MSU RETURNING MATERIALS: Place in book drop to LIBRARIES remove th1’ s checkout from “ your record. FINES will be charged if book is returned after the date stamped below. THE SYNTHESIS, CHARACTERIZATION, AND REACTIVITY OF MOLYBDENUM AND TUNGSTEN - IRON - SULFUR COMPLEXES By Harry Craig Silvis A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1981 ABSTRACT THE SYNTHESIS, CHARACTERIZATION, AND REACTIVITY OF MOLYBDENUM AND TUNGSTEN — IRON - SULFUR COMPLEXES By Harry Craig Silvis The binuclear complexes [FeMSuX2]2- (M = Mo, w; X = SAr, OAr) have been prepared by ligand exchange reactions with [FeMSuC1212- and either triethylamine/aryl thiol or sodium thiolate or phenolate. The thiolate and phenolate complexes react with benzoyl chloride to form [FeMSuCl2]2-. The X-ray crystal structures of the tetraethylammonium salts of [FeMoSu(SPh)2]2—, [FeWSu(SPh)2]2-, and [FeMoSu— (OPh)2]2-°MeCN have been determined. All structures contain a bimetallic unit bridged by two sulfides, with terminal sulfides on molybdenum or tungsten and thiolate or pheno- late ligands on iron. The Optical and infrared spectra of these complexes have been obtained. Electrochemical measure- ments demonstrate a single irreversible reduction between -l.O and -2.0 volts. Mossbauer spectra have been recorded at “.2 K in zero applied magnetic field and show simple quadrupole doublets. Magnetic susceptibility measurements Harry Craig Silvis at room temperature on all binuclear complexes yield COI‘I’ “eff Temperature-dependent magnetic susceptibility results and = A.9 - 5.1 BM, consistent with an S = 2 ground state. magnetic Mossbauer data on [FeMoSu(SPh)2]2- suggest an Fe(II) — Mo(VI) description of these complexes. Proton NMR spectra of these complexes show isotropically shifted resonances for the aromatic ring protons which increase with decreasing temperature. The pattern of these shifts indicate that they arise from dominant contact interactions. Evidence suggests that [FeMoSu(SPh)2]2' and [FeMoSuc12j2' are converted to [Fe(MoSu)2]3' under reducing conditions. The reactivity of the FeSZMo core towards a variety of other reagents has been examined. The trinuclear complex [Fe2MoSllJ3- has been synthesized. Its Optical, infrared, and Mossbauer spectra suggest that it possesses a core structure analogous to that of [Fe2MoS6- (S-p-Tol)2]3-. Reactions of [FeuSuXMJZ- (X = SPh, S-E-Bu, and Cl) with M083- have been examined. When X = S-E-Bu or C1, the primary product is [Fe(MoSu)2]3-. When X = SPh, the product is [Fe2MoS6(SPh)2]3-. The reactions of the complexes [Mosgj2-, [Mo2(SPh)9]3-, [MoO(SPh)u]—, and [MoCl6J3- with a variety of iron species have been investigated in an attempt to synthesize new cluster complexes. To Cindy, Harry, Louise, and Grace. Thank you. ii ACKNOWLEDGMENTS I would like to gratefully acknowledge Professor Bruce A. Averill for his patient guidance throughout the course of the research described herein. His insights concerning synthetic inorganic chemistry and the interpretation of physical data have been invaluable. [Furthermore, I would like to thank the following in— dividuals for their contributions to this work: Drs. E. Mfinck, T. A. Kent, and B. H. Huynh for obtaining the Moss- bauer spectra; Drs. B.-K. Teo and D. Ward, and M. R. Antonio for crystal structures of various compounds; W. E. Cleland for obtaining the NMR spectra; Dr. J. V. Waszczak for providing temperature-dependent magnetic susceptibility data on (EtuN)2[FeMoSu(SPh)2]; K. Guyer for helpful discussions concerning electrochemistry. I would like to thank my coworkers, Walter Cleland, Robert Tieckelmann, Jim Davis, and Mark Antonio for their friendship and advice on various matters throughout my graduate career. Also, thanks to Scott Sandholm and Ken Guyer for companionship on various trout streams and bass lakes. I would like to acknowledge the Dow Chemical Company for a one-year fellowship and the General Electric Company for a summer fellowship. Also, I would like to thank the iii Department of Chemistry for providing a teaching assist- antship and specifically Professors Brubaker, Dye, and Chang for serving on my Committee. Finally, special thanks are extended to Mrs. P. Warstler for her expert typing of this thesis and to B. Adams for assistance with the graphics. iv TABLE OF CONTENTS Chapter Page LIST OF TABLES. . . . . . . . . . . . . . . . . . . x LIST OF FIGURES . . . . . . . . . . . . . . . . . . xi LIST OF ABBREVIATIONS . . . . . . . . . . . . . . . xv I. INTRODUCTION. . . . . . . . . . . . . . . . . . 1 II. EXPERIMENTAL . . . . . . . . . . . . . . . . . 5 A. Materials and Methods . . . . . . . . . . . 5 B. Preparation of (EtuN)2- [FeMSu(SAr)2] 8 1. [FeMoSu(SPh)2]2_ Method A 8 2. Method B. 8 3. [FeMoSu(S-prTol)2]2- Method B 9 A. Method C. 9 5. [FeWSu(SPh)2]2- Method B . . . . . . . 10 6. [Fewsu(s-9¢To1)2]2‘ Method B. . . . . . 10 C. Preparation of (EtuN)2[FeMSu(OAr)2] (M = M0, W) . . . . . . . . . . . . . . . . 10 1. [FeMoSu(oph)2]2-. . . . . . . . . . . . 10 2- [FGMOSM(O—peTol)2]2-. . . . . . . . . . ll 3. [FeWSu(OPh)2]2- . . . . . . . . . . . . 11 u. [FeWSu(O-p-Tol)2]2- . . . . . . . . . . 12 D. Preparation of (EtuN)3— [Fe2MoSll]. . . . . . . . . . . . . . . . . 12 Chapter Page 1. (EtuN)2[Fe2Sl2] . . . . . . . . . . . . 12 2. (EtuN)3[Fe2MOSll] . . . . . . . . . . . l3, Interconversion Chemistry . . . . . . . . . 1A 1. [FeMoSu(sPh)2]2- + [FeMoSuCl2JZ— . . . . . . . . . . . . . 1A 2. [FeMoSu(SPh)2]2- + [FeM08932—- . . . . . . . . . . . . . . 1A 3. [FeMoSu(OPh)2]2’ i [FeMoSuCIBJ2- . . . . . . . . . . . . . 1A A. [FeMoSu(OPh)2]2' + [FeMoSu(SPh)2]2-. . . . . . . . . . . . 15 5. Tungsten Analog Interconversions. . . . 15 Reaction of [FeMoSu(SPh)2]2- with Acenaphthalene Radical Anion . . . . . . . . . . . . . . . . . . . 15 Reaction of [FeMoSu(SPh)2]2- with Ferric Chloride and Thiophenolate . . . . . . . . . . . . . . . 16 Reactions of (EtuN)2[FeMoSu012] . . . . . . 17 1. Methyl Iodide . . . . . . . . . . . . . 17 2. Benzoyl Chloride. . . . . . . . . . . . 18 3. Methyl Magnesium Chloride . . . . . . . 18 A. Sodium Diethyldithio- carbamate . . . . . . . . . . . . . . . l9 5. Sodium benzylthiolate . . . . . . . . . 19 Reactions of [Feusuxuj2' (x = s-teBu, SPh, Cl) . . . . . . . . . . . 20 vi Chapter Page I. X = S—t-Butyl . . . . . . . . . . . . . 20 2. X = S-t-Butyl with Trimethyl- pyridine hydrochloride. . . . . . . . . 2O 3. X SPhenyl . . . . . . . . . . . . . . 21 A. X Chloride. . . . . . . . . . . . . . 21 . 2- J. Reactions of [M089] . . . . . . . . . . . 21 l. Triethylammonium Thiophenolate. . . . . 21 2. Ferrous Chloride. . . . . . . . . . . . 22 3. Sodium Ethylthiolate and Ferrous Chloride. . . . . 22 A. Ferric Chloride and Sodium Thiophenolate . . . . . . . . . . . . . 23 5. (EtuN)2[FeuSu(SPh)u]. . . . . . . . . . 2a 6. Na2[Fe(CO)u] ° 1-1.5 Dioxane. . . . . . 2A K. Reactions of [Mo01613‘. . . . . . . . . . . 25 1. (EtuN)2[Fe282(SPh)u]. . . . . . . . . . 25 2. Ferric Chloride and Thio- phenolate . . . . . . . . . . . . . . . 25 3. Thiophenolate . . . . . . . . . . . . . 26 L. Reactions of [M02(SPh)9]3_. . . . . . . . . 26 1. (EtuN)2[Fe282(SPh)u]. . . . . . . . . . 26 2. (EtuN)2[Fe2812] . . . . . . . . . . . . 27 3. (EtAN)2[FeASA01UJ . . . . . . . . . . . 27 M. Reaction of Ferrous Chloride with [MoO(SPh)u]-. . . . . . . . . . . . . . . . 28 III. RESULTS AND DISCUSSION. . . . . . . . . . . . 29 A. [FeMSu(SR)2]2- ( M = M0, w; R = Ph, peTol). . . . . . . . . . . . . . . 29 vii Chapter v Page 1. Synthesis and Intercon- versions. . . . . . . . . . . . . . . . 29 2. Structure . . . . . . . . . . . . . . . 33 3. Optical and Infrared Spectra. . . . . . A1 A. Electrochemistry. . . . . . . . . . . . 51 5. Magnetic Susceptibility . . . . . . . . 56 6. Massbauer Spectra . . . . . . . . . . . 50 7. Proton Magnetic Resonance Spectra . . . . . . . . . . . . . . . . 69 8. Summary . . . . . . . . . . . . . . . . 76 B. [FeMSu(OR)2]2- (M = Mo, w; R = Ph, peTol). . . . . . . . . . . . . . . 81 1. Synthesis and Intercon— versions. . . . . . . . . . . . . . . . 81 2. Structure . . . . . . . . . . . . . . . 85 3. Optical and Infrared Spectra. . . . . . 89 A. Electrochemistry. . . . . . . . . . . . 9“ 5. Magnetic Susceptibility . . . . . . . . 96 6. MESsbauer Spectra . . . . . . . . . . . 96 7. Nuclear Magnetic Resonance Spectra . . . . . . . . . . . . . . . . 100 8. Summary . . . . . . . . . . . . . . . . 110 C. Comparisons between Phenolate and Thiolate Ligation . . . . . . . . . . . 111 3- D. [Fe2M0811] . . . . . . . . . . . . . . . . 113 1. Synthesis . . . . . . . . . . . . . . . 113 2. Optical and Infrared Spectra. . . . . . 11“ viii Chapter Page 3. Electrochemistry. . . . . . . . . . . . 118 A. Magnetic Susceptibility . . . . . . . . 119 5. Mossbauer Spectrum. . . . . . . . . . . 121 6. Electron Paramagnetic Resonance Spectra . . . . . . . . . . . . . . . . 12A 7. Conclusions . . . . . . . . . . . . . . 125 E. Reactions of [FeMoSu(SPh)2]2- . . . . . . . 126 1. Acenaphthalene Radical Anion . . . . . . . . . . . . . . . . . 126 2. Ferric Chloride and Thio- phenolate . . . . . . . . . . . . . . . 127 F. Reactions of [FeMoSuC12]2_. . . . . . . . . 128 1. Methyl Iodide . . . . . . . . . . . . . 128 2. Benzoyl Chloride. . . . . . . . . . . . 129 3. Methylmagnesium Chloride. . . . . . . . 129 A. Sodium Benzylthiolate . . . . . . . . . 130 5. Sodium Diethyldi- thiocarbamate . . . . . . . . . . . . . 131 . 2- G. Reactions of [FeASAXA] (x = S-t-Bu, SPh, 01) . . . . . . . . . . . 131 1. X = thutyl Thiolate. . . . . . . . . . 131 2. X = Thiophenolate . . . . . . . . . . . 132 3. X = Chloride. . . . . . . . . . . . . . 13A H. Reactions of [Mosg]2_ . . . . . . . . . . . 135 I. Reactions of Molybdenum(III) and Molybdenum(V) Compounds . . . . . . . . . . 135 IV. CONCLUSIONS. . . . . . . . . . . . . . . . . . 139 REFERENCES. . . . . . . . . . . . . . . . . . . . . 1A2 Table II III IV VI LIST OF TABLES Selected Interatomic Distances (A) with Esd's for (EtuN)2[FeMSu- (SPh)2] (M = Mo, W). Selected Bond Angles (deg) with Esd's for (EtuN)2[FeMSu(SPh)2] (M = Mo, W). . . . . Electronic and Infrared Spectral Features, and Isotropic Shifts of Ligand Protons of (EtuN)2[Femqu2] (M = Mo, W; X = SAr, OAr) and (EtuN)3[Fe2MoSll]- Electrochemical Data, Magnetic Moments, and Méssbauer Parameters for (EtuN)2[FeMSuX2] (M = Mo, w; x = SAr, OAr) and (EtuN)3[Fe2MOsllj. Selected Interatomic Distances (A) and Angles (deg) for (EtuN)2[FeMoSu- (OPh)2]'MeCN . . . . . . Comparison of Relative Isotropic Shifts for Various Metal-Sulfur Centers in CD3CN Solution. . . . . . X Page 36 37 AA 52 88 . 108 Figure LIST OF FIGURES Syntheses and reactivity of [FeMSu(SPh)2]2- (M = M0, W). . X—ray Crystal structure of (EtuN)2- [FeMoSu(SPh)2]. Ellipsoids repre- sent fifty percent probability X-ray crystal structure of (EtuN)2- [FeWSu(SPh)2]. Ellipsoids represent fifty percent probability. Electronic spectra of (EtuN)2- [FeMSu(SPh)2] (M = Mo, W) in aceto- nitrile solution at 23°C Near—infrared spectra of (EtuN)2- [FeMoSu(SPh)2] (A), (EtuN)2[FeMoSu- (OPh)2] (B), and (EtuN)2[FeMoSuCl2] (C) in dimethylsulfoxide solution at 23°C. Magnetic susceptibility of (EtuN)2- [FeMoSu(SPh)2] as a function of tem- perature in the solid state (C) and in N,N-dimethylformamide glass (0 ), obtained in 0.16-T applied xi Page 32 35 A0 A3 A9 Figure 10 Page field. The left vertical axis refers to the solid, and the right to the frozen solution data. . . . . . . . 59 M6ssbauer spectrum of (EtuN)2— [FeMoSu(SPh)2] in frozen N,N-dimethyl- formamide solution recorded at A.2 K with zero applied field. Vertical bar indicates one percent absorption . . . 63 Mossbauer spectrum of (EtuN)2_ [FeWSu(SPh)2] in frozen N,N-dimethyl- acetamide solution recorded at A.2 K with zero applied field. Vertical bar indicates one percent absorption . . . . 65 Correlation diagram relating isomer shift to formal oxidation state for various iron-sulfur centers153 . . . . . 67 NMR spectra of (EtuN)2[FeMoSu- (SPh)2] in CD SOCD at 69 (A) 3 3 and 17°C (B), along with NMR spectra of (EtuN)2[FeMoSu(S—p—Tol)2] in CD CN at -A5 (E), 25 (D), and 3 75°C (C). Small amounts of impurity are denoted by X. Chemical shifts are in parts per million from MGMSI. . . . . o . . . . . . . . . o . . o . 71 xii Figure 11 12 13 1A 15 16 Temperature dependence of me, 9—H, and p—CH resonances of (EtuN)2- 3 [FeMoSu(SPh)2] in CD SOCD3 (O ) 3 and CDBCN (O ), and of (EtAN)2‘ [FeD’IoSu(S-_p_-Tol)2] in CD CN (0) 3 solution . . . . . . . . . . . . . Temperature dependence of m—H, 9-H, and prH3 resonances of (EtuN)2[FeWSu(SPh)2] in CD3CN (o) and of (EtuN)2[FeWSu(S-p-Tol)2] in CD3CN (D ). Synthesis and reactivity of [FeMSu- (OPh)2]2‘ (M = Mo, W). . . . . X-ray crystal structure of (EtuN)2— [FeMoSu(OPh)2]-MeCN. Ellipsoids represent fifty percent probability. Electronic spectra of (EtuN)2- [FeMSu(OPh)2] (M = Mo, W) in aceto- nitrile solution at 23°C Massbauer spectrum of (EtuN)2- [FeMoSu(OPh)2]‘MeCN in the solid state diluted with boron nitride at A.2 K with zero applied field. Vertical bar indicates one per- cent absorption. . . . . . . . xiii Page 78 80 8A 87 91 98 Figure 17 18 19 20 21 Page NMR spectra of (EtuN)2[FeMoSu— (OPh)2] in CD3CN at 50°C (A) and of (EtuN)2[FeMoSu(0-prol)2] in CD3CN at 50°C (B). Small amounts of impurity are denoted by X. Chemical shifts are in parts per million from MeuSi . . . . . . . . . . . . 102 Temperature dependence of m-H, E—H, and p—CH resonances of (EtuN)2- 3 [FeMoSu(OPh)2] in CD CN (0) and of 3 (EtuN)2[FeMoSu(0-p-Tol)2] in CD3CN (D’) . . . . . . . . . . . . . . . . . . . 105 Temperature dependence of m—H, p-H, and p-CH3 resonances of (EtuN)2[FeWSu(OPh)2] in CD3CN (o) and of (EtuN)2[FeWSu(O-p-Tol)2] in CD CN (0 ).... . . . . . . . . . . . . 107 3 Electronic spectrum of (EtuN)3— [Fe2M0811] in acetonitrile solu- tion at 23°C . . . . . . . . . . . . . . . 116 Mdssbauer spectrum of (EtuN)3- [Fe2M0811] in the solid state diluted with boron nitride at A.2 K in a 600 G field. Vertical bar in- dicates one percent absorption . . . . . . 123 xiv 1' .-~ nAv—o .V..- a... ’9'“ 'l-A ‘tv , - 1.... ~ h. 1‘ ‘ ,_ . " H ’|O ‘— .5 .‘ - v \v . .Iu ".‘ 1 v Solvents: MeCN MeOH EtOH Et2O THF NMF DMF DMA DMSO HMPA Reagents: PhCOCl MeI MeMgCl Et3N Me3py-HC1 ACN LIST OF ABBREVIATIONS acetonitrile methanol ethanol diethyl ether tetrahydrofuran N-methylformamide N,N—dimethylformamide N,N—dimethylacetamide dimethylsulfoxide hexamethylphosphoramide benzoyl chloride methyl iodide methylmagnesium chloride triethylamine trimethylpyridine hydrochloride acenaphthalene XV Miscellaneous: EtuN MeuN PhuAs Ph p—tol Bz Me Et BM SCE tetraethylammonium tetramethylammonium tetraphenylarsonium phenyl paraetolyl benzyl methyl ethyl Bohr magneton standard calomel electrode xvi e. v. r“ l. L. ... a; .; v. z . ...n a. H. .. a .3 S. I. INTRODUCTION In recent years, there has been an increased interest in the structure and function of inorganic metal complexes within biological systems. Actual modeling of these metal centers outside their native protein environments has met with mixed success. The most notable example of this to date has been the synthesis of one, two, and four iron- sulfur complexes, which mimic those found in the ferredoxin and rubredoxin electron-transfer proteins.1 Lately, much attention has been focused on the enzyme, nitrogenase, which is responsible for the catalytic reduc- tion of dinitrogen to ammonia in aerobic and anaerobic micro-organisms. The enzyme itself consists of two com- ponents. The first is an iron protein (MW m60,000), con— taining four iron atoms and four acid-labile sulfur atoms (343;, sulfide) per molecule. Spectroscopic and cluster displacement evidence indicate that the iron and sulfur are arranged in a normal FeASA cluster complex.2 The second component is a molybdenum—iron protein (MW m220,000), which contains ca. 32 iron atoms, 32 sulfides, and 2 2 Roughly half of the iron molybden atoms per molecule. and sulfide a e arranged in Fens“ centers, where one of the iron atoms in each cluster is distinguishable from a. the other three by Mossbauer spectroscopy.3 The Massbauer parameters of the unique iron atom resemble those of high- spin iron (II) in a tetrahedral sulfur environment.Ll The remaining iron, sulfur, and molybdenum are contained within a low molecular weight cofactor,5 which can be separated from the rest of the iron-molybdenum protein. This co- factor contains six to eight iron atoms and ca. six sul- fides per molybdenum atom in a unique cluster arrangement. Relevant spectroscopic methods employed to characterize the cofactor include electron paramagnetic resonance (EPR), Massbauer, and extended X-ray absorption fine structure (EXAFS) spectroscopy. The low—temperature (6-20 K) EPR 6 with spectrum of the cofactor displays an axial signal slight rhombic distortion (g-values 2.0, 3.8, and A.3). This spectrum corresponds to one of the Kramers doublets of an S 3/2 system.6’7 Cofactor from organisms grown on 95 Mo (I 5/2) shows no detectable hyperfine broadening in the EPR spectrum, in contrast to results from organisms grown on 57Fe (I = 1/2).8 Molybdenum K-edge EXAFS results indicate two9 or three10 iron atoms at ca. 2.71 A and four or five sulfur atoms at ca. 2.35 A as nearest neighbors to the molybdenum. The actual reaction sequence which results in di- nitrogen reduction involves reductive dephosphorylation. For every two electrons transferred to substrate, at least four ATP molecules are hydrolyzed to ADP.ll Studies :r‘ ..J n 'h indicate that electrons are transferred from external reductant (reduced ferredoxin) to the iron protein, then to the molybdenum—iron protein, and ultimately to sub— 2 strate. Furthermore, recent results suggest that oxygen ligands may be coordinated to the iron atoms within the cofactor.12-lu Due to the extreme oxygen sensitivity and other dif- ficulties in purification of the cofactor, resolution of the structure by crystallographic techniques in the near future seems unlikely. Therefore, in order to ascertain the structure and functioning of this unprecedented ar- rangement of iron, molybdenum, and sulfur, it seemed worth- while to synthesize polynuclear metal complexes containing these elements, characterize them, and investigate their reactivity. Intensive synthetic efforts by several groups have resulted in two basic categories of Fe-Mo—S compounds. The first includes the so-called "cubane" complexes such as [Fe6Mo2SB(SR)9]3-, [Fe6Mo289(SR)8]3—, [Fe7M0288(SR)12]3_’u-, 15 and and [FeuMoSu(SR)3(C6HuO2)3]3-, synthesized by Holm Garner.l6 They result from "self-assembly" reactions in- volving simple reagents analogous to the strategy used to prepare the ferredoxin analogs. The common feature of all these complexes is the Fe3MoSu core unit. However, the stoichiometry and EXAFS results for these complexes indi— cate that they are not true synthetic models for the .— \ ’ - ‘a ‘- .1 'c- '..... v . . I '— . . - r0. 1 .s ...a~ .. y- .. v... ‘ r -A._ er. -.._ --. cofactor.15a The second class of Fe-Mo-S complexes are those pos- sessing the FeS2Mo core unit. They include heteronuclear dimers of the type [FeMoSuX2J2- (X = SAr, OAr, C1 or X2 = 20,21 85)]‘7—19 and various linear trimers such as [FeZMoS6X213- (X = SAr; X2 = SS)’ [Fe(MoSu)213-, and [Fe2MoSuC12]2_. A linear tetramer of formula [Fe2M02Sloju has also been prepared.22 The research described herein deals primarily with the synthesis, characterization, and selected reactivity of the heteronuclear dimer complexes, [FeMSuX2J2- (X = SAr, OAr) and the trimeric complex, [Fe2M081133-. In addition, tungsten analogs of the dimeric complexes have been pre- pared, and their properties compared to their molybdenum counter parts. II. EXPERIMENTAL A. Materials and Methods All Operations were performed in an atmosphere of di- nitrogen prepurified by sequential passage over hot BASF R3-ll catalyst to remove oxygen and through supported phosphorus pentoxide (Aquasorb) to remove water. Solutions and solvents were degassed prior to use by repeated evacua- tion and flushing with purified dinitrogen. MeCN, MeOH, and THF were distilled from calcium hydride, magnesium methoxide, and sodium/benzophenone or lithium aluminum hydride, respectively. NMF and HMPA were vacuum distilled from barium oxide. PhCOCl and Mel were distilled from barium oxide and A A molecular sieves, respectively, at room temperature by using a liquid nitrogen trap. Spectro- grade DMA (Aldrich Chemical Co., Gold Label), along with Et3N’ were stored over A A molecular sieves. PhOH and p- TOlOH were vacuum sublimed and distilled, respectively, by Walter Cleland. ACN was recrystallized from MeOH. Thiols and all other reagents were of commercial reagent grade and used without further purification. Ammonium23 and tetraalkylammonium15a salts of tetrathiomolybdate and tetrathiotungstate, (EtuN)2[FeMoSuC12],l7 (EtAN)2[FeASu' A (Sthl.2 (MeuN)2[FeuSu(S—E§Bu)u],2” (EtuN)3[Mo2(SPh)9],25 (EtuN)[MOO(SPh)u],25 and (EtuN)2[M089]26 were prepared by published procedures. Ferrous chloride,27 (EtAN)2' [Fe2SZ(SPh)u2]28 and (EtuN)2[FeASAC1A]29 were prepared by Robert Tieckelmann as described. (EtuN)3[M0016] was prepared by electrolyzing a solution of molybdenum trioxide dissolved in 6N hydrochloric acid for four hours at -2.2 volts. Three equivalents of EtuNCl were added and the solution was saturated with hydrogen chloride gas. Addition of EtOH and subsequent cooling to —20°C yielded the product, (EtuN)3[MoC16], as red plate-like crystals in 16% yield. Tetraethylammonium thiophenoxide was prepared by adding PhSH to a solution containing one equivalent of NaOMe in MeOH, and then adding this solution to one con- taining one equivalent of EtuNC1 in isopropanol. Sodium chloride precipitated immediately and was filtered off. Volume reduction of the filtrate caused separation of the product as white crystals in 80% yield. These were fil- tered, washed with hexane, and dried in vacuo. An NMR spectrum indicated the absence of water and other solvents. Sodium salts of thiophenols and phenols were prepared by reaction with one equivalent Of NaOMe in MeOH, evapora- tion to dryness, evaporation twice with MeCN and drying in vacuo. Stock solutions of sodium or lithium acenaphthal- enide were generated by reaction of the metal with ACN in DMA. The concentration of ACN radical anion was determined by titration with a standard solution Of iodine in DMA. Methylmagnesium chloride was purchased from Aldrich Chemi— cal CO. as a THF solution and was assayed according to the method of Gilman.3O Optical spectra were recorded by means of a Cary Model 17 or Cary 219 spectrophotometer. Infrared spectra were Ob- tained with a Perkin-Elmer Model A57 grating spectro- photometer by using Nujol mulls between polyethylene plates. Proton NMR spectra were taken by means of a Bruker Model WH-180 or WM-250 Fourier transform spectrometer. Room- temperature magnetic susceptibility measurements were per- formed with an Alpha Faraday balance by using Hg[Co(SCN)u] as a calibrant. Variable temperature data were obtained at Bell Laboratories by J. V. Waszczak using a Faraday-type apparatus already described.31 Electrochemical measurements were made with a PAR Model l7AA polarographic analyzer em- ploying either do polarography (dropping mercury electrode) or cyclic voltammetry (platinum electrode). EPR spectra were recorded by means Of a Varian E-A spectrometer Operat- ing at 9.1 GHz and equipped with an Oxford cryostat for Operation from A to 300° K. The Mossbauer spectra were Obtained by T. A. Kent and B. H. Huynh at the Gray Freshwater Biology Laboratory, University of Minnesota, Navarre, Minnesota. The MOssbauer spectrometer was Of the constant acceleration type and has been described 32 elsewhere. The velocity scale of the instrument was calibrated with iron foil at room temperature. Melting points were Obtained in sealed capillaries in vacuo and are uncorrected. Elemental analyses were performed by Spang Microanalytical Laboratory, Eagle Harbor, Michigan, and by Galbraith Laboratories, Inc., Knoxville, Tennessee. B. Preparation of (EtuN12fFeMSu(SR)2] (M = Mo, w; R = Ph, peTol) 1. [FeMOSu(SPh)2]2-. Method A. To 2.A5 g of the compound formulated as (EtuN)3_ [Fe(MoSu)2]20a was added a solution Of 37 mL of PhSH and 50 mL Of Et N (m300 eq. of (Et NH)(SPh)) in A0 mL of MeCN. 3 3 The solution changed from red—violet to red-orange upon stirring for 1-2 h. The product precipitated as brick red microcrystals upon the addition of 100 mL Of THF. It was filtered, washed with THF, and then redissolved in an MeCN/THF mixture (m3zl) at room temperature, filtered, and cooled to -20°C. The product crystallized as 1-2 mm red— black prisms in 3 50% yield (mp 126-128°C dec). Anal. Calcd for C28H S6FeMoz C, AA.32; H, 6.6A; N, 3.69; S, 25.35; 5oN2 Fe, 7.36; MO, 12.6A. Found: C, AA.20; H, 6.A2; N, 3.86; S, 25.50; Fe, 7.55; Mo, 13.02. 2. Method B. To 0.20 g (0.33 mmol) Of (EtuN)2[FeMoSu012] was added a solution containing 0.13 mL of PhSH and 0.18 mL of Et3N (1.3 mmol Of (Et3NH)(SPh)) in 15 mL Of MeCN. The re- action was instantaneous as indicated by an immediate color change from yellow-brown to red-orange. The crude product was precipitated as microcrystals by the slow addition Of A0 mL of THF, filtered, washed with THF, and recrystallized as above (yield mA5%). 3. [FeMO§A(S—p-Tol)912-. Method A. This compound was prepared by Method A above and ob- tained as red—black prisms (mp 131—13A°C dec) after re- crystallization from MeCN/THF. Anal. Calcd for C3OH5AN2_ S6FeMo: C, A5.78; H, 6.92; N, 3.56. Found: C, A5.77; H, 6.80; N, 3.58. Method B also afforded this compound if a large excess (N100 eq.) of PhSH/Et3N was used. A. Method C To 0.86 g (1.Al mmol) Of (EtuN)2[FeMoSu012] was added 0.AA g (3.01 mmol) of NaS-prOl. The solids were slurried in A0 mL of MeCN and allowed to stir overnight, whereupon the color Of the solution changed from yellow-brown to red— orange with precipitation of NaCl. The solution was re- duced in volume by m25% in vacuo and then filtered. Upon the slow addition of THF, red microcrystals precipitated. These were filtered, washed with THF, and recrystallized as described above (yield m80%). 10 5. [FeWSu(SPh)2]2-. Method B. This compound was prepared by Method B above and afforded dark red prisms (mp 129—131°C dec) after re- crystallization from MeCN/THF. Anal. Calcd for 028H50N286FeW: C, 39.71; H, 5.95; N, 3.31; S, 22.72; Fe. 6.60; W, 21.71. Found: C, 39.59; H, 6.10; N, 3.3A; S, 22.88; Fe, 6.66; W, 21.A7. 6. [FeWSu(S-p-Tol)p]2-. Method B. This compound was likewise prepared by Method B and obtained as dark red prisms (mp 169-172°C dec) after re- crystallization from MeCN/THF. Anal. Calcd for C30H5u- N286FeW: C, Al.18; H, 6.22; N, 3.20. Found: C, Al.5A; H, 6.22; N, 3.30. C. Preparation of (Et N) [FeMS,(OR) ] (M = Mo, W; R = Ph, A——2 4 2 prol). 1. [BeMosu(OPh)212‘ A mixture of 5.7A g (9.39 mmol) Of (EtuN)2[FeMoSuC12] and 2.A5 g (21.1 mmol) Of anhydrous NaOPh were taken up in 250 mL of MeCN. Upon being stirred overnight, the color of the solution changed from yellow-brown to red-orange ac- companied by precipitation of finely divided NaCl. The volume was reduced to N150 mL in vacuo and the mixture 11 was filtered. With the addition of 200 mL Of anhydrous EtZO, the product precipitated as small, brick red plates, which were filtered, washed with Et2o, and vacuum dried. The product was dissolved in a minimum of MeCN/Et2O (m2zl), filtered, and cooled to -20°C, whereupon the product crystal- lized as 1-2 mm dark red plates in 65% yield (mp 87-90°C dec). Anal. Calcd for 028H50N2O2SuFeMo: C, A6.27; H, 6.9A; N, 3.86; S, 17.65. Found: C, A6.8A; H, 6.70; N, 3.56; S, 16.98. When saturated solutions of this compound were allowed to cool very slowly, the material crystallized as an MeCN solvate (1:1). These crystals were 3-A mm blades which were more suitable for X-ray analysis than the un— solvated material. 2. [FeMOSu(O—E§Tol)7]2-. This compound was prepared in the same manner as the -OPh analog described above. After recrystallization from MeCN/EtZO, the product was obtained as ca. 1 mm dark red plates in 60% yield (mp 112-115°C dec). Anal. Calcd for C3OH5MN2O2SuFeMoz C, A7.73; H, 7.21; N, 3.71. Found: c, A8.Ao; H, 7.u8; N, 3.50. 3. [FeWSu(OPh)2l?- This compound was likewise prepared as described above, using (EtuN)2[FeWSuC12]33 and NaOPh. Upon recrystallization 12 from MeCN/THF (“2:1), the product was obtained as well- formed 3-A mm golden blades (mp 126—130°C dec). Elemental analysis results indicated the product to be a 1:1 MeCN solvate. This was confirmed by NMR methods as well. Anal. Calcd for C3OH53N3O2SuFeW: C, A2.11; H, 6.2A; N, A.91; S, 1A.99. Found: C, A2.A9; H, 6.36; N, A.77; S, 1A.63. u. [FeWSu(O-peTol)2]2- This compound was synthesized by the same method used to prepare the -0Ph analog. Recrystallization from MeCN/ THF afforded the product as 1-2 mm gold-brown plates in A5% yield (mp 138-1A0°C dec). Anal. Calcd for C3OH54N2- o2SuFeW: C, A2.76; H, 6.A6; N, 3.32. Found: C, A3.A3; H, 6.66; N, 3.29. D. Preparation of (EtuN13lFe2MoSl}l 1° SEEuEl2L§92§12l A solution of 3.1 g (18.5 mmol) of anhydrous FeCl3 in 15 mL of MeCN was quickly filtered and added to one containing 17.7 g (7A.0 mmol) Of (EtuN)(SPh) in 150 mL of MeCN. The resultant purple-brown mixture was added to A.8 g (1A8.0 mmol) of elemental sulfur. Upon being stirred overnight, the mixture turned deep red-brown and much of the product precipitated as black microcrystals. Upon addition of 75 mL of THF, the remainder Of the product separated. After filtration, the solid was dissolved in a minimum Of warm MeCN/DMSO (m9zl, v/v), filtered, and allowed to cool to -20°C. The product crystallized as small black prisms; the optical spectrum in MeCN cor— responded exactly to that described by Coucouvanis and co- 3A workers, who report an alternate synthesis. 2. LEEQN13LE§2MQ§111 A solution of 1.7 g (2.3 mmol) Of (EtuN)2[Fe2812] and 1.1 g (2.3 mmol) Of (EtA)2MOSA in 150 mL Of MeCN was allowed to stir until the optical spectrum of the reaction solution became constant (NA days). The product was pre- cipitated in microcrystalline form by the slow addition Of Et20, filtered, washed with Et 0, and dried in vacuo. 2 It was then dissolved in a minimum of MeCN/Et2O (m2zl) at room temperature, filtered, and allowed to cool slowly to -20°C, whereupon the product precipitated in microcrystal- line form. It was filtered, washed with Etzo, and vacuum dried. The yield was ca. 50% (mp 205-207°C dec). Anal. Calcd for CZAHOONBSllFeZMO: C, 30.31; H, 6.36; N, A.A2; S, 37.09. Found: C, 30.07; H, 6.06; N, A.22; S, 38.02. Efforts to prepare the tungsten compound by an analogous route failed. 1A E. Interconversion Chemistry 1. [FeMoSu(SPh)2]2- +LFeMOSu01213: TO 8.0 mL of a 11.0 mM solution Of (EtuN)2[FeMOSu- (sph)2] in MeCN was added 210 uL of a 0.86 M solution Of PhCOCl in MeCN. The color Of the reaction solution changed from red-orange to yellow-brown within five minutes. An optical spectrum Of the solution confirmed quantitative formation of [FeMoSuC12J2-. 2. [FeMOSu(SPh)212- + [FeM08913_ Reaction of 0.30 g (0.28 mmol) Of (PhuAs)2[FeMOSu- (SPh)2] with 0.0A g (1.25 mmol) of elemental sulfur in A0 mL of MeCN for two days resulted in a gradual color change from red-orange to red-brown. The Optical spectrum of the reaction solution corresponded exactly to that of [FeM089JZ- prepared by a different procedure.19a 3. [FGMOSH(OPh)212- + IFeMoSUQIZIII To 5.0 mL of an 8.6 mM solution of (EtuN)2[FeMoSu- (OPh)2] in MeCN was added 100 ”L of an 0.86 M solution Of PhCOCl in MeCN. The color of the solution quickly changed from red-orange to yellow-brown and the Optical spectrum indicated quantitative formation of [FeMoSuC12]2_. 15 A. LFeMOSu(OPh)212- + [FeMOSu(SPh)2]2— To ca. 3 mL Of an 8.6 mM solution of (EtuN)2[FeMOSu- (OPh)2] in MeCN was added ca. 1 mL Of MeCN containing 50 BL (m20 eq.) of PhSH. The solution color immediately changed from red to red-orange. The optical spectrum of the solu- tion ShOWEd that [FGMOSQ(SPh)212_ had been formed quanti- tatively. 5. Tungsten Analog Interconversions The analogous interconversions outlined in parts 1-A were Observed for the tungsten analogs, [FeWSu(SPh)2]2’ and [FeWSu(OPh)2]2-. F. Reaction of [FeMOSQ(SPh)2]2- with Acenaphthalene Radical Anion TO a solution containing 3.5A mmol of (PhuAs)2[FeMOSu- (SPh)2] in A0 mL of DMA was added 3.2A mL (1.1 eq.) of a 1.2 M solution Of Na+ACN‘ in THF. The color of the solu- tion changed from red-orange to brown. A black, amorphous solid precipitated upon the slow addition of Et20. This solid was filtered, washed with Et2O, and dried in vacuo. Extraction of the solid with THF gave a brown solution whose optical spectrum was simply a rising absorbance. All attempts to crystallize the brown material met with failure as it slowly decomposed in solution. The remaining 16 solid was soluble in DMSO and only sparingly soluble in DMA or MeCN. Optical spectra of the solid in DMSO/MeCN or DMA resembled those spectra20a Obtained for [Fe(MOSu)2]3_, with peak absorbances near 590, 510, 3A0, and 290 nm. Also, upon reaction with excess Et3N/PhSH, the compound formed [FeMoSu(SPh)212-, as does [Fe(MOSu)2]3-. A11 attempts to recrystallize the material from various solvent systems were unsuccessful. G. Reaction Of_[FeMOSu(SPh)2]2- with Ferric Chloride and Thiophenoxide To a filtered solution of 0.80 g (A.9 mmol) of anhydrous FeCl3 in 30 mL of MeCN was added a solution containing 1.9 mL (18.8 mmol) Of PhSH and 2.7 mL (19.5 mmol) of Et3N in 10 mL of MeCN. A dark green slurry was formed, which became an intense brown solution upon the addition of 1.2 g (1.6 mmol) of (EtuN)2[FeMoSu(SPh)2] in 60 mL of MeCN. The solu- tion was stirred overnight, reduced in volume by ca. 30%, and then filtered. The slow addition of 50 mL of THF to the filtrate precipitated black microcrystals which were filtered Off, washed with THF, and dried in vacuo. They were then redissolved in a minimum of MeCN/THF (mAzl), filtered, and allowed to cool slowly to -20°C, whereupon small black plates crystallized. These were filtered, washed with THF, and vacuum dried. The optical spectrum of the product in MeCN contained shoulders at ca. 350 and 17 A50 nm on a rising absorbance. The product was recrystal— lized once more for elemental analysis. Calcd for (EtuN)3- [Fe6MO2S8(SPh)91, C78H105N3817Fe6Mo2: c, A3.AA; H, u.91; N, 1.95; S, 25.27; Fe, 15.5A; Mo, 8.90. Found: C, AA.O6; H, 5.22; N, 2.06; S, 23.A5; Fe, 15.10; MO, 8.5A. An NMR spectrum of the compound in d6-DMSO demonstrated broad peaks at -13.0, -5.5, and +3.5 ppm relative to tetramethylsilane for the protons Of the phenyl rings. The NMR and analytical data identified the compound as (EtuN)3[Fe6MO2SB(SPh)9], which had been previously synthesized by Garner and Co- 16 workers. H. Reactions Of (Etuy12LfleMgéu9121 1. Methyl Iodide A solution containing 2.33 g (9.60 mmol) of Et MeNI 3 and 0.30 g (0.96 mmol) Of (EtuN)2[FeMOSu012] in 20 mL of MeCN was prepared. The Optical spectrum of the solution was monitored for ca. A h to verify that iodide was not substituting for chloride in the complex ion. To this solution was added one containing 0.12 mL (1.92 mmol) of Mel in ca. 5 mL Of MeCN. A brown solid precipitated immediately and was filtered, washed with MeCN and dried in vacuo. The solid displayed no solubility in common solvents but did dissolve in 6N nitric acid. The optical spectrum of the filtrate was identical to that 18 12‘ reported20b for [C12FeS MOS FeCl . Addition of 0.20 g 2 2 2 (0.A1 mmol) of (EtuN)2MoSu in DMF resulted in quantitative conversion to [FeMOSuC1212-. 2. Benzoyl Chloride A solution of 90 uL (0.78 mmol) of PhCOCl in ca. 2 mL of MeCN was slowly added to 22 mL of a 17.7 mM solution (0.39 mmol) of (EtuN)2[FeM084012] in MeCN. A brown solid precipitated during the addition, leaving a nearly color— less filtrate. The filtered solid was insoluble in MeCN, DMSO, and toluene. 3. Methyl Magnesium Chloride (EtuN)2[FeMOSuC12], 0.11 g (0.18 mmol), was dissolved in 25 mL Of HMPA by prolonged stirring. TO this solution was added 150 uL (0.53 mmol) of a 3.6 M solution of MeMgCl in THF in dropwise fashion. The color of the solution gradually turned from yellow-brown to red-brown upon stirring, and some solid material precipitated. This was filtered, washed with THF, and vacuum dried. The product was a mixture Of light and dark solids. The light material gave a positive chloride test and was most likely magnesium chloride. The dark solid dissolved in MeCN and gave an optical spectrum consistent with [Fe(MOSu)2]3-. The optical spectrum of the pale filtrate showed only a rising absorbance at short wavelengths. 19 A. Sodium Diethyldithiocarbamate (EtuN)2[FeMOSu012], 0.12 g (0.20 mmol), and 0.05 g (0.22 mmol) Of NaEt2NCS2 were slurried in 30 mL of MeCN. Within fifteen minutes, the color of the solution changed from yellow—brown to deep red-violet. An Optical spectrum of the reaction mixture revealed formation Of [Fe(MoSu)2]3- 5. Sodium Benzylthiolate To 1.0 g (2.0 mmol) Of (MeuN)2[FeMoSuCl2] was added 0.61 g (A.l7 mmol) of NaSCH Ph. Both solids were taken 2 up in 50 mL Of MeCN and stirred overnight. A significant amount of amorphous solid precipitated and was filtered, washed with MeCN, and vacuum dried. The solid dissolved only in DMSO and water and demonstrated a spectrum cor- responding to MOSfi-. The filtrate was reduced in volume by ca. 50%, filtered, and combined with an equal volume of THF. A small amount (W20 mg) Of black solid precipi- tated; its optical Spectrum in MeCN corresponded to mixture of [Fe(MOSu)2]3- and MOS2-. 20 1. Reactions of [FeABAEAJ2- (X = S-thu, SPh, and C1 1. x = S-teButyl A solution containing 0.10 g (0.12 mmol) Of (MeuN)2- [FeuSu(S-teBu)u] and 0.23 g (0.A8 mmol) of (EtuN)2MoSu in 50 mL of MeCN was prepared. Optical spectra of the reaction solution taken at intervals up to four days in- dicated no reaction had taken place. 2. x = S-thutyl with Me3py'HC1 TO a flask containing 0.68 g (0.79 mmol) of solid (MeuN)2[FeuSu(S—t—Bu)u] was added a solution of 1.18 g (3.18 mmol) Of (MeuN)2MOSu dissolved in A0 mL of MeCN and 30 mL of NMF. To this deep orange—brown solution was added a solution containing 0.50 g (3.18 mmol) Of Me3py-HCI in 50 mL of MeCN. Upon addition of this latter solution, the reaction mixture turned an intense red-violet. After stirring overnight, no evidence Of decomposition was noted. THF (”100 mL) was then added to the solution and a violet- black, amorphous solid precipitated; it was filtered, washed with THF, and dried in vacuo. Stirring of the amorphous solid in MeCN overnight resulted in conversion to microcrystals which were filtered, washed with MeCN, and vacuum dried. Anal. Calcd for CA8H120N6S20FGAMOA: C, 17.02; H, A.30; N, A.96; S, 37.87; Fe, 13.19; Mo, 22.66 21 Found: C, 16.9A; H, A.A6; N, A.8l; S, 37.65; Fe, 12.90; Mo, 22.32. 3. X = SPhenyl A solution of 1.5A g (3.17 mmol) of (EtuN)2MoSu in 100 mL of warm (W30°C) MeCN. This solution was added to a flask containing 0.83 g (0.79 mmOl) of solid (EtuN)2[FeASA‘ (SPh)u], which dissolved upon stirring overnight. After several days of stirring, a red—black of microcrystalline material precipitated, which was filtered, washed with THF, and dried in vacuo. The optical Spectrum corresponded 21 exactly to that of [Fe2MoS6(SPh)2J3‘. A. X = Chloride To a flask containing 0.AO g (0.53 mmol) of (EtuN)2— [FeASAClA] solid was added a solution of 1.03 g (2.12 mmol) of (EtuN)2MoSu in 70 mL of MeCN. After being stirred over- night, the optical spectrum of the reaction mixture re- vealed formation of [Fe(MOSA)2]3- in high yield. J. Reactions of [M089]2- 1. Triethylammonium Thiophenoxide A solution Of 0.30 mL (3.00 mmol) Of PhSH and 0.A2 mL (3.00 mmol) Of Et3N in 15 mL Of MeCN was added to 0.16 g 22 (0.25 mmol) Of solid (EtuN)2[M089]. The solid dissolved slowly upon being stirred overnight, and the optical spectrum of the reaction mixture indicated quantitative formation 2.. of MOS“ . 2. Ferrous Chloride TO a flask containing 0.0A g (0.32 mmol) of anhydrous ferrous chloride and 0.20 g (0.32 mmol) of (EtuN)2[Mosg] was added 20 mL of MeCN and 2 mL of DMSO. The solution was stirred for ca. two days, at which point a small amount of solid precipitated. The optical spectrum of the supernate was that of [FeMOSuC12]2-. 3. Sodium Ethylthiolate and Ferrous Chloride TO a slurry of 0.55 g (A.35 mmol) Of ferrous chloride in 50 mL of MeCN was added a solution of l.A6 g (l7.A mmol) of NaSEt in 30 mL Of MeOH. The mixture became intensely red within ca. 5 minutes. The color gradually changed to green—brown upon the addition Of a slurry of 0.9A g (1.A5 mmol) of (EtuN)2[M089] in MeCN. The optical spectrum of the reaction solution after being stirred for A8h contained maxima at 59A, 509, 396, and 321 nm, and roughly corres- ponded to the spectrum reported22 14. for the [Fe2M02SlO] ion. The filtrate was reduced in volume in vacuo, and ca. 150 mL of Et2O was then slowly introduced. This resulted 23 in precipitation Of an oily black material that slowly solidified upon prolonged stirring. The solid was filtered, washed with Et20, and vacuum dried. The Optical spectrum of the solid in MeCN was the same as that of the reaction mixture; indicative Of [Fe2MO2810JM- formation. The spec- trum Of the filtrate more closely resembled that of [Fe- (MOSu)2]3-. An attempt was made to recrystallize the solid from MeCN, but it slowly decomposed upon standing in solu- tion. A. Ferric Chloride and Sodium Thiophenolate To a solution containing 0.A0 g (l7.A mmol) of sodium metal in 15 mL of MeOH was added 1.8 mL (17.6 mmol) of PhSH. To this was added a filtered solution of 0.71 g (A.35 mmol) of anhydrous ferric chloride in 20 mL of MeCN, resulting in a deep purple—brown slurry. Addition of 0.93 g (l.A5 mmol) of (EtuN)2[M089] and 0.75 g (A.53 mmol) Of EtuNC1 in 30 mL of MeCN caused an immediate color change to an intense red-brown. After being stirred overnight, a white solid (NaCl) had precipitated from the reaction mixture. The volume Of the solution was reduced in vacuo by ca. 50%, and the sodium chloride was filtered off. Slow addition of Et20 to the filtrate resulted in precipitation of a black Oil that partially solidified upon stirring. The supernatant liquid was separated by filtration, and the pro- duct was extracted with ca. 60 mL of MeCN and filtered. 2A The optical spectrum of the filtered solution showed es- sentially a rising absorbance with features at 580, 510, AAO, and 350 nm. Slow cooling Of the solution to -20°C resulted in precipitation of an amorphous solid that would not redissolve upon warming. This slow decomposition of the material prevented further purification. 5- ifiihfllgLE§h§h1§Efllul A mixture of 0.30 g (0.29 mmol) of (EtuN)2[FeuSu(SPh)u] and 0.37 g (0.58 mmol) of (EtuN)2[M039] was dissolved in A0 mL Of MeCN and allowed to stir for four days. Slow addition Of Et2O resulted in separation of a dark oil that quickly solidified. The material was filtered, washed with Et20, and vacuum dried. The Optical spectrum Of the solid in MeCN corresponded exactly to that of [Fe2MOSll]3'. 6. Na2[Fe(CQ)u1'l—l.5 Dioxane A solution of 0.50 g (l.A6 mmol) of Na2[Fe(C0)u]- l-l.5 dioxane in 30 mL of DMA was added to a solution of 0.A7 g (0.73 mmol) of (EtuN)2[M089] in 100 mL of DMA. Upon stirring overnight, the color of the solution changed from brown to red and a gray solid precipitated; which, after filtration, was found to be soluble only in 6N hydrochloric acid with evolution of hydrogen sulfide gas. The optical spectrum of the filtrate was analogous to that of Mosfi‘ in DMA. 25 K. Reactions of [Mo01613‘ 1. (gtuN)2£332§2(SPh)ul A solution containing 0.027 g (0.038 mmol) Of (EtuN)3- [MOC16] in 10 mL of MeCN was added to a slurry Of 0.10 g (0.12 mmol) of (EtuN)2[Fe2S2(SPh)u] in MeCN. The mixture was allowed to stir for several days and periodic Observa- tion Of the Optical spectrum showed only slow conversion of [Fe2S2(SPh)u]2- to [FeuSu(SPh)u]2-. 2. Ferric Chloride and Thiophenolate A solution containing 2.6 mL (25.7 mmol) of PhSH and 3.6 mL (25.7 mmol) of Et3N in 10 mL of MeCN was added to a solution containing 1.0 g (l.A3 mmol) of (EtuN)3[MOC16] dissolved in 50 mL of MeCN. This solution was then added to a filtered solution of 0.70 g (A.28 mmol) Of anhydrous ferric chloride in 20 mL Of MeCN. After being stirred for 2A h, 1.0 g (31.2 mmol) of solid sulfur was added and upon stirring overnight, the solution became an intense red- brown color. At this point, the solution was reduced in volume by ca. 50% in vacuo, and then filtered. Addition of ca. 100 mL of THF resulted in precipitation of a light red solid whose Optical spectrum in MeCN, after filtration, matched that of [M0C16J3—. As a control, the analogous reaction was performed with the omission of (EtuN)3[MoCl6] and identical results were obtained. The product remaining 26 in solution was subsequently identified3u as [Fe281212- 3. Thiophenolate A solution containing 0.15 g (0.21 mmol) of (EtuN)3- [MOC16], 0.36 mL (2.60 mmol) Of Et3N, and 0.26 mL (2.55 mmol) Of PhSH in 25 mL of MeCN was allowed to reflux over- night. The color Of the solution changed from pink to deep red-violet and the Optical spectrum of the reaction mixture contained a maximum at 550 nm. Brief exposure Of an aliquot of the reaction solution to air produced an intense blue solution whose Optical spectrum contained a 25 maximum at 590 nm. These Observations identified the initial product as [Mo2(SPh)9]3‘, which reacts with oxygen35 to form [MOO(SPh)u]-. L. Reactions of [MO2(SPh)9]3- l ' fluflgflgg§2i§mul To a flask containing 0.A2 g (0.A8 mmol) Of (EtuN)2- [Fe282(SPh)u] slurried in 100 mL Of MeCN, was added 9.A mL (0.A8 mmol) of a 0.05 M solution of (EtuN)3[MO2(SPh)9]. After stirring overnight, there was no apparent color change in the solution and most Of the iron dimer remained un- dissolved. The optical spectrum of the supernatant liquid was simply that of [M02(SPh)9]3-. Heating of the reaction mixture for ca. A0 h at A5-50°C resulted in dissolution 27 of the solid and a change in the optical spectrum tO a rising absorbance. Subsequent slow cooling Of the solution to -20°C resulted in precipitation of an oily black solid that would not redissolve upon warming. 2' SEEAE12LEE2§121 TO a solution containing 5.9 mL (0.A2 mmol) of 0.07 M (EtuN)3[MO2(SPh)9] in 20 mL Of MeCN was added a solution containing 0.63 g (0.83 mmol) of (EtuN)2[Fe2Sl2] dissolved in 100 mL of MeCN. After being stirred overnight, THF was slowly added to the reaction mixture and a black oil immediately separated out. The supernatant liquid was decanted, and the oil was taken into a minimum of MeCN/THF, filtered, and slowly cooled to -20°C. This precipitated an amor- phous black solid that would not redissolve upon warming. This insoluble material was filtered off, and the filtrate was recooled to -20°C, yielding more insoluble residue. 3. (EtuN)2[FeuSuClu1 A solution containing 0.23 g (0.31 mmol) of (EtuN)2- [Fe S Cl 1 dissolved in 50 mL of MeCN was prepared. To A A A this was added 11.9 mL (0.61 mmol) of a 0.05 M stock solu- tion of (EtuN)3[M02(SPh)9] in MeCN. The reaction mixture was allowed to stir until the optical spectrum remained constant, at which point THF was slowly introduced. A dark Oil was deposited on the walls of the flask and did 28 not solidify upon prolonged stirring. The supernatant liquid was decanted, and the oil was dried in vacuo. The optical spectrum of the oil in MeCN was a rising absorbance with a slight maximum at A50 nm, possibly corresponding to [FEASA(SPh)A]2_' The filtrate decomposed upon standing. M. Reaction Of Ferrous Chloride with [MOO(SPh)ul: Addition of a solution of 0.70 g (0.103 mmol) of (EtuN)- [MOO(SPh)u] dissolved in 70 mL of MeCN to a slurry of 0.13 g (0.103 mmol) Of anhydrous ferrous chloride in MeCN resulted in a gradual (W3h) color change of the reaction solution from deep blue to deep orange-brown. Addition Of nonpolar solvents such as THF, Et20, and toluene to small aliquots of the reaction mixture all resulted in separation of a red Oil, but addition of the reaction solution to an excess of isopropanol precipitated the product as an orange, amorphous powder. These results are exactly analogous to those ob- tained by Dance, Wedd, and coworkers, who examined36 the reaction between [MoO(SPh)u]- and FeCl3 in MeCN. They formulated the product as (EtuN)[Mo202(SPh)6Cl]. III. RESULTS AND DISCUSSION A. [FeMSu(SR)p]2- (M = Mo, W; R = Ph, peTol) 1. Synthesis and Interconversions The aryl thiol complexes, (EtuN)2[FeMSu(SR)2], can be prepared in high yield by a variety of pathways: (1) by reaction of (EtuNkflFeMSu0123 with 2-A equivalents of RS“; (11) by treatment Of (EtuN)3[Fe(MSu)2] with a large excess Of RSH/Et3N; (iii) by reaction19b of MSi- with [Fe(SR)u12-. Route (i) can be viewed as a simple ligand exchange Of thiolate for chloride and can be accomplished with either RSH/Et3N or NaSR. Use of the latter reagent results in a 20-30% higher yield of the thiolate complex, probably due to precipitation Of sodium chloride, which drives the re- action to completion. Pathway (ii) demonstrates the degra- dation of a trinuclear species to a dinuclear one under extreme conditions, i.e., reaction with ca. 300-A00 equiva- lents of RSH/Et3N. This reaction is also accompanied by oxidation of the metal centers. Method (iii) represents substitution of two thiolate ligands by the tetrathio- metallate, which acts as a bidentate ligand. Once formed, the species [FeMSu(SR)2]2- can be con- verted into other dimeric complex ions by reactions with 29 30 simple reagents. Treatment of the thiolate complex with two equivalents of PhCOCl in MeCN gives rise to the chloro complex, [FeMSuC12]2-, in quantitative yield. The reaction presumably proceeds via thioester formation: 2_ MeCN 2_ [FeMSu(SR)2] + 2PhCOCl——.[FeMSuC12] + 2PhCOSR This reactivity is identical with that Observed29 for the FEASA and Fe2S2 thiolate complexes towards PhCOCl. Use of excess reagent results in immediate decomposition, which may result from acylation of terminal sulfide ligands. Similarly, reaction of [FeMSu(SR)2]2_ with excess sul- fur in MeCN results in gradual formation of [FeM89]2-, which contains the 8%” ion as a bidentate ligand on the iron atom. Presumably, this occurs by partial reduction Of elemental sulfur with concomitant oxidation Of thiolate. The complex ion, [FeM89]2-, was first synthesized by 19a Coucouvanis and coworkers by reaction of [FeMSu(sR)2]2- with five equivalents of benzyl trisulfide. The S?- 1igand, although not common, has been reported before in37 [Pt(85)3]2-, [Ti(Cp)2(SS)],38 and the dimeric complex,3u [Fe2812]2-. The reaction scheme given in Figure 1 sum- marizes the synthesis and reactivity of the [FeMSu(SR)2]2- complexes. 31 Figur e 1 . Synth eses (M = Mo and rea ° , W). ctiVlty f o [F eM Su(SPh)212- 32 H osswflm an: + maoos :m: + Immaflcfiummoc mam mze mo OHOHmczOO mpOHLw moozm mam.ml .mw.ml "mOHOBIQ mmoél among www.ml «OH.>I "mmHOBIO mom.sI "mmcd .moocosogos capocwwemfio MM nonficm mcofipSHOm zomao CH0 .HIEO CH moomm pm nonmao ocoaazpozfioa coozpon mHHSE HOnsz :Hp .AmIoH x HIEo HIzv s: 2H Adv woe‘ moomm ME: oowA omocp Loo COHpSHOm omzm we: oomv pm mopdpmom Lou soapsfiom zoo: ch I AmaoImewma- is.mmveam .Am.mavawm aoeImIo I a .AmIeva.meI .AmIovee o .Aw.mvoem .Am.sveae .Aeavmm: as I z I Am.smvosm I AmInvm.ma .Am.HmVsmm .Ae.mvmmm .Am.mvmae ago I x .AmIsvs.maI .AmIovm: mmm .sms .mm: .Aenvoms .Asa.ovows .Aenvosoa A: I z ompmflnm OHOOLpOmH moszpmom cosmsmcH mmOLSpmom n Hmsuooom OficosuOOHm .OOBCAQCOQ .HHH canoe A6 They reflect splitting of the degenerate S+Mo or S+W charge-transfer transitions in the free tetrathiometallate into components at both higher and lower energy due to lowering of symmetry upon coordination to iron. The spectra also exhibit other features, and since this region will un- doubtedly contain thiolate— and sulfide-tO-iron charge- transfer bands as we11,28’29’u5 exact assignments of op- tical transitions between 300—500 nm would only be specu- lation. Each feature in the spectrum of [FeWSu(SPh)2]2- is shifted to higher energy by ca. 30 nm with respect to that of [FeMoSu(SPh)2]2-, much as the absorptions of WSE' are blue-shifted relative to those of MOSE-. In View of the expected higher orbital energy difference between sulfur and tungsten 1s sulfur and molybdenum, this behavior seems reasonable, at least for those charge-transfer transitions within the tetrathiometallate unit. Substitution of chloride for thiolate on iron causes a slight shift (m10-20 nm) to higher energy in this region. Qualitatively, one might expect such behavior due to the slightly greater orbital energy difference between chlorine and iron 1s sulfur and iron. Below 290 nm, very intense absorption bands are Observed. These undoubtedly contain internal ligand (n + N*) excitation bands when the terminal ligand is thiolate. Other transitions due to charge-transfer between metal centers or to d-d transitions localized on one metal center are possible. High spin iron(II) in A7 tetrahedral symmetry is expected to have a characteristic 5 5 E + T2 transition, with the energy of this absorption equal to the tetrahedral crystal field splitting, At. This band has been noteduu for the complex [Fe(SPh)u]2- and 1 -1). occurs at 1700 nm (e mlOO M- cm For the complexes [N(82PS)2]FE and [HC(R2PS)2]Fe, this band is Observedus near 2700 nm (e mlOO M-lcm-l). Similarly, Holm and co- workers report“0 a broad band for this transition at l 1800 nm (e M120 M- cm-l) for [Fe(82—s-Xyl)2]2_. The asym— metry of this latter band is attributed to a small split- ting of the t2 orbitals due to deviations from ideal tetra- hedral symmetry. NO such absorptions are observed for [FeMOSu(SPh)2]2- and [FeWSu(SPh)2]2— at such low energies, at least out to ca. 2200 nm. Instead, broad, asymmetrical 1 absorptions between 900 and 1100 nm (e m100 M— cm_l) are Observed. These are illustrated in Figure 5. If these are indeed E + T2 transitions, they reflect a nearly two— fglg increase in the magnitude Of At relative to [Fe(SPh)A]2- l 1 and [Fe(s2-o—Xy1)2]2' (t10,000 cm“ is 5700 cm' ); this seems unlikely. Also, the calculated value of At’ based A A2 A on the energy Of the A2 + T1(P) transition, for [Co(WSu)2]2- is smaller than that reported“0 for [Co- (SPh)u]2_ (m2A00 cm"1 is mAOOO cm-l), implying that tetrathiometallates are not strong crystal field splitting ligands relative to aryl thiolates. Noting these trends, 5E + 5T d—d transition it seems unlikely that a simple 2 A8 Figure 5. Near-infrared spectra of (EtuN)2[FeMoSu- (SPh)2] (A), (EtuN)2[FeMOSu(OPh)2] (B), and (EtuN)2[FeMOSu0121 (C) in dimethylsulfoxide solution at 23°C. e (M"cm") 50 ISO A9 I I I I I I l 800 IOOO IZOO I400 '600 '800 2000 A(NM) 50 is responsible for the Observed near-IR absorptions. An alternative assignment would be an intervalence charge-transfer between the metal centers, i.e., Fe(II)— h Mo(VI) .3 Fe(III)-MO(V). Most intervalence bands have large half-widthsu6 (m2000—6000 cm-l) and extinction co- efficients ranging from 500-5000 M.1 cm—l. Also, they typically Obey (i10%) the relation, A = ( 2310)1/2 at 300 K,“7 VIT where VIT is the frequency of the absorption maximum in cm-1 and A is the half-width of the absorption. The approximate half-Widths of the near—IR bands of [FeMoSu- (SPh)2]2- and [FeWSu(SPh)2]2- are ca. 5800 and 5200 cm-1, respectively, and are within 15% of the calculated values. The asymmetric nature Of these bands, i.e., a peak with a shoulder, suggests that they really consist of two com- ponents which makes the determination of vIT and A difficult. It is possible that they represent an intervalence charge- transfer and a thiolate-tO-molybdenum or tungsten charge- transfer which are very close in energy. In support of this, the analogous band Observed for [FeMOSuC12J2- is nearly symmetrical. Although these absorptions resemble intervalence bands in position and half-width, their extinction coefficients are very low, ca. 110 M-1 cm-l. Therefore, actual assess- ment of the nature of these near-IR absorptions will have to await measurements on a variety of other binuclear systems containing tetrathiometallate centers. 51 The infrared spectra of the molybdenum and tungsten thiolate dimers, in the region of 250-600 cm’l, demonstrate the typical splitting pattern observed for tetrathiometal- late anions coordinated to a metal atom.39’u2 Terminal sulfur-metal stretches are shifted to higher energy and bridging sulfur-metal stretches to lower energy, relative to the degenerate vibrations of the free tetrathiometal- late. These data appear in Table III. M. Electrochemistry The electrochemical behavior of the complexes, [FeMSu- (SR)2]2— (M = Mo, W and R = Ph, p—Tol), has been examined by using polarographic and cyclic voltammetric techniques. Significant data are presented in Table IV. All measure- ments were made in MeCN solutions containing 0.05 M [EtuN][C10u] as supporting electrolyte and recorded rela- tive to a standard calomel reference electrode. Due to problems with adsorption of these sulfur-rich complexes onto conventional electrode surfaces such as platinum and glassy carbon, attempts to use cyclic voltammetry to study electrochemical behavior met with little success. The cathodic wave was of reasonable shape and magnitude, but the anodic wave was featureless and indicated a sharp decrease in current. In the case of platinum, the elec- trode surface was usually discolored after a single scan. Polarography, employing a dropping mercury electrode, 52 e c we s.m :N.Hu Hoermwm H m p no.m ww m.m mm.H| smo H m a a aw m.m 2.? 3.7mm“ .u.. m mAmm.HVHm.o om.: Hmafi m.m H:.Hu smm H m Amm.ovmm.o "m Ame pmav MAH:.HV::.O “a om.m gm: s.m mm.au ummfifimozmmmi o e as s.m mm.H- Hoequm H m vom.HVam.o mo.m we o.m Hm.H- mwm M m o e we m.m om.Hu HOBnWmm H m mhmo.avss.o mm.: mm w.m mm.fi- WWW H m m\ee .Amav mH nzm .pwww: >5 .mQOHm .mmmoflmwan m> .m\Hw .mHHmozmmmimA23pmv cam Apao .pH mfian 53 .AEE m.mv coupwppcoocoo swam .oouhoumao mefixmz .SOHHSHom E m mmcmom mum mo moSHm> omwpm>w ohm Ham mmom MN comm um coau3H0m zoo: :Hw .omscfipcoo .>H magma 5“ proved to be the most useful method for studying these complexes. This system permits continual generation of a fresh electrode surface two times a second. Well-formed polarographic waves were obtained that permitted a more quantitative analysis. Significant electrochemical data appear in Table IV. Examination of the electrochemistry of these complexes over the potential range 0.0 to —2.0 V yielded only one electrochemical reduction between -l.0 and -2.0 V in each case. At potentials lower than -2.0 V, very high currents were observed, presumably corresponding to ligand reduc- tion. Plots of log [i/(iD-i)]y§_voltage for each complex gave slopes much higher than the theoretical value (59 mV) for a single, one-electron reversible reduction. These results are similar to those29 obtained for the dimeric species, [Fe282xu]2' (X = halide), which also demonstrate irreversible reductions over the same potential range. However, they are contrary to those obtained28 for [Fe282- (SR)u]2-, which undergoes two reversible one-electron re- duction processes. In order to further examine the irreversible nature of the reduction of the [FeMSu(SR)2]2‘ complexes, attempts were made to generate a trianion with a chemical reductant, namely sodium acenaphthalenide (Na+ACN;). This radical anion has a reduction potential of -l.6 V vs SCE, and is well suited as a one-electron reductant. Addition of 55 successive, one equivalent aliquots of‘ Na+ACN‘ in THF to solutions of [FeMoSu(SPh)2]2- in DMA, followed by rapid freezing and investigation by EPR spectroscopy over a range of temperatures and microwave powers showed no evidence for the formation of a stable, half-integral spin species. Likewise, examination of a frozen DMA solution of [FeMoSu- (SPh)2]2- containing one equivalent of ACN; by Mossbauer spectroscopy at A.2 K revealed a complex spectrum attribut— able to a mixture of products. Also, a large scale reaction of a DMA solution of [FeMoSu(SPh)2]2- with 1-3 equivalents of ACN; yielded a product whose optical spectrum resembled that of [Fe(MoSu)2]3-. This latter result is in accord with those obtained upon exposing [FeMoSuClZJZ— to other potential reducing agents, as will be discussed in detail later. Therefore, spectroscopic and electrochemical, as well as chemical, evidence suggests that the reduced SDGCieS, [FeMoSu(SPh)2]3‘, does not possess any reasonable stability. The reduction potentials of [FeMoSu(SPh)2]2- and [Fewsu(SPh)2]2- are -l.28 V and -l.Ul V, respectively; substitution of tungsten for molybdenum thus makes the complex harder to reduce. This behavior is analogous to results“8 obtained for the bis(tetrathiometallate) anions, [Ni(MSu)2]2-, [Pd(MSu)2]2-, and [Pt(MSu)]2- (M = Mo, W). These latter complexes also exhibit more negative reduc- tion potentials when molybdenum is replaced by tungsten. 56 This effect may be due to the lower electron-with- drawing (n-acceptor) properties of WSE’ relative to MoSi', as predicted by theoretical (SCC EH) calculations,“9 on the [FeMoSu01212- and [FeWSuCl2J2- ions. This effect also manifests itself in the 57Fe Mossbauer results on the two thiolate dimers. These results will be discussed in detail in a later section, but briefly they indicate greater "ferric" character for the iron atom in the molybdenum complex, which correlates well with the observed greater ease of reduction for the molybdenum dimer. Substitution of p—TolS for PhS results in a slight (10-20 mV) decrease in the reduction potentials of both the molybdenum and tungsten dimers. This is in accord with results obtained for the Fens“ and Fe282 thiolate 28,U3 and can be explained by simple inductive complexes, effects of the electron-releasing methyl groups on the phenyl rings. 5. Magnetic Susceptibility The room temperature magnetic susceptibilities of the compounds, (EtuN)2[FeMSu(SPh)2] (M = Mo, W), have been determined by the Faraday method. These data give rise to effective magnetic moments of “.89 BM for M = Mo and “.90 BM for M = w. These values have been corrected for diamagnetic contributions to the susceptibility from the 57 ligands and cations by use of Pascal's constants. These effective magnetic moments are consistent with the pres- ence of four unpaired electrons in an S = 2 ground state in these molecules. This configuration could conceivably correspond to either a high-spin Fe(II)—Mo(VI) system or to a high-spin Fe(III)-Mo(V) system with appreciable anti— ferromagnetic coupling between iron (d5) and molybdenum (d1) ions at room temperature. In order to examine this question further, the tempera- ture-dependent susceptibility of a polycrystalline sample of (EtuN)2[FeMoSu(SPh)2] was measured from A.2 to 300 K. The sample was found to be free of ferromagnetic impurities by noting no change in the susceptibility as a function of applied field at 300 K. The temperature-dependent sus- ceptibility datawere found to conform to the Curie-Weiss formula, x = C/(T+0) + x0, over the temperature range 30-300 K. A least—squares analysis provided 0 = “.2 K, C = 3.187 emu'K/mol, and x0 = -5.96 x 10")"l emu/mol. The resulting plot of XM vs T is shown in Figure 6. Deviation from Curie-Weiss Law was noted below 30 K, with a maximum in the plot at m8 K followed by a rapid decrease in the susceptibility down to A.2 K. The antiferromagnetic coup- ling noted could arise from two sources: (1) an intra- molecular coupling between high spin Fe(III) and Mo(V); (ii) an intermolecular interaction between neighboring ions in the crystalline lattice. In order to distinguish Figure 6. 58 Magnetic susceptibility of (EtuN)2[FeMoSu- (SPh)2] as a function of temperature in the solid state (0) and in N,N-dimethylformamide glass (0), obtained in 0.16-T applied field. The left vertical axis refers to the solid, and the right to the frozen solution. 59 om om. 0% 0% oxen: oov omv own ovw ’0/ 4L 1,0, r—a-i—v—"f" _ A 2 _._1_ _ 0.? 0m ON— Om. 0x09 OON OVN 0mm ONM 60 between these two alternatives, the temperature-dependent susceptibility of a magnetically "dilute" frozen solution of the same compound in DMF was measured. In this case, no antiferromagnetic interaction was observed down to A.2 K; this plot is also presented in Figure 6. These data suggest that the antiferromagnetic interaction is intermolecular in nature and is simply a function of the crystalline lattice. Also, if the coupled Fe(III)-Mo(V) description were correct, the energy difference between S = 2 and S > 2 would have to be much larger than thermal energy even at 300 K, and this seems unlikely. 6. Mossbauer spectra The successful interpretation of 57Fe Mossbauer spectra can provide useful insight towards assigning formal oxida- tion state(s) to iron atom(s) in complex molecules. Since magnetic susceptibility measurements cannot completely rule out the possibility of an Fe(III)-Mo(V) interaction, Moss- bauer Spectra of frozen DMF solutions of (EtuN)2[FeMSu- (SPh)2] (M = Mo, W) were recorded both in the absence and presence of an applied magnetic field. Solution sam— ples were free of intermolecular magnetic interactions, thus supplying more easily interpreted spectra compared to solid samples. However, spectra of polycrystalline samples diluted with boron nitride were also obtained. 61 The isomer shift measured for the solid samples was es- sentially the same as that obtained for the solution samples while the quadrupole splitting of the former was ca. 0.10 mm/s lower than the latter. The rigorous theoretical treatment of zero—field or magnetic Mossbauer phenomenon is very complex and will not be presented here. Instead, semi-quantitative interpreta- tions of the theory and data obtained will be presented. In order to assess the formal oxidation state of an iron atom in a molecule based on the 5 and AEQ values, one must compare the observed values to those of complexes with similar coordination environments and known oxidation states. Unfortunately, the relative lack of well—charac— terized Mo/Fe/S and W/Fe/S systems, from a Mossbauer stand- point, prevents such a direct comparison. Instead, one must resort to the Mossbauer data reported for the mono- meric, dimeric, and tetrameric, iron-sulfur centers which formally contain Fe2+, Fe3+, 2'5+ 2'25+, 15 and Fe or Fe respectively. The “.2 K Mossbauer spectra of the molybdenum and tungsten thiolate dimers are presented in Figures 7 and 8, respectively, while relevant parameters are listed in Table IV. The plot relating isomer shift to formal oxida- 15 tion state is given in Figure 9. The isomer shifts (quad- rupole splitting) of the molybdenum and tungsten thiolate dimers are 0.A7 mm/s (1.63 mm/s) and 0.51 mm/s (1.92 mm/s), 62 Figure 7. Mossbauer spectrum of (EtuN)2[FeMoSu(SPh)2] in frozen N,N-dimethylformamide solution re- corded at A.2 K with zero applied field. Vertical bar indicates one percent absorption. 63 m3: 2. t._oo._m> e m N . o _. m- n- T . . . . _ _ . _ _ _..... __ __. _ _ . _ . __ . _ . __ .. .. H _. _ . . ._. __. __. .3. __ c... .....!......=.._ .5... - Fur _ . . at._.....__{.,._._i......._..__....__...2 . . . . . L 64 Figure 8. Mossbauer spectrum of (EtuN)2[FeWSu(SPh)2] in frozen N,N-dimethylacetamide solution re- corded at “.2 K with zero applied field. Vertical bar indicates one percent absorp- tion. 65 m\22 z. >._._oo.._m> 1 .1 .1 66 Figure 9. Correlation diagram relating isomer shift to formal oxidation state for various iron-sulfur centers.15a 67 m._.<._.m 202K038 .325“. abut +Nmu . endow — _ Nd nd ed 0.0 0.0 5.0 'IVLBW 9.:l '3‘ (on/WW) .Ldll-IS HSWOSI 68 respectively. The analogous data for monomeric high-spin iron(II) and iron(III) in a tetrahedral sulfur environ- ment are 0.6 mm/s (3.28 mm/s) and 0.13 mm/s (0.57 mm/s), respectively. Comparison of the isomer shifts observed for [FeMoSu(SPh)2]2- and [FeWSu(SPh)2]2- to those of the documented iron-sulfur predicts a formal oxidation state for iron of +2.6 in the molybdenum dimer and +2.5 in the tungsten dimer. Regardless of whether the absolute magnitudes of these calculated values are correct, their relative order correlates well with electrochemical results, i.e., greater "ferrous" character for iron in [FeWSu— (SPh)2]2'. Likewise, it follows the generally observed trend that Mo(VI) complexes are easier to reduce than their W(VI) counterparts. Because the dependence of the isomer shift on the formal oxidation state of iron in the iron-sulfur culsters may not be the same as for the molybdenum and tungsten thiolate dimers, the spectrum of (EtuN)2[FeMoSu(SPh)2] in applied magnetic fields was examined. These spectra are complex, but can be simulated50 very well by spec— tral calculations based on a tetrahedral, high—spin, ferrous model. Also, the hyperfine parameters obtained from the magnetic Mossbauer spectrum indicate anisotropy in at least one direction. Qualitatively, this result is expected for a high-spin ferrous system where a non- symmetrical electronic field gradient prevails. 69 Indeed, the magnetic Mossbauer spectrum of reduced (Fe(II)) 51 52 rubredoxin, an iron-sulfur protein with an FeSu center, also demonstrates magnetic anisotropy in its hyperfine parameters, while the oxidized form of the protein (Fe(III)) shows very little hyperfine anisotrOpy. In conclusion, Mossbauer results on the molybdenum and tungsten thiolate dimers suggest that they are best des- cribed as an Fe(II)-M(VI) system with slight delocalization of electron density through bridging sulfide ligands to either molybdenum or tungsten. 7. Proton Magnetic Resonance Sppctra The isotropic shifts for the protons of the aromatic thiolate ligands in the complexes, (EtuN)2[FeMSu(SR)2] (M = Mo, w; R = Ph, peTol), have been measured as a function of temperature in deuterated MeCN or DMSO. Selected Spectra of the molybdenum complexes are shown in Figure 10 and room temperature chemical shifts for all complexes are listed in Table III. Before a discussion of the actual data is given, a brief investigation into the factors contributing to the observed shifts is in order. The observed isotropic shift is really the sum of two separate elements:53 The contact shift and the dipolar shift. This can be expressed as follows: Figure 10. 7O NMR spectra of (EtuN)2[FeMoSu(SPh)2] in CD3SOCD3 at 69 (A) and 17°C (B), along with NMR spectra of (EtuN)2[FeMoSu(S-peTol)2] in CD3CN at —U5 (E), 25 (D), and 75°C (C). Small amounts of impurity are denoted by X. Chemical shifts are in parts per million from MeuSi. 71 TMS A Et‘Ntr 69'C \ m-H -339 -30 B I7-c —JIL¥ -394 C 75'C p-CH3 m-H x X -42.7 £513 D 25'c Aid/Iii -501 -381 E -45'C «(L A -666 -489 Q-H sés I 40.2 \ms 72 isotropic AH contact AH dipolar = (T?) + (fif) The contact term results from direct unpaired Spin de- localization onto the magnetic nucleus being observed. The magnitude of this shift is given by the equation,53 AH contact (__) _ _ AYe gBS(S+l) H 3 YH 3kT where A is the proton-electron coupling constant; YH and Ye are the gyromagnetic ratios of the proton and electron respectively; g is the nuclear g factor and B is the Bohr magneton. The key point is that the contact term is dir- ectly proportional to the susceptibility of the system, inversely proportional to the temperature, and not a func— tion of distance from the paramagnetic center. 0n the other hand, the dipolar term is given by the relation,53 (Afl)dipolar = -(X - X H II _|_ I, 3 2 ) [3c0330-lJ where XII and XJ are the axial components of the magnetic susceptibility; 6 is the angle between the magnetic axis 73 and the nucleus in question; and r is the distance between the nucleus and the paramagnetic center. Without single crystal magnetic anisotropy measurements, the dipolar term cannot be assessed, but its effects, particularly the l/r3 dependence, can be observed in the NMR spectrum. For ligands with aromatic ring systems, the easiest way to distinguish which of these terms is dominant is to substi— tute methyl groups for hydrogen atoms. The coupling constant for the interaction between unpaired spin density on carbon with the hydrogen nucleus is opposite in sign 53 to that between two carbon atoms. The net result is an isotropic shift in the opposite direction and of comparable magnitude for the protons of the methyl group relative to that of the replaced proton. The usual convention for reporting chemical shifts in paramagnetic complexes designates those Shifts which are downfield of the diamagnetic reference as negative and those upfield as positive. This convention will be ad- hered to in subsequent discussions. Resonances for the tetraethylammonium cations were broadened significantly, but were only slightly shifted and will not be discussed further. At 297 K, isotropic shifts for the phenyl ring proton resonances for [FeMoSu(SPh)2]2- are observed at +96.6 ppm, ca. +39 ppm, and -3l.2 ppm relative to free PhSH in CD3CN. The resonance at ca. +39 ppm is very broad and is not 7U observable at low temperatures. Since dipolar linewidth 6 dependence,53 this broadening is expected to have a l/r resonance has been assigned to the pgphp protons. Upon substitution of the proton in the papa position by a methyl group, the resonance at +U6.6 ppm disappears and a reson- ance with three times the intensity appears at -M8.7 ppm. This phenomenon signifies that the contact term in the isotropic shift expression is dominant. The smaller isotropic shift for the ppphp proton may reflect some contribution from the dipolar term in the opposite (downfield) direction. Since an asymmetrical system such as this would be expected to experience some degree of magnetic anisotropy, the dipolar term must be considered. And, inasmuch as this component is a function of l/r3, the ppphp position would certainly be affected the most. This trend is observed for other metal-sulfur centers as well and the relative ratios of isotropic shifts for a variety of complexes is presented in Table VI. The analogous isotropic shifts for the [FeWSu(SPh)2]2- complex occur at +UU.3 ppm, ca. +37 ppm, and -30.1 ppm. The broad resonance at ca. +37 ppm is again attributable to the QEEDQ protons. As with the molybdenum complex, replacement of the para_proton with a methyl group results in the appearance of a new resonance at -U6.2 ppm and the disappearance of the +AU.3 ppm resonance, again suggesting that dominant contact interactions are operative. 75 In comparing the magnitudes of the shifts of [FeMoSu_ (SR)2J2' to those of [Fewsu(sa)2]2', one finds less than 1-2 ppm difference for any position on the phenyl ring at comparable temperatures. This suggests that the mechanism of spin delocalization responsible for the observed dom- inant contact shifts is relatively localized at iron and is virtually unaffected by substitution of tungsten for molybdenum. This is in contrast to electrochemical and Mossbauer results already discussed, which demonstrate sensitivity to such an interchange. This observation is reasonable when one considers a qualitative approachSu to the delocalization mechanism for contact shifts. The iron atom contains populated d orbitals, most of which are only half-filled. Antiparallel charge—transfer from sulfur to iron would leave unpaired spin on sulfur, which could be delocalized around the phenyl ring via n molecular orbitals: ng‘.~.©._..fs Spin density of one Sign would be placed at the ortho and para positions and by correlation effects, opposite spin would be imparted to the meta and pfmethyl positions. In 76 effect, this mechanism would liken the thiolate ligand to a benzyl radical in its delocalization pattern.5u The isotropically shifted ligand proton resonances are also quite temperature dependent. Plots of the isotropic shift as a function of temperature for the molybdenum and tungsten thiolate dimers are presented in Figures 11 and 12, respectively. The magnitudes of the shifts for all ring positions increases with decreasing temperature. 53 This is in accord with the simple Curie-Weiss behavior of the magnetic susceptibility (X) measured for the molybdenum complex and the direct proportionality between X and the magnitude of the contact term in the net iso- tropic shift expression. In particular, no evidence for deviations from Curie-Weiss behavior at high temperatures (m1000C) due to antiferromagnetic coupling, i.e., Fe(III)- Mo(V), are observed in the NMR Spectra. 8. Summar In conclusion, the [FeMSu(SR)2]2- (M = Mo, w; R = ph, peTol) complexes can be readily prepared and converted 2- 5 point, the tetraethylammonium salts are isomorphous and into chloro and S analogs. From a structural stand- nearly isostructural. Their optical and infrared spectra reflect the expected differences based on those properties of the parent tetrathiometallate anions. Magnetic sus- ceptibility measurements and proton NMR results for these 77 Figure 11. Temperature dependence of meH, p—H, and 2‘ CH3 resonances of (EtuN)2[FeMoSu(SPh)2] in CD3SOCD3 (O) and CD3CN (O ), and of (EtuN)2[FeMoSu(S-peTol)2] in CD3CN ([1) solution. 78 _ _ if _ _ T l s _ _ a. 1. o 1 . .3 I. H x e , Io .. .p. u .5. a I Fm. /. ms 1m .. \w n... \p‘ I». .\. m. V 0.01:. I \ (I. / 13. \ // /. L. I\.\ /.//.o / L0 o. _ _ 5_ fl/ 0 o m m. m 9,6 M. 2 _ _/ l... 3 2 a. m w w m AER: 3:5 28:92 79 Figure 12. Temperature dependence of mgH, p—H, and p—CH3 resonances of (EtuN)2[FeWSu(SPh)2] in CD3CN (o) and of (EtuN)2[Fewsu(s-peTol)2l in CD3CN (D ). 80 233 denim 0.858. 60 80 4O 20 -20 -40 T(°C) 81 complexes are nearly independent of metal substitution. Significant differences between M = Mo and W are observed in electrochemical measurements and Mossbauer spectra, and are consistent with greater "ferrous" character in the tungsten complexes due to less electron withdrawal by tungsten relative to molybdenum. B. [FeMSu(OR)2]2- (M = Mo, W; R = Ph,prol) 1. Synthesis and Interconversions The phenolate complexes, (EtuN)2[FeMSu(0R)2], can be synthesized by reaction of (EtuN)2[FeMSuCl2] with the an- hydrous sodium salt of the appropriate phenol. This represents a simple ligand exchange reaction where con— comitant precipitation of sodium chloride from MeCN drives the reaction to completion in a short time. Attempts to prepare and isolate the molybdenum complex by addition of an excess of PhOH/Et3N to an MeCN solution of [FeMoSu01212- met with failure. Optical spectra of the reaction mixture indicated quantitative formation of the phenolate complex under these conditions, but when nonpolar solvent is intro- duced to precipitate the product, only unreacted (EtuN)2- [FeMoSuCl2] separates out. This result is in contrast to the preparation of the aryl thiol complexes previously discussed, which can be isolated via this latter route. This is most likely related to the greater solubility of 82 the phenolate complex and does not necessarily suggest that thiolate is a much better ligand than phenolate for iron (II) in these dimeric complexes. Unlike their thiolate analogs, the tetraethylammonium salts of [FeMSu(OPh)2]2— are isolated as 1:1 MeCN solvates when saturated solutions are allowed to cool very slowly. The presence of solvent molecules has been established by elemental analysis, X-ray crystallography, and NMR tech- niques, and suggests that the phenolate complexes may not be isostructural with their thiolate analogs. The complexes can be obtained in unsolvated form by rapid precipitation from MeCN with THF or EtZO. Once prepared and isolated, the phenolate complexes are quite stable in the absence of oxygen and moisture. They demonstrate interconversion chemistry analogous to the thiolate complexes already discussed. For example, treatment with two equivalents of PhCOCl in MeCN results in a rapid conversion to the corresponding dimeric chloro complex. Presumably, this occurs via the same mechanism that has been documented for the FeuSu and Fe282 thiolate complexes,29 i.e., benzoylation of the coordinated ligand. Reaction of the phenolate complexes with a slight ex- cess of PhSH in MeCN results in quantitative formation of the corresponding thiolate species. This can be viewed as simple protonation of the phenolate ligand by the stronger acid, PhSH. The overall reactivity of the 83 2 a i l n 8H [FeMSu(OAr)2]2- ArSH NaOAr PhCOCl Et3N/ArSH [FeMSuCl2J2- 2-«e— [FeMSu(SAr)2] :, PhCOCl 0 A _ _ 2- [Fe(SAr)u]2 + MSu2 FeCl2 + MS“ Figure 13 85 phenolate complexes is illustrated in Figure 13. It is apparent that the phenolate dimers are analogous to their thiolate counterparts in that they possess reason- able stability and undergo reactivity with retention of the FeSZM core. 2. Structure The crystal structure (EtuN)2[FeMoSu(OPh)2]'MeCN has been solved and is illustrated in Figure 1“. Important bond distances and angles are listed in Table V. The struc- ture contains discrete cations, anions, and MeCN molecules in a 2:1:1 ratio. The structure is not disordered in any way, but due to near perfect alignment of the solvent mole- cules on a major crystal symmetry axis, the cations possess a pseudosymmetry that is different from that of the anions. This complication has resulted in a current refinement factor of 1U%. Use of anisotropic thermal parameters for all atoms and non-standard refinement techniques are cur- rently being employed to resolve this problem. Unfortun— ately, this difficulty prevents any meaningful comparisons with the thiolate analog structure, which has been more ex- tensively refined. At present, nearly all core bond distances in the phenolate complex structure are m0.03 A longer than corresponding ones in the thiolate Species. In view of the fact that bond distances usually shorten as the level of refinement increases, this discrepancy 86 Figure 1A. X-ray crystal structure of (EtAN)2[FeMOSu’ (OPh)2]'MeCN. Ellipsoids represent fifty percent probability. 87 Figure 1A 88 Table V. Selected Interatomic Distances (A) and Angles (deg) for (EtuN)2[FeMoSu(OPh)2]'MeCN. Distances Mo-Fe 2.797 Fe-S2 2.308 MO-Sl 2.279 Fe-05 1.861 MO-S2 2.266 Fe—O6 1.932 Mo-S3 2.107 05-01 1.362 Mo-SU 2.1M“ O6-C7 1.206 Fe—Sl 2.273 Angles Mo-Sl—Fe 75.81 Sl-Fe-S2 103.83 Mo—S2-Fe 75.38 Sl-Fe-05 113.23 Sl-Mo-S2 104.98 Sl-Fe-O6 111.8u Sl-Mo-S3 108.53 s2-Fe—05 111.82 Sl—MO-SU 109.69 S2—Fe-O6 114.0“ S2-Mo-S3 110.71 05—Fe—O6 102.93 S2—MO-SU 108.91 Fe-OS-Cl 133.09 S3-Mo-Su 113.70 Fe-O6-C7 135.U3 89 may become negligible. Certain gross structural differences that will not be altered Significantly by further refinement can be men- tioned. The most obvious of these are the relative orien- tations of the phenyl rings in either structure. Since both structures have been portrayed using nearly the same perspective, one can see that the phenyl rings are nearly coplanar in [FeMoSu(OPh)2]2_, but close to perpendicular in [FeMoSu(SPh)2]2-. Analysis of the iron-oxygen distances is difficult due to the lack of structurally characterized tetrahedral iron(II)-oxygen complexes and the relatively low level of refinement in the structure. The current average Fe-O distance is 1.90 A, which is m0.l A shorter than the average terminal Fe—O distance55 in [Fe(CSH7o2)2]2, The correspond- 56 ing average distance in [Fe(C5H7O2)3] is 1.95 A. As the Fe-O distance in [FeMoSu(OPh)2]2- will undoubtedly shorten slightly upon further refinement, it may indicate a rather strong interaction between iron and oxygen in these complexes. Further evidence of this is presented in the following sections. 3. Optical and Infrared Spectra The optical spectra of the [FeMSu(OPh)2]2- (M = Mo, W) are shown in Figure 15, while significant data are presented in Table III. As with their thiolate analogs, 90 Figure 15. Electronic spectra of (EtuN)2[FeMSu(OPh)2] (M = Mo, W) in acetonitrile solution at 23°C. 91 000.0 .. é 000.0N 1 T1 Lem 2, «mouNSeé Alvtflamosumoueaea; 92 these complexes demonstrate intense absorptions between 300 and 500 nm. This result is again attributed to splitting of degenerate sulfur + molybdenum or tungsten charge-transfer transitions upon coordination to iron. Also, the optical spectrum of [FeWSu(OPh)212_ is shifted to higher energy relative to [FeMoSu(OPh)2]2-, analogous to the trend ob— served for the thiolate complexes, and presumably occurs for reasons already discussed in Section III.A.3. Below 300 nm, both complexes demonstrate very intense absorptions that are undoubtedly due, in part, to ligand n + w* transitions. It was initially hoped that spectra of these complexes might aid in assigning the terminal ligand-+ iron charge— transfer transition(s), Since the only physical perturba— tion is substitution of oxygen for sulfur. The spectra exhibit a variety of differences with respect to those of the thiolate analogs, preventing such qualitative analysis. The most obvious of these is that the spectra are blue- shifted relative to those of the corresponding thiolate complexes. These results are similar to those observed upon substitution of chloride for thiolate; they seem to correlate to the relative electronegativities of the ligand atoms. Unlike the thiolate complexes, inductive effects observed when peTolO- is substituted for PhO' cause sig- nificant perturbation (m5-10 nm) of the absorption maxima. Broad, asymmetrical absorptions are observed in the near- 1 infrared region at ca. 750 nm (e = 170 M- cm-l) and 93 1 780 (e = 1A0 M— cm—l) for [FeMoSu(oph)2]2- and [FeWSu— OPh)2]2-, respectively. The Spectrum of the molybdenum complex is shown in Figure 5. These near-infrared spectra are blue-shifted by more than 100 nm from the correspond- ing absorpances recorded for the thiolate analogs. For 5 5 57 comparison, the T2g + Eg d-d transition + hedral [Fe(OH2)6]2 occurs at ca. 1000 nm. Owing to for octa— the greater magnitude of octahedral ye tetrahedral crystal field splitting energies, it is doubtful that these higher— energy, near-IR absorptions of the phenolate complexes can 5E + 5T d-d transi- 2 tions. Intervalence charge-transfer between metal centers, be satisfactorally explained by simple as discussed in Section III.A.3. could account for these absorptions, but such an assignment will have to await further characterization of known tetrathiometallate com- plexes. The phenolate dimers represent the first synthetic ex- amples of molybdenum-iron—sulfur compounds containing oxygen ligation on iron. Because of this, direct comparisons to other similar systems are not possible. Fortunately, pre- liminary results58 indicate that oxygen analogs of the one, two, and four iron-sulfur centers possess reasonable stability and are synthetically accessible. In particular, spectralmeasurements on [Fe(OPh)u]2— and [F8282(0Ph)u]2- should prove useful in interpreting optical data for the molybdenum and tungsten phenolate dimers. 9“ Infrared Spectra for both [FeMoSu(OPh)2]2- and [FeWSu- (OPh)2]2— between 250—600 cm.1 show splitting of the single metal—sulfur stretch observed for the parent tetrathio metallate into components at slightly higher and lower energies upon coordination to iron, analogous to spectra obtained for the thiolate complexes. A. Electrochemistpy Electrochemical measurements on the complexes, (EtuN)2- [FeMSu(OR)é](M = Mo, W; R = Ph, peTol), have been performed by using dc polarography and cyclic voltammetry. All pertinent information is given in Table IV, along with that obtained for the thiolate analogs obtained under identical conditions. As mentioned before, results obtained by using polarography with a dropping mercury electrode were superior to those acquired using cyclic voltammetry. Electrochemical measurements on the phenolate complexes from 0.0 V to —2.0 V (ye SCE) indicate only one electro- chemically irreversible reduction. The irreversible nature of this Single reduction is based on Slopes (72—77 mV) derived from plots of log [i/(iD-i)] XE voltage. These values are Significantly greater than the theoretical value of 59 mV for a one-electron reversible reduction. The reduction potentials of the molybdenum and tung- sten phenolate dimers are -l.52 V and -l.68 V, respectively. As with the thiolate analogs, the tungsten complex is more 95 difficult to reduce (by N160 mV) than its molybdenum counter- part. This result agrees with the observed trend that tung- sten complexes generally possess more negative reduction potentials than corresponding molybdenum complexes. As mentioned, the reduction potentials of the pheno— late dimers are significantly more negative (N250 mV) than those of the corresponding thiolate complexes. These 17 for the results are in contrast to those obtained [FeMoSuC12J2- ion, which suggests that substitution of a more electronegative ligand atom, i.e., chlorine, results in a complex that is easier to reduce (mlO mV). However, d59 for the they are in agreement with those obtaine [FeuSu(OPh)u]2- complex, which indicates that phenolate ligation to iron in these metal-sulfur systems results in a complex that is more difficult to reduce, relative to its thiolate analog. This trend implies that oxygen ligation tends to stabilize the oxidized form of these complexes to a greater extent than sulfur. Massbauer and NMR results, to be discussed later, reinforce the observed trend in electrochemistry. Furthermore, greater average negative shifts in reduc— tion potential are observed when prolO- replaces PhO", than when prTolS- is substituted for PhS- (N50 mV ye m15 mV). The direction of the shift can be explained as before by a simple inductive effect of the electron— releasing prmethyl group. 96 5. Magnetic Susceptibilipy The room temperature magnetic susceptibilities of the complexes, (EtuN)2[FeMSu(OPh)21'MeCN (M = Mo, W), have been measured by using the Faraday method. Data obtained have been corrected for diamagnetic contributions due to ligands, cations, and solvent molecules using Pascal's constants. These data appear in Table IV, along with that for the thiolate complexes. The experimental values of the mag- netic moments for the molybdenum and tungsten phenolate dimers are 5.08 BM and 5.07 BM, respectively, and are in accord with an S = 2 ground state, i.e., four unpaired electrons. They also compare well with the values ob- tained for the thiolate analogs, within experimental un- certainty. Again, the question of a high-spin Fe(III)- MO(V) or a high-spin Fe(II)-Mo(VI) interaction arises. Mossbauer results, to be discussed next, suggest that the latter is a better description of the formal oxidation states. 6. M6ssbauer Spectra The zero-field Massbauer spectrum of a polycrystal- line sample of (EtuN)2[FeMoSu(OPh)2]'MeCN diluted with boron nitride has been recorded at ”.2 K. The spectrum is shown in Figure 16 and relevant parameters are listed in Table IV. The observed isomer shift is 0.57 mm/S and Figure 16. 97 MESSbauer spectrum of (EtuN)2[FeMoSu(OPh)2]- MeCN in the solid state diluted with boron nitride at A.2 K with zero applied field. Vertical bar indicates one percent absorp— tion. 98 m\s_s_ z. >._._00:_w> itai iii-truifiz‘.’ - i. Iii-233.1511... l........=..l.1..r.t=. p 99 the quadrupole splitting is 1.20 mm/s. The asymmetric nature of the quadrupole doublet is probably due to a slightly non-random distribution of crystal orientations in the sample. To date, no solution M6ssbauer data on [FeMoSu(OPh)2]2— have been obtained. Also, none are avail- able for the tungsten analog, thus preventing comparisons to the molybdenum complex. Fortunately, solid sample data for [FeMoSu(SPh)2]2- have been obtained. The measured isomer shift is 0.A7 mm/s and quadrupole splitting is 1.73 mm/s. The isomer shift for the phenolate complex, 0.57 mm/s, is significantly higher than that of the thiolate analog, and, together with the change in quadrupole splitting (0.53 mm/s), reflects a significant rearrangement of charge around the iron nu- cleus in the phenolate complex. Comparing the observed isomer shift for the phenolate dimer with tabulated values for iron in various tetra- 15a yields a formal oxidation hedral sulfur environments state on iron of +2.3 (see Figure 9). This represents a significant increase in ferrous character (0.10 mm/s) for the phenolate complex relative to its thiolate analog, and certainly would not be predicted by simple electro- negativity arguments. These results parallel those ob- tained17 for the [FeMoSuCl2J2- ion, which yields an isomer shift of 0.60 mm/S- For comparison, substitution of phenolate for thiolate on the FeuSLI core increase559 the 100 isomer shift by ca. 0.1 mm/s. The higher ferrous character of the iron in the pheno- late dimer is in accord with observed electrochemical results already discussed. Together, they suggest a significant increase in the d electron density around the iron nucleus which has the net effect of screening the nucleus from existing s electron density, thereby in- creasing the isomer shift. One possible explanation for the above phenomenon would be a notable degree of co- valency in the iron-oxygen interaction. Proton NMR results (Vida infra) also support such an interpretation. Based on results obtained when tungsten is substituted for molybdenum in the thiolate dimers, the isomer shift of the [FeWSu(OPh)2]2- complex should be mo.6 mm/s. Veri_ fication of this prediction will have to await future meas— urements. 7. Nuclear Megnetic Resonance Spectra Proton NMR spectra of the complexes, (EtuN)2[FeMSA’ (OR)2]°MeCN (M = Mo, W; R = Ph, prol), have been recorded as a function of temperature in CD3CN; isotropic chemical shifts are observed for the ligand protons. Typical spectra for [FeMoSu(OR)2]2- complexes are presented in Figure 17 and the isotropic shifts at room temperature are listed in Table III. Plots of chemical shift ye tempera- ture for both the molybdenum and tungsten phenolate dimers Figure 17. 101 NMR spectra of (EtuN)2[FeMoSu(OPh)2] in CD3CN at 50°C (A) and of (EtAN)2[FeMOSA‘ (O-p—Tol)2] in CD3CN at 50°C (B). Small amounts of impurity are denoted by X. Chemical shifts are in parts per million from MeuSi. 102 I '3'" 2'" -46|.. 34' l3. 0 4 B Figure 17 103 appear in Figures 18 and 19, respectively. As before, those shifts downfield of tetramethylsilane are negative in Sign; those upfield are positive. Likewise, all Shifts are reported relative to their diamagnetic references. At 297 K, the measured isotropic shifts for [FeMoSu- (OPh)2]2_ are +A7.6 ppm, ca. +A5 ppm, and -u3.5 ppm. The resonance at +U5 ppm is very broad and has been assigned to the ppre hydrogens for reasons already discussed in Section III.A.7. Upon substitution of a methyl group in the peee position, the resonance at +A7.6 ppm disappears and a new resonance at three times the intensity appears downfield at -50.2 ppm (at 293 K). This effect is again indicative of a dominant contact term in the net iso- tropic shift expression, similar to results observed for the thiolate complexes. Magnitudes of isotropic shifts, relative to those of the meEe protons, are listed in Table VI. The general trend of peye > 23229 > meEe is obeyed, but the relative shifts for 23229 and peye are less than corresponding values for [FeMoSu(sph)2]2-. This effect is due to a greater increase in the meee shift (W12 ppm), relative to 23322 and peye (m6 and N1 ppm, reSpec- tively), and may represent an enhanced contribution by the dipolar term in the opposite direction, due to closer proximity to the paranagnetic center. Similar results are noted for the [FeWSu(oPh)2]2- ion, whose isotropic shifts at 293 K are +A5.3 ppm, ca. +42 ppm, 10H Figure 18. Temperature dependence of me, peH, and p-CH3 resonances of (EtuN)2[FeMoSu(OPh)2] in CD3CN (O) and of (EtAN)2[FeMOSu(O-E- Tol)21 in CD3CN (0). 105 b . if _ Au .0 nu Ru - nu 4 w . .52: mtim 0.858. _ nu 4 -50 .— -60 ' -7C)P -20 20 4O 60 80 T(°C) -40 Figure 18 106 Figure 19. Temperature dependence of EfH: p—H, and p—CH3 resonances of (EtuN)2[FeWSu(OPh)2] in CD3CN (O) and of (EtuN)2[FeWSu(O-p-Tol)2] in CD3CN (D). 107 60" . 0 :u _ nu 4 1’7 n.u 1.. O . .52.. «Kim 2.1.058. _ o .4 -6&)-. -7o - -20 20 4O 60 80 TN” -40 Figure 19 108 Table VI. Comparison of Relative Isotropic Shiftsa for Various Metal-Sulfur Centers in CD CN Solution. 3 Complex ortho meta para 2... [FeMoSu(SPh)2] 1.25 1.00 l.A9 [FeMoSu(0Ph)2]2- 1.03 1.00 1.09 2- [Fe2S2(SPh)u] 1.09 1.00 1.81 _c [Fe 3 (SPh) )2 1.3A 1.00 1.96 A A A d [Fe S (OPh) 12' 1.0A 1.00 1.29 A A A e -f,h [Fe6Mo288(SPh)9]3 g 1.00 1.73 aRelative to meee shift; 22-2500. bReference 67. CReference 69. dReference 59. 8Reference 68. fReference 70. gObscured. solution. 109 and -A2.7 ppm for the para, ortho, and meta protons, respectively. The resonance for the pemethyl protons shifts downfield to -A8.5 ppm, again demonstrating the typical pattern for a dominant contact interaction. Little change is noted in the magnitude of the isotropic shifts at all ring positions when tungsten is substituted for molybdenum, again suggesting that the spin delocalization mechanism is not sensitive to changes at the other end of the molecule. Qualitatively, one might expect significant changes in the absolute magnitudes of corresponding chemical shifts when thiolate is replaced by phenolate. Experimentally, the shifts for the phenolate complexes are significantly larger than those of the thiolate complexes, particularly for the eypEe and meEe positions (A5 ye 39 ppm and -A3.5 ye -3l.2 ppm, respectively, for the molybdenum complexes). This occurs for the peye resonance as well, but to a lesser extent (A7.6 ye A6.6 ppm). These data are similar to NMR results59 obtained for [FeASA(OPh)A]2- ye [FeuSu(SPh)u]2-, and clearly indicates more delocalization of unpaired spin density when phenolate replaces thiolate as a ligand on iron. The actual explanation for this effect may be due to a combination of several factors: (1) the anomalously short iron-oxygen distance; (ii) some degree of covalency in the iron-oxygen interaction; (iii) the improved orbital energy match between oxygen and carbon ye sulfur and 110 carbon, which should enhance the delocalization of unpaired spin density into the remainder of the phenyl ring. The temperature dependence of the observed isotropic shifts parallels that obtained for the thiolate analogs. The relative magnitudes of all shifts increases with de- creasing temperature. These data are indicative of simple Curie-Weiss behavior of the magnetic susceptibility, similar to that observed for (EtuN)2[FeMoSu(SPh)21- Quantitative treatment of this will have to await the results of magnetic susceptibility measurements as a function of temperature. 8. Summary Both the molybdenum and tungsten phenolate dimers can be readily prepared and isolated in crystalline form as their tetraethylammonium salts. They can be readily con- verted into the thiolate and chloro analogs by treatment with PhSH and PhCOCl, respectively. Dissimilarities in optical and infrared spectra can be related to analogous differences in these properties for the free tetrathio- metallates. Magnetic susceptibility and proton NMR results are essentially insensitive to replacement of molybdenum by tungsten. Electrochemical measurements show a more negative reduction potential for the tungsten dimer, in accord with the suggested reasons in Section III.B.A. 111 C. Comparisons Between Phenolate and ThiOlate Ligation Both classes of dinuclear complexes are most readily accessible by ligand exchange reactions via combination of the [FeMSuCl212— (M = Mo, W) ion with the anhydrous sodium salt of the appropriate ligand. The [FeMSu(0Ph)2]2- (M = Mo, W) species can be isolated as a 1:1 MeCN solvate when allowed to crystallize slowly. Once obtained as their tetraethylammonium salts, all complexes react slowly with air and moisture in the solid form, but are much more sensitive to these reagents in solution. Both the thiolate and phenolate dimers react with two equivalents of PhCOCl in MeCN to yield the [FeMSuCl212- complex in quantitative fashion. Direct structural comparisons between the [FeMoSu— SPh)2]2_ and [FeMoSu(OPh)2]2— ions are difficult due to the dissimilarity in their respective refinement factors. Average FeS2MS2 core bond distances for the phenolate com— plex are currently ca. 0.03 A larger than corresponding distances in the thiolate complex. The only major dif- ference between the two structures is the relative orien— tation of the phenyl rings. The optical Spectra of the phenolate complexes are blue- shifted and bear little resemblance to those of the thiolate analogs, thus preventing direct assignment of the terminal ligand + iron charge-transfer transition(s). All com- plexes demonstrate low intensity bands between 750 nm and 112 1200 nm, which are most likely ascribable to intervalence charge-transfer transitions. Room temperature magnetic susceptibility measurements on all complexes are consistent with an S = 2 ground state. Proton NMR results indicate simple Curie-Weiss magnetic behavior down to 233 K for the phenolate and thiolate dimers. The following trends are observed for Mossbauer param- eters, electrochemical measurements, and proton NMR results, respectively: (1) The isomer shift is larger for the phenolate analog, indicating that the iron atom in that complex more closely resembles iron(II). (ii) The phenolate complexes are more difficult to reduce than their thiolate analogs. (iii) Isotropic shifts for the phenolate complexes are significantly larger than corresponding shifts for the thiolate species. In conclusion, these latter comparisons demonstrate that a simple electrostatic model of these systems based on the relative electronegativities of sulfur and oxygen is not adequate and that the iron-oxygen interaction possesses notable covalent character. 113 13' D. [Fe2MoSll 1. Synthesis The trinuclear species, (EtAN)3[Fe2MOSllJ’Can be pre- pared by reaction of (EtuN)2MoSu with (EtuN)2[Fe2812] in MeCN. Unlike the syntheses of the dimeric complexes already discussed, which involves ligand displacement, the reaction to form [Fe2MoSllj3- involves an oxidation—reduc— tion process. In the course of the reaction, the FeS2Fe core is reduced by one electron, accompanied by presumed oxidation of the displaced SE- ligand. This is analogous21 to the reaction sequence yielding the [Fe2MoS6(SPh)213- complex ion from a mixture of M083. and [Fe2SZ(SPh)u]2- in MeCN. The reaction to form [Fe2MosllJ3- proceeds at a substantially faster rate than the reaction that yields [Fe2MoS6(SPh)2]3_ (3—A d ye 2 wk). In attempting to explain this difference in relative rates, one would cer- tainly have to take into account the relative lability and ease of oxidation of bidentate S§_ _§ two thiophenolate ligands. The fact that reduction of the Fe282 core does occur is in accord with the high affinity of MoSE- for iron(II) rather than iron(III). Once prepared, (EtuN)3[Fe2MoSll] is much more sen- sitive to oxygen and protic solvents than the dimeric complexes. This behavior seems reasonable because the complex can be likened to reduced FeSZFe core 11A stabilized by MoSi-. To date, the reduced dimeric Species, [Fe282(SR)u]3_, has only been generated in solution60 and is extremely susceptible to oxidation, analogous to the reduced tetrameric complex,61 [FeASA(SPh)A]3-' 3.. ll] , by an identical procedure failed. The optical spectrum of Attempts to prepare the tungsten analog, [Fe2WS the reaction mixture after ca. two weeks at room tempera- ture was that of the unreacted starting materials. This parallels the observation62 that [Fe2WS6(SPh)2]3_ cannot 12- and WSE'. be prepared from [Fe2S2(SPh)u A comment should be made concerning the preparation of the starting material, (EtuN)2[Fe2812]. The published pro- cedure3u involves reaction of [Fe(SPh)u]2- with five equiva- lents of benzyl trisulfide. It was discovered that this complex can be easily prepared by reaction of less sophis- ticated reagents, namely ferric chloride, (EtuN)(SPh), and excess elemental sulfur. Indeed, any Fe-S or Fe—Mo-S synthesis reported to date involving benzyl trisulfide can also be carried out with an excess of sulfur. 2. Optical and Infrared Spectra The optical spectrum of (EtuN)3[Fe2MoSll] in MeCN has been obtained and is shown in Figure 20. Intense absorptions between 700 and 300 nm are observed, which is consistent with results on the other tetrathiometallate systems already discussed, and again reflects splitting of the A75 nm band 115 Figure 20. Electronic spectrum of (EtuN)3[Fe2Mosll] in acetonitrile solution at 23°C. 116 A85 x 08 com CON 08 00m 00.? 00m .1 a d A d _ - 1 000.m 1 000.0_ 57.57% w - 80.9 117 of free M083- into bands at higher and lower energy as the degeneracy of the sulfide ligands is broken by coordination to another metal center. Other features in addition to S + Mo charge-transfer absorptions are observed in the Spectrum, and undoubtedly arise from S + Fe charge-transfer, as well as internal polysulfide ligand transitions. Maxima in the spectrum of [Fe2MoSll]3- occur at nearly the same positions as those21 of the thiophenolate analog, [FezMoS6(SPh)2]3—. Indeed, the optical spectrum of the former complex resembles that of the latter set on a gradually rising absorbance. This phenomenon is exactly analogous to that observedlga with [FeMong2' and [FeMoSu(SPh)212-. From these two cases, it would appear that little perturbation of the electronic energy levels within the metal-sulfur core occurs when bidentate SE- replaces two thiophenolate ligands. This is in contrast, 3A however, to the observation that the optical spectrum of [Fe2812]2_ differs markedly from that of its thiolate analog,28 [Fe282(SPh)u]2-. The Spectrum of the former complex has absorption maxima at 370 and AA5 nm, while the latter complex displays maxima at 270, 333, and A90 nm. The infrared spectrum of (EtAN)3[Fe2MOSll] exhibits the splitting pattern typical of tetrathiomolybdate coordinated to a first-row transition metal. Absorbances assignable to molybdenum-sulfur stretches are observed at A97, A80, and AA2 cm_l, compared with the single band for free MOSE- at A72 cm"l 118 3. Electrochemistry The electrochemical behavior of (EtAN)3[Fe2M0811] in MeCN solution has been examined between 0.0 and -2.0 V by using dc polarography with a dropping mercury elec— trode. Only one obvious reduction process occurs within this range, with a half-wave potential of —l.25 V ye SCE. The current associated with this process is 2.7 uA per mM iron and agrees well with the measured average value for the dimeric complexes already discussed (2.8 uA per mM Fe). The polarographic wave is distorted at its maximum and beyond by what appears to be another ill—defined, multi- step reduction process. This process could correspond to further reduction of the metal-sulfur core, but more likely to adsorption or reduction of the polysulfide ligand. This hypothesis depends on the observation28 that the [Fe282(SR)u]2_ complexes demonstrate two electrochemical reductions, while the [Fe281232- ion ShOWS3u no well-defined reduction characteristics, possibly for the same reason, i.e., polysulfide ligand reduction. Plots of log [i/(iD-i)] ye voltage yield an average lepe of A8 mV, which is substantially lower than the theoretical value of 59 mV for an electrochemically reversible one—electron process. It is unlikely, how- ever, that the observed reduction process is actually reversible in nature, since the polarographic wave 119 is severely distorted near the maximum, which would cer- tainly affect the calculated value of the slope. Also, the thiOphenolate analog, [Fe2MoS6(SPh)2]3', exhibits only one electrochemically irreversible reduction at —l.l V ye SCE. A. Magnetic Susceptibility The room-temperature magnetic susceptibility of (EtuN)3- [Fe2MoSll] has been determined by the Faraday method. After correcting for diamagnetism associated with the cations and sulfur atoms, the calculated magnetic moment is 2.50 BM per iron atom (3.5A BM per formula unit). This value is higher than the spin-only value of 1.73 BM which might be expected for a system containing high-Spin iron(III) and high-spin iron(II) in a coupled arrangement to yield a net S = 1/2 state. This discrepancy could arise from three sources: (1) a significant amount of orbital contribution to the observed magnetic moment; (ii) some population of higher spin states at room tempera— ture; (iii) a considerable amount of simple paramagnetic (hs Fe3+) impurity. In view of the results obtained for the iron-sulfur centers, the second alternative seems most likely. The room-temperature magnetic moment for the anti- ferromagnetically coupled system, [FeuSu(SPh)u]2-, is 1.08 BM per iron atom.61 Upon reduction by one electron, the same complex yields a magnetic moment of 2.18 BM per iron. 120 The coupled dimeric complex, [Fe2S2(SPh)u]2', has a magnetic moment of l.AO BM per iron (1.98 BM per mole) at lA°C,28 demonstrating some population of a higher spin state at 1). this temperature (J m150 cm- Since the complex, [FezMoSll]3-, is essentially a reduced FeS2Fe core stabi- lized by MoSE', one might expect its effective magnetic moment per iron to be higher than l.AO BM (1.98 BM per mole). In contrast, the effective magnetic moment of (EtuN)3- [Fe2MoS6(SPh)2] is 1.6 BM per iron atom (2.2 BM per 62 mole). This value is 0.9 BM per iron (1.3 BM per mole) lower than that obtained for (EtuN)3[Fe2Mos ]. This dis- ll crepancy could arise from intermolecular interactions at room temperature similar to the magnetic behavior of a solid sample of (EtuN)2[FeMoSu(SPh)2] at low temperature. Also, the proton NMR spectra of [Fe2M086(SPh)2]3- Show isotropic shifts for the ligand protons that decrease with increasing temperature (25 -75°C), which is opposite in behavior to 28 similar measurements on [Fe282(SPh)u]2—. If the Fe Mo 2 unit is an antiferromagnetically coupled system, the NMR data imply that its Neel temperature is lower than room temperature, unlike the FeS2Fe unit, whose magnetic moment and isotrOpic shifts are still increasing at room tempera- ture. It is clear that in order to quantitatively assess the magnetic behavior of these trimeric complexes in relation 121 to observed results for the FeSZFe dimer, both solid and solution temperature-dependent susceptibility data will have to be obtained, inrnfllshould be straightforward once the recently-purchased SQUID susceptometer is operational. 5. M6ssbauer Spectrum The Mossbauer spectrum of a polycrystalline sample of (EtAN)3[Fe2M0811] diluted with boron nitride was obtained at A.2 K in a weak applied magnetic field of 600 G. This spectrum is shown in Figure 21. Two Slightly asymmetrical quadrupole doublets are observed instead of a magnetic spectrum, which implies that the complex experiences rapid electronic relaxation at A.2 K. The outer quadrupole doublet, iron A, has an isomer shift of 0.AA mm/S and a quadrupole Splitting of l.Al mm/s- These parameters are similar to those for the iron site in [FeMoSu(SPh)2]2- (IS = 0.A7; QS = 1.63 mm/s), and correspond to a formal oxidation state of +2.7 for iron A. In view of this, it seems logi- cal that iron A is that coordinated to tetrathiomolybdate. These parameters also agree well with those21 of one iron site in [Fe2MoS6(spn)213‘ (IS = 0.A2; os = 1.u1 mm/s). The second site, iron B, possesses an isomer shift of 0.32 mm/s and a quadrupole splitting of 0.92 mm/s. This isomer shift corresponds to a formal oxidation state of +2.9 for iron B (see Figure 9), and again compares well with the second iron site in [Fe2MoS6(SPh)2]3— 122 Figure 21. Massbauer spectrum of (EtuN)3[Fe2MoSll] in the solid state diluted with boron nitride at A.2 K in a 600 G applied field. Vertical bar indicates one percent absorption. 123 «£22 2. >._._00._m> V m N . 0 T N- n: c... - _ d — - _ fl _ .3. ’— __k . . _. .3 ._ . . a _ . . _ ._ . _ . _ _ _ _ ._ _ _ _ _ _ __ :5... _. . _. H ..__ __ ______ J... .__.__.__.__. __.._._._._..._ 3:1. :_.}___.2__._.S.{=_a i..ii§::!?£.__ {3.1.1 5... p b n b — p n - 12A (IS = 0.30 mm/s). The quadrupole splitting observed for iron B is 0.2A mm/s greater than that measured for the corresponding iron site in the thiolate analog, and sug- gests a high degree of asymmetry in the electric field gradient experienced by the iron nucleus at site B in [Fe2MosllJ3-. This effect is probably a consequence of the difference in terminal ligands on iron B, i.e., Sg' ye thiolate. Unfortunately, the Méssbauer parameters of [Fe2Sl2J2- have not been reported, thus preventing a direct comparison to its thiolate analog, [Fe282(SPh)u]2- (IS = 0.17; QS = 0.32 mm/s). It has only been stated that 3A they are "similar". The key point is the resolution of two distinct iron sites for [Fe2MoS J3- by Méssbauer spectroscopy, which 11 offers further evidence for a core structure analogous to that62 of [Fe2MoS6(S-p-Tol)2]3-. 6. Electron Paramagnetic Resonance Spectra A frozen solution of (EtuN)3[Fe2MoSll] in DMA has been examined by ESR Spectroscopy with an applied microwave frequency of 9.089 GHz and a power of 10 mW. At 10 K, an intense, isotropic signal is observed at g = 2.00A3, and is analogous to the observed21 g value of 2.005 for the [FeZMoS6(SPh)2]3_ complex. Exact reasons for the isotropic nature of the signal at g = 2.00A3 are unclear but may result from a high 125 degree of delocalization of the unpaired spin density over the metal centers, which would have the effect of decreasing the spin—orbit coupling constant. It can be said that the unpaired spin is not delocalized onto the molybdenum atom to any great extent, since no perturbations due to 95Mo (1 = 5/2) and 97Mo (1 = 5/2) nuclear hyperfine coupling are observed in the spectra. 7. Conclusions 1113-, can be prepared 2— by direct reaction of [Fe2812j2- and M08“ The polysulfide complex [Fe2MoS and isolated as the tetraethylammonium salt. The air-sensitive complex demonstrates optical and infrared properties in accord with bidentate tetrathiomolybdate coordination to iron. M6ssbauer, ESR, and magnetic susceptibility results are qualitatively similar to those of the [Fe2MoS6(SPh)213- complex, and in conjunction with elemental analysis data, support the proposed structure: \ _=SFe/\S/Fe\s/M°\S L_ .- 126 E. Reactions of [FeMoSu(SPh)2]2- l. Acenaphthalene Radical Anion Previous electrochemical results demonstrated the oc- currence of an irreversible reduction process for the [FeMoSu(SPh)212- anion. In order to examine this phenomenon on a larger scale, several attempts were made to reduce [FeMoSu(SPh)2]2- in DMA with one to three equivalents of sodium or lithium acenaphthalenide in THF. In all cases, tfluaproduct was a black, amorphous solid. All attempts at crystallization from a variety of solvent systems failed. The product was only soluble in high dielectric solvents such as DMSO, NMF, and to a Slight extent, DMA. Complete dissolution of the product in these solvents was not pos— sible indicating that some decomposition had taken place. Also, no distinguishable difference in solubility was noted when lithium instead of sodium acenaphthalenide was em- ployed, suggesting that the product was not a salt of either of these metals. The optical Spectrum of the solid in DMA/MeCN contained strong absorbances at 590, 510, 3A0, and 290 nm, strongly resembling the spectrum of the [Fe(MoSu)2]3- ion,2oa and would represent not only reduction, since the [Fe(MoSu)2]3— complex formally contains either molybdenum (V) or iron(I), but also substantial rearrangement of the starting material, [FeMoSu(SPh)212-. Similar results have 127 been observed upon treatment of [FeMoSuC1212- with other potential reductants, such as MeMgCl and NaEtZNCS2 (to be discussed later), as well as (EtuN)2S.62 The product also reacts with an excess of PhSH/Et3N to yield [FeMoSu(SPh)2]2-, as does a genuine sample of [Fe(MoSu)2]3-. Examination of frozen solutions of [FeMoSu(SPh)212_ containing one to three equivalents of sodium acenaphthal- enide by EPR spectroscopy did not show evidence for forma— tion of any distinct half-integer Spin systems. Therefore, chemical and spectroscopic evidence do not support the existence of [FeMoSu(SPh)2]3- as a stable species. 2. Ferric Chloride and Thiophenolate A reaction analogous to that16 leading to formation of [Fe6M0288(SPh)9]3- was performed vfiifll an equimolar amount of [FeMoSu(SPh)2]2- in place of MoSfi'. It was hoped that the product might be a higher order cluster complex con- taining an even higher iron to molybdenum ratio than [Fe6Mo2SB(SPh)9]3—, and thereby be a better approximation to the observed stoichiometry for the nitrogenase co- factor (%6:l). The reaction product was readily purified by recrystallization. Unfortunately, elemental analysis and proton NMR results identified the product as simply (EtuN)3[Fe6Mo2S8(SPh)9]. This reaction is Significant in that it demonstrates that the FeS Mo core of the thiolate 2 dimer complex can be an intermediate in self-assembly 128 F. Reactions of [FeMoSugl212- 1. Methyl Iodide The reactivity of the terminal sulfur atoms on molyb- denum in the [FeMoSuCl2J2- complex has been explored. These sulfur atoms possess some net negative charge and might be expected to be susceptible to electrophilic attack by a reagent such as MeI. Reaction of [FeMoSuCl2]2_ with two equivalents of Mel in MeCN resulted in precipitation of an amorphous brown solid, which demonstrated no solubility in common solvents but which did dissolve in dilute nitric acid. The solid is most likely a simple molybdenum sulfide such as M08 or M082, judging from its solubility properties. 3 The optical spectrum of the filtrate was identical with that reported for [Fe2MoSuClu12-. This latter complex dis- sociates in more polar solvents into FeCl + [FeMoSuC12]2-, 2 and this was observed in the optical Spectrum upon addition of DMF to the solution. Quantitative conversion to [FeMo- suCl212— was accomplished upon addition of M082". Judging from these results, it appears that alkylation of a terminal sulfur atom does occur, but that the result- ing product iS unstable and immediately decomposes into an insoluble molybdenum sulfide and free FeCl2, which can then coordinate to remaining [FeMoSuClZJ2- to yield [Fe2MoSuC1u12—. This latter complex contains no terminal sulfur atoms and should not be as susceptible to attack 129 by Mel as the [FeMoSuC1212- complex. 2. Benzoyl Chloride In addition to the attempted alkylation described above, the reactivity of the terminal sulfur atoms towards acyla— tion was examined. Reaction of [FeMoSuCl2JZ- with two equivalents of PhCOCl in MeCN resulted in immediate pre— cipitation of a dark solid, leaving behind a colorless filtrate. The solid displayed no solubility in common sol- vents, but did dissolve in dilute nitric acid. The solid was most likely M08 or M082, indicating that decomposition 3 had taken place. The reactionvmmsnot investigated further. 3. Methylmagnesium Chloride In view of the successful ligand substitution reactions involving treatment of [FeMoSuCl2J2— with sodium phenolate or thiolate, an attempt to form an organometallic analog was performed. The reaction of ca. three equivalents of MeMgCl in THF with one equivalent of [FeMoSuCl2J2- in HMPA resulted in the formation of a purple—black amorphous solid mixed with magnesium chloride. The dark solid was Slightly soluble in MeCN and gave an optical Spectrum in that solvent identical with that reported20a for [Fe(MoSu)2]3-. This qualitatively suggests that reduction is preferred over simple carbon-iron sigma bond formation in this 130 system. The reductive nature of this reaction is in accord with earlier results (Section III.B.l), which also suggest that [Fe(MoSu)2]3- is the preferred product when iron- molybdenum dimers are exposed to reducing conditions. A. Sodium Benzylthiolate In considering the fact that the FeASA center co- ordinates benzyl, alkyl, and aryl thiols, the reaction between [FeMoSuC1212— and the anhydrous sodium salt of benzyl thiol was attempted. Reaction conditions analogous to those used to prepare the aryl thiolate dimers were employed. Upon stirring a mixture of (EtuN)2[FeMoSuCl21 and two equivalents of NaSCH Pb in MeCN, an amorphous red- 2 brown solid precipitated; its optical spectrum corresponded to that of MoSi_. Subsequent workup of the filtrate re- vealed formation of a small amount of [Fe(MoSu)2]3-. The above results suggest that benzyl thiolate, and probably alkyl thiolates as well, are too reducing in nature to stabilize the FeS2M082 core. This is evidenced by the formation of the reduced species, [Fe(MoSu)2]3_, in small yield. However, attempts to prepare the [FeMoSu- S,S-e-Xy1)]2- complex have apparently been successful.614 These results are in accord with those observed28 for the FeS2Fe core, which is only stabilized by aryl or e- xylyldithiolate ligands. 131 5. Sodium Diethyldithiocarbamate In order to further ascertain the variety of ligands that the FeS2MoS2 core will tolerate, a ligand exchange reaction between [FeMoSuCl212- and two equivalents of the anhydrous sodium salt of diethyldithiocarbamic acid in MeCN was explored. Reaction was immediate, producing a deep violet solution whose optical spectrum contained features at 580, 510, 3A0, and 290 nm, characteristic of [Fe(MoSu)2]3- formation. It is very likely that the reductant in this case is the dithiocarbamate ligand, which can be easily oxidized to a thiuramdisulfide. Indeed, Newton and co— workers65 prepare [Fe(MoSu)2]3- from [Fe(Et2NC82)21 and MoSE-. The above experimental results represent yet another example of formation of [Fe(MoSu)2]3- via exposure of an iron-molybdenum dimer complex to reducing conditions. 0. Reactions of [FeuSuK 32. (X = S—e-Bu, SPh, Cl) l. e—Butyl Thiolate Considering the fact that the FeASA core readily ac- commodates a variety of ligand types, such as aryl and 59 and halides,29 the pos- alkyl thiols,2u aryl phenols, Sibility of coordinating tetrathiomolybdate to the FeuSu center was examined. Optical spectra of an MeCN solution of [Feusu(S‘E‘BU)u]2- and four equivalents of MoSE- in- dicated no reaction had taken place after ca. four days at 132 room temperature. When four equivalents of a proton donor such as trimethylpyridinium hydrochloride was added, however, the reaction mixture immediately became deep red-violet in color. Subsequent workup gave a product whose optical 20a of [Fe(MoSu)213—, spectrum strongly resembled that even though elemental analysis results agreed with the formulation, (MeuN)3n[Fe2Mo2SlO]n. Also, reaction of the product with an excess of PhSH/Et3N in MeCN gave [FeMoSu- (SPh)2]2-, analogous to the results observed for a genuine sample of [Fe(MoSu)2]3-. Likewise, X-ray diffraction H H ’ powder patterns of (EtuN)3n[Fe2Mo2310]n and authentic (EtuN)3[Fe(MoSu)2] were very similar. When a sample of I "(EtuN)3n[Fe2MQ was further purified and reanalyzed, 2SlOJn' the molybdenum-to-iron ratio rose to 1.3. These analytical and chemical data strongly suggest that "(RuN)3n[Fe2M02- S101n", as synthesized here, is really an impure form of (RAN)3[Fe(MoSu)2]. A variety of potential reductants exist in the reaction mixture, such as alkyl thiol, sulfide, and iron(II), which could account for formation of the reduced product, [Fe(MoSu)213-. 2. Thiophenolate The reaction between [FeuSu(SPh)u]2- and four equivalents 2- . of M08“ in MeCN was also examined. In this case, the reac- tion mixture turned deep maroon, and a dark microcrystal- line solid precipitated after being stirred for ca. four 133 days at room temperature. The optical spectrum of the crystalline solid in MeCN corresponded exactly to that reported21 for the trinuclear complex, [Fe2MoS6(SPh)2]3-. This reaction is significant in that it indicates that aryl FeuSu centers fragment in the presence of tetrathio- metallates into Fe2S2 centers, in contrast to the results observed for an alkyl FeuSu center both with and without a proton donor. The explanation for this difference may in- volve electronic effects of aryl ye alkyl thiols towards molybdenum-iron-sulfur cores. To date, the Fe2MOS6 core has not been prepared with alkyl thiol ligands.62 The fragmentation of FeASA to Fe2S2 may be due to the affinity of MoSi- for lower formal oxidation states of iron. The FeASA core is best described66 as a delocalized system containing four iron (+2.5) atoms (rather than a localized one with two iron(II) + two iron(III) atoms). Subsequent coordination by MoSi- may stabilize the ferrous iron sites, thereby disrupting the delocalized core. This perturbation may be enough to initiate fragmentation of the Fens“ center. Like the [Fe2MoSllJ3- ion, [Fe2MoS6- (SPh)2]3- essentially represents a partially reduced FeS2Fe core stabilized by MoSE—. Why the reaction does not con- tinue and form [Fe(MoSu)2]3- may be related to the de— creased reducing ability of aryl ye alkyl thiols towards 28 the metal-sulfur core, and is supported by the fact that only aryl or e—xylyldithiolate complexes of the FeS Fe 2 13A core, which formally contains iron(III), have been prepared to date. 3. Chloride The reaction between [FeASAClAJ2- and MoSfi- in MeCN was carried out in order to determine whether the trinuclear species, [Fe2MoSGC12J3-, was indeed stable, or whether the FeASA core would degrade into the mononuclear iron complex, [Fe(MoSu)2]3—. Optical spectra of a reaction solution con- taining [FeASAClAjz- along with four equivalents of M082- indicated the formation of [Fe(MoSu)2]3- in high yield. Attempts to prepare the [Fe2MoS6C12JB’ complex from reaction of [Fe2MoS6(SPh)2]3- with two equivalents of PhCOCl in MeCN have also failed.62 In View of the results discussed in the preceeding sections, the preferred product obtained upon reaction of various FeASA centers with MoSi- is [Fe(MoSu)2]3_, which represents total degradation of a tetranuclear iron center into a mononuclear one stabilized by MoS§-. The exception is the reaction between [FeuSu(SPh)u]2- and MoSi-, which stOps at a dinuclear iron Species, [F92M036(8Ph)2]3', stabilized by NOSE- and thiophenolate. 135 H. Reactions of [MoSO]2- 7 One of the major handicaps in trying to synthesize new iron—molybdenum—sulfur complexes has been the relative lack of well characterized molybdenum-sulfur starting materials. Therefore, the reactivity of the recently reported26 [MOSQJ2- complex has been examined. In reactions with potential reductants such as PhS— and [Fe(CO)u]2_, the [M08912— complex was simply reduced to MoSfi— in near quantitative yield. In reactions with ferrous chloride and [FeuSu(SPh)u]2-, the [M03912- complex behaved like MoSi’, 1113-’ respectively. The excess of sulfur present in the latter reaction un- yielding [FeMoSuCl2]2- and [Fe2MoS doubtedly prompts formation of [FeZMoS 13- instead of 11 [Fe2MoS6(SPh)213-. Under reaction conditions that would normally give rise to "double cubane" complexes of the type, [Fe6M0288(SR)9]3-, if MoSi- were employed, very