated temperature dependent magnetic behavior for Ir(IV), this is sometimes not very notice-able due to the inherent small magnitude of the magnetic suscep-t i b i l i t y i t s e l f . However, antiferromagnetic coupling is prevalent in this system, and [Ir-Hal 6] 2~ complexes,with Hal=Cl, Br,have low moments of about 1.5 yg at room temperature 1 9 7 . It was of interest to see i f the replacement of a halide ion with a S O 3 F group would lead to a drastic increase in the magnetic moment of the species, as was observed in the palladium system. Therefore, the investigation into iridium-fluorosulfate is' not so much in terms of solution studies, but rather as a study into the magnetic behavior of fluorosulfates containing a transi-tion metal in a high oxidation state and to collect more informa-tion as to the coordination properties of the fluorosulfate group. 200 6.B EXPERIMENTAL 6.B.l SYNTHESIS OF IRIDIUM TRIS(FLUOROSULFATE) Ir + 2S 20 6F 2 ^ 4 Ir(S0 3FK (6.1) Ir(S0 3F) 4 — I r ( S 0 3 F ) 3 + JsS205F2 + %02 (6.2) In a typical reaction, iridium metal (158 mg, 0.822 mmol) was reacted with an excess of S 20 6F 2/HS0 3F (-10 mL) at ~140°C. Parts of the metal ini t ia l ly dissolved to form a brown solution containing IrfSOsF)^; after 3 days of heating, a deep purple sol-ution together with unreacted metal was formed. After the removal of most of the S 20sF 2 and oxygen formed from the reaction, more S20gF2 (~3 mL) was added to the mixture which was subsequently reheated at ~140°C. After 3 days, all metal had dissolved and the solution again changed back into having a deep purple color. All volatile materials were removed in vacuo and a solid which analyzed as Ir(S0 3F) 3 was obtained (398 mg, 0.813 mmol). Ir(S0 3F) 3 can also be obtained by the thermal decomposition of ^(SOsF)^ at ~120°C in vacuo according to equation (6.2). In a typical pyro-lysis experiment, Ir(S0 3F) 4 (294 mg, 0.500 mmol) was converted into Ir(S0 3F) 3 (239 mg, 0.488 mmol) after 12 hours of heating. Ir(S0 3F) 3 is a very deep bluish purple, hygroscopic solid. It is soluble in HS03F and decomposes at ~200°C with the evolu-tion of a colorless gas which was not identified. Analysis Ir S F Calculated % 39.27 19.66 11.65 Found % 39.10 19.46 11.48 201 6.B.2 SYNTHESIS OF IRIDIUM TETRAKIS(FLUOROSULFATE) Ir(S0 3F) 3 + *sS 20 6F 2 • Ir(S0 3F) 4 (6.3) ^(SOsF)^ can be prepared in a reaction similar to that dis-cussed for Ir(S0 3F) 3. Iridium metal (198 mg, 1.030 mmol), was reacted with S 20 6F 2/HS0 3F (~4 mL) and was taken through the same cycles of S 20 6F 2 addition. When a l l the metal had dissolved and most of the S 20 5F 2 had been removed, S 20 6F 2 (~2 mL) was added to the solution of essentially Ir(S0 3F) 3 in HS03F. This was heated at ~60°C for 6 hours, after which time a deep brown solution had formed. The removal of a l l volatile materials by the usual method yielded Ir(S0 3F) 4 with traces of the blue Ir(S0 3F) 3 (the weight of the product was slightly lower than expected for Ir(SO3F)i+ also). More S 20 6F 2 (~2 mL) was added to the impure product which was heated at ~60°C for another 6 hours. The removal of a l l excess S 20 6F 2 (and traces of S 20sF 2 and oxygen) at room tem-perature yielded pure ^(SOsF)^ (607 mg, 1.032 mmol). ^ ( S O ^ K is a deep brown, almost black, hygroscopic solid which is very soluble in HS03F. It decomposes into Ir(S0 3F) 3 according to equation (6.2) at ~100°C in vacuo but is stable up ~150°C in an atmosphere of N2. Analysis Ir S F Calculated % 32.66 21.80 12.92 Found % 32.50 21.60 12.69 202 6.B.3 SYNTHESIS OF [Ir(S0 3F) 6] 2~ COMPLEXES  6.B.3.1 PREPARATION OF (C10 2) 2[Ir(S0 3F) 6] " 2C102S03F + Ir + 2S 20 6F 2 (C10 2) 2 [Ir(S0 3F) 6] (6.4) A mixture of C102S03F (~1 mL), and S 20 6F 2/HS0 3F (~6 mL) was allowed to react with iridium metal (289 mg, 1.505 mmol) at ~120°C for 3 days. All the metal by that time had dissolved and the solution took on a dark brown color. The removal of a l l volatile materials at ~70°C yielded a compound which analyzed as (C10 2) 2-[Ir(S0 3F) 6] (1.362 g, 1.478 mmol). (C10 2) 2[Ir(S0 3F) 6] can also be prepared by: a) replacing C102S03F with C1S03F in equation (6.4) and per-forming the reaction under the same conditions, or b) a reaction between I r C l 3 and S 20 6F 2/HS0 3F at ~120°C for 4 days. In both these alternative synthetic routes, the required C102S03F is generated by an attack of C1S03F (formed from the reaction of I r C l 3 with S 20 6F 2 in (b)) on the glass reactor. (C102)2.IIr(S03F)6] is a dark orange-brown, hygroscopic solid. It is soluble in HS03F and melts with decomposition at ~190°C to form a dark brown liquid. Analysis Cl Ir F Calculated % 7.69 20,86 12.37 Found % 7.51 2.1.04 12.44 203 6.B.3.2 PREPARATION OF Cs 2 [Ir(S0 3F) 6] CsCl + HSO3F y CsCl + HC1 (6.5) 2CsS03F + Ir + 2S 20 6F 2 Cs 2[Ir(S0 3F) 6] (6.6) Cs 2[Ir(S0 3F) e] was prepared by the oxidation of iridium metal (398 mg, 2.071 mmol) with S 20 6F 2/HS0 3F (~10 mL) in the presence of CsCl (798 mg, 4.146 mmol) at ~150°C for 10 days. The removal of al l volatile materials at ~70°C yielded a product with a weight corresponding to that expected for Cs 2[Ir(S0 3F) 6], (2.121 g, 2.015 mmol). Cs 2 [Ir(S0 3F) 6] is a pale orange, hygroscopic solid. It is soluble in HS03F and melts with decomposition at ~150°C. 6.B.3.3 PREPARATION OF Ba [Ir(S0 3F) 6] BaCl 2 + 2HS03F • Ba(S0 3F) 2 (6.7) Ba(S0 3F) 2 + Ir + 2S 20 6F 2 Ba [Ir(S0 3F) 6] (6.8) A mixture of dried BaCl 2 (198 mg, 0.951 mmol) and iridium metal (183 mg, 0.952 mmol) was reacted with S 20 6F 2/HS0 3F (~10 mL) at ~100°C for 2 weeks and, after reDlenishing the largely decom-posed S 20 6F 2, at ~160°C for another 2 weeks. The removal of a l l volatile materials at ~70°C yielded a product with a weight sug-gesting the composition of Ba [Ir(S0 3F) 6] (788 mg, 0.853 mmol). The low yield is due to the visually evident etching of the glass reactor by the solution. Ba[Ir(S0 3F) 6] is a light orange, hygroscopic solid. It appears to be almost insoluble in HS03F, and decomposes at ~200°C. 204 6.C DISCUSSION 6.C.1 SYNTHESIS AND GENERAL DISCUSSION 6.C.1.1 BINARY FLUOROSULFATES Iridium was found to be much less reactive than platinum towards fluorosulfonation, whether the starting material is the metal i t s e l f or I r C l 3 . Although, as pointed out in previous chapters, these noble metal chlorides are not susceptible to dis-placement reactions in H S O 3 F , they are usually quite reactive towards S 20 6 F2/HS0 3F. All the preparative reactions described could take up to 1 month for completion, while some had actually been aborted even-tually. The preparation of the binary fluorosulfates has one of the most serious d i f f i c u l t i e s that a synthetic reaction could encounter the decomposition of both the reactant (S 20 6F 2) and the product (Ir(S0 3 FK). Since S 20 6F 2 is not known to decompose into S 20 5F 2 and 0 2 at a temperature of only ~120°C, the path that i t takes must be via IrCSOsF)^, in a catalytic manner. A higher reaction temperature was not feasible since i t was found that the decomposition took place much faster than the conversion of the metal to either Ir(III) or Ir(IV). The S 20 5F 2 formed, identified by i . r . spectroscopy, had to be removed before the S 20 6F 2 could be replenished because i t s presence would reduce the ionizing abil-ity of the solvent. Because of the extremely dark colors of the solutions, the addition of the oxidant was usually repeated at least twice after a l l the metal had reacted, to ensure a com-plete reaction. The decomposition of IrCSOsF)^ to form S 20 5F 2 and oxygen is unlike that found for Pd 2(S0 3F) 6, Au(S0 3F) 3 or Pt(S0 3F) l t. While the latter two compounds break down completely to form S0 2F 2, the decomposition of palladium tris(fluorosulfate) giving S 20 6F 2 has been used as an evidence for Pd(IV)'s strong oxidizing a b i l i -ty. It therefore appears that Ir(IV) is not a very strong oxidi-zing agent. Although the compound is not very thermally stable, i t is one of the few examples of binary Ir(IV) compounds, another example of which is Ir02 1 9 9 -S.C.1.2 [I r ( S 0 3 F ) 6 ] 2 - COMPLEXES As with the synthesis of the binary fluorosulfates of iridium, these reactions are very time consuming. Although the thermal s t a b i l i t y of the anionic Ir(IV) complexes appears to be greater than that of iKSOsF)^, the formation of the complexes, for some inexplicable reasons, is very slow. An attempt to synthesize Na2.[IrCS03F)6] was terminated after six months, at which time the reactor had become badly etched. The above reaction was actually thought to proceed quickly by the following reasonings. a) The decomposition of [ I r ( S 0 3 F ) 6 ] 2 " is faster when i t is solvated, and for the almost insoluble Ba[Ir(S0 3F) 6], a higher reaction temperature of ~160°C could be used, but, because of the insolubility of the product, the reaction 206 proceeded slowly, b) An inspection of the solubility of the various alkali metal hexachloroplatinates(IV) reveals that Na 2[PtCl6] has a very low solubility and Na 2[Ir(S03F) 6] may be expected to behave similarly. This was later found to be the case during the course of the reaction, but the formation of the complex was exceedingly slow, and i t more than compen-sated for the presumably slower decomposition of the pro-duct. 6.C.2 VIBRATIONAL SPECTRA For both the binary fluorosulfates, vibrational spectroscopy had been of very limited help in their investigation. Being deeply colored compounds, no Raman spectra could be obtained, even at ~80 K. The i . r . spectra obtained were of extremely poor resolution. Using a variety of mulling agents such as fluorolube oi l , hexachlorobutadiene and silicone o i l , separate regions of the spectrum could be 'pieced' together for Ir(S03F) 3. The i . r . fre-quencies of Ir(S03F) 3 and Ir(S0 3 FK are listed in Table 6.1 and compared to those of PtCSOsF)^. All three spectra are very simi-lar, especially in the region of ~800 - 1200 cm - 1, where the large number of bands probably leads to overlaps and thus the observed unresolvability. This suggests that the two iridium 207 TABLE 6.1 Ir. FREQUENCIES OF Ir(S0 3F) 3 AND Ir(S0 3F) H Ir(S0 3F) 3 Ir(S0 3FK Pt(S0 3FK 1400 s 1390 s ~1400 vs,b 1210 s 1210 s -1225 vs -1150 s,vb 1100 vs, 1070 s,sh -1070 vs,vb 1020 vs,sh 1020 s,b -1000 vs.vb 950 s,vb -900 vs,vb 870 s,b 800 s,vb 820 s,b -820 s,vb -670 s,b 630 s 640 m,b -640 s,b 575 580 m 580 s 540 540 m 540 m 450 208 fluorosulfates are extensively S03F-bridged species. Because of the increased complexity of the i . r . spectrum of Ir(S03F) 3 over that of Au(S03F)3, the former may not be square planar, but rather octahedral. This may be resolved with the synthesis of anionic complexes of either [Ir(S0 3F)iJ~ or [ I r ( S 0 3 F ) 5 ] 3 ~ , a l -though this was not persued. In the halide system, the trihalides contain octahedrally coordinated Ir(III) with, as expected, a low spin d 6 electronic configuration. For the [ I r ( S 0 3 F ) 6 ] 2 " complexes, i . r . spectra of reasonable resolution could be obtained, but due to the samples' deep yellow color, Raman spectra of only relatively poor quality was recorded at ~80 K. The vibrational frequencies of these compounds are listed in Table 6.2 in comparison with the Raman frequencies of Cs 2[Pt(S0 3F) 6]. The three iridium complexes a l l have spectra that are similar to that of Cs 2[Pt(S0 3F) 6], suggesting a similar type of metal-ligand interaction and a monodentate coordination of the S0 3F groups. The positions of the three vibrational modes of C10 2 + at (1310, 1295), 1060 and 522 cm'1 in the i . r . spectrum also indicate an ionic formulation for the compound can be made. In summary, i t appears that a l l the iridium species involved in this study have octahedral coordination for iridium, with bi-dentate S03F groups evident for the two binary fluorosulfates. TABLE 6.2 VIBRATIONAL FREQUENCIES OF [ I r ( S 0 3 F ) 6 ] 2 -(C£0 2) 2[Ir(S0 3F) e] Cs 2[Ir(S0 3F) 6] Ba[Ir(S0 3F) 6] Cs 2[Pt(S0 3F) 6] IR R + IR R IR R 1400s,b 1400,1380w 1390sb 1415w 1390s 1416w,sh 1410m 1310s* 1295s* 1250vs 1255m 1250vs 1200s 1200w 1195s,b 1212w 1200s 1219m 1050w* 1047* 1057m 1043s 950vs,vb 950vs,b 978w 950s,b 1010m 920vs,b 950w 900s,b 820s 840,81Ow ~800s,b ~802w,b 820s ~800w,b 660s 640w 645m 630w 630w 634vs 590s 580w 570s 57 5w 575m 579vw 550s 548w 540s 550w 540m 550vw 522m* 520w* 460w 460 vw 460w,sh 44 3w 440m 440vw 447m 440m 286m 287w 280vs 268m * + denotes bands due to C£0 2 obtained at -80K 210 6.C.3 MAGNETIC SUSCEPTIBILITY According to Figgis, the magnetic susceptibility of Ir(IV), a d 5 low spin system with a 2T 2g electronic ground state can be represented by the following equation 1 4 1 : 2 _ 8 + (3x - 8)exp(-3x/2) q )  y x[2 + exp(-3x/2)] [ 0'^' where x = x/kl X = -E, = -5000 cm"1 for Ir(IV) 1 9 6 k = Boltzmann's constant = 0.6950 cm"1 Therefore, as x —*- -» (for large | x | or low temperature), y —> /3, and as x —• 0, y —* /5. In the temperature range accessible in this study (~100 - 330 K), y should have values from 1.76 yg at 100 K to 1.84 yB at 330 K. For the hexachloro- and hexabromo-complexes of Ir(IV), y e f f was found to generally decrease with temperature, but in most cases they were lower than /3, the mini-mum value predicted by equation (6.9). The moments also showed marked drops at temperatures below "150 K to ~1.3 yg; this had been attributed to antiferromagnetic couplings in the system 1 9 7 . The magnetic susceptibility results on (C10 2)2tIr(S0 3F) 6], Cs 2[Ir(S0 3F) 6] and Ir(S0 3FK are listed in Tables 6.3 to 6.5, respectively. The temperature dependence of their y e f f is shown in Fig 6.1. In a l l three cases, the observed magnetic moments are lower than the expected values (the estimated uncertainty in y is about ±0.02 yg, contributions from temperature and weighing errors con-sidered). y e f f ranges from 1.55 to 1.71 yg for Cs 2[Ir(S0 3F) 6], 211 TABLE 6.3 MAGNETIC PROPERTIES OF ( C J c 0 2 ) 2 [ M S 0 3 F ) 6 ] T c xm l / x mC (calc'd) y e f f u e f f (calc'd) 78 4.348xl0-3 230 234 1.64 1.63 108 3.236x10"3 309 314 1.67 1.66 130 2.653xlO-3 377 371 1.66 1.67 155 2.294x10"3 436 436 1.68 1.68 178 2.000x10"3 500 498 1.69 1.69 203 1.770xl0-3 565 563 1.70 1.70 227 1.585x10"3 631 626 1.70 1.70 251 1.447x10"3 691 689 1.70 1.71 276 1.326x10-3 754 755 1.71 1.71 300 1.238x10"3 808 816 1.72 1.71 K cgs cgs cgs yB y B units l / x m C (calc'd) = (T + 12)/0.382 TABLE 6.4 MAGNETIC PROPERTIES OF Cs 2[Ir (S0 3F) 6] T c xm l / x ^ (calc) y e f f v e f f (c< 78 3.831x10" 3 261 269 1.55 1.52 106 3.003x10" 3 333 339 1.59 1.58 129 2.494x10" 3 401 399 1.60 1.61 154 2.141x10" 3 467 462 1.62 1.63 179 1.880x10" 3 532 524 1.64 1.65 203 1.681x10" 3 595 586 1.65 1.66 229 1.536x10" 3 651 651 1.68 1.68 252 1.410x10" 3 709 709 1.68 1.68 277 1.290x10" 3 775 773 1.69 1.69 302 1.214x10" 3 824 836 1.71 1.70 K cgs cgs cgs yB UB l / X m C (calc) = (T + 28)/0.395 212 230 TABLE 6.5 MAGNETIC PROPERTIES OF Ir(S0 3F) 1 ( xm c l / v c l / v c (calc) y e f f (calc) 284 9.06X10-1* 1104 1155 1.43 1.40 254 9.63X10"4 1039 1040 1.40 1.40 1.028x10-3 972.9 948.9 1.38 1.39 205 1.102xl0-3 907.7 853.5 1.34 1.39 78 K cgs l/xm° (calc) veff 1.40 1.38 1.34 757.5 746.7 1.37 642.8 659.0 1.38 569.1 567.5 1.35 489.3 479.8 1.32 337.5 369.1 1.36 cgs cgs yB 177 1.320x10-3 1.38 154 1.556x10-3 642.8 659.0 1.38 .37 130 1.758x10-3 1.35 107 2.044x10-3 1.34 2.963x10-3 1.30 l / x C (calc) = (T + 19 )/0.262 213 FIG 6.1 MAGNETIC MOMENTS OF (C1Q ?) 2 [Ir(S0 3F) 6] AND Cs 2 [Ir(S0 3F) 6] 214 1.64 to 1.72 y B for (C10 2) 2[Ir(S0 3F) 6] and 1.32 to 1.43 vBfor IriSO^F)^. At high temperature, the moments remain relatively constant and appear to obey the Curie-Weiss Law. This suggests either the presence of antiferromagnetic coupling in these systems, or perhaps, a slight departure of the coordination sphere around Ir(IV) from cubic symmetry. Jahn-Teller distortion in a low-spin d 5 system is predicted from theory, and results in an axial elongation of the coordina-tion sphere. However, since the splitting occurs in the orbitals not directly facing the ligands, a distortion of this type is usually very small compared to that present in species such as Cu 2 +, d 9, which can give rise to severe axial elongations leading to the formation of square planar complexes. For T i 3 + , d 1, which has a 2T 2g electronic ground state but with a positive spin-orbi-tal coupling constant, X, (resulting in a drop of y e f f from /5 as x —* 0 to zero as x —> », i.e., a larger range of variation in the moment as compared to a d 5 system), the magnetic behavior of CsTi (S0i +)*12H 2 0 has been interpreted in terms of a large distor-tion of about 3.6 times the magnitude of the spin-orbital coup-ling l l f l. Because the effect of low symmetry on the magnetic moment of a compound is inversely proportional to X, the spin orbital coupling constant, Jahn-Teller distortion may not be ex-p l i c i t in Ir(IV), which has a x 0 value almost 33 times that of Ti(III). In any case, a distortion would lead to an increase of the magnetic moment, at the temperature of this study, to about 215 1.8 yg, and this is certainly not observed here. Besides, the vibrational spectra do not offer any suggestion that a large dis-tortion from 0 n symmetry occurs. Therefore, i t must be concluded that Jahn-Teller distortion, although i t should be present in the iridium compounds studied here, does not appear to contribute to the observed lowering of the magnetic moment to below 1.73 yg. The other li k e l y alternative antiferromagnetic coupling, was therefore investigated. As the vibrational spectra of the [Ir(S03F) 6] 2~- and [ P t ( S 0 3 F ) 6 ] 2 _ - complexes show great similari-ties, and i t is not really expected that they would be structural-ly very different, a dilution experiment involving a mixture of ( C 1 0 2 ) 2 L M ( S 0 3 F ) 6 ] , M=Ir, Pt, was undertaken. Starting with a mixture of metal powders containing 8.85 mole % of iridium, the mixed chloronium salt was prepared by reacting the metals with C10 2S0 2F and S 20 6F 2/HS0 3F. The result-ing product corresponded to the expected weight and had the appearance of ( C 1 0 2 ) 2 [ P t ( S 0 3 F ) 6 ] . The result of a bulk magnetic susceptibility measurement on the mixture is listed in Table 6.6 and shown in Fig 6 .2 . The C102-containing complexes were used because, in the case of [ I r ( S 0 3 F ) 6 ] 2 " , i t is the only iridium complex that can be pre-pared comparatively easily. The presence of an excess reactant C10 2S0 3F, also ensure the formation of a product of high puri-ty. The diamagnetic correction of ( 2 5 0 ± 1 5 ) x l 0 - 6 cgs units, ob-tained for (C10 2) 2[Pt(S0 3F) 6] over a temperature range of 81 to Table 6.6 MAGNETIC PROPERTIES OF (C10 2) 2 [Ir(S0 3F) 6] DILUTED IN (C10 2) 2[Pt(S0 3F) 6] (8.85 moU Ir) T Xm i/xS 1/Xm (calc) fe f f veff (calc) 303 1.506 664 728 1.91 1.82 277 1.515 660 675 1.83 1.81 251 1.435 697 621 1.70 1.80 229 1.616 619 575 1.72 1.78 204 1.911 523 523 1.77 1.76 179 2.201 454 472 1.78 1.74 154 2.416 414 420 1.73 1.71 129 2.745 364 369 1.68 1.67 106 3.073 325 320 1.61 1.62 78 4.049 247 263 1.59 1.54 K 10 - 3cgs cgs cgs 1/X m(calc) = (T + 49)/0.484 217 FIG 6.2 MAGNETIC MOMENT OF ( C l 0 2 ) 2 [Ir(SQ 3F) 6] DILUTED IN (C10 2) 2[Pt(S0 3F)6] (8.85 mo1% Ir) 218 300 K, was used for both the complexes. Any Temperature Indepen-dent Paramagnetism present in these systems, estimated to be of the order of only 5 0 x l 0 - 5 cgs units l h l , should be compensated for by this procedure. Although the accuracy of the measurements is greatly reduced by the introduction of (C10 2)2[Pt(S03F) 6], a diamagnetic species, as the majority constituent (the estimated uncertainty, arising mainly from weight measurements, ranges from ±0.12 yg at ~300 K to ±0.05 yg a t ~ 8 0 K)> i t c a n D e s e e n that the y e f f obtained agree quite well with those calculated from equation (6.9), and in general, are consistently higher than those of pure (0102)2-[Ir(S0 3F) 6]. This shows that antiferromagnetic coupling does occur in the Ir(IV) fluorosulfate system, and is the most l i k e l y cause contributing to a lowering of y e f f in these compounds. A similar explanation has been invoked for the hexahalo-complexes of Ir(IV), although in this latter class of compounds, a much more extensive coupling must be present as their magnetic moments are much lower than what is found here. This is consistent with the conclusion arrived at in Chapter 3 on the paramagnetic Pd(II) compounds that SO3F is a much poorer medium for the superexchange to occur than monatomic ligands such as the halide ions. For Ir(S03F)3, diamagnetism is observed, with xm of -(80+30) x l O " 6 cgs units, compared to a sum of diamagnetic correction of - 1 5 4 x l 0 - 6 cgs units. It therefore appears that Ir(III) in the fluorosulfate is a spin-paired, d 8 system, not surprising in 219 view of the high 10 Dq expected for a 5d metal in the +3 oxida-tion state. The coordination for Ir(III) in Ir(S03F) 3 can either be octahedral or square planar in order to explain the diamag-netism, but there appears to be no reason why a square planar geometry should prevail over an octahedral one for a d 6 system, unless by stereochemical restrictions which should not be present here. Therefore, the ligand configuration for Ir(S03F)3 is most likely octahedral, or conversely, a l l fluorosulfate groups pre-sent in the compound must be bidentate, leading to the great complexity observed in the S-0 stretching region in the i . r . spectrum. 6.C.4 SOLUTION STUDIES IN HS03F 6.C.4.1 CONDUCTIVITY MEASUREMENTS Iridium-fluorosulfate was not intended to be investigated for its possible use as a superacid system in HSO3F, although there exist indications that i t may be one, at least for Ir(S0 3F) 4. The conductivity measurements of solutions of (Cl0 2)2[Ir(S0 3-F) 6] in HS03F were undertaken to further establish its similarity to the corresponding platinum complex, and to provide evidence for the existence of [Ir(S0 3F) 6] 2"-containing species in solution. The results are shown in Fig 6.3 and listed in Table 6.7. From FIG 6.3 CONDUCTIVITY OF (C10 2) 2 [Ir(S0 3F) 6] IN HSO3F TABLE 6.7 CONDUCTIVITY OF (C£0 2) 2[Ir(S0 3F) 6] IN HS03F* m K 0.00 0.280 0.20 3.801 0.40 7.581 0.60 9.885 0.80 12.20 1.00 14.92 1.20 17.56 1.40 19.60 1.60 21.50 1.80 23.60 2.00 25.63 2.20 27.49 2.40 30.06 2.60 32.22 10- 2mol.kg - 1 * interpolated 222 the plot, i t becomes evident that the ionization of (0102)2[1^-(S0 3F) 6] in HSO3F is very similar to that of the other [M(S0 3F) 6] 2 complexes, M=Pd, Pt, Ir, Sn, and a similar solution behavior must be present for a l l these solutes, i.e., [ I r ( S 0 3 F ) 6 ] 2 " undergoes a basic dissociation in HS03F according to: [ I r ( S 0 3 F ) 6 ] 2 _ + HS03F > H[Ir(S0 3F) 6]" + S03F" , (6.10) with an equilibrium constant of the order of 10"3, by comparison with the conductivity of the other hexafluorosulfato-complexes, described in detail in Chapters 5 and 9. 6.C.4.2 ELECTRONIC SOLUTION SPECTRA All the compounds, with the exception of the insoluble Ba-[Ir(S0 3F) 6], form deeply colored solutions in HS03F. Ir(S0 3F) 3 dissolves to form a bluish purple solution with an absorption maximum at 563 nm, e - l x l O 4 M^cm"1. Both Ir(S0 3FK and (C10 2) 2-[Ir(S0 3F) 6], although they are differently colored in the solid state, form solutions with an identical A m a x at 200 nm, e ~ l x l 0 4 M " 1 ™ " 1 , suggesting the presence of similar species in solution. The high E values imply that these absorptions must be due to charge transfer, most likely from L —*• M. 223 6.D CONCLUSION The Ir(IV)-fluorosulfate system has been shown to be a para-magnetic analogue of the Pt(IV)-fluorosulfate system, although no attempts were made to test the iridium fluorosulfates 1 possible acidity in H S O 3 F . The d 5 low-spin electronic configuration of Ir(IV) resulted in some very interesting magnetic properties. From a dilution experiment involving the presumably isomorphous compounds of (C10 2 )2[M(S0 3F) 6], M=Ir, Pt, antiferromagnetic ex-change has been shown to be present in (C10 2 ) 2[Ir(S03F) 6], a l -though to a lesser extent than in the halide system. The conten-tion that the superexchange mechanism cannot operate effectively through a fluorosulfate ligand, f i r s t suggested in the Pd(II)-fluorosulfate system, is supported by the results here. Ir(S0 3F) 3 is diamagnetic, and therefore, most lik e l y octahe-dral. This system was not investigated in much detail as i t was not the original intention of studying iridium chemistry. 224 CHAPTER 7 MOLYDENUM(VI)-FLUOROSULFATE 7.A INTRODUCTION The investigation into the molydenum-fluorosulfate system is guided mostly by interests on synthetic and structural aspects. For a binary or ternary metal fluorosulfate, a l l previous work indicates that the highest oxidation state obtainable by the metal appears to be +4, although the coordination sphere around the central metal ion may be octahedral. This may be due to the limited oxidizing power of S 20 6F 2, the strongest oxidant in the fluorosulfate system, or due to the steric crowding in the coor-dination sphere i f a highly charged M(V) or M(VI) species is pre-sent. For molydenum, as well as tungsten, the +6 oxidation state is readily attainable, even when relatively weak oxidizing agents are used 1 4 0 . Mo(VI) and W(VI) are found in binary halides, MX6, and their complexes, as well as in oxyhalides mainly of the two types MOX^  and M02X2, with X=F and Cl. With the possible exception of Mo02F2 and MF 8 2 _, which are octa-coordinated, and MF7", which is seven-coordinated, these compounds have octahedral structures. In most of the oxyhalides, an octahedral geometry is obtained by using bridging oxygen to form highly polymerized compounds, whereas MoOF4 and WOF^  appear to have bridging fluorine 2 0 0 - 2 0 1 . Complexes containing terminal M=0 bonds generally show 225 strong Raman and i . r . absorptions in the region of 900-1100 cm - 1, which can be assigned to metal oxygen stretching 2 0 0 , 2 0 5 , 2 0 6 . With the absence of detailed crystallographic data, vibrational spectra of the oxyhalides have been used frequently to elucidate structures. The substitution of a fluorosulfate group for a halide ligand in the oxyhalides introduces a potential bridging group and diminishes the importance of oxygen bridging. The reason for this expectation is the larger S03F ligand would in-crease the distance between intermolecular M=0 groups and make bridging via the smaller oxygen atoms more d i f f i c u l t . Vibration-al spectroscopy should be very helpful in this case because v M-0 is readily detected and the conformation of the S O 3 F group is also evident from vibrational spectroscopy. Previously published work in this area is represented by the synthesis of W0(S03F)4 and a report on it s 19F-n.m.r. and Raman spe c t r a 2 1 1 . The compound is a clear liquid at room temperature and appears to be polymeric with both bridging and terminal S O 3 F groups; bands at ~960 cm - 1 in it s Raman spectrum has been assigned to the terminal W=0 stretch. The reaction of Mo or Mo(C0)6 with S 20 6F 2 is reported to give rise to two products: Mo0 2(S0 3F) 2, a white solid, and a red liquid that analyzed as MoO(S03F)2 9 0 . Since the chemistries of Mo(VI) and W(VI) are very similar, i t was decided to characterize the two previously reported molydenum oxyfluorosulfates and to attempt complex formation reactions on these compounds. 226 7.B EXPERIMENTAL 7.B.l PREPARATION OF Mo0 2(S0 3F) 2 Mo + xsS 20 6 F 2 • Mo0 2(S0 3F) 2 + MoO(S03F)Lf + S 2 0 5 F 2 (7.1) M o 0 2 ( S 0 3 F ) 2 was prepared by a method based on the original synthesis described by Shreeve and Cady. Molydenum metal (~0.5 g) was allowed to react with an excess of S 2 0 6 F 2 (~5 mL) in the tem-perature range from ~25 to ~70°C, resulting in the formation of a white solid and a red liquid. The mixture was vacuum filtered and washed with S 2 0 6 F 2 . The removal of a l l volatile materials from the solid residue yielded M o 0 2 ( S 0 3 F ) 2 . Mo0 2(S0 3F) 2 is an extremely hygroscopic white solid which decomposes at ~230°C into a blue solid. Analysis Mo Calculated % 29.42 Found % 29.55 Mo0 2(S0 3F) 2 can also be conveniently prepared by reacting Mo metal with C1S03F in a quartz reactor at ~40°C. In a typical ' reaction, Mo metal powder (386 mg, 4.02 mmol) was reacted with ClS0 3F (~4 mL) for 2 weeks, giving Mo0 2(S0 3F) 2 (1.315 g, 4.03 mmol) without contamination from MoO(S0 3F) 4. Mo + 6C1S03F -> Mo0 2(S0 3F) 2 + 2S 20 6 F 2 + 3C12 . (7.2) S 2 06F2 was detected by its i . r . spectrum. 227 7.B.2 PREPARATION OF Mo0(S03FK The red f i l t r a t e from the reaction of Mo with S 20 6F 2 was subjected to a dynamic vacuum with the sample held at ~40°C. After 12 hours, an orange-red liquid of very high viscosity was obtained. The sample analyzed as MoOCSOsF)^. MoO^O^)^ appears to show a large range of miscibility with S 20 5F 2 at room temperature. Upon heating at ~100°C, MoOCSOaF)^ decomposes according to: MoO(S03FK y Mo0 2(S0 3F) 2 + S 20 5F 2 , (7.3) with S 20 5F 2 detected by its i . r . spectrum. Analysis Mo S F Calculated % 18.88 25.24 14.95 Found % 19.13 24.82 15.13 228 7.C DISCUSSION 7.C.1 SYNTHESIS AND GENERAL DISCUSSION The reaction of molydenum metal with S 20 6F 2 resulted in the formation of two Mo(VI)-containing compounds — a white solid, and a red liquid which analyzed as MoOCSOsF)^, instead of the previ-ously reported Mo(IV)-containing species. Different ratios of the two oxyfluorosulfates seem to be obtainable under different reaction temperatures. If the reaction is performed at room tem-perature, most of the product is MoO(S03F)i+, but i f the reaction temperature is raised to ~70°C, Mo0 2(S0 3F) 2 is the main product. Furthermore, the thermal decomposition of MoO(S03F)it giving Mo02-(S0 3F) 2, although this does not constitute a viable synthetic route, because of relatively high v o l a t i l i t y of MoOCSOsF)^, also suggests a similar thermal stability trend. A scheme of formation of the two compounds, consistent with the above observations, may be presented: Mo + 3S 20 6F 2 'Mo(S03F)6' (7.4) ,Mo(S0 3F) 6 1 MoO(S03F)1+ + S 20 5F 2 (7.5) Mo0(S03FK Mo0 2(S0 3F) 2 + S 20 5F 2 . (7.6) The formation of molydenum hexakis(fluorosulfate) as the i n i t i a l oxidation product is followed by the elimination of S 20 5F 2, giving the two oxyfluorosulfates. The energetically favorable formation of a Mo=0 bond, coupled with the resulting reduction in the steric crowding around Mo(VI),are most lik e l y the contributing factors. This is also consistent with the general reaction 229 chemistry of S 2 O 6 F 2 , where the i n i t i a l reaction of '0S02F radicals leads to the formation of a fluorosulfate. The a b i l i t y of moly-denum to form Mo=0 bonds with p^-d^ donation from f i l l e d oxygen orbitals probably promotes the decomposition. Whereas W0(S0 3F)L, was reported to be obtainable by the reac-tion of the metal with S 20 6F 2 at ~100°C, i t appeared to be the sole product 2 7 . When the synthesis was repeated in this study, the removal of a l l volatile materials yielded a small amount of white precipitate. This could possibly be W0 2(S0 3F) 2, but no attempts were made to identify this solid or to increase i t s yield. In order to test the complex formation a b i l i t y of the Mo(VI)-oxyfluorosulfate, a slight excess of C102S03F was added to a mix-ture of Mo0 2(S0 3F) 2 and MoCKSOsF)^ in S 20 6F 2. The removal of a l l volatile materials at ~50°C did not lead to any weight increase and therefore, no complex formation was evident. If a complex was formed, i t must have been of a rather low thermal st a b i l i t y . 7.C.2 VIBRATIONAL SPECTRA  7.C.2.1 Mo02(S03F)2 The vibrational spectra of Mo0 2(S0 3F) 2 show a distinct pair of sharp peaks at ~994 and 955 cm - 1, with inverted intensities in the Raman and i . r . . These vibrations are assigned to stretching 230 vibrations of the 0=M=0 group, and their positions compare well with those found for liquid or gaseous Mo0 2Cl 2, in which terminal Mo=0 bonds have been postulated 2 0 ° . Therefore, i t appears that terminal Mo=0 bonds are^resent, and the relatively high inten-sity of the symmetric mode at 994 cm - 1 in the i . r . spectrum is consistent with a non linear 0=Mo=0 arrangement. The vibrational frequencies for Mo0 2(S0 3F) 2 are shown in Table 7.1 in comparison with those of SnCl 2(S0 3F) 2 3 3 . There are three possible geometries for the compound: a) tetrahedral with monodentate S03F groups, b) trigonal bipyramidal with one monodentate-and one bridging S03F group, or, c) octahedral with only bridging S03F groups and cis-M=0 bonds. The vibrational spectra indicate the absence of monodentate S0 3F groups, thus ruling out the f i r s t two possi b i l i t i e s . The observed spectra are most consistent with the presence of bidentate S03F groups in the compound, as indicated by the characteristically strong asymmetric S0 2 stretch at ~1170 cm - 1 in the i . r . spectrum. Strong vibrational couplings within the Mo0 2(S0 3F) 2 moiety and best reflected in the S0 2 stretching vibrations at ~1000 and ~1200 cm - 1, and between v M-0 and 6 0-S02F vibrations producing Raman intensity enhancements at ~650, ~450 and ~280 cm - 1, as found in anions of the type [M(S0 3F) n] m -, seem to be absent. The position of v SF is found at 873-885 cm - 1, indicating a strong 231 Table 7.1 VIBRATIONAL FREQUENCIES OF Mo0 2(S0 3F) 2 Mo0 2(S0 3F) 2 SnCl 2(S0 3F) 2 3 3 R. I.R. R. Assignment 1420 w cn 1394 w 1390 s,b 1389 s v a s b U 3 1244 v w 1247 1175 m 1170 s,b 1125 w n 1088 w 1089 vs v s 5 U 3 1038 w 1040 s,b 1061 m v a s s o 3 994 vs 990 m vs0=Mo=0 955 s 955s vas0=Mo=0 873 w 850 s,b 870 s vS-F 630 w 630 m 632 s 60=Mo=0 596 w 590 m 587 m «.n 561 vw 555 m 552 m 6 a s i U 3 ~450 vw,sh 445 m 442 s 60=Mo=0 402 m 410 w 424 m 6 rock 356 vs v sSnCl 312 s <5Sn-(0S02FK + def S-0 260 w SMo-COSOzF)^ 232 Mo(VI)-0S02F interaction 9 2 . In general, the vibrational bands assigned to the S03F group correspond closely to a pattern shown by SnCl 2(S03F) 2, also listed in Table 7.1; however, in the latter case, a linear Cl-Sn-Cl grouping is evident from both vibrational and 1 1 9Sn Mb'ssbauer spectroscopy. For this compound, the presence of bidentate bridging S O 3 F groups has been proposed 3 3 . For Mo02(S03F)2, an arrangement consisting of two oxygen atoms on one side of the molecule and four bridging S03F groups on the other side of the distorted octahedron can be proposed here. 7.C.2.2 MoCKSCVK The vibrational frequencies of MoOCSOsF)^ are listed in com-parison with Raman data of WCKSOsF)^ 2 1 1 in Table 7.2 (the rela-tive intensities of the Raman bands of W0(S03F)4 were not given in the reference). As can be seen, a good correspondence can be made between the spectra of the two compounds, suggesting struc-tural similarities. The terminal Mo=0 stretch is assigned to a sharp Raman band at 933 cm - 1. The presence of both bridging and terminal S0 3F groups is indicated by the band positions. The band at ~700 cm - 1, present also in the spectrum of W0(S03F)[+, is presumably due to bridging S0 3F groups in a specific comformation-al mode. Again, strong vibrational couplings are not observed for MoO(S03F)tt. Table 7.2 VIBRATIONAL FREQUENCIES OF MoO(S03F)lt MoO(S03FK I.R. 1460 m 1410 m 1243 vs 1165 w,b 1045 s,sh 1028 s 933 s 1450 s,sh 1400 vs,b 1200 vs,b 1025 vs 950 s 850 vw,b 840 s 735 m 720 s 636 m 630 m 582 vw 580 m,sh 550 m 555 s 451 w 275 w W0(S03FK 2 1 1 R. 1464 1416 1240 1163 1044 964 944 873 853 826 801 707 700 644 552 455 421 280 273 234 7.C.3 19F-n.m.r. SPECTRA OF MoO($03F)tt Since MoOfSOsF^ is very soluble in S 20 5F 2, solution 1 9 F -n.m.r. spectra of the compound were obtained. The spectrum of a ~10 mol% solution in S 20 6F 2 consists of a single broad peak at -42.3 ppm (Fig 7.1). The peak is highly asymmetric, with a shoulder definitely present on the. low-field side. When the concentration of MoO(S03F)Lt is increased to almost 100%, with only a trace of S 20 6F 2 as the reference, two definite resonances become evident at -42.0 and -41.4 ppm, although they are extreme-ly broad and overlapping. FIG. 7.1 F-19 nmr spectra of MoO(SO,F)4 ~10 mol % MoO (S0 3 F ) 4 in S ^ F , MoOtSOjF^ -42.3 ppm 4 ppm 4 ppm 235 A dilution of MoCKSOsF)^ with S 20 6F 2 is expected to lead to a break down of the polymer and thus a decrease in the number of bridging S03F groups. Therefore, the resonance at -42.0 ppm can be assigned as due to monodentate S0 3F groups while the one at -41.4 ppm must be due to bridging S O 3 F groups. The extreme broadness of the peaks indicates an exchange process involving the inter-convertion of the two types of S03F groups or possibly a random polymerization process. The broadness of the peaks also prevented any attempts at quantitative measurements on the relative abundance of the two types of SO3F groups, but i t appears that in almost pure MoO(S03F)i+, the resonance due to the bridging species is slightly more promi-nent. If the coordination around Mo(VI) is octahedral, the ratio of bridging to terminal S03F groups should be 1:3. The observa-tion here implies an almost equal population for both types of ligands, or an increase in the coordination number. Octa-coordi-nation for Mo(VI) has been suggested for Mo02F2 2 0 0 and MoF 8 2 - 2 0 1 . It is also li k e l y , from the broadness of the resonance and the lack of appreciable vibrational couplings between v Mo-0 and v S03F, that dynamic S D e c i e s with possibly various coordination numbers from six to eight, or even higher, may be present. 236 7.D CONCLUSION Two Mo(VI) oxyfluorosulfates have been characterized mainly by vibrational spectroscopy. For Mo0 2(S0 3F) 2, a white solid, hexa-coordination of molydenum with a cis-0=Mo=0 bond and bridging S03F groups is suggested. For MoOCSOsF)^, both bridging and ter-minal S O 3 F groups are present, with a terminal Mo=0 bond evident. From 19F-n.m.r., MoO(S03F)i+ appears to undergo intermolecular exchange reactions between the two types of differently bonded S0 3F ligands, possible resulting in the compound's existence as a liquid at room temperature. The presence of Mo(S0 3F) 6 as a possible intermediate in the synthesis of these two compounds is suggested, although i t could not be isolated. Neither oxyfluorosulfates appear to have any tendency towards accepting S03F" to form anionic complexes. 237 CHAPTER 8 GERMANIUM- AND TIN- FLUOROSULFATE 8.A INTRODUCTION The group IV halides are potential halide ion acceptors because of the possibility of the expansion of their coordination spheres to six (carbon is an exception due to its lack of avail-able low energy d-orbitals). The covalency of the bond between the element and electronegative atoms such as oxygen and the ha-logens increases as the element becomes more non-metallic, with the result that the lighter halides are volatile whereas PbFi* is a high-melting crystal. The fluorosulfates of C(IV) 7 9 , Sn(IV) 3 3 and Pb(IV) 2 0 7 , and the adduct of Si(S03F)(+.2CH3CN 9 6 have been reported, but only Sn(S0 3F) l t is known to form anionic complexes 3 2 . The other fluorosulfates are either unstable at room temperature or dissociate in HS03F solutions. During the investigation into the palladium-fluorosulfate system, described in Chapter 3, i t became apparent that SnfSOsF)^ could have a S0 3F-accepting a b i l i t y comparable to that of Pt(S0 3-F)it. Unfortunately, Sn(S0 3F) l t reportedly is insoluble in HS03F 3 3 , thus ruling out i t s usefulness as a superacid in that solvent system. As PbfSOsF)^ was reported to exhibit basic behavior in HS03F 2 0 7 , i t was decided to investigate the hitherto unknown ger-manium-fluorosulfate system as possibly an easily obtainable super-acid in H S O 3 F . Sn(S0 3F)i t is similar to Pt(S0 3F) l t in that in both systems, 238 the hexakis(fluorosulfato)-complexes can be synthesized. A com-plex of the empirical formula of CsPt ( S 0 3F) 5 was used as an i l l u s -tration of the polymerization of the platinum-fluorosulfate system in solution, although l i t t l e structural information could be ob-tained from vibrational spectra analysis other than that i t was polymeric in the solid state. By using 1 1 9 S n Mossbauer spectro-scopy, which had been applied to the study of SnCSC^FK and [Sn-(SOsF)^] 2-, i t may be possible to provide more insights into the structure of the analogous [Sn ( S 0 3F)5] x x~ complex, i f i t can be synthesized. The solution behavior of the tin-fluorosulfate sys-tem will also be investigated and compared to those of other M(IV)-fluorosulfates. 239 8.B EXPERIMENTAL 8.B.l Sn(IV)-FLUOROSULFATE 8.B.1.1 PREPARATION OF Sn(S0 3F) 4 Sn + 2S 20 6F 2 Sn(S0 3F) 4 (8.1) In a typical reaction, tin metal granules (121 mg, 1.019 mmol) were reacted with a mixture of S 20 6F 2 (~3 mL) and HS03F (~2 mL) at room temperature. A rapid reaction occurred and after 3 days, a l l of the metal had reacted and a white precipitate had formed. The removal of a l l volatile materials yielded a product identified by vibrational spectroscopy as Sn(S0 3F) l t (530 mg, 1.029 mmol). 8.B.1.2 PREPARATION OF KSn(S0 3F) 5 KC1 + HS03F • KS03F + HC1 (8.2) KS03F + Sn + 2S 20 6F 2 ^ 4 KSn(S0 3F) 5 (8.3) Tin metal granules (402 mg, 3.387 mmol) were added to a reac-tor containing solid KS03F.obtained from the solvolysis of KC1 (251 mg, 3.366 mmol) in HS03F. After a mixture of S 20 6F 2/HS0 3F ("3 mL) was d i s t i l l e d into the reactor, a violent reaction occur-red at room temperature. The reaction mixture was subsequently kept at ~50°C for 12 hours to ensure completion. The removal of al l volatile materials yielded a glass-like material which has a weight corresponding to that of KSn(S0 3F) 5 (2.201 g, 3.370 mmol). KSn(S0 3F) 5 is a clear, hygroscopic crystalline solid which 240 may be a supercooled glass at room temperature. It is very solu-ble in HSO3F. 8.B.1.3 PREPARATION OF CsSn(S0 3F) 5 CsCl + HSO3F > CsS03F + HC1 (8.4) CsS03F + Sn + 2S 20 6F 2 ^ 4 CsSn(S0 3F) 5 (8.5) In a manner analogous to the preparation of KSn(S0 3F) 5, tin metal granules (566 mg, 4.769 mmol) were allowed to react with CsS03F formed from the solvolysis of CsCl (802 mg, 4.764 mmol) in HSO3F, giving a white, crystalline, hygroscopic product which ana-lyzed as CsSn(S0 3F) 5 (3.492 g, 4.675 mmol). CsSn(S0 3F) 5 is soluble in HS03F. Analysis S F Calculated % 21.47 12.72 Found % 21.11 12.39 8.B.1.4 PREPARATION OF Cs 2[Sn(S0 3F) 6] 2CsCl + Sn + 3S 20 6F 2 ^ £ Cs 2[Sn(S0 3F) 6] + C l 2 (8.6) A mixture of dried CsCl (839 mg, 4.983 mmol) and tin metal granules (296 mg, 2.494 mmol) is reacted with S 20 6F 2/HS0 3F (~5 mL) at ~90°C for 1 day. After the removal of a l l volatile materials from the solution at 90°C, a white solid which corresponds to Cs 2[Sn(S0 3F) 6] (2.444 g, 2.497 mmol) was obtained. K 2[Sn(S0 3F) 6] can be prepared similarly. 241 8.B.2 GERMANIUM-FLUOROSULFATE 8.B.2.1 PREPARATION OF GeF 2(S0 3F) 2 Ge + 2S 20 6F 2 ^ 4 GeF 2(S0 3F) 2 + 2S03 (8.7) Germanium metal powder (129 mg, 1.777 mmol) was reacted with a mixture of S 20 6F 2/HS0 3F (~2 mL) at ~80°C for 3 days. A white sublimate formed on the cooler parts of the reactor and the pro-duct was vacuum f i l t e r e d , washed with S 20 6F 2 and dried by evacua-tion at room temperature for 30 minutes. A white powder which analyzed as GeF 2(S0 3F) 2 (428 mg, 1.386 mmol) was obtained; the yield was very low due to the loss of the product by removal in vacuo. GeF 2(S0 3F) 2 can also be obtained by the reaction of germanium metal powder with S 20 6F 2 at ~150°C for 1 week. This method, a l -though takes longer to complete, eliminates the f i l t r a t i o n process. GeF 2(S0 3F) 2 is a white, hygroscopic, flaky solid which is insoluble in HS03F. It is sublimable at room temperature in vacuo. Analysis Ge S F Calculated % 23.52 20.77 24.62 Found % 21.06 18.68 22.12 Although the analytical result appears to be low, the calculated mole ratio for Ge:S:F is 1.000:2.009:4.014, an excellent compari-son with the expected 1:2:4 ratio. 242 8.B.2.2 PREPARATION OF (C10 2) 2 [Ge(S0 3F) 6] 2C102S03F + Ge + 2S 20 6F 2 (C10 2) 2 [Ge(S0 3F) 6] (8.8) A mixture of C102S03F (~1 mL) and S 20 6F 2 (~4 mL) was d i s t i l l -ed into a reactor containing germanium metal powder (130 mg, 1.791 mmol). An exothermic reaction occurred at room temperature, and the metal dissolved into the deep red C102S03F layer. After keeping the reaction mixture at ~50°C for 1 day to ensure a com-plete reaction,the excess liquid reactants were removed by pump-ing at ~80°C for 12 hours. A light yellow solid, which analyzed as (C10 2) 2[Ge(S0 3F) 6] (126 mg, 0.157 mmol) was obtained; again, the low yield was caused by the v o l a t i l i t y of the product. (C10 2) 2[Ge(S0 3F) 5] is a light yellow, hygroscopic solid. It is sublimable in vacuo at room temperature, but melts at ~145°C under an atmosphere of nitrogen, giving off a C10 2S0 3F-like gas. Analysis Ge F Calculated % 9.05 14.22 Found % 8.75 14.22 243 8.C DISCUSSION 8.C.1 SYNTHESIS AND GENERAL DISCUSSION The synthesis of Sn(S0 3F)i +, using the medium of S 20 6F 2/HS0 3F, is an improvement over the original synthesis using SnCl^ 3 3 , both in terms of speed and purity of the product. Because of the possibility of HS03F being reduced by tin metal, a larger than normal ratio of S 20 6F 2 (~75 %) was used as a precaution, and the reaction mixture had to be shaken from time to time to ensure no local depletion of S 20 6F 2. When S 20 6F 2 was omitted from the reac-tion, the glass reactor became badly etched and the solution took on a blue color, most lik e l y due to sulfur cations 2 5 . For the same reason, the solvolysis of CsCl or KC1 in HS03F cannot be per-formed in the presence of metallic t i n , which must be added to the alkali-metal fluorosulfate in the dry-box. Although Sn(S0 3FK is insoluble in HS03F, the reactivity of tin metal towards S 20 6F 2 is greatly enhanced by the addition of the acid, suggesting a very slight solubility, combined with the energetically highly favorable formation of SnCSQ^F^ from Sn. CsSn(S0 3F) 5 does not precipitate Sn(S03F)(j in solutions in HS03F; this argues against the complex as being a mixture of Sn-(S0 3F) 4 and Cs 2[Sn(S0 3F) 6]. Furthermore, highly concentrated solutions can be obtained at room temperature (^ 1.3 mol/kg). KSn(S0 3F) 5 is extremely soluble in HS03F, and at high concentra-tions, results in the formation of a glass-like material which 244 loses H S O 3 F very slowly. The attempted synthesis of GeCSOaF)^ resulted in the forma-tion of a compound that analyzed as GeF 2(S03F) 2. The systematic lowering of the analytical results for the compound seems to sug-gest the presence of an inert impurity in the sample; small frag-ments of glass in the germanium metal powder could be the source of such an error. Mixed halide-fluorosulfate compounds of Sn(IV) such as SnF 2(S0 3F) 2 2 0 8 and SnCl 2(S0 3F) 3 3 are known; 1 1 9 S n Moss-, bauer and vibrational spectroscopy suggest the presence of termi-nal halide ligands and bridging S0 3F groups in these compounds. SnF 2(S03F) 2 can be prepared by the replacement of chlorine in SnF 2Cl 2 by S 20 6F 2 or C1S03F 2 0 8 and SnCl 2(S0 3F) 2 can be obtained by a ligand redistribution reaction involving SnClit and SnCSOsF)^ 3 3 . The existence of GeF 2(S03F) 2 instead of GeCSQ^FK as a stable compound at room temperature provides a transition between the Si-S0 3F and the Sn-S03F systems. Due to the preferential forma-tion of a Si-F covalent bond over a Si-0S0 2F bond, the product obtained from the fluorosulfonation of silicon compounds is in-variably SiFt+, and the compound Si(S03F)it'2CH3CN requires stabi-lization by the two coordinated n i t r i l e s 6 6 . SnCSOsF)^, on the other hand,is a crystalline solid stable up to 216°C 3 3 . In view of this, and the expectation that the electronegativity of ger-manium is between that of silicon and t i n , i t is not surprising to find the observed st a b i l i t y of the Ge-0S02F bond with respect to replacement by a Ge-F bond. 245 Ge(S03F)i+, however, can be stabilized by the formation of [Ge(S0 3F) 5] 2 _ complexes, but these also exhibit quite low thermal s t a b i l i t i e s , subliming or decomposing in vacuo at room tempera-ture. Attempts to prepare Cs 2[Ge(S0 3F) 6] resulted in the forma-tion of a CsS0 3F-rich product, as shown by Raman and i . r . spec-troscopy. (C10 2)2[Ge(S0 3F) 6] could be obtained as a pure product because the C10 2S0 3F formed from the decomposition is also vola-t i l e , but in order to ensure the complete removal of the excess C102S03F used in the reaction, a large quantity of the desired product was also lost. An alternative method may be to use f i l -tration instead of evacuation as a means of removing excess reac-tants, followed by washing with liquid S0 2 which can be evapora-ted at room temperature without the application of a vacuum. 8.C.2 VIBRATIONAL SPECTRA 8.C.2.1 CsSn(S0 3F) 5 As mentioned in Chapter 5, the vibrational spectra of CsSn-(S0 3F) 5 are very similar to those of KRu(S0 3F) 5 2 0 9 and CsPt(S0 3F) 5, suggesting structural similarities. Only a relatively poorly re-solved Raman spectrum could be obtained for CsSn(S0 3F) 5 due to the sample's fluorescence in the laser radiation. The Raman frequencies of CsSn(S0 3F) 5 are compared to the literature values Table 8.1 RAMAN FREQUENCIES OF CsSn(S0 3F) 5 CsSn(S0 3F) 5 1430 w 1406 w 1265 m 1220 w 1130 w 1112 w 1060 w 1055 w 1000 w 850 w 820 w 630 m 605 vw 585 w 560 w 430 w Sn(S0 3FK 3 3 1431 s 1420 s,sh 1233 s 1124 s 1075 s 986 w 911 m 845 m 827 m 640 w,sh 632 s 589 s 552 m 427. s 320 w K 2[Sn(S0 3F) 6] 1407 m 1390 m 1278 ms 1228 m 1208 w 1096 s 1002 m 859 m 836 m 823 m 625 s 582 m 560 m 435 m 416 m 360 m 266 w 247 for Sn(S0 3FK 3 3 and K 2[Sn(S0 3F) 6] 3 2 in Table 8.1 It appears that the Raman spectrum of CsSn(S03F)s contains bands from both the other two compounds' spectra, but some very intense bands have been shifted, indicating the existence of not merely a mixture of the two known compounds. For example, the strong S0 2 symmetric stretch is shifted to 1220 cm - 1; similarly, a downward shift of the band at ~1060 cm - 1 is also evident. The similarities between the three spectra are not unexpected because [Sn(S0 3F) 6] 2 _ contains only monodentate S03F groups, Sn-(S03F)i+ has an equal number of monodentate and bidentate S03F groups,and for CsSn(S0 3F) 5, there should be one bidentate S03F group for every four monodentate S0 3F groups. The presence of the diagnotic band at ~1130 cm - 1 in both the i . r . and Raman spec-tra of CsSn(S0 3F)6 confirms the existence of bidentate bridging S03F groups. 8.C.2.2 GeF 2(S0 3F) 2 AND (C10 2) 2[Ge(S0 3F) 6] Again because of the fluorescence of the sample when i t is irradiated by the laser, only a rather poorly resolved Raman spec-trum could be obtained for GeF 2(S0 3F) 2. The i . r . spectrum of the compound has a broad absorption from ~1000 to ~1150 cm - 1, indica-ting the presence of bridging S0 3F ligands. The vibrational fre-quencies of GeF 2(S0 3F) 2 are listed in Table 8.2 and compared to the i . r . frequencies of SnF 2(S0 3F) 2. The absence of the strongly Raman active symmetric S0 3 Table 8.2 VIBRATIONAL FREQUENCIES OF GeF 2(S0 3F) 2 GeF 2(S0 3F) 2 SnF 2(S0 3F) 2 2 0 8 I.R. R. I.R. Assignment 1390 vs 1426 w 1420 m,sh -1050 vs,vb 1064 w 1070 s,b vSOc 1405 vs,b V O U 3 1115 vs,b <-n -1100 vs,sh 1095 w 1103 s,sh V O U 3 vSOc 1055 w V O U 3 880 vs 880 vs 855 vs vSF 770 vs vGe-F asym 691 s vSn-F 660 s 656 w,sh 628 ms S0 3 bend 648 m vGe-F sym 570 s 576 w 590 s S0 3 bend 555s 560 w 548 vs S0 3 bend 495 s 490 w vGe-OS02F 420 m 437 w 430 ms S0 3 rock ~350 m vSn-0S02F 280 w 249 stretch at ~1220 cm - 1 in the spectrum of GeF 2(S0 3F) 2 denotes the absence of monodentate, tridentate and ionic S03F groups. There-fore, together with the feature noted in the S0 2 stretching re-gion in the i . r . spectrum, i t can be concluded that the fluoro-sulfate is present in a bidentate form. A good comparison can be made with the i . r . spectrum of SnF 2(S0 3F) 2 2 0 8 , suggesting simi-l a r i t i e s in the structure and bonding of the two compounds. A strong band at 770 cm - 1 in the i . r . spectrum of GeF 2(S03F) 2, not observed in i t s Raman spectrum, is assigned to the asymmetric stretching of the Ge-F bonds. A relatively strong Raman band at 648 cm - 1 is most likely the vGe-F sym (this assignment is based on the fact the band at ~650 cm"1 is usually not very strongly Raman active relative to the bands at ~450 and ~280 cm - 1). For the tin compound, the analogous vibrations are found at 691 and 612 cm-1. The mutual exclusion of the two Ge-F stretching vib-rations indicates a linear or nearly linear configuration for the [F-Ge-F] unit; a similar conclusion has been reached for the chlorine 3 3 , fluorine 2 0 8 and methyl 9 6 derivatives of tin(IV), thus indicating that a l l four compounds are closely related struc-turally. Assuming a linearity in the [Hal-M-Hal] bonds, a comparison can be made on the approximate stretching force constants of GeF 2(S0 3F) 2, SnF 2(S0 3F) 2 and SnCl 2(S0 3F) 2 2 0 6 (stretch-bend inter-actions are ignored): V l ( A l g ) = f ^ + M) 1 5.889xl0"7 (8.9) v 2 ( A 2 u ) = ' (y H +2y m)(k + ki) (8.10) 5.889xl0"7 where v is the frequency in cm - 1 , y is the reciprocal of the atomic masses of the halogen or metal atoms in a.m.u., and k, k-j are the stretching and stretch-stretch interaction force constants in m dynes/K, respectively. By using vx(Ge-F) = 648 cm - 1, v2(Ge-F) = 770 cm - 1, V l(Sn-F) = 612 cm - 1, v 2(Sn-F) = 691 cm"1 2 0 8 , vi(Sn-Cl) = 356 cm - 1, v 2(Sn-Cl) = 411 cm - 1 3 3 , the following force constants are obtained: k(Ge-F) = 4.53 ki(Ge-F) = 0.17 • k(Sn-F) = 4.12 ki(Sn-F) = 0.07 k(Sn-Cl) = 2.43 kj(Sn-Cl) = 0.22 Stretch-stretch interaction can also arise from the involve-ment of IT orbitals in the bonding; as one M-Hal bond stretches, lessening the interaction of the d^ orbitals on the metal to the halogen, these TT orbitals will become more available to the halo-gen on the other side, and TT bonding to them will become stronger. Consequently, this would result in kn* having a positive sign 2 0 6 . The degree of ir-bonding in the chlorine-containing compound is large, as can be seen from the relative values of kj and k. This is consistent with chlorine having d-orbitals of lower energy 251 than fluorine has. In spite of the increase in ir-bonding, the Sn-Cl bond is weaker than the Sn-F bond, presumably due to an increase in covalent character in the latter. For the same rea-son, the Ge-F bond is shown by these calculation to be much stronger than the Sn-F bond, and supports the contention that this may be an important factor contribution to the preferential formation of GeF 2(S0 3F) 2 instead of GeCSOsF)^. By extending the force constant calculations to the literature values for the v i -brational frequencies of the tetrahedral fluorides of CF^, SiF^ and GeF4 2 1 ° , using the formulae given in reference 206, a com-parison can be made in Table 8.3. Table 8.3 STRETCHING FORCE C0NSTANTSo0F SOME GROUP IV ELEMENT-FLUORINE BONDS (in mdynes/A) v(Ai g) v(F 2) v(A 2 u) k ki CF 4 2 1 0 908 1281 — 6.73 0.83 Si F 4 2 1 0 800 1010 — 6.29 0.29 GeFi+ 2 1 0 738 800 — 5.50 0.20 G e F 2 ( S 0 3 F ) 2 648 — 770 4.53 0.17 SnF 2(S0 3F) 2 2 0 8 612 — 691 4.12 0.07 It can be seen that the element-fluorine bond strengthens as the element becomes more electronegative. The substitution of a fluorine atom with two bridging fluorosulfate groups tends to weaken the covalent Ge-F bond by possibly a delocalization of the electron density away from germanium. The increase in the stretch-and-stretch interaction for CF 4 is most likely a result of steric 252 effects in the small molecule. For example, the stretching of one C-F bond would decrease the crowding for the remaining three fluorine atoms, leading to an increase in the strength of the other C-F bonds. In the vibrational spectrum of GeF 2(S03F) 2, vS-F is found at 880 cm-1, an unusually high frequency, suggesting the presence of a strong Ge-0 bond 9 2 . This is not unexpected since the Ge-F bond is strong and the electronegativities of oxygen and fluorine are similarly high. (C10 2) 2[Ge(S03F) 6] is an excellent Raman scatterer, and a good comparison can be made with the Raman frequencies of (C10 2) 2-[Sn(S0 3F) 6] 3 2 in Table 8. 3. The ionic nature of the complex is shown by the presence of the three vibrational modes for C10 2 + at ~1300, 1063 and 526 cm - 1 in the Raman spectrum. The rest of the spectrum is essentially that of a monodentate S03F group coordinated in a manner similar to other hexakis(fluorosulfato) complexes. It appears that the coordination of six S03F groups reduces the polarizing a b i l i t y of Ge(IV), as can be shown by a lowering of the stretching frequency of the Ge-0S02F vibration to 461 cm"1 from 495 cm'1 in GeF 2(S0 3F) 2. Furthermore, vS-F is lowered to the normal range at ~820 to ~850 cm - 1 from 380 cirr 1 in GeF 2(_S03F)2. These observations are not surprising in view of the increased delocalizing a b i l i t y of a S03F group as compared to a fluorine atom. Table 8.4 VIBRATIONAL FREQUENCIES OF (C102)2[Ge(S03F)6] R. 1414 w 1390 m 1380 m 1307 w 1295 vw 1272 m 1215 s 1205 s 1098 m 1063 vs 1001 w 985 vw 965 vw 847 w 821 w 631 m 591 w 567 w 554 w 526 m 461 w 436 m 274 m 255 m (C102)2[Ge(S03F)6] I.R. 1385 vs 1305 s 1295 s ~1200 vs,b 1020 vs,b 820 vs,b 640 s 580 s 550 s 520 m 432 w,sh (C102)2[Sn(S03F)6] R. 1412 w 1386 s 1308 m 1295 w 1265 ms 1206 s 1092 m 1062 vw 1030 s 990 sh 842 m 820 m 624 s 582 m 555 m 526 m 429 m 399 w 348 m 264 m 32 Assignment vS02 asym. v 3(C10 2 +) vS02 sym. vi (C10 2 + ) vO-S02F vS-F S02 bend S02 bend S02 bend v 2(C10 2 +) vGe-0S02F S02 rock vSn-0S02F 254 8.C.3 CONDUCTOMETRIC STUDIES IN HSQ3F Neither GeF 2 ( S 0 3F) 2 nor SnCSOsF)^ 3 3 show any appreciable solubility in HSO3F, and thus solution studies on the systems are limited to the anionic complexes. In the previous chapters, the conductivities of solutions of [ M ( S 0 3 F ) 6 ] 2 - , where M=Pd, Pt, Ir, Sn, have been found to be almost identical, suggesting a similar ionization scheme for a l l the complexes. From solution studies of the platinum(IV)-fluorosulfate system, [ P t ( S 0 3 F ) G ] 2 ~ has been shown to undergo a basic dissociation in HS0 3F. A conductometric titration of KSn(S03F)5 with KSO3F was un-dertaken to investigate the acid-base behavior of the complex in HSO3F. KSn(S0 3F) 5 is potentially a SO3F acceptor according to: KSn(S0 3F) 5 + KSO3F • K 2[Sn(S0 3F) 6] . (8 . 11) But by analogy with [Pt(S0 3F) 6] 2~, the reverse reaction may be appreciable. The result of such a titration is shown in Fig 8 . 1 and Table 8.5. The procedure involved in the calculation is described in detail in Appendix A but will be briefly explained here. Since there is no reason for the polymeric units of KSn(S0 3F) 5 to remain intact in the relatively dilute solution in which the titration was performed in , the species most l i k e l y present under such a condition is the solvated monomeric species, H[Sn(S0 3F) 6] . The titration is therefore the neutralization of this 'acid' by SO3F" according to: H[Sn(S0 3f) 6r + S03F" J i U [Sn(S0 3F) 6] 2" + HS03F . (8 . 1 2 ) FIG 8.1 80 10 V (J2~/cm) 60 The condjctornetric titration of K Sn(S03F) 5 with K S 0 3 F in HSO^F 4 0 i 20H (K/Sn-Table 8.5 CONDUCTOMETRIC TITRATION OF KSn(S0 3F) 5 WITH KS0 3F IN HS0 3F * <(obs) K(calc) I(calc) 0.0 48.52 48.52 0.090 0.1 46.69 43.90 0.100 0.2 45.99 43.12 0.112 0.3 45.76 42.83 0.123 0.4 45.98 42.73 0.133 0.5 46.35 42.80 0.143 0.6 46.86 43.09 0.152 0.7 47.46 43.72 0.160 0.8 48.13 44.94 0.167 0.9 49.16 47.29 0.173 1.0 51.50 51.50 0.178 l O^o^cm - 1 l O - ^ s r ^ n r 1 mol-kg - 1 units * Calculated using the following parameters: xJ(H 2 S0 3F +) = 320 X*(S0 3F") = 227 X*(K+) = 29 X*(H[Sn(S0 3F) 6]-) = 28 X*([Sn (S0 3F) 6] 2-) = 29 Ki = 5 . 1 x l 0 " 5 mol-kg-1 257 By fixing the i n i t i a l (R=0), and the last points (R=l) in the t i -tration curve, x£([Sn (S0 3F) 6] 2 _) and Ki can be obtained for each set of the other four variables listed in Table 8 .5 . In addition, the mobility of H 2 S 0 3F + is fixed at 320 2 h and those of S0 3F" and K+ are related 2 1 . Because the system is only weakly acidic, the autoprotolysis of HS0 3F must also be taken into account. A rea-sonable f i t with the experimental curve is obtained with the parameters listed in Table 8 .5 . The ionic mobilities and the value for Ki compare quite well with those obtained for the plat-inum -fluorosulfate system, described in Chapter 5. Again, [Sn-( S 0 3 F ) 6 ] 2 ~ has about the same ionic mobility as K + and H[Sn (S0 3F) 6]~, and can be attributed to the opposing effects of i t being a doubly charged ion and the resulting increased solvation. The equilibrium constant for the solvolysis of [Sn ( S 0 3F) 6] 2~ in HS0 3F according to: [Sn (S0 3F) f i] 2" + HS0 3F >• H[Sn(S0 3F) f i]" + S0 3F~ , (8.13) is about 7 . 5x l0 _ 1 + mol/kg, indicating a weakly basic behavior for the solute. In general, the two curves in Fig 8.1 deviates from each other by less than 8 %, this can be caused by: a) the dependence of X* on I used in the calculation is not a good approximation, and b) the average concentration of a l l ionic species during the titration is about 0.1 mol/kg, which cannot be s t r i c t l y defined as being dilute. Polymerization according to: 258 2H[Sn(S0 3F) 6]" > [Sn 2(S0 3F) 1 0] 2~ + 2HS03F (8.14) and 2[Sn(S0 3F) 6] 2- >• [Sn 2(S0 3F) 1 0] 2- + 2S03F", (8.15) etc, would serve to increase the basicity of the solution, and since H[Sn(S0 3F) 6]~ is only a very weak acid, the con-ductivity for the most part of the titration (from R slightly greater than 0) would be increased. The validity of the calculated Kj and X 0 values is tested by calculating the conductivity of CsSn(S0 3F) 5 and K 2[Sn(S0 3F) 6l 3 2 solutions in HS03F, shown in Fig 8.2. The concentration range * covered is quite large, and the approximation in X is not expect-ed to adequately describe the real system, especially when the x's were extrapolated from concentrated solutions in which solutions do not behave ideally. In any case, the calculated curves f i t the experimental curves better at higher concentrations, and no * further attempts were made to modify the dependence of X on I. In summary, the conductometric studies on solutions of KSn-(S0 3F) 5 in HS03F suggest a complicated behavior, involving both polymer-formation and weak proton-accepting properties. N.m.r. results, to be discussed later, also support such a hypothesis. The series of [M(S0 3F) 6] 2 -, where M=Sn, Pd, Pt, Ir, a l l appear to ionize in a similar manner in HS03F, forming weakly basic solutions. (C10 2) 2 [Ge(S0 3F) 6] dissolves in HS03F to give light yellow solutions, suggesting the presence of solvated C10 2 + 3 2 . The FIG 8.2 The electrical conductivity of KSn(SCLF) 5 in HS0 3 F at 25°C 0 .01 .02 .03 .04 m 260 conductivity of (C10 2)2[Ge(S0 3F) 6] in HS03F is shown in Fig. 8.3 and Table 8.6. In comparing these results with those of the cor-responding platinum species, i t is evident that the latter system has a higher conductivity of about 30%. The high conductivity of [Pt(S0 3F) 6] 2" and related solutes in HS03F has been attributed to solvolysis reactions, giving rise to S0 3F _, which, like HS0 3F +, has a high mobility in the solvent. It may also be inferred from this that in the solutions containing [Ge(S0 3F) 6] 2 -, the protona-tion according to: [Ge(S0 3F) 5]2- + HS03F * H[Ge(S0 3F) 6]- + S03F", (8.16) occurs to a lesser extent. By assuming X*( [Ge(S0 3F) 6] 2 _) = 30, and X*(C10 2 +) = 25, for a 0.01 mol/kg solution of (C10 2) 2[Ge(S0 3F) 6] that undergoes a dissociation giving rise to only C10 2 + and [Ge(S03-F e ] 2 - , the conductance of the solution should be about 9 x 10"4 j r ^ c n r 1 , when solvent conductivity is included. With x*(S03F") = 230, i t can be calculated that the equilibrium constant for equation (8.16) is about 25% that of the other [M(S0 3F) 6]2- complexes in solutions, or of the order of the solvent dissociation constant. Therefore, solutions of [Ge(S0 3F) 6] 2~ are anomalous when compared to those of the other [M(S0 3F) 6]2 _ complexes investi-gated in this study, displaying a much less conductivity. It is also not expected that the other [M(S0 3F) 6] 2 - complexes, where M=Pd, Pt, Ir, Sn, should have identical equilibrium constants in solution but slight variations in them are masked by the autopro-tolysis of HS03F because the amount of S03F~ derived from both FIG 8.3 CONDUCTIVITY OF (C10 2) 2[Ge(S0 3F) 6] IN HS03F Table 8.6 CONDUCTIVITY OF (C10 2) 2[Ge(S0 3F) 6] IN HS03F m 0.0000 0.0020 0.0040 0.0060 0.0080 0.0100 0.0120 0.0140 0.0160 0.0180 0.0200 0.0220 mol•kg-I. 162 3.699 5.799 7.854 9.806 II. 70 13.56 15.35 17.04 18.69 20.34 21.96 l O - ^ c n T 1 * interpolated 263 equilibria are of about the same order of magnitude. 8.C.4 N.M.R. STUDIES IN HS03F Both 1 1 9Sn- and 19F-n.m.r. spectroscopy were applied to the study of the Sn(IV)-fluorosulfate system in HS03F solution. The 119 Sn-n.m.r. spectra of CsSn(S0 3F) 5 and Cs 2[Sn(S0 3F) 6] are shown in Fig 8.4. Of al l the naturally occurring isotopes of t i n , two have nuclear spins of *§, they are 1 1 7 S n with an abundance of 7.61% and 1 1 9Sn with an abundance of 8.58%. 1 1 9Sn- n.m.r. has a rela-tive sensitivity of 5.18xl0~ 2 compared to ^-n.m.r., and for a 80 MHz spectrometer, the resonance for 1 1 9 S n is at about 29.88 MHz. Both of the 119Sn-n.m.r. spectra consist of a single reso-nance, with that of CsSn(S03F)5 shifted 41 Hz down-field from that of Cs 2[Sn(S0 3F) 6] (the frequency indicated on the spectra is simply for use as a reference). It becomes evident that the resonance for CsSn(S03F)5 is broad (~25 Hz at half-height as compared to ~9 Hz for Cs 2[Sn(S0 3F) 6]). The same is also found for the 1 9 F -n.m.r. spectra; even at a concentration of 0.12 m, much lower than used for the 119Sn-n.m.r. spectra, a single relatively broad peak occurs at -41.00 ppm. As an excess of KS03F was added to the solution, two sharp resonances appeared, a low intensity one at -41.75 ppm assigned to [Sn-S03F] species and a solvent reso-264 FIG 8.4 119Sn-N.M.R. SPECTRA IN HSC^F 0.63m C s 2 S n 2 ( S O j F ) 1 0 -5104.4Hz 0.68n Cs 2Sn(SOjF) g -5145.6Hz 4 2 0 ppm 265 nance at -40.75ppm. The solvent peak is hardly changed from that of pure HSO3F, and the solute peak compares well with that reported for [Sn(S0 3F) 6] 2~ solutions 3 2 . For more concentrated solutions of KSn(S03F)5, such as at 0.88 mol/kg, the single resonance is shifted to -41.42 ppm (close to that for other [Sn-S03F] species) and the resonance is noticeably broadened (~24 Hz at half-height). A 'titration' of KSn(S0 3F) 5 monitored by 19F-n.m.r. is shown in Fig.8.5; the sharpening and splitting of the resonance upon the addition of KSO3F are well illustrated . It therefore appears that an exchange process, involving SO3F groups, must be present in MSn(S0 3F) 5, where M=K, Cs, as proton exchange in species such as H[Sn(S0 3F) 6]" is not expected to give rise to the observed broadening in the spectra. The d i -merization of H[Sn(S03F) 6]" according to: 2H[Sn(S0 3F) 6]~ >• [ S n 2 ( S 0 3 F ) 1 0 ] 2 " + 2HS03F , (8.17) could produce such an effect. The dimerization of [Sn(S0 3F) 6] 2~ according to: [Sn(S0 3F) 6] 2- - > [ S n 2 ( S 0 3 F ) 1 0 ] 2 " + 2S03F" , (8.18) may not occur as readily because of i t s strong dependence on the acidity of the solution. This is especially true when an excess of KSO3F is added, resulting in the reduction of the solvolysis of [Sn(S0 3F) 6] 2" to form H[Sn(S0 3F) 6]. The small difference in the chemical environments of the fluorine atoms when the SO3F group is bound as a terminal or a bridging ligand can also lead to the broadening in the 19F-n.m.r. spectrum of KSn(S0 3F) 5. FIG. 8.5 F-19 nmr of [Sn (S0 3 F) x : f i n HS0 3 F -41.00 ppm [K Sn(S03F) 5] oonc.= 0.88 m -41.42 ppm 10 ppm -I h 10 ppm 267 In summary, it appears that the species present in HSO3F solutions of M*Sn(S03F)5 are dynamically exchanging S03F-contain-ing moieties with the solvent, leading to a broadening and coales-cence of the 1 9 F resonance; this is supported also by the 1 1 9 S n -n.m.r. results. Some differences exist between the n.m.r. studies described here and those involving platinum: a) There was no Sn-F coupling observed. For the 1 9 F spectra, the low natural abundance of 1 1 7 S n and 1 1 9 S n may be a fac-tor here. For the 1 1 9 S n spectra, in which a septet is ex-pected for [Sn(S0 3 F) 6 ] 2 _ , the relatively sharp resonance suggests that either no Sn-F coupling occurs, or J(Sn-F) is very small ( s5 Hz). b) In the 1 9 F-n.m.r. spectra of the platinum compounds, no shifting or broadening in the resonance was observed. This is most likely due to the large separation between the solvent resonance and the [Pt-S03F] resonance, and does not rule out the occurrence of exchange reactions. 268 8.D CONCLUSION Complexes of the type of M ISn(S0 3F) 5, where M=Cs, K, have been prepared and characterized by vibrational spectroscopy and solution studies involving conductometry and n.m.r. spectroscopy. These complexes are similar to those containing Pt(IV); the ana-logy also extends to the hexakis(fluorosulfato)-complexes of Pt(IV), Ir(IV) and Pd(IV) in terms of their structures and solu-tion behaviors. Unlike K[Au(S03F)iJ , KSn(S0 3F) 5 is weakly acidic, and K 2[Sn(S0 3F) 6] is weakly basic, when dissolved in HS03F. Using n.m.r. spectroscopy, exchange reactions involving the transfer of S0 3F groups seem to be prevalent in solutions of KSn(S0 3F) 5; this has been attributed to a dimerization or polymerization in solu-tion. Unfortunately, no 1 1 9 S n Mossbauer studies could be per-formed on the complexes, as a suitable source was not available. This study confirms the conclusion arrived in the palladium-fluorosulfate system that SnCSOsF)^ is a strong fluorosulfate acceptor, and i t would have been a practical superacid in the HS03F solvent system i f not for i t s insolubility. In the germanium-fluorosulfate system, the tetrakis(fluoro-sulfate) was not obtainable; instead, GeF 2(S0 3F) 2 was obtained as the product from the oxidation of germanium metal with S 20eF 2. From vibrational spectroscopy, the fluoride-fluorosulfate seems to be isostructural with SnF 2(S0 3F) 2. Although GeCSOsF)^ could not be successfully prepared, an anionic complex containing [Ge-(S0 3 F ) 6 ] 2 ~ could be isolated, but i t too displays limited thermal 269 s t a b i l i t y in vacuo. The [Ge-S03F] system bridged a gap between the highly unstable fluorosulfates of carbon and s i l i c o n , and the easily formed SnCSOsF)^. 270 CHAPTER 9 CONCLUSION From a synthetic point of view, this study has provided a very versatile preparative reagent for the synthesis of metal fluorosulfates and their anionic complexes. The mixture of S 20 6F 2/ HSO3F combines the strong oxidizing a b i l i t y of S 20 6F 2 with the ionic solvatlng property of HSO3F, and can be used from below room temperature up to almost 200°C. The usefulness of this oxidizing mixture is in i t s adaptability to most of the synthetic reactions described in this thesis, giving rise to compounds con-taining the highest oxidation state of the metal in the fluorosul-fate system. Fluorosulfonation reactions using S 20 6F 2/HS0 3F do not result 1n the formation of a reactive by-product, like the halogen-monofluorosulfates do, and 1n cases where by-products are formed, they usually come from the decomposition of the metal fluorosulfate. Of the systems investigated here, Au(S03F)3 and PtCSOsF)^ have been found to be superacids 1n HS03F, with acidities compar-able to that of SbF 2(S0 3F) 3, the strongest superacid in the sol-vent system. Novel polybromine cations, B r 3 + and B r 5 + can be stabilized by [Au(S0 3F)tJ", illustrating the low basicity and low oxidizing a b i l i t y of the anion. The two ansolvo acids are polymerized 1n the solid state, and, in the platinum system, also in solution, as Illustrated by conductivity results and the syn-thesis of CsPt(S0 3F) 5-type complexes. 271 Although being a strong acceptor of SO3F- is the f i r s t requirement for a superacid, too strong an acceptor property also seems to lead to its demise. This can be illustrated in the Sn-SO3F system, for which evidence from electronic spectra of Pd(II)-fluorosulfato complexes seems to suggest an extremely strong S O 3 F -accepting property. It is perhaps because of this reason that Sn(S0 3F)i + is highly polymeric and insoluble in HS0 3F. The study into the palladium system provided the f i r s t interpretation of the ligand f i e l d parameters of Pd(II) in an octahedral ligand f i e l d . The magnetic susceptibility of Pd(II) in various fluorosulfato complexes was found to have Curie-Weiss Law behavior in magnetically dilute systems. The iridium-fluoro-sulfate system is essentially a paramagnetic analogue of platinum's. SUGGESTIONS FOR FURTHER WORK IN THIS AREA a) A quantitative measure of acidity, the Hammett coefficients, should be applied to H [Au(S0 3FK] and H 2[Pt ( S 0 3F) 6]. It may also be useful to investigate the n.m.r. technique of mea-surement 5 1 » 5 2 . b) The application of the two superacids is a definite possibi-l i t y , especially in view of their high acidity, high thermal stability and low oxidizing power. The catalytic polymeri-zation of small organic molecules such as CHi+ should serve 272 as a comparison with the HS03F-SbF5 superacid system. c) Since [Au(S03F)i+]" can stabilize B r 5 + , other similarly large novel cations may be synthesized, containing the same counter-anion. d) The application of 1 1 9 S n Mossbauer to the study of [Sn(S0 3F) 5] x", and 1 9 5 Pt-n.m.r. to the platinum-fluorosulfate system should provide insights into the polymerization and bonding in these compounds. e) The group IVB halides such as TiCli+, TiF^, are also strong Lewis acids; their corresponding fluorosulfates may provide economical alternatives to the noble metal superacids. f) The group VB and VIB oxyfluorosulfates appear to be complica-ted, multicomponent systems; some preliminary results from this study seem to indicate exchange reactions resulting in the formation of metal-fluorine bonds and weakly acidic beha-vior in the tantalum system. These may merit further investi-gations. g) The HSO3CF3 solvent system is more tolerant to water than HS0 3 F is . Although Au(S0 3 CF 3 ) 3 is insoluble in HS0 3 CF 3 2 1 2 , and PtCSOsCFs)^ appears to behave likewise, this may be the next solvent system to expand into, especially since it is also a very strongly acidic medium. 273 REFERENCES 1) T.E. Thorpe and W. Kirman, J. Chem. Soc. London, 921 (1892). 2) W. Traube, Chem. Ber., 4 6 , 2513 (1913). 3) A.A. Woolf, J. Chem. Soc, 2840 (1954). 4) E. Hayek and A. Czaloun, Monatsh, 87, 790 (1956). 5) R.C. Paul, K.K. Paul, and K.C. Malhotra, Inorg. Nucl. Chem. Lett., 5, 689 (1969). 6) W. Traube and E. Reubke, Chem. Ber., 45B, 1618 (1921). 7) W. Lange, Z. Anorg. Allg. Chem., 215, 321 (1933). 8) R.J. Gillespie and E.A. Robinson, Can. J. Chem., 40, 644 (1962). 9) R.J. Gillespie, J.B. Milne, and J.B. Senior, Inorg. Chem., 5_, 1233 (1966). 10) W. Lange, Fluorine Chemistry, 1, 167 (1950). 11) G.J. Gillespie, Acc. Chem. Res.J, 202 (1968). 12) R.J. Gillespie and T.E. Peel, Adv. Phys. Org. Chem., 9, 1 (1971). 13) R.C. Thompson, Inorg. Sulfur Chem., Edited by G. Nickless, Elsevier, Amsterdam (1968). 14) A.W. Jache, Adv. Inorg. Chem. Radiochem., 16, 177 (1975). 15) S. Natarajan and A.W. Jache, Chem. of Non Aqueous Solvents, VB, 53 (1978). 16) G. Olah and J. Sommer, Rlcherche, 10, 624 (1979). 17) G.C. Cady, Adv. Inorg. Chem. Radiochem., 2, 123 (1960). 18) S.M. Williamson, Prog. Inorg. Chem., 7_, 39 (1968). 19) R.A. De Marco and J.M. Shreeve, Adv. Inorg. Chem. Radiochem., 16, 109 (1974). 20) A. Engelbrecht, Angew Chem. Int., Ed., 4, 641 (1965). 274 21) J. Barr, R.J. Gillespie, and R.C. Thompson, Inorg. Chem., 3_, 1149 (1964). 22) A. Commeyras and G.A. Olah, J. Am. Chem. Soc, 91_, 2929 (1969). 23) G.A. Olah, Chem. Eng. News., 45, 76 (March 27, 1967). 24) R.C. Thompson, J. Barr, R.J. Gillespie, J.B. Milne, and R.A. Rothenbury, Inorg. Chem 4, 1641 (1965). 25) R.J. Gillespie and J. Passmore, Adv. Inorg. Chem. Radiochem., 17, 49 (1975). 26) W.W. Wilson and F. Aubke, J. Fluorine Chem., 13, 431 (1979). 27) R.E. Noftle and G.H. Cady, J. Inorg. Nucl. Chem., 29, 969 (1967). 28) R.J. Gillespie and M.J. Morton, .Inorg. Chem., 11, 586 (1972). 29) R.J. Gillespie and M.J. Morton, Inorg. Chem., 11, 591 (1972). 30) R.J. Gillespie and E.A. Robinson, Can. J. Chem., 39, 2189 (1961). 31) P.A. Yeats, B. Landa, and F. Aubke, Inorg. Chem., 15_, 1452 (1976). 32) P.A. Yeats, J.R. Sams, and F. Aubke, Inorg. Chem., 12, 328 (1973). 33) P.A. Yeats, B.L. Poh, B.F.E. Ford, J.R. Sams, and F. Aubke, J. Chem. Soc. A, 2188 (1970). 34) M. Lustig and G.H. Cady, Inorg. Chem., _1, 714 (1962). 35) H.A. Carter, S.P.L. Jones, and F. Aubke, Inorg. Chem., 9, 2485 (1970). 36) R.J. Gillespie and J.B. Milne, Inorg. Chem., 5_, 1236 (1966). 37) F. Aubke and G.H. Cady, Inorg. Chem., 4, 269 (1965). 38) F. Aubke and R.J. Gillespie, Inorg. Chem., ]_, 599 (1968). 39) R.J. Gillespie, Inorganic Sulfur Chem., Edited by G. Nickless, Elsevier, Amsterdam (1968). 40) Handbook of Chemistry and Physics, 53rd Edition., Chemical Rubber Company Press, Cleveland (1973). 275 41) P.A.W. Dean and R.J. Gillespie, J. Am. Chem. Soc, 92, 2362 (1970). 42) R.J. Gillespie and T. Birchall, Can. J, Chem., 41, 148 (1963). 43) R.J. Gillespie and E.A. Robinson, Can. J. Chem., 40, 675 (1962). 44) G.A. Olah, Chem. in Brit., 8, 281 (1972). 45) G.A. Olah and C.W. McFarland, Inorg. Chem., U, 845 (1972). 46) R.J. Gillespie and E.A. Robinson, Can. J. Chem., 39, 2171 (1961). 47) A. Engelbrecht and E. Tschager, Z. Anorg. ATlg. Chem., 433, 19 (1977). 48) R.J. Gillespie and T.E. Peel, J. Am. Chem. Soc, 95, 5173 (1973). 49) R.J. Gillespie, T.E. Peel, and E.A. Robinson, J. Am. Chem. Soc, 93, 5083 (1971). 50) R.J. GillesDie, K. Ouchi, and G. P. Pez, Inorg. Chem., 8, 63 (1969). 51) J. Sommer, P. Rimmelin, and T. Drakenberq, J. Am. Chem. Soc, 98, 2671 (1976). 52) J. Sommer, S. Schwartz, P. Rimmelin, and P. Cam"vet, J. Am. Chem. Soc, 100, 2576 (1978). 53) G.M. Krammer, J. Org. Chem., 40, 298 (1975). 54) G.M. Krammer, J. Org. Chem., 40, 302 (1975). 55) R.J. Gillespie, 1973 I.U.P.A.C. Symposium on Non Metal Chem. 56) R.J. Gillespie and J.B. Milne, Inorg. Chem., 5_, 1577 (1966). 57) R.J. Gillespie and K.C. Malhotra, Inorg. Chem., 8, 1751 (1969). 58) R.J. Gillespie and M.J. Morton, Chem. Comm., 24, 1565 (1968). 59) G.A. Olah and M.B. Comisarow, J. Am. Chem. Soc, 90, 5033 (1968). 60) R.J. Gillespie and M.J. Morton, Inorg. Chem., 9, 811 (1970). 276 61) D.A.Edwards, M.J. S t i f f , and A.A. Woolf, Inorg. Nucl. Chem. Lett., 3, 427 (1967). 62) J. Goubeau and J.B. Milne, Can. J. Chem., 45, 2321 (1967). 63) 0. Ruff, Chem. Ber., 47, 656 (1914). 64) J.N. Brazier and A.A. Woolf. J. Chem. Soc. A., 99, (1967). 65) R.J. Gillespie and R.A. Rothenburg, Can. J. Chem., 42, 416 (1964). 66) E. Hayek, A. Czaloun, and B. Krismer, Monatsh, 87_, 741 (1956). 67) E.L. Muetterties and D.D. Coffmann, J. Am. Chem. Soc, 80, 5914 (1958). 68) F.B. Dudley, J. Chem. Soc, 3407 (1963). 69) S.D. Brown and G.L. Gard, Inora. Nucl. Chem. Lett., 11_, 19 (1975). 70) F.B. Dudley and G.H. Cady, J. Am. Chem. Soc, 79, 513 (1957). 71) J.M. Shreeve and G.H. Cady, Inorg. Syntheses, 7, 124 (1963). 72) F.B. Dudley and G.H. Cady, J. Am. Chem. Soc, 85_, 3375 (1963). 73) E. Castellano, R. Gatti, J.E. Sicre, and H.J. Schumacher, Z. Phys. Chem. (Frankfurt am Main), 42, 174 (1964). 74) H. Imoto and F. Aubke, J. Fluorine Chem., 15, 59 (1980). 75) R. Dev, W.M. Johnson , and G.H. Cady, Inorq. Chem., 11., 2259 (1972). 76) A. Storr, P,A. Yeats, and F. Aubke, Can. J. Chem., 50, 452 (1972). 77) W.P. Gil breath and G.H. Cady, Inorg. Chem., 2, 496 (1963). 78) J.E. Roberts and G.H. Cady, J. Am. Chem. Soc, 82, 352 (1960). 79) D.D. DesMarteau, Inorg. Chem., 7, 434 (1968). 80) C. Chung and G.H. Cady, Z. Anorg. Allg. Chem., 385, 18 (1971). 81) W.M. Johnson and G.H. Cady, Inorg. Chem., _12, 2481 (1973). 277 82) W.M. Johnson, R. Dev, and G.H. Cady, Inorg. Chem., 11, 2260 (1972). 83) P.C. Leung and F. Aubke, Inorg. Nucl. Chem. Lett., 13_, 263 (1977). 84) P.C. Leung and F. Aubke, Inorg. Chem., 17, 1765 (1978). 85) W.W. Wilson and F. Aubke, Inorg. Chem., 13, 326 (1974). 86) P.A. Yeats, W.W. Wilson, and F. Aubke, Inorg. Nucl. Chem. Lett., 9, 209 (1973). 87) C.S. Alleyne, K. 0'Sullivan-Mailer, and R.C. Thompson, Can. J. Chem., 52, 336 (1974). 88) G.C. Kleinkoff and J.M. Shreeve, Inorg. Chem., 3_, 607 (1964). 89) J.E. Roberts and G.H. Cady, J. Am. Chem. Soc, 82, 353 (1960). 90) J.M. Shreeve and G. H. Cady, J. Am. Chem. Soc, 83, 4521 (1961). 91) W.W. Wilson, Ph. D. Thesis, U.B.C., 1975. 92) D.W.J. Cruickshank and B.C. Webster, Inorganic Sulfur Chemistry, Ed. by G. Nickless, Elsevier, Amsterdam, 1968. 93) K. O'Sullivan, R.C. Thompson, and J. Trotter, J. Chem. Soc A, 2024 (1967). 94) A.M. Qureshi, H.A. Carter, and F. Aubke, Can. J. Chem., 49, 35 (1971). 95) A.M. Qureshi, L.E. Levchuk, and F. Aubke, Can. J. Chem., 49, 2544 (1971). 96) P.A. Yeats, B.F.E. Ford, J.R. Sams, and F. Aubke, J. Chem. Soc, Chem. Comm., 791 (1969). 97) F.A. Allen, J. Lerbscher, and J. Trotter, J. Chem. Soc. A., 2507 (1971). 98) K.C. Lee and F. Aubke, Can. J. Chem., 57, 2058 (1979). 99) J.M. Taylor and R.C. Thompson, Can. J. Chem., 49, 511 (1971). 100) J.R. Dalziel, R.D. Klett, P.A. Yeats and F. Aubke, Can. J. Chem., 52, 231 (1974). 278 101) A. Vogel, Quantitative Inorg. Analysis, 3rd Ed., J. Wiley and sons, New York (1961). 102) J.E. Lind, J.J. Zwolenik, and R.M. Fuoss, J. Am. Chem. Soc, 81, 1557 (1959). 103) R.J. Gillespie, J.B. Milne, and R.C. Thompson, Inorq. Chem., 5., 468 (1966). 104) N. Bartlett and R. Maitland, Acta Crystallogr., 747 (1958). 105) H.C. Clark and R.J. O'Brien, Can. J. Chem., 39, 1030 (1961). 106) B.N. Figgis and R.S. Nyholm, J. Chem. Soc, 4190 (1958). 107) Landolt-BBrnstein, Numerical Data and Functional Relation-ships in Science and Technology, Vol. 2 and Sup. 2, Springer-Verlag, Berlin (1966,1976). 108) N. Bartlett and P.R. Rao, Proc Chem. Soc, 393 (1964). 109) H.A. Carter, A.M. Qureshi, and F. Aubke, J. Chem. Soc, Chem. Comm., 1461 (1968). 110) F.A. Cotton and G. Wilkinson, Adv. Inorg. Chem., 3rd Ed., p. 924, Interscience, New York (1972). 111) W.E. Falconer, F.J. DiSalvo, A.J. Edwards, J.E. Gr i f f i t h s , W.A. Sunders, and M.J. Vasile, J. Inorg. Nucl. Chem. Sup-plement, 59 (1976). 112) L.F. Warren and M.F. Hawthorne, J. Am. Chem. Soc, 92, 1157 (1970). 113) N. Bartlett and M.A. Hepworth, Chem. and Industry, 1425 (1957). 114) E.L. Wagner and D.F. Hornig, J. Chem. Phys., 18, 296 (1950). 115) D. Babel, in Structure and Bonding, 3_» Springer-Verlag (1967). 116) D. Paus and R. Hoppe, Z. Anorg. Allg. Chem., 431, 207 (1977). 117) N. Bartlett and J.W. Quail, J. Chem. Soc, 3728 (1961). 118) Brauer, Handbook of Prep. Inorg. Chem., Academic Press, N.Y.,(1963). 119) Ref. 110, p. 1031. 279 120) M. Wilhelm and R. Hoppe, Z. Anorg. Allg. Chem., 424, 5 (1976). 121) R. Mattes, Z. Anorg. Allg. Chem., 364, 290 (1969). 122) N. Bartlett, Angew. Chem., Int. Ed., 7_, 433 (1968). 123) A.G. Sharpe, J. Chem. Soc, 3444 (1950). 124) R.S. Nyholm and A.G. Sharpe, J. Chem. Soc, 3579 (1952). 125) A. Tressaud, M. Wintenberger, N. Bartlett, and P. Hagenmuller, CR. Herbd. Seances Acad. Sci. Ser. C , 282, 1069 (1976). 126) H. Henkel and R. Hoppe, Z. Anorg. Allg. Chem., 359, 160 (1968). 127) P.R. Rao, A. Tressaud, and N. Bartlett, Inorg. Nucl. Chem. Lett, 23 (1976). 128) A.F. Wright, B.E.F. Fender, N. Bartlett, and K. Leary, Inorg. Chem., 17, 746 (1978). 129) R. Hoppe and W. Klemm, Z. Anorg. Allg. Chem., 268, 364 (1952). 130) A.G. Sharpe, J. Chem. Soc, 197 (1953). 131) L. Graham, Lawrence Berkeley Lab. Report, 1978, LBL-8088. 132) K.C. Lee and F. Aubke, Can. J. Chem., 55, 2473 (1977). 133) A. MacCragh and W.S. Koski, J. Am. Chem. Soc, 87, 2496 (1965). 134) H.A. Carter, Ph. D. Thesis, 1970, U.B.C 135) D.D. DesMarteau and M. Eisenberg, Inorg. Chem., _11, 2641 (1972). 136) M. Wechberg, P.A. Bulliner, F. 0. Sladky, R. Mews and N. Bartlett, Inorg. Chem., IJ., 3063 (1972). 137) C.S. Alleyne, M. Sc. Thesis, 1968, U.B.C. 138) K.O. Christe, C. J. Schack, D. Pillpovich, and W. Sawodny, Inorg. Chem., 8, 2489 (1969). 139) A.B.P. Lever, Inorganic Electronic Spectroscopy, Elsevier, Amsterdam, (1968). 140) Ref. 110, p. 998. 280 141) B.N. Figgis, Introduction to Ligand Fields, Interscience, N.Y. (1967). 142) H.B. Gray and C.J. Ballhausen, J. Am. Chem. Soc, 85_, 260 (1963). 143) Ref. 110, p. 561.*""" 144) P. Kohler, W. Massa, D. Relnen, B. Hofmann, and R. Hoppe, Z. Anorg. Allg. Chem., 446, 131 (1978). 145) T.M. Dunn, Trans. Far. Soc, 48, 1441 (1961). 146) V. Halpen, Proc. Roy. Soc, 291A, 113 (1966). 147) F.O. Sladky and N. Bartlett, J. Chem. Soc. A., 2188 (1969). 148) R.C. Thompson, unpublished results. 149) W. Rudorff, J. Kandler, and D. Babel, Z. Anorg. Allg. Chem., 317, 261 (1962). 150) A. Westland, R. Hoppe, and S. Kaseno, Z. Anorg. Allg. Chem., 338, 319 (1965). 151) R.P. Rao, Ph. D. Thesis, U.B.C. 1965. 152) J.H. Waters and H.B. Gray, J. Am. Chem. Soc, 87, 3534 (1965). 153) K. Leary and N. Bartlett, J. Chem. Soc, Chem. Comm., 903 (1972). 154) K. Leary, A. Zalkin, and N. Bartlett, J. Chem. Soc, Chem. Comm.,131 (1973). 155) M.J. Vasile, T. J. Richardson, F.A. Stevie, and W.E. Falconer, J. Chem. Soc, Dalton 351 (1976). 156) B. Armer and H. Schmidbaur, Angew. Chem. Int., Ed., % 101 (1970). 157) B.F.G. Johnson, Gold Bulletin Johannesberg, 4, 9 (1971). 158) H. Schmidbauer, Angew. Chem. Int., Ed., 15_, 728 (1976). 159) R.J. Puddephatt, The Chemistry of Gold, Elsevier, Amsterdam, 1978. 160) F.W.B. Einstein, P.R. Rao, J. Trotter, and N. Bartlett, J. Chem. Soc. (A.), 478 (1967). 281 161) A.J. Edwards and G.R. Jones, J. Chem. Soc. A., 1936 (1969). 162) CD. Garner and S.C. Wallwork, J. Chem. Soc. A., 3092 (1970). 163) A.M. Qureshl and F. Aubke, Inorg. Chem., 10, 1116 (1971). 164) A.A. Woolf, J. Chem. Soc, 433 (1955). 165) K.C. Lee and F. Aubke, Inorg. Chem., 18, 389 (1979). 166) K.C. Lee and F. Aubke, Inorg. Chem., 19, 119 (1980). 167) R.S. Nyholm and A.G. Sharpe, J. Chem. Soc, 3579 (1952). 168) L.B. Asprey, F.H. Kruse, K.H. Jack, and R. Maitland, Inorg. Chem., 3, 602 (1964). 169) D.W.A. Sharp and J. Thorley, J. Chem. Soc, 3557 (1963) and references therein. 170) 0. Glemser and A. Smalc, Angew. Chem. Int., Ed., 8, 517 (1969). 171) A. Smalc, Abstract, 4th European Symposium on Fluorine Chemistry, 1972. 172) W.W. Wilson, J.M. Winfield and F. Aubke, Inorg. Chem., 1_1, 2260 (1972). 173) D.J. Merryman, P.A. Edwards, J.D. Corbett and R.E. MacCarley, J. Chem. Soc, Chem. Comm., 779 (1972). 174) J.W. Moore, J.W. Baird, and H.B. Miller, J. Am. Chem. Soc, 90, 1359 (1968). 175) CS. Alleyne and R.C. Thompson, Can. J. Chem., 52_, 3218 (1974). 176) W.W. Wilson, B. Landa, and F. Aubke, Inorq. Nucl. Chem. Lett., 11, 529 (1975). 177) CC. Addison, G.S. Brownlee, and N. Logan, J. Chem. Soc, Dal ton, 1440 (1972). 178) P. Gallezot, D. Weigel, and M. Prettre, Acta Crystallogr., 22, 699 (1967). 179) K.O. Christe, and C.J. Schack, Adv. Inorg. Chem. Radiochem., 18, 319 (1976). 281 161) A.J. Edwards and G.R. Jones, J. Chem. Soc. A., 1936 (1969). 162) CD. Garner and S.C. Wallwork, J. Chem. Soc. A., 3092 (1970). 163) A.M. Qureshi and F. Aubke, Inorg. Chem., 10, 1116 (1971). 164) A.A. Woolf, J. Chem. Soc, 433 (1955). 165) K.C Lee and F. Aubke, Inorg. Chem., 18, 389 (1979). 166) K.C. Lee and F. Aubke, Inorg. Chem., 19, 119 (1980). 167) R.S. Nyholm and A.G. Sharpe, J. Chem. Soc, 3579 (1952). 168) L.B. Asprey, F.H. Kruse, K.H. Jack, and R. Maitland, Inorg. Chem., 3, 602 (1964). 169) D.W.A. Sharp and J. Thorley, J. Chem. Soc, 3557 (1963) and references therein. 170) 0. Glemser and A. Smalc, Angew. Chem. Int., Ed., 8_, 517 (1969). 171) A. Smalc, Abstract, 4th European Symposium on Fluorine Chemistry, 1972. 172) W.W. Wilson, J.M. Winfield and F. Aubke, Inorg. Chem., U_, 2260 (1972). 173) D.J. Merryman, P.A. Edwards, J.D. Corbett and R.E. MacCarley, J. Chem. Soc, Chem. Comm., 779 (1972). 174) J.W. Moore, J.W. Baird, and H.B. Miller, J. Am. Chem. Soc, 90, 1359 (1968). 175) C.S. Alleyne and R.C Thompson, Can. J. Chem., 52, 3218 (1974). 176) W.W. Wilson, B. Landa, and F. Aubke, Inorq. Nucl. Chem. Lett., 11, 529 (1975). 177) CC. Addison, G.S. Brownlee, and N. Loqan, J. Chem. Soc, Dalton, 1440 (1972). 178) P. Gallezot, D. Weigel, and M. Prettre, Acta Crystallogr., 22, 699 (1967). 179) K.O. Chrlste, and C J . Schack, Adv. Inorg. Chem. Radiochem., 18, 319 (1976). 282 180) P.A. Yeats, J.R. Sams and F. Aubke, Inorg. Chem., 11, 2634 (1972). 181) R.J. Gillespie and E.A. Robinson, Can. J. Chem., 39, 2179 (1961). 182) A.J. Edwards and G.R. Jones, J. Chem. Soc. A., 1467 (1969). 183) W.W. Wilson, J.M. Winfield, and F. Aubke, J. Fluorine Chem., 7, 245 (1976). 184) W,J, Moore, Physical Chemistry, 4th Ed., Prentice-Hall Inc., New Jersey, 1972. 185) R.J. Gillespie and R.F.M. White, Can. J. Chem., 38_, 1371 (1960). 186) R, Savoie and P.A. Gignere, Can. J. Chem., 42_, 277 (1964). 187) CD. Desjardins and J. Passmore, J. Fluorine Chem., 6_, 379 (1975). 188) E.E. Aynsley, R.D. Peacock, and P.L. Robinson, Chem. Ind., 1117 (1951). 189) L.H. Vogt, J.L. Katz, and S.E. Wiberley, Inorq. Chem., 4_, 1157 (1965). 190) Ref. 110, p. 998. 191) A. Tressaud, F. Pintchovski, L. Lozano, A. Wold, and P. Hagenmuller, Mat. Res. Bull., 11, 689 (1976). 192) N. Bartlett and D.H. Lohmann, J. Chem. Soc,, 619 (1964). 193) N. Bartlett and D.H. Lohmann, J. Chem. Soc, 5253 (1962). 194) K.O. Christe, Inorg. Chem., 12_, 1580 (1973). 195) D.F. Evans and G.K. Turner, J. Chem. Soc, Dalton, 1238 (1975). 196) W.P G r i f f i t h , The Chemistry of the Rarer Platinum Metals, Interscience-Wiley, New York (1968). 197) V. Norman and J.C Morrow III, J. Chem. Phys., 2U, 455 (1959). 198) Ref. 110, p. 772. 283 Ref. 110, p.1018. I.R. Beattle, K.M.S. Livingston, D.J. Reynolds and G.A. Ozin, J. Chem. Soc. A., 1210 (1970). E.I. S t i e f e l , Prog. Inorg. Chem,, 22, 1 (1977). 0. Jarchow, F. Schroder, and H. Schulz, Z. Anorg. Allg. Chem., 363, 58 (1968). D.M. Adams and R.G. Churchill, J. Chem. Soc. (A.), 2310 (1968). D. L. Kepert, The Early Transition Metals, Academic Press, 1972. C. G. Barraclough, J. Lewis, and R.S. Nyholm, J. Chem. Soc, 3552 (1959). E. O. Fisher, N.Q. Dao, and W.R. Wagner, Angew. Chem. Int., Ed. 17_, 50 (1978). H.A. Carter, CA. Milne, and F. Aubke, J. Inorg. Nucl. Chem., 37, 282 (1975). L.E. Levchuk, J.R. Sams, and F. Aubke, Inorg. Chem., 11, 43 (1972). P.C Leung, Ph. D. Thesis, U.B.C, 1979. K. Nakamoto, I.R. Spectra of Inorg. and Coord. Compounds, J. Wiley and Sons, New York (1963). F. B. Dudley and G.H. Cady, J. Am. Chem. Soc, 85, 3375 (1963). P.C. Leung, K.C. Lee, and F. Aubke, Can. J. Chem., 57_, 326 (1979). R.E. Noftle and G.H. Cady, Inorg. Chem., 4, 1010 (1965). D. D. DesMarteau, J. Am. Chem. Soc, 100, 340 (1968). 284 Appendix A CONDUCTIVITY CALCULATIONS INTRODUCTION In this study, a theoretical treatment of the conductivity results is applied to three systems in HS0 3F — A u ( S 0 3 F ) 3 , PtfSOsF)^ and KSn(S0 3F) 5. The purpose of these calculations is to attempt to describe the solution behavior of these species in a semi-quantitative manner, to have a comparison of their acidities with other systems, and to obtain ionic mobility values . for ions hitherto unknown. As with any treatment of this type, a few assumptions have been made. a) The equilibrium constants are set to be invariant through-out the range of a particular experiment. This is not entirely valid as they do have a dependence on the ionic strength of the solution. b) The autoprotolysis constant for HS0 3F, "1.e., [ H 2 S 0 3F +]-[S0 3F"]is 3 . 8 x l 0 - 8 mol-kg"1, although this has only been approximately determined 1 0 3 . This is not expected to have a large contribution to the conductivity of solutions in HS0 3F except at or near neutrality conditions. c) The molal conductivity of H 2 S 0 3F + at inf i n i t e dilution, x £ ( H 2 S 0 3F +), is set at 320. Also the sum of A*(S0 3F") and A*(K + ) 1s set at 256, using the value of KS0 3F 2 1 . d) A dependence of X* on the Ionic strength, I, is approxi-mated as follows: 285 A * = A * x exp(-1.567 x I). (A.l) The Limiting Law predicts a decrease of A * with respect to I*5, but this is seldom obeyed in solutions with concentrations greater than ~10"3 mol«kg - 1. An empirical f i t was attempted on the published data on solutions of N^SC^F and M I I(S0 3F) 2, where M I =Li, Na, K, NH4, Rb, and M I I =Sr, Ba 1 1 . Although polynomials, usually of less than the 3rd degree, were found to provide the best least-square f i t within the range of the data, as is usually found for curve f i t t i n g of this type, extrapolations outside the limits are extremely unreliable. Furthermore, the coefficients of the polynomials obtained for different solutions vary from solution to solution. The best compromise between the a b i l i t y to extrapolate (required for the concentrated solutions in the KSn-(S0 3F)s system), and the degree of f i t with the experimental data was obtained for an exponential relationship of the form of equa-tion (A.l). The coefficient of -1.567 was obtained by taking an average of a l l the different solutions' values (these range from -1.4 to -1.7). In general, a difference of less than 2% was encountered for solutions of concentrations higher than ~0.01 rnol'kg"1, when equation (A.l) was applied to the solutions in reference 21. While i t may be an over-simplification to expect other ions to behave as the alkali and alkali earth cations, and S0 3F _, equat1on(A.1) does provide a crude means of adjustment, and i t Improves the f i t of the calculation performed here. It must also be stated that the equation has no theoretical impli-286 cations, and is simply the result of a quantitative f i t . To illustrate the degree of variation in the ionic strength during a t i t r a t i o n , i t is included in the appropriate tables. A.l HS03F-Au(S03F)3 H[Au(S03F)iJ is a strong acid in HS03F, as can be illustrated by its titration with KS03F. The autoprotolysis of HS03F can be ignored for the most part of the calculation except at the end of the t i t r a t i o n , at which point the solution is essentially neutral. In order to adjust for the solvent conductance without involving the autoprotolysis equilibrium constant into the calculation, a quantity of l.Oxl-O"1* ST1 cm - 1 is subtracted from the conductance at the end point. For the t i t r a t i o n , the calculation for a strong acid-strong base is applied. The titration can be represented by: H[Au(S0 3FK] + HS03F -f±-+ HoS0 3F + + [Au(S0 3FK]" (A.2) m-x-z x x+z = (x)x(x+z) (m-x-z) (A.3) and K = X*H x(x) + X* A u x(x+z) + X * K x(z) (A.4) where m=total concentration of a l l Au-species x=dissociation of H[Au(S0 3F) u] 287 z=amount of KS03F added, z=rxm r=K/Au xV^*(H 2S0 3F +) x \ r * * ( r A u ( S 0 3 F K r ) A * K = A * ( K + ) a) For each value of x\, X * AU 1S obtained from K at r=l and converted Into X * Au. b) X*'s are found at r=0 by letting I=m. x is obtained from < at r=0, z=0. I is obtained from x. x*'s are recalculated using the new I. This procedure is repeated a total of 3 times to get a good approximation for I. Ki is calculated from equation (A.3). c) For each value of r, m, and z are calculated (m changes because of dilution). d) By solving equation (A.3) for x, the following can be obtained — I, X * , K . e) Likewise, the conductivity of solutions of pure Au(S0 3F) 3 can be found by doing procedure d at r=z=0 for the required concentration, m. 288 A.2 HSO3F- KSn(S0 3F) 5 The titration of KSn(S0 3F) 5 with KS03F indicates the former's weak acicity 1n HS03F, and therefore the autoprotolysis of HS03F has to be taken into account. For reasons discussed in Chapter 8, polymerization in solution is ignored in a f i r s t approximation. The titration can be represented by the following two equilibria and equations: H[Sn(S0 3F) 6]" + HS03F y^—> H 2S0 3F + + [Sn(S0 3F) 6] 2" (A.5) m-x-z x+y x+z 2HS03F T ^ 2 — * H 2S0 3F + + S0 3F" (A.6) x+y y Ki = (*+y)«(*+z) (A.7) (m-x-z) K2 = (x+y)x(y) , and (A.8) K = X*Hx(x+y) + X*Sx(y) + X*HSnx(m-x-z) + X*Snx(x+z) (A.9) where, most of the terms have similar definitions as in Appendix (A.l) and y = dissociation of HS03F X*S = x*(S03F") X*HSn = X*(HfSn(S0 3F) G]-) X*Sn = X*([Sn(S0 3F) 6] 2~) r = KS0 3F/KSn(S0 3F) 5 a) For each set of X*K and x*HSn, Ki is found at r=0 and r=l, using K at r=0 and K at r=l, respectively, and a test value for x*Sn. Using the Secant method, the two Kx values (which should be identical), are 289 made to converge by solving for: Kilr=o * h\r=i - 1 = 0 (A.10) As I is unknown at the begining of this procedure, i t is found by successive approximation by resubstitution simi-lar to procedure b) in Appendix (A,l),X*Sn and Kl are obtained. b) In equation (A.7), x is substituted by y using equation (A.8). A polynomial of 3rd order in y, for which the roots can be found, is obtained. The titration is calcu-lated for increasing values of r, therefore x should be-come progressively smaller. The value of x which is closest to, but less than, the last x is chosen. c) I can then be obtained, and subsequently, K . d) For the conductivity of KSn(S0 3F) 5 solutions at the above set of X* values, for increasing concentrations, using procedures b) and c), the results of x,I and < are calcu-lated by letting z=0. e) Similarly, procedures b) and c) can be used to calculate the conductivity of K 2[Sn(S0 3F) 6] solutions by letting z=m. Because of the choose-x procedure in b), the calculation begins at the highest concentration. 2 9 0 A.3 HS0 3F-Pt (S0 3FK This system is a combination of the two previous ones, and the titration consists of both a strong and a weak acid with a strong base. Two assumptions are made: the conductivity ar r=0 is essentially that of the strong acid, and at r=0, the conduc-t i v i t y is due solely to the dissociations of [Pt ( S 0 3F) 6] z~ and HSO3F. Again, no polymerization is taken into account. The titration can be represented by the following three equilibria and equations. H 2[Pt ( S 0 3F) 6] + HS0 3F 7^-> H[Pt (S0 3F) 6]- + H 2 S 0 3F + (A.11) m-w w-x w+x+y-z H[Pt ( S 0 3F) 6]" + H S O 3 F 7^ 2-> [ P t ( S 0 3 F ) 6 ] 2 _ + H 2 S 0 3F + (A.12) w-x x w+x+y-z H S ° 3 F 7^-> H 2 S 0 3F + + S0 3F" (A. 13) w+x+y-z y K i = (w-x) x (w+x+y-z) ( A 1 4 ) (m-w) K 2 = <x? * ( w + x + y z ) (A.15) (w-x) K 3 = (w+x+y-z) x (y) (A.16) and = X*Hx(w+x+y-z) + A*Sx(y) + x*Kx(z) + X*HPtx(w-x) + X*Ptx(x) where w = dissociation of H 2[Pt ( S 0 3F) 6] x = dissociation of H[Pt (S0 3F) 6] y = dissociation of HS0 3F (A.17) 291 z = amount of KS03F, z= r*m r = K/Pt X*HPt = X*(H[Pt(S0 3F) 6]-) X*Pt = X*([Pt(S0 3F) 6] 2-) a) For each value of r, m, and z are calculated. b) For each set of X*K, X*HPt and X*Pt, Ki 1s obtained from K at r=0 and by letting x=y=z=0, K2 is obtained from K at r=2 and by letting w=m. A check at the end of the calculation shows that both approximations are valid. c) Equations (A.14) and (A.15) are combined into one contain-ing only x and w as the variables. Ki(m-w) x (w-x) - K2x x (w-x) = 0 (A.18) w can be calculated given x. d) By solving equation (A.16), y can be calculated given x and w. e) By solving equation (A.14), Ki can be calculated given x, w, and y. f) For each value of r, x is found by applying the Secant Method to solve for: Kj * Kx - 1 = 0 (A.19) where K* is the value of Ki calculated for each test value of x using procedures c), d), and e). Although Kx can be substituted by K2 in equation (A.19), a higher degree of accuracy is obtained in calculating Kx because KX>K2. When 292 K2 was used, the function described by equation (A.19) also became quite ill-conditioned and was found not to converge in certain cases. g) After finding the value of x (therefore w and y) that satisfies equation (A.18). < can be calculated. ic h) For a l l the calculations involving \ and an unknown quantity of I, the latter is found by the successive approximation method similar to that described in proce-dure b) in Appendix (A.l). i) The conductivity of pure HS0 3F-Pt(S0 3F)i t is found by using equation (A.14), letting x=y=z=0. 293 APPENDIX B GOLD-TRIFLUOROMETHYLSULFATE B.l INTRODUCTION The go!d-trifluoromethylsulfate system was investigated as a possible ansolvo-acid in the H S O 3 C F 3 solvent system. While the acidity of H S O 3 C F 3 is only slightly lower than that of H S O 3 F h l , the former has the advantage of being stable in the presence of water. If Au(S0 3CF 3) 3 is found to be a strong acid in H S O 3 C F 3 and i f i t is also compatible with the aqueous system, then the scope of application of this superacid system could con-ceivably be wider than that of the corresponding fluorosulfate system. Ag(S0 3CF 3) 2 was also studied in this laboratory 2 0 9 , and the results of both investigations have been published in refer-ence 212; the gold-S0 3CF 3 system will be discussed briefly here as i t is related to a degree to the topic of this thesis. For a more detailed discussion, the reader is asked to refer to refer-ence 212. B.2 DISCUSSION The number of known synthetic routes to trifluoromethylsul-fates is very limited, mainly due to the in s t a b i l i t i e s of oxi-dizing agents such as (CF 3S0 3) 2 2 1 3 and CF3S03C1 2 1 4 , the ana-logues of S 20 6F 2 and ClS0 3F in the fluorosulfate system. Most 294 known syntheses involve the solvolysis of salts of other acids in HSO3CF3. Au(S0 3CF 3) 3 and Cs[Au(S0 3CF 3)iJ are prepared by the f o l -lowing reactions: Au(S0 3F) 3 + HSO3CF3 v Au(S0 3CF 3) 3 + HS0 3F (B.l) Cs[Au(S0 3F) l t] + HSO3CF3 y Cs[Au(S0 3CF 3)iJ + HS0 3F , (B.2) the HSO3F formed presumably decomposed by reactions with HSO3CF3. These are most lik e l y mass-action type reactions, since the rever-sal of both reactions was found to occur when HSO3F was added to the trifluoromethylsulfates. Neither compounds were found to be very soluble in HSO3CF3, thus limiting the possiblity of Au(S0 3CF 3)3 being a useful ansol-vo-acid in the solvent system. The formation of Cs [Au(S0 3CF 3)i t], however, does indicate that Au(S0 3CF 3) 3 may act as a tr i f l u o r o -methyl sul fate ion acceptor. The vibrational spectra of these two diamagnetic compounds suggest structural similarities to the fluorosulfates square planar coordination for Au(III), the presence of monodentate SO3CF3 groups in Cs [Au(S0 3CF 3)iJ and the presence of both mono-dentate and bidentate SO3CF3 groups in Au(S03CF 3) 3. B.3 CONCLUSION Au(S0 3CF 3) 3 was found to exhibit a lower sensitivity to moisture than Au(S0 3F) 3, but i t is s t i l l very hygroscopic. 295 Furthermore, the former's low solubility in the parent acid lim-ited its use as an ansolvo-acid in the system. Lewis acidity towards S O 3 C F 3 - , however, was established for the compound by the synthesis of the complex Cs[Au(S0 3CF 3) 4]. The solvolysis of a fluorosulfate in H S O 3 C F 3 should find use as a synthetic route to trifluoromethylsulfates containing transition metals in rela-tively high oxidation states. 296 APPENDIX C LIST OF ABBREVIATIONS Maqnetochemistry and Electronic Spectroscopy B interelectronic repulsion parameter, unit=cm-1 B° B of free ion e B/B° Dq ligand field splitting parameter, unit=cm_1 A 0 octahedral ligand field splitting energy, A o=10 Dq k electronic derealization factor E x dia s u m °^ a ^ diamagnetic susceptibilities in c.g.s. units xm molar magnetic susceptibility x£ xm corrected for Sx^,-, m  dia e Weiss constant TIP Temperature Independent Paramagnetism y g Bohr Magneton effective magnetic moment, unit=yg vQff7 ^eff c o r r e c t e d f ° r TIP X spin orbital coupling constant, unit=cm_1 Vibrational Spectroscopy s strong u $ symmetric stretch m medium u _ asymmetric stretch as w weak p bending mode v very 6 deformation mode b broad Conductometry HQ Hammett acidity function Kap autoprotolysis equilibrium constant mmol 10" 3 mole K specific conductance, un i t^^cnf 1 R base/acid ratio in a titration A + molar conductivity of the cation A~ molar conductivity of the anion 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items