DALTON FULL PAPER J. Chem. Soc. Dalton Trans. 1999 301–310 301 Synthesis and characterization of ReV ReVI and ReVII complexes of the [·2-P2W17O61]102 isomer † Anne Venturelli,a Mark J. Nilges,b Alex Smirnov,b R. L. Belford b and Lynn C. Francesconi *a a Department of Chemistry Hunter College of the City University of New York New York NY 10021 USA b Illinois EPR Research Center College of Medicine University of Illinois-Champaign/Urbana Urbana IL 61801 USA Received 8th July 1998 Accepted 13th November 1998 Rhenium-(V) -(VI) and -(VII) complexes of the [a2-P2W17O61]102 isomer a mono-lacunary derivative of the [a-P2W18O62]62 (Wells–Dawson) ion have been prepared and characterized by multinuclear NMR spectroscopy electrospray mass spectrometry and electron paramagnetic resonance spectroscopy among other techniques.The molecules have the formulation [a2-ReOP2W17O61]n2 where n = 7,6,5 for ReV ReVI and ReVII respectively. 183W NMR spectroscopy for the ReV and ReVII analogs shows that the molecules have Cs symmetry as expected for substitution in the a2 site. Simulations of the X-band and Q-band EPR spectra of the ReVI analog using Cs symmetry allow determination of the g hyperfine and quadrupole coupling tensors. X-Band Q-band and W-band EPR spectroscopy show extreme variations in linewidths due to random strains or distortions of the complex. Introduction The [a2-P2W17O61]102 isomer shown in Fig. 1 may be a useful ligand for stabilizing high valent transition and rare-earth (actinide and lanthanide) metal ions.1 The [a2-P2W17O61]102 isomer is obtained by base degradation of the parent Wells– Dawson molecule [a-P2W18O62]62 (Fig.1); eVectively removal of a [WO]41 unit from a “cap” WO6 polyhedron of the parent [a-P2W18O62]62 results in the defect or lacunary structure of Cs symmetry.1 Well characterized stable complexants may find a number of applications in biochemistry and medicine catalysis as well as separation of high valent metal ions from waste streams. In the medicinal area for example a number of families of polyoxoanions have been shown to interact with enzymes specifically suggesting that these compounds may be valuable for study of the structure and function of enzymes and proteins.2 Polyoxoanions including derivatives of the [a2-P2W17O61]102 isomer have also been noted for anti-viral activity. In the area of catalysis while there are few reports of the catalytic properties of derivatives of the Wells–Dawson molecule “Keggin” derivatives are well known thermal catalysts and recently photocatalysts for oxidative transformation of various organic molecules in solution.3,4 Derivatives of the Keggin and Wells–Dawson molecules appear to be stable under conditions found in catalytic reactions.A useful property of polyoxoanions is their resistance to decomposition by selfoxidation. Complexes can be rendered soluble in aqueous or organic solution by an appropriate choice of counter cation(s). Transition metal derivatives of isomerically pure [a2-P2W17- O61]102 have been subjected to oxidation catalysis studies.5 We are investigating the chemistry of nonoxidizable lacunary tungstate [a2-P2W17O61]102 with rare earth ions 6,7 and technetium and rhenium 8 for a variety of applications.While using rhenium as a “surrogate” for technetium,9 we found that the chemistry of high oxidation state rhenium incorporated into † Supplementary data available cyclic voltammogram of 1; 31P NMR spectra of 1a 3 and 2; 183W NMR spectrum of 3. Available from BLDSC (No. SUP 57462 5 pp.) or the RSC Library. See Instructions for Authors 1999 Issue 1 (http://www.rsc.org/dalton). the [a2-P2W17O61]102 polyoxoanion framework is quite rich similar to that found by Pope for Re incorporated into two other “cage” systems.10,11 In that work complexes of Re(V) Re(VI) and Re(VII) have been isolated for both the Keggin and [W10O32]42 families of compounds establishing that high valent rhenium in oxidation states V VI VII can be stabilized within polyoxoanion frameworks.We add to this database by providing for the first time multinuclear NMR data and electrospray mass spectrometry data for the structural characterization of Re complexes of [a2-P2W17O61]102. The reaction of ReCl6 22 and [a2-P2W17O61]102 has been reported by Charreton and Meunier,12 however neither a detailed procedure nor characterization of the material was presented. The details of the preparation characterization and chemistry of Re(V) Re(VI) and Re(VII) complexes of the [a2-P2W17O61]102 polyoxoanion constitute this report. There have been other reports of the preparation of transition metal complexes of the [a2-P2W17O61]102 isomer where characterization has included multinuclear NMR techniques. For example both 31P and 183W NMR spectroscopy provided proof that transition metal derivatives of mixtures of the a2 and a1 isomers of [P2W17O61]102 can be prepared by base degradation of the parent [a/b-P2W18O62]62.13 Isomerically pure [a2- P2W17O61]102 was used to prepare a variety of transition metal Fig.1 Ball and stick model of A [a-P2W18O62]62 and B the lacunary [a2-P2W17O61]102 ligand. 302 J. Chem. Soc. Dalton Trans. 1999 301–310 derivatives. These compounds characterized by a variety of techniques including both 31P and 183W NMR spectroscopy constitute the first isolated monosubstituted Wells–Dawson complexes free of the a1 isomer as an impurity.14 In another study 2-D 183W NMR spectroscopy was used to completely assign all of the 183W resonances and connectivities of the lacunary [a2-P2W17O61]102 ligand and the VV]] O (d0) derivative [a2-VP2W17O62]72.15 High valent oxorhenium complexes and Re(VII) oxides have experienced interest for their unique redox properties and potential for oxidation catalysts.16–19 There has been increasing interest in methylrhenium trioxide 20 as a potential epoxidation catalyst.We report herein the preparation and isolation of Re complexes of the [a2-P2W17O61]102 isomer and characterization by multinuclear NMR spectroscopy EPR spectroscopy electrospray ionization mass spectrometry among other techniques. In this work we use the early report by Charreton and Meunier 12 as a starting point to prepare the K7[a2-ReVOP2W17O61] species 1 in excellent yield. Oxidation of this species aVords the ReVI 2 and ReVII 3 analogs. Experimental General comments Most common laboratory chemicals were reagent grade purchased from commercial sources and used without further purification.Deionized distilled water was used throughout. Preparative operations unless stated otherwise were carried out under atmospheric conditions. Standard Schlenk techniques under an atmosphere of nitrogen were also utilized. K10[a2- P2W17O61] was prepared according to the method of Finke; 14,21 this is important to insure that the resulting rhenium complexes are isomerically pure. NBu4Br (Fisher) AgO3SCF3 (Aldrich) and H2SO4 (Fisher) were purchased and used without further purification. Elemental analysis was performed by the University of Illinois Microanalytical Laboratory. Negative-ion electrospray mass spectra were recorded on a VG Quattro at the University of Illinois School of Chemical Sciences Mass Spectrometry Laboratory.Electrochemical analyses were accomplished with the use of a Bioanalytical Systems BAS-100 electrochemical analyzer. Collection of NMR data NMR spectra were obtained on a JEOL GX-400 spectrophotometer. 31P NMR spectra at 161.8 MHz were acquired using a broad band decoupler coil of a 5 mm reverse detection probe. 183W NMR spectra recorded at 16.7 MHz utilized a 10 mm low-frequency broad band probe. Typical acquisition parameters for 31P spectra included the following spectral width 10,000 Hz; acquisition time 0.8 s; pulse delay 1.0 s; pulse width 15 ns (508 tip angle). For 183W spectra typical conditions included the following spectral width 10,000 Hz; acquisition time 1.6 s; pulse delay 0.5 s; pulse width 50 ns (458 tip angle). For all spectra the temperature was controlled to ±0.2 8C.31P spectra were referenced to 85% H3PO4 and 183W spectra were referenced to 2.0 M Na2WO4. For both 31P and 183W chemical shifts the convention used is that the more negative chemical shifts denote upfield resonances. Collection of EPR data Electron paramagnetic resonance experiments were performed on Varian E-122 X-band (9.27 GHz) Varian E-15 Q-band (34.63 GHz) and the IERC Mark II W-band (94.43 GHz)22 spectrometers. A microwave power of 2.0 mW was used with a modulation amplitude of 5.0 G for X-band and 10.0 G for Q- and W-band at 100 kHz. Samples were run at 77 K (at X- or Q-band) and 20 K (at W-band) as glasses of 0.01 M 50 50 acetonitrile–toluene solutions. The magnetic field was calibrated with a Varian NMR Gauss meter (X-band) and a Metrolab Teslameter (Q-band and W-band) and the microwave frequency was determined with an EIP frequency meter with extended frequency option (to 100 GHz).Because of the very large magnetic field scans used X-band and Q-band field scans were corrected using a second and third order respectively least squares fit of the field calibration data with a resulting rms field error of 0.4 and 2.1 G respectively. Simulation and fitting of electron paramagnetic resonance spectra used the automated simulation program SIMPIP,23 which is based upon QPOW.24 The spin Hamiltonian is solved by exact diagonalization followed by a fourth-order fieldfrequency perturbation to transform the energy spectrum into a field-swept spectrum. The spin Hamiltonian parameters are varied using the SIMPLEX method to minimize the rms deviation between the experimental and calculated spectrum.25 Because of the large nonlinear eVects in rhenium spectra due to the large hyperfine and nuclear quadrupole interaction the “forbidden,” DMI = ±1 ±2 ±3 ±4 and ±5 transitions were all included.H = {bB?g?S 2bngnB?I} 1 hS?A?I 1 hI?P?I where g is the electronic Zeeman matrix gn is the nuclear Zeeman g-factor A is the hyperfine matrix P is the nuclear quadrupole coupling tensor B and bn are the Bohr and nuclear magnetons respectively and S and I are the electron and nuclear spin operators respectively. For this rhenium system the linewidth is nearly completely dominated by strain eVects and simple MI and MI 2 dependant linewidth expressions are deemed inappropriate for spectra which are clearly not first-order.Strain eVects were included by calculating exact analytical gradients of the eigenvalues of the spin Hamiltonian transformed to the field domain through the fourth-order frequency perturbation. Rhenium has two naturally occurring isotopes 187Re (62.6%) and 185Re (37.4%). Because the ratio of moments is 0.9899 1 no splitting due to the isotopes is resolved but in order to properly evaluate hyperfine and quadrupole strain eVects we had to include the spectra of both isotopes explicitly in the simulations. Preparation of compounds K7[·2-ReVOP2W17O61] 1. A 150 mL round-bottom flask was charged with K2ReCl6 (0.839g 1.76 mmol) K10[a2-P2W17O61] (6.67g 1.47 mmol) and 80 mL of deionized H2O resulting in a bright blue slurry. This mixture was refluxed for 30 minutes during which time the color of the solution became a dark purple-black color.The solvent was removed by rotary evaporation at 50 8C. To remove any impurities the sample was dissolved in a minimum amount of hot H2O (80 8C) (ª9 mL) and placed in a refrigerator to precipitate the product. The purple-brown solid was filtered oV through a medium glass frit and washed with 3 × 20 mL of EtOH followed by 3 × 30 mL of diethyl ether to give 6.13 g of 1 (M of anion = 4635) in 90% yield. In order to increase solubility for 183W NMR the potassium salt was partially lithiated by a metathesis reaction wherein a stoichiometric amount of LiClO4 was added to an aqueous solution of 1 precipitating KClO4. 31P NMR (D2O) d 211.86 (s) 212.86 (s). 183W NMR (D2O) d 2134 (s 2W) 2161 (s 1W) 2182 (s 2W) 2191 (s 2W) 2210 (s 2W) 2271 (s 2W) 2273 (s 2W) 2280 (s 2W) 2479 (s 2W).Negative ESI-MS ([ReP2W17O62]72 = 4365) 872 (m/5) 1091 (m/4) (Calc. for K7O62P2ReW17?2 H2O K 5.58; Cl 0.0; P 1.26; Re 3.79; W 63.67. Found K 5.15; Cl 0.04; P 1.16; Re 4.02; W 59.43%). Infrared (KBr cm21) 1612.8s 1090.9s 1019.2 (sh) 953.5s 905.2s 777.4s (br) 597.4w 564.1w 523.1m. J. Chem. Soc. Dalton Trans. 1999 301–310 303 [NBu4]xK7 2 x[·2-ReVOP2W17O61] 1a. A 50 mL Schlenk flask was charged with 5 mL of deoxygenated water and 0.100 g (0.022 mmol) of K7[a2-ReVOP2W17O61]. To this brown solution was added 0.069 g (0.216 mmol) of solid NBu4Br under a flow of N2. The pH was maintained between 4 and 5 by the addition of 0.18 M H2SO4. After stirring for a few minutes the solvent was removed in vacuo resulting in a brown solid.31P NMR (CD3CN) d 212.26 (s) 213.19 (s) and 212.70 (small amount of [NBu4]6[P2W18O62] impurity). The sample is air sensitive. [NBu4]6[·2-ReVIOP2W17O61] 2. In air a solution of K7[a2- ReOP2W17O61] (0.500 g 0.107 mmol) in 100 mL of deionized water (initial pH 3.7) was treated with small portions of NBu4Br (0.347 g 1.07 mmol). The pH was maintained between 4 and 5 via the addition of 0.18 M H2SO4. This results in the formation of a brown solid which was filtered and washed with 3 × 30 mL of water followed by 3 × 30 mL of diethyl ether. As the solid was drying under vacuum in the glass frit the color of the solid changed from dark brown to a blue-green (over the course of two hours). Yield 0.464 g (78%) (Calc. for C96H216N6O62P2ReW17 C 19.81; H 3.74; N 1.44; P 1.06; Re 3.20; W 53.70.Found C 20.35; H 3.93; N 1.30; P 0.68; Re 3.30; W 53.64%). 31P NMR (CD3CN) d 27.60 (br s) 212.39 (s). Negative ESI-MS 22 charge state [NBu4]4[a2-ReVIOP2W17O61] :2666.9 [NBu4]3H1[a2-ReVIOP2W17O61] 2546.4; 32 charge state [NBu4]3[a2-ReVIOP2W17O61] 1697.3 (NBu4)2 H1[a2-ReVIOP2W17O61] 1616.9; 42 charge state [NBu4]H1- [a2-ReVIOP2W17O61] 1152.2; 2 H1[a2-ReVIOP2W17O61] 1091.9 52 charge state H1[a2-ReVIOP2W17O61] 873.3. [NBu4]5[·2-ReVIIOP2W17O61] 3. To a 100 mL Schlenk flask charged with AgO3SCF3 (0.047 g 0.18 mmol) and 5 mL of deoxygenated acetonitrile was added a deep green solution of [NBu4]6[a2-ReVIOP2W17O61] in 20 mL of CH3CN by cannula filtration. This results in a green solution with a gray solid (Ag0). After stirring for two hours the mixture was exposed to air and filtered through a nylon disk filter to remove the elemental silver.The solvent was removed by rotary evaporation leaving a pale green solid. The sample was redissolved in 2 mL of hot CH3CN and filtered to remove any insoluble species. With stirring ca. 4 mL of diethyl ether was added dropwise to precipitate a pale yellow solid. The solid was collected on a glass frit and washed with 3 × 30 mL of ether to give 0.606 g of 3 (74%). 31P NMR (CD3CN) d 211.70 (s) 213.33 (s). 183W NMR (CD3CN) d 285.60 (1) 2114.56 (2) 2127.78 (2) 2130.42 (2) 2146.57 (2) 2147.94 (2) 2154.38 (2) 2157.80 (2) 2195.91 (2) (Calc. for C80H180N5O62P2ReW17 C 17.23; H 3.25; N 1.26; P 1.11; Re 3.34; W 56.04. Found C 17.09; H 3.38; N 1.09; P 1.11; Re 3.24; W 56.65). Results and discussion Synthesis of complexes K7[·2-ReOP2W17O61].The reaction of an aqueous solution of K2ReCl6 and K10[a2-P2W17O61] gives the purple-brown compound K7[a2-ReOP2W17O61] in 90% yield. This product was purified by recrystallization from hot H2O. This reaction was reported in 1974 but no procedure or characterization of the product was provided.12 We had attempted to prepare Re(V) complexes of [a2-P2W17O61]102 using ReV starting materials; however our eVorts resulted in mixtures of unidentifiable products presumably a mixture of Re oxidation states. The reaction of the ReIV reagent produces the desired product in excellent yield. Occasionally in preparations of this compound we observed a peak in the 31P NMR spectrum at ca. d 10.2 (<2%) signalling another species perhaps a paramagnetic Re(VI) species. This impurity could be removed by recrystallization.All of the data given below (elemental analysis 31P and 183W NMR spectroscopy) and lack of EPR signal (X-Band 9.5 GHz) over a wide magnetic range in the solid state (293 K) or a frozen water–ethylene glycol mixture (77 K) are consistent with a diamagnetic ReV]] O product. Apparently the rhenium abstracts an oxo ligand from either the aqueous media or the lacunary tungstate cluster. Elemental analysis clearly shows that there are no chloride atoms present and gives a reasonable analysis for K7O62P2- ReW17. We and others have found that elemental analysis is often unreliable for analysis of polyoxoanions especially for the tungsten analysis.14 We find that other techniques including multinuclear NMR spectroscopy are often more sensitive and accurate. The infrared spectrum not surprisingly is similar to the parent P2W18O62 62 species.In the previous study reported by Pope10 the IR spectra of the ReV ReVI and ReVII substituted lacunary [XW11O39]n2 (X = ReV ReVI ReVII) anion were indistinguishable from those of the corresponding [XW12O40]n2 species. A similar oxidation process has been reported in the reaction of RuIIICl3?xH2O with [a2-P2W17O61]102 to give [O{RuIVCl(a2- P2W17O61)}2].26 We and others have found that the [a2-P2W17O61] framework is oxophilic stabilizing ions with high charge size ratio 27 such as tetravalent Ce.28 In addition preliminary reports in the Russian literature show stabilization of the quadrivalent oxidation state in transuranic (TRU) ions like Am and Cf by complexation with the lacunary heteropolyanion [P2W17O61]102 in aqueous solutions.29 In that context the oxidation of ReIV to ReV in this reaction is not surprising.The oxidation of ReIV complexes upon incorporation into polyoxoanion frameworks may be an appropriate synthetic method to prepare rhenium complexes of other metal oxides. We used isomerically pure [a2-P2W17O61]102 as the starting material in this preparation. This is important as the preparation of metal complexes using base degradation of the parent [a2-P2W17O61]102 isomer often results in mixtures of the a2- and a1-[P2W17O61]102 isomers.13,14 Examination of the 31P and 183W NMR data given below are consistent with the production of isomerically pure K7[a2-ReOP2W17O61]. The 31P NMR spectrum of a D2O solution of the Re(V) product K7[a2-ReOP2W17O61] consists of two sharp singlets at P(1) d 211.86 and P(2) d 212.86 indicative of two inequivalent phosphorus centers (Fig.2A). P(1) is the phosphorous atom closest to the site of substitution. P(2) is that remote from the subsitution site. These resonances are diVerent from the lacunary K10[a2-P2W17O61] and further the NMR indicates that the complex is >99% isomerically pure. For comparison the positions of the two resonances in the parent [a2-P2W17O61]102 and diamagnetic lanthanide complexes in water are at d 7–8 for P(1) and at ca. 14 for P(2). There is a paucity of literature comparing the 31P resonances of transition metal complexes incorporated into the [a2-P2W17O61]102 framework especially second or third Fig. 2 A 31P NMR spectrum of K7[a2-ReVOP2W17O61] 1 taken in D2O 20 mM. B 183W NMR spectrum of LixK7 2 x[a2-ReVOP2W17O61] 0.2 M.304 J. Chem. Soc. Dalton Trans. 1999 301–310 row metal ions. In one study the [a2-MoP2W17O62]62 species showed chemical shifts in water of d 211.9 and 212.6 close to our complex.30 The 31P NMR spectrum of the paramagnetic Ru(IV) dimeric species discussed above identifies the broad resonance at d 10.2 as the P(1) resonance and the P(2) resonance that at d 213.2.26 Complexation of a ReO unit to the lacunary K10[a2-P2W17- O61] generates a plane of symmetry (Cs) through the molecule. The 183W NMR spectrum of the partially lithiated LixK7 2 x[a2- ReOP2W17O61] (0.2 M) in D2O shows the expected nine line pattern in which eight peaks integrate to two and one resonance integrates to one and is assigned to the unique tungsten (Fig. 2B). The tungsten resonance shifted upfield d 2479 is out of the range we usually observe for diamagnetic transition metal or lanthanide complexes of the [a2-P2W17O61]102 isomer.6,7 The two tungsten atoms represented by this resonance are shielded relative to the other tungsten atoms. This shielding may result from strong magnetic anisotropy of the Re]] O bond. This resonance may be tentatively assigned to the two tungsten atoms in the “cap” region connected to the Re]] O unit via “edge bonding” each to two oxygen atoms in the cap. In other studies of [ReV]] O]31 complexes of organic ligands we have observed unusual upfield shifts for protons bound to carbon atoms in close proximity to the [ReV]] O]31 (or [TcV]] O]31) site.31 Cyclic voltammetric studies on aqueous solutions of K7[a2- ReOP2W17O61] (0.5 M LiO2CCH3 pH 4.7) reveal four redox waves with chemical potentials at 1823 1476 251 and 2194 mV (platinum working electrode Ag–AgCl reference electrode Pt wire auxiliary electrode) (see SUP 57462 S1).The chemical potentials are assigned to Re71/Re61 (1823 mV) Re61/Re51 (1476 mV) and Re51/Re41 (251 mV). It is diYcult to definitively assign the electrochemical potential located at 2194 mV this potential can be assigned to either the Re41/Re31 couple or a reduction at the tungsten sites. The ia/ic ratios of non-unity and the DEpp chemical potential separations of over 56 mV (ranging from 70–120 mV) over scan rates of 50 mV s21 to 500 mV s21 suggest electrochemical irreversibility. We intend to study the aqueous electrochemistry of 1 at variable pH. The aqueous electrochemistry of both [W9ReVO32]52 and [SiW11ReVO40]52 show significant dependence on pH.10,11 In those studies reduction peaks assigned to the Re5/4 and Re4/3 couples were highly pH dependent.For both compounds two one electron reductions were observed from pH 6 to pH 4 and one two electron process was observed below pH 4. [NBu4]xK7 2 x[·2-ReVOP2W17O61] 1a. The oxidative conversion of [NBu4]xK7 2 x[a2-ReVOP2W17O61] to its Re61 2 analog can be prevented by treating an aqueous solution of K7[a2- ReVOP2W17O61] with NBu4Br under strict anaerobic conditions. This reaction results in the formation of an aqueous-insoluble brown solid. SUP 57462 S2 shows the comparison of the 31P NMR spectra of the tetrabutylammonium salts of the ReV ReVI and ReVII complexes of [a2-ReOP2W17O61]n2. The chemical shift values are in line with those observed by Lyon for the tetrabutylammonium salts of a variety of transition metal derivatives of [a2-P2W17O61]102 taken in CH3CN–CD3CN.14 A small percentage of [P2W18O62]62 (d 212.70) is often observed as an impurity.The 31P NMR spectrum of an acetonitrile solution of the NBu4 1 salt of 1 consists of two phosphorus chemical shifts at d 212.26 and 213.19 (SUP 57462 S2C). [NBu4]6[·2-ReVIOP2W17O61] 2. Performing the metathesis reaction of K7[a2-ReVOP2W17O61] in air with NBu4Br resulted in an oxidation of the Re(V) to Re(VI). Treatment of an aqueous solution of K7[a2-ReVOP2W17O61] with NBu4Br while maintaining the pH at 4–5 results in the formation of an aqueousinsoluble brown solid. As the brown solid dried under vacuum in a glass frit the solid’s color changed to blue-green.The molecule is paramagnetic; EPR data will be discussed later. A similar process in which a metathesis reaction resulted in oxidation was demonstrated recently by Pope in which conversion of [CN3H6]5[W9ReVO32] to the NBu4 1 salt resulted in an aerial oxidation of ReV to ReVI.11 The 31P NMR spectrum of a CD3CN solution of [NBu4]6[a2-ReVIOP2W17O61] 2 shows a broad peak at d 27.60 due to the paramagnetism of the Re(VI) (d1) and a sharp resonance at d 212.39 (SUP 57462 S2B). Cyclic voltammetry of an acetonitrile solution of this blue complex showed four reversible redox features (glassy carbon working electrode Pt auxiliary electrode Ag–AgCl reference electrode 0.1 M NBu4PF6) at 1261 (ReVII/ReVI) 2292 (ReVI/ ReV) 21151 (ReV/ReIV) and 21363 mV (ReIV/ReIII or tungsten reduction) (Fig.3). The reduction processes were reversible over scan rates of 50 mV s21 to 450 mV s21. Electrochemical studies of [ReVOPW11O39]42 showed four oxidative/reductive processes which were assigned to rhenium in oxidation states ranging from III to VII.9 The technetium analog [TcVOPW11O39]42 however showed a single reversible oxidation and a quasireversible reduction.32 [NBu4]5[·2-ReVIIOP2W17O61] 3. Treatment of a green acetonitrile solution of [NBu4]6[a2-ReVIOP2W17O61] with an excess of AgO3SCF3 (1.2 equivalents) leads to the oxidation of the Re from 16 to 17 and a solution color change from deep green to a pale yellow-green. The 31P NMR spectrum of this species consists of two sharp singlets at d 211.70 and 213.33 (SUP 57462 52A). Elemental analysis of this yellow solid confirms the formulation of [NBu4]5[a2-ReVIIOP2W17O61] (C80H180N5- O62P2ReW17).Electrochemical studies of this highly oxidized species show the same CV as seen for the Re(VI) complex. These data are consistent with the complexes having the same structure. The 183W NMR spectrum of [NBu4]5[a2-ReVIIOP2W17O61] (SUP 57462 S3) also showed nine resonances eight integrating for two and one integrating for one. This pattern confirms a molecule of Cs symmetry as would be found by substitution in the “cap” position of the [a2-P2W17O61]102 isomer. The resonances for this species are diVerent from the lacunary species the ReV species are within the range usually observed for diamagnetic metal complexes of the [a2-P2W17O61]102 isomer. Electrospray mass spectrometry Fast-atom bombardment mass spectrometry of K7[a2-ReOP2- W17O61] was uninformative due to substantial cluster fragmentation.We found that negative-ion electrospray mass spectrometry provided data with minimal fragmentation. Electrospray ionization is a softer ionization technique than FAB therefore it is commonly used for large and/or labile proteins and polymers.33 Other electrospray ionization mass spectral analyses conducted on molybdenum and tungsten “cage” compounds of the Lindquist Keggin octamolybdate and decatungstate structures showed the polyoxoanions as base peaks with minimal degree of fragmentation in organic Fig. 3 Cyclic voltammogram of [NBu4]6[a2-ReVIOP2W17O61] 2 taken in CD3CN (glassy carbon working electrode Pt auxiliary electrode Ag–AgCl reference electrode 0.1 M NBu4PF6) 100 mV s21. J.Chem. Soc. Dalton Trans. 1999 301–310 305 Fig. 4 Electrospray mass spectrum of K7[a2-ReVOP2W17O61] 1. Inset shows expansion of m/z = 850 to 1650 region. aqueous organic and aqueous solutions.34a,b One study displayed the potential for ES-MS to monitor equilibrium processes of aqueous polyoxoanion mixtures with time.34b Polyoxoanions are highly negatively charged molecules and form adducts with cations such as K1 and H1 in this case. Thus diVerent charge states can be formed. The ES mass spectrum for K7[a2-ReOP2W17O61] taken in water displayed in Fig. 4 shows the 22 32 42 and 52 charge regions. The inset shows an expansion of the 32 42 and 52 charge regions. Fig. 4 and Table 1 show the corresponding m/z values for the various charge states of the molecule with variable amounts of potassium and proton adducts.The electrospray mass spectrum for [NBu4]6[a2-ReVIOP2W17O61] species taken in acetonitrile also is consistent with the theoretical values for the 22 32 42 and 52 charge states with variable amounts of tetrabutylammonium cations adducts. One must be cautious however as the electrospray mass spectral technique may not distinguish oxoanions of diVerent Re oxidation states well. In these cases the oxoanions may diVer by only one proton adduct and thus would have the same or close to the same m/z value. EPR studies of [NBu4]6[·2-ReVIOP2W17O61] X-Band Q-band and W-band EPR spectra shown in Fig. 5 6 and 7 S4 respectively show remarkable changes especially in linewidth. For the X-band spectrum Fig. 5 the linewidth varies from 25 G at lowest fields to more than 120 G at highest fields.In the Q-band spectrum Fig. 6 the linewidth varies from 80 to 280 G and in the W-band spectrum Fig. 7 the linewidth varies from 250 to 700 G. This linewidth variation can be readily ascribed to random strains or distortions of the rhenium ion complex with a resultant distribution of spin Hamiltonian parameters particulary in the g-tensor. The values of the g- A- and P- tensors can be found in Table 2 and comparative EPR parameters for d1 systems in heteropolytungstates can be found in Table 3. While some of the perpendicular (or ‘z’) features can be readily assigned in the spectra at all three frequencies the perpendicular (or ‘x’ and ‘y’) features are poorly resolved at the higher frequencies. The “perpendicular” X-band cannot be assigned to any first-order splitting pattern and analysis of the X-band spectrum shows that putatively “forbidden” DMI = ±1 ±2 .. . transitions completely dominate this part of the spectrum and that they have become “allowed” while the putatively “allowed” DMI = 0 transitions have become “forbidden.” The observation of such strong forbidden transitions can be ascribed to state-mixing of the nuclear and electronic levels the Fig. 5 X-Band (9.27 GHz) electron paramagnetic resonance spectrum of [NBu4]6[a2-ReVIOP2W17O61] 2 in a CH3CN–C7H8 glass at 77 K. Experimental (——) simulated (–––). The asterisks mark forbidden MI = 5/3 to 3/2 and MI = 3/2 to 5/2 transitions. See text. 306 J. Chem. Soc. Dalton Trans. 1999 301–310 splittings of which are of comparable strength at X-band. The two lowest field transitions of the “parallel” part of the X-band spectrum which can be assigned to allowed MI = 5/2 and 3/2 transitions appear to be split (Fig.5). These extra peaks marked with asterisks are actually “forbidden” mI = 5/2 to 3/2 and mI = 3/2 to 5/2 transitions. The spacing between these two peaks is equal in first-order to 8QD/geffb where QD=3/2Pz. From this splitting we obtain QD = 203 MHz which compares well with the value of 209 MHz obtained with spectral simulation Table 3. This is one of the best examples to date of how state-mixing can allow the direct measurement of the nuclear quadrupole coupling from a powder spectrum.35 Fig. 6 Q-Band (34.63 GHz) electron paramagnetic resonance spectrum of [NBu4][a2-ReVIOP2W17O61] 2 in a CH3CN–C7H8 glass at 77 K. Experimental (——) simulated (–––).Table 1 Electrospray mass spectral data for K7[ReO(a2-P2W17O61)] 1 a and [N(Bu)4]6[ReO(a2-P2W17O61)] 2 b K7[ReVO(a2-P2W17O61)] Charge state 22 22 22 22 22 32 32 32 32 32 42 42 42 42 52 52 52 Species K5[ReO(a2-P2W17O61)]22 K4H[ReO(a2-P2W17O61)]22 K3H2[ReO(a2-P2W17O61)]22 K2H3[ReO(a2-P2W17O61)]22 KH4[ReO(a2-P2W17O61)]22 K4[ReO(a2-P2W17O61)]32 K3H[ReO(a2-P2W17O61)]32 K2H2[ReO(a2-P2W17O61)]32 KH3[ReO(a2-P2W17O61)]32 H4[ReO(a2-P2W17O61)]32 K3[ReO(a2-P2W17O61)]42 K2H[ReO(a2-P2W17O61)]42 KH2[ReO(a2-P2W17O61)]42 H3[ReO(a2-P2W17O61)]42 K2[ReO(a2-P2W17O61)]52 KH[ReO(a2-P2W17O61)]52 H2[ReO(a2-P2W17O61)]52 m/z 2280.3 2261.2 2242.2 2223.1 2204.1 1507.1 1494.3 1481.7 1469.0 1456.3 1120.6 1111.1 1101.5 1092.0 888.6 881.0 873.4 [NBu4]6[ReVIO(a2-P2W17O61)] Charge state 22 22 32 32 42 42 52 Species [NBu4]4[ReO(a2-P2W17O61)]22 [NBu4]3H[ReO(a2-P2W17O61)]22 [NBu4]3[ReO(a2-P2W17O61)]32 [NBu4]2H[ReO(a2-P2W17O61)]32 [NBu4]2H[ReO(a2-P2W17O61)]42 H2[ReO(a2-P2W17O61)]42 H[ReO(a2-P2W17O61)]52 m/z 2666.9 2546.1 1697.3 1616.9 1152.2 1091.9 873.3 a [ReV]] O(a2-P2W17O61)]72 M = 4365 sample dissolved in water.b [ReVI]] O(a2-P2W17O61)]62 M = 4365 sample dissolved in acetonitrile. While the “parallel” portion of the X-band spectrum appears to be first-order it is not. The peak at highest field which would be nominally assigned as the parallel feature of the “allowed” MI = 25/2 to 25/2 transition is actually an oV-axis extremum peak or angular anomaly. The actual “parallel” MI = 25/2 feature occurs 100 G downfield. Because almost all peaks above 3800 G can be attributed to oV-axis peaks in the X-band spectrum and the same above 15000 kG in the Q-band spectrum the high-field end of these spectra are highly sensitive to noncoincidence between the z-axes of the g- A- and P-tensors.Non-coincidence of the tensor axes is found only in the yz plane allowing us to assign the yz plane to the symmetry plane of this Cs system. The non-coincidence is greatest between the g- and A-tensors (15.28) and smallest between the A- and P-tensors (5.08) as expected. The axes of the g-tensor are determined primarily by orientations of the ground and excited states whereas the axes of A are determined by the orientation of the ground state orbital and the axes of P are determined by the orientation of the ligands.36 By carefully simulating the X-band and Q-band spectra we can obtain very accurate values of the g hyperfine and quadrupole coupling tensors which are reported in Table 2.We know of only two other d1 systems in heteropolytungstates for which values of all three tensors g A and P were determined.37,38 The results of all three systems V(IV) Mo(V) and Re(VI) are reported in Table 3. While for all three systems the geometry can be described as octahedral with axial and trigonal distortions and with one short terminal oxygen and five shared oxygens there are expected diVerences. The structure of the a site in [PMoW11O40]42 should be quite similar to Fig. 7 W-Band (94.43 GHz) electron paramagnetic resonance spectrum of [a2-ReVIOP2W17O61]62 2 in a CH3CN/C7H8 glass at 20 K. Experimental (——) simulated (–––). Table 2 EPR spin Hamiltonian and strain parameters for [ReP2W17- O62]62 in CH3CN–C7H8 glass a gx gy gz Ax b Ay Az qgA c Px b Py Pz qgP c 1.7011(7) 1.6646(9) 1.8039(9) 21034(2) 21138(2) 22104(2) 115.28(5) 264.0(5) 275.4(5) 1139.4(9) 110.28(5) sgx sgy sgz sAx sAy sAz sq sPx sPy sPz sq 20.011(1) 20.015(1) 20.008(1) 28(3) 216(3) 0(3) 21.18 11.8(9) 20.1(9) 21.7(9) 21.18 [1.037] [1.044] [1.042] [1.008] [1.014] [1.000] [1.072] [2.028] [1.001] [2.012] a Hyperfine and nuclear quadrupole principal values in MHz (for units of 1024 cm21 divide by 3).Fractional strain is given in square brackets (for g Dg(Dg)/Dg is given); correlation parameter of 0.99 used between g A and P strains. b 187Re A(185Re) = A(187Re) × 0.9899 P(185Re) = P(187Re) × 1.05. c Angle of non-coincidence in the yz plane between g and A or P. J. Chem. Soc. Dalton Trans.1999 301–310 307 Table 3 Comparative EPR parameters for d1 systems in heteropolytungstates a Oxyanion [VW5O19]42c [a-PMoW11O40]42d [a2-P2ReW17O62]62e gx 1.967 1.9368 1.7011 gy 1.967 1.9340 1.6646 gz 1.944 1.9108 1.8039 Ax 176.1 96.5 21034 Ay 176.1 96.5 21138 Az 495.6 221.2 22104 QDb 20.29 9.0 209.1 a Hyperfine and nuclear quadrupole couplings are in units of MHz. Hyperfine couplings are given for V51 Mo95 and Re187 isotopes. b QD = 3/2 Pz; values are given for V51 Mo97 and Re187 isotopes. c Ref. 38. d Ref. 37. e This work. the a2 site in [a2-P2ReW17O62]62 since both have two shared corners and two shared edges while the vanadium in the [VW5O19]42 has no shared corners and four shared edges. In addition the site symmetry is expected to be C4v for the vanadium system and Cs for the rhenium and molybdenum systems.Comparison of the g values in Table 3 show two trends. One is the increase in shift from the free-electron g-factor for average value of the g-tensor i.e. the isotropic g-factor as one goes down the periodic table; the other is the increase in rhombic character and a corresponding decrease in the axial character of the g-tensor. The increase in g-shift is most noticeable between Re and Mo the ratio of isotropic g-shifts being 3.71. This increase can be readily explained by a large spin–orbit coupling (SOC) for Re which is expected to be 4.25 times larger than for Mo.39 For a system with Cs symmetry the ground state orbital wavefunction will be hybridized from dx2 2 y2 dz2 dyz and ligand orbitals the x axis being normal to the symmetry plane and bisecting the O–Re–O bonds and the z-axis lying approximately along the terminal Re–O bond and normal to the plane of the four in-plane oxygen atoms.This wavefunction can be written as y0 = a(a|x2 2 y2Ò 1 b|z2Ò 1 c|yzÒ) 2 a9|L0Ò (1) with excited states y1 = b |xyÒ 2 b9|L1Ò (2) y2 = g |xzÒ 2 g|L2Ò (3) y3 = d(a |yzÒ 2 c |x2 2 y2Ò)/÷a2 1 c2 2 d9|L3Ò (4) y4 = h((a2 1 c2) |z2Ò 2 ab |x2 2 y2Ò 2 bc |yzÒ)/ ÷a2 1 c2 2 h9|L4Ò (5) where |LoÒ |L1Ò etc. are appropriate symmetry adapted linear combinations of ligand orbitals. Expressions for g and A can be derived for Cs symmetry similarly to those for C2v symmetry.40 Assuming b and c are much smaller than a and terms in b2 c2 and bc are ignored the resulting equations for g are Dgzz = 28a2Dxy (6) Dgyz = 22ac[2Dxy 1 Dxz] (7) Dgyy = 22a(a 2 2÷3b)Dxz (8) Dgx = 22a(a 1 2÷3b)Dyz (9) where Dgpq = gpq 2 ge.To first order Dxy = a2b2 ·xÒ/DExy Dxz = a2g2·xÒ/DExz and Dyz = a2d2·xÒ/DEyz; where DEpq are the d– d excitation energies from the ground state to the corresponding dpq state and the rhenium spin–orbit interaction x(r) varies approximately as Zeff/r3. Overlap spin-polarization eVects and contributions from bonding orbitals (hole excitations) can also be lumped into Dxy Dxz and Dyz. The SOC for the oxygen ligands can be readily neglected but cannot be neglected for neighboring tungsten atoms. In order to relate the observed angle of non-coincidence between gz and Az qgA to the angle of non-coincidence between gz and the z-axis of the reference system qg we need use the corresponding equations for the hyperfine tensor components Azz = 2P[(k0 1 4/7 a2)a2 2 Dgzz 2 1 (3a 2 ÷3 – b) 14 (a 1 ÷3 – b) Dgx 2 1 (3a 1 ÷3 – b) 14 (a 2 ÷3 – b) Dgyy] (10) Ayz =2P[6 7 – aca2 2Dgyz 1 3 14 c (a 2 ÷3 – b) Dgyy 2 3c 28a Dgzz] (11) Ayy = 2P[(k0 2 2 7 – a(a 1 2÷3 – b))a2 2 Dgyy 1 1 (3a 2 ÷3 – b) 14 (a 1 ÷3 – b) Dgx 1 ÷3 – b 14 a Dgzz] (12) Ax = 2P[(k0 2 2 7 – a(a 2 2÷3 – b))a2 2 Dgx 1 1 (3a 1 ÷3 – b) 14 (a 2 ÷3 –b) Dgyy 2 ÷3 – b 14 a Dgzz] (13) with overlap and ligand terms neglected as well as taking the Fermi contact term k as being simply equal to k0 times a2.We assume Dxz = Dyz which is a good approximation since the hyperfine and nuclear quadrupole coupling tensors are nearly axial. Using this assumption and P = 1157 MHz,40 we find b = 20.014 c = 0.060 a2 = 0.912 k0 = 0.95 qA = 24.68 and qg = 10.68. The angle of non-coincidence between g and A for molybdenum in [PMoW11O40]42 was found to be 188,37 which when using P = 2150.7 MHz41 gives b = 20.013 c = 0.072 a2 = 0.89 k0 = 0.84 qA = 24.48 and qg = 222.38.The predicted asymmetry in A (Ax 2 Ay)/(Az 2 Aiso) equal to 20.09 for the rhenium system accounts for about half of the observed asymmetry of 20.15. The remaining diVerence 20.06 can be attributed to overlap and ligand contributions to ·r23Ò. For the molybdenum system an asymmetry of 20.04 is expected although the A-tensor was assumed to be axial. EPR data have been reported for [PVW11O40]5238 and [P2VW17O62]7242 but the spectra were interpreted assuming C4v symmetry. The hybridization coeYcients b and c are small for both rhenium and molybdenum systems showing that the groundstate wavefunction is primarily dx2 2 y2 (a = 0.99).The values of c 0.060 and 0.072 correspond to a tilting of the groundstate dx2 2 y2 orbital from the xy plane by 23.48 and 24.28. While the z axis of the hyperfine tensor is orientated very close to the normal of this tilted dx2 2 y2 orbital that of the g-tensor is not. Also the tilt in the g-tensor z-axis for the rhenium system is opposite in direction to that for the molybdenum system. The angle of non-coincidence can be related to the elements of g tan(2qg) = 2Dgyz Dgyy 2 Dgzz >23 ac Dgiso Dgyy 2 Dgzz (14) 308 J. Chem. Soc. Dalton Trans. 1999 301–310 and shows the sensitivity of qg to the diVerence between Dgyy and Dgzz. If Dgyy is very close to Dgzz qg can switch from 1458 to 2458. On the other hand if (Dxz 1 Dyz)/2 is close to Dxy Dgzz will be about eight times Dgyy and the z-axes of the g- and A-tensors will be nearly coincident and aligned with the normal to the tilted dx2 2 y2 orbital.In the rhenium system the ratio of (Dxz 1 Dyz)/2 to Dxy is six while in the molybdenum system it is three and in the vanadium system it is two and a half. The decrease in axial character in the g-tensor as one goes down Table 3 can be explained in terms of an increase in in-plane s-bonding and a weakening of the axial or terminal M]] O bond. With the increase in in-plane metal– ligand covalency there will be a decrease in the MO coeYcient b a decrease in ·xÒ due to increased electron screening (the twoelectron part of SOC) an increase in overlap and added contributions from bonding orbitals all of which will contribute to the quenching of the in-plane angular momentum and to a decrease in Dxy.This eVect is strong enough for the rhenium system to make Dgz smaller than Dgx and Dgy. The nuclear quadrupole tensor represents the total electric field gradients at the metal nucleus and includes contributions from the paired bonding electrons as well as that from the unpaired electron. The principal nuclear quadrupole coupling constant can be expressed as 4I(2I 2 1) 3 QD = e2Qa2q5d(1 2 R) 1 e2Qqov(1 2 R) 1 e2Qqlig(1 2 g) where q5d is equal to ·r23Òd qov and qlig are the overlap and ligand contributions to ·r23Ò and R and g are the Sternheimer shielding and antishielding constants. The first term is determined primarily by the unpaired electron density giving a similar contribution to the hyperfine tensor while the third term is dominated by contributions from the paired electron density and surrounding nuclear charges.While the third term is equal to zero for octahedral geometry it is very sensitive to deviations from octahedral symmetry. Rhenium is expected to have a large value of QD (3/2 Pz) since QD scales as the product of nuclear quadrupole moment and ·1/r3Ò. Q(Re187/Mo97) is equal to 11.1 and the average value of 1/r3 is 2.19 times larger for rhenium than molybdenum.41 Assuming the ligand fields are similar and neglecting diVerences in eVective charge and relativistic eVects one expects that the value of QD should be 24.2 times larger for the rhenium system than for the molybdenum system (Table 3). This value compares extremely well to the observed ratio of 23.4 1. The corresponding scaling factor between V51 and Mo97 is expected to be 211.3 1 although a larger ratio 231.5 1 is actually observed between the vanadium and molybdenum systems.The poor scaling between the vanadium and molybdenum systems can be in part attributed to the diVerence in structure between the vanadium system and that of the molybdenum and rhenium systems but also to smaller in-plane metal–ligand covalency expected for vanadium. These diVerences show up more in the nuclear quadrupole coupling than the hyperfine interaction because of nuclear quadrupole coupling’s sensitivity to geometric changes in the ligand field. The angle of non-coincidence between the z-axes of the A-tensor and the rhenium nuclear quadrupole coupling tensor is found to be 5.08 and corresponds to a tilt of 0.48 from the z-axis of the nuclear quadrupole tensor from the reference z-axis.If only the first term in the above equation is considered and the contribution from bonding valence electrons neglected this angle would be 23.58 the same as the tilt angle of the hybrid d orbital. The alignment of the nuclear quadrupole tensor by the ligand geometry rather than the ground-state orbital is due to the importance of the ligand field contributions. The asymmetry parameter (Px 2 Py)/Pz expected for our hybrid d orbital will be equal to 20.04 but instead a value of 10.08 is observed. The ligand field will polarize the d electrons producing electric field gradients that oppose those of the ligands. Because the quadrupole coupling is much more sensitive to the ligand field gradient than is the hyperfine tensor the asymmetry parameter for the hyperfine will be reversed from that of the nuclear quadrupole coupling.The EPR linewidth of the rhenium center in [ReOP2- W17O61]62 is dominated by a highly correlated g-strain The strain in the three principal values of g A and P were assumed then to be totally correlated; that is e = ±1. The overall strain due to g DWg A DWA and P DWP is then convoluted to give the net Gaussian linewidth DW DW2 = DWr 2 1 DWg 2 1 DWA 2 1 DWP 2 1 2(egADWgDWA 1 egPDWgDWP 1 eAPDWADWP) where DW is the residual or intrinsic linewidth which was found to have an isotropic value of 22.5 G and egA is the correlation between g and A strains etc. The g-strain along all three axes is roughly proportional [D(Dg)/Dg ranges from 10.037 to 10.044] to the shift from the free electron g-factor ge = 2.00236.An increase in Dg corresponds to an increase in Dxy Dxz and Dyz and an increase in unquenched angular momentum due to a decrease in the ligand field. A highly correlated strain is also found for the hyperfine tensor ega = 0.99 but is both relatively smaller and quite anisotropic as compared to the strain in g. This is the opposite of the behavior seen for copper d9 systems where the A-strain is proportionally larger than the g-strain.43 The small A-strain for rhenium can be explained by diVerentiating eqn. (10)–(13) with b = 20.014 c = 0.060 a2 = 0.912 and k0 = 0.95. DAzz = 2P [0.74Da 2 D(Dgzz) 2 0.22D(Dgx) 2 0.21D(Dgyy) 1 0.08Db] (15) DAyz = 2P [0.02Da 2 D(Dgyz) 1 0.01D(Dgyy) 2 0.01D(Dgzz) 2 8.04Dc] (16) DAyy = 2P [0.35Da 2 D(Dgyy) 1 0.22D(Dgx) 2 0.00D(Dgzz) 2 0.89Db] (17) DAx = 2P [0.34Da 2 D(Dgx) 1 0.21D(Dgyy) 1 0.00D(Dgzz) 1 0.89Db] (18) If Da is small the strain in A will be dominated by the strain in g and changes in hybridization.The A-strain in d1 systems is expected to be much smaller than in d9 systems because Dg for d1 systems is opposite in sign from that in d9 systems where it adds in phase to the Da term. The changes in hybridization as determined from the g-strain and the correlated part of the angular strain non-coincidence between g and A (20.888 e = 0.81) are Db = 20.0011 Dc = 20.0048. While these changes are small they make a large impact on the A-strain in particular in the oV-diagonal element. Using P = 1157 MHz and Da = 20.01 gives DAzz = 26.5 MHz DAyy = 29.4 MHz DAx = 24.1 MHz and a very large DAyz = 42.7 MHz.In the frame in which A is diagonal these values become DAz = 10.4 MHz DAy = 216.3 MHz and DAx = 24.1 MHz which compare very well to the observed values of 0 216 and 28 respectively. The strain in the components of the nuclear quadrupole coupling tensor is also correlated (egP = 0.95) to the A-strain but in an opposite phase. Along y the strain is largest in A and yet smallest in P. While P and the anisotropic part of A are both measures of the electric field gradient at the nucleus the nuclear quadrupole tensor does not include contributions from the gtensor. Considering only the dipolar contributions in the above equations for the hyperfine coupling and using Da = 20.01 gives DPz = 20.9 MHz DPy = 10.8 MHz and DPx = 10.1 J. Chem. Soc. Dalton Trans. 1999 301–310 309 MHz.The strain along z is underpredicted but we have included only the unpaired electron contribution to P. The strains along x and y are reversed. However it has been noted that the equilibrium asymmetry parameter of the P-tensor (10.08) is opposite in direction from that of the A-tensor (20.15) so that for both P and A an increase in the magnitude of asymmetry along with a corresponding increase in Db/b and a decrease in Dc/c is correlated with the increase in Dg. In addition to the strain-induced variation in the angle of non-coincidence between g and A in the yz plane of 21.18 an angular deviation of 2.58 of the z-axis of the g-tensor away from that of the yz plane was needed to simulate the EPR spectra. Unlike the in-plane angular variation which is well correlated to the other strain parameters (e = 0.81) the out-ofplane variation is not and is the only strain for which the planar symmetry is broken.Thus there are probably two diVerent types of strains one in which the rhenium and terminal oxygen are pulled away from the polyoxoanion framework and another in which the rhenium and terminal oxygen are bent out of the plane of symmetry. In the former additional stretching of the Re]] O bond would be needed to give an increase in the g-shift along all three directions. Along with the increase in g-shifts is a decrease in Dc/c an increase in Db/b and a decrease in DQD/QD. The increase in Db/b corresponds to an increase in rhombic character which is opposite to what would be predicted from our model and probably results from increased directed solvent interaction with the Re]] O group.The X-band and Q-band spectra reported for [RePW11O40]52 and [ReSiW11O40]62 are qualitatively similar to the rhenium spectra reported in this paper.10 However the earlier analyses did not include the eVects of the rather large nuclear quadrupole coupling or take into account the possible noncoincidence between the tensors and thus it is diYcult to quantitatively compare the EPR parameters.44 Conclusion We have prepared K7[a2-ReVOP2W17O61] by reaction of ReIVCl6 22 with the [a2-P2W17O61]102 isomer following the example of Charreton and Meunier. This complex was characterized using electrospray mass spectrometry and multinuclear NMR techniques among other techniques. Cyclic voltammetry in aqueous solution shows four non-reversible oxidation–reduction processes.Multiple charge states as well as varying amounts of cation adducts to the multiply charged clusters are observed in the negatively charged electrospray mass spectrometry. 183W NMR spectroscopy confirms the Cs symmetry of the molecule. The NBu4 1 analog of this complex [NBu4]xK7 2 x[a2-ReVOP2- W17O61] is extremely air sensitive oxidizing to the ReVI analog upon exposure to air. The ReVI analog [NBu4]6[a2-ReVIOP2- W17O61] can be prepared simply by metathesis of the ReV material with tetrabutylammonium ion in air. The 31P NMR EPR and mass spectral data are consistent with a paramagnetic ReVI species. Cyclic voltammetry of an acetonitrile solution of the ReVI species shows four reversible redox features. The X-band Q-band and W-band EPR spectra of 2 show extreme variations in linewidths due to random strains or distortions of the complex.Simulations of the X-band and Q-band EPR spectra allow determination of the g hyperfine and quadrupole coupling tensors; the molecule has rhombic character due to an increase in in-plane s bonding and a weakening of the axial M]] O bond. [NBu4]6[a2-ReVIOP2W17O61] can be oxidized by AgO3SCF3 to give the Re(VII) analog 3. Acknowledgements We acknowledge the following sources of support for this research Faculty Research Award Program of the City University of New York Eugene Lang Faculty Development Award NSF-CHE9309001 NSF-CHE9502213 (L. C. F.) NIH-Research Centers in Minority Institutions Grant RR03037-08S2 and NSF Grant PCM8111745 for the purchase of the 400 MHz Spectrometer. NIH Research Resources Grant P-41-RR01811 is acknowledged for the use of IERC resources.References 1 M. T. Pope Heteropoly and Isopoly Oxometalates Springer New York 1983; M. T. Pope and A. Muller Angew. Chem. Int. Ed. Engl. 1991 30 34. 2 D. C. Crans Comments Inorg. Chem. 1994 16 35 and refs. therein; D. C. Crans Polyoxometalates From Platonic Solids to Anti- Retroviral Activity eds. A. Muller and M. T. Pope Kluwer Academic Publishers The Netherlands 1993 pp. 399–406; C. L. Hill M. S. Weeks and R. F. Schinazi J. Med. Chem. 1990 33 2767; M. S. Weeks C. L. Hill and R. F. Schinazi J. Med. Chem. 1992 35 1216 and refs. therein; G.-S. Kim D. A. Judd C. L. Hill and R. F. Schinazi J. Med. Chem. 1994 37 816; N. Yamamoto D. Schols E. deClercq Z. Debyser R. Pauwels J. Balzarini H. Nakashima M. Baba M. Hosoya R. Snoeck J. Neyts G. Andrei B. A.Murrer B. Theobald G. Bossard G. Henson M. Abrams and D. Picker Mol. Pharmacol. 1992 42 1109; Y. Inouye Y. Tokutake T. Yoshida Y. Seto H. Hujita K. Dan A. Yamamoto S. Nishiya T. Yamase and S. Nakamura Antiviral Res. 1993 20 317; D. A. Judd R. F. Schinazi and C. L. Hill Antiviral Chem. Chemother. 1994 5 410. 3 For recent reviews see C. L. Hill and C. M. Prosser-McCartha Coord. Chem. Rev. 1995 143 407 and refs. therein; N. Mizuno and M. Misono J. Mol. Catal. 1994 86 319; I. V. Kozhevnikov Chem. Rev. 1998 98 171; N. Mizuno and M. Misono Chem. Rev. 1998 98 199; M. Sadakane and E. Steckhan Chem. Rev. 1998 98 219. 4 R. Neumann and M. Dahan J. Chem. Soc. Chem. Commun. 1995 1995; C. L. Hill and R. B. Brown Jr. J. Am. Chem. Soc. 1986 108 536; R. Neumann and M. Dahan Nature (London) 1997 388 353.5 D. Mansuy J.-F. Bartoli P. Battioni D. K. Lyon and R. G. Finke J. Am. Chem. Soc. 1991 113 7222. 6 J. Bartis S. Sukal E. Kraft R. Kronzon M. Dankova M. Blumenstein and L. C. Francesconi J. Chem. Soc. Dalton Trans. 1997 1937; Q.-H. Luo J. Bartis M. Dankova N. Ullah V. Young and L. C. Francesconi manuscript in preparation. 7 J. Bartis M. Dankova M. Blumenstein and L. C. Francesconi J. Alloys Comp. 1997 249 56. 8 A. Venturelli and L. C. Francesconi American Chemical Society Meeting San Francisco April 1997. 9 E. Deutsch K. Libson and J.-L. Vanderheyden in Technetium and Rhenium in Chemistry and Nuclear Medicine eds. M. Nicolini G. Bandoli and U. Mazzi Cortina International Verona Raven Press New York 1990. In some situations Re is a suitable replacement for Tc. This is seen in the vast literature on Tc and Re small coordination complexes.However the kinetic inertness of Re compared to Tc as well as the higher reduction potentials of Re compared to Tc may result in very diVerent behavior of the two elements in some matrices such as metal oxides. 10 F. Ortéga and M. T. Pope Inorg. Chem. 1984 23 3292. 11 F. Ortéga M. T. Pope and H. T. Evans Inorg. Chem. 1997 36 2166. 12 B. Charreton and R. C. R. Meunier C.R. Hebd. Seances Acad. Sci. Ser. C 1974 275 945. 13 T. L. Jorris M. Kozik N. Casan-Pastr P. J. Domaille R. G. Finke W. K. Miller and L. C. W. Baker J. Am. Chem. Soc. 1987 109 7402. 14 D. K. Lyon W. K. Miller T. Novet P. J. Domaille E. Evitt D. C. Johnson and R. G. Finke J. Am. Chem. Soc. 1991 113 7209. 15 M. Kozik R. Acerete C. F. Hammer and L. C. W. Baker Inorg.Chem. 1991 30 4429. 16 K. P. Gable Adv. Organomet. Chem. 1997 41 127. 17 W. A. Herrmann and F. E. Kuhn Acc. Chem. Res. 1997 30 169. 18 W. A. Herrman Angew. Chem. Int. Ed. Engl. 1988 27 1297. 19 A. M. Al-Ajlouni and J. H. Espenson J. Am. Chem. Soc. 1995 117 9243. 20 A. K. Yudin and K. B. Sharpless J. Am. Chem. Soc. 1997 119 11536. 21 W. J. Randall M. W. Droege N. Mizuna K. Nomiya T. J. R. Weakley and R. G. Finke Inorg. Synth. 1997 167. 22 R. B. Clarkson W. Wang M. J. Nilges and R. L. Belford Influence of Organic Sulfur in Very High Frequency EPR of Coal in Processing and Utilization of High-Sulfur Coal eds. R. Markuszewski and T. D. Wheelock Elsevier Amsterdam 1990 pp. 67–77. 23 M. J. Nilges SIMPIP Illinois EPR Center University of Illinois 1998. 310 J. Chem. Soc. Dalton Trans.1999 301–310 24 M. J. Nilges PhD Thesis University of Illinois Urbana- Champaign 1979. 25 K. Mattson PhD Thesis University of Illinois Urbana- Champaign 1991. 26 W. J. Randall T. J. R. Weakley and R. G. Finke Inorg. Chem. 1993 32 1068. 27 M. R. Antonio L. Soderholm G. Jennings L. C. Francesconi M. Dankova and J. Bartis J. Alloys Compds. 1998 275–277 827. 28 N. Haraguchi Y. Okaue T. Isobe and Y. Matsuda Inorg. Chem. 1994 33 1015. 29 D. A. Malikov M. S. Milyukova E. V. Kuzovkina and B. F. Myasoedov Sov. Radiochem. (Engl. Transl.) 1992 34 106; A. S. Saprykin V. P. Shilov V. I. Spitsyn and N. N. Krot Dokl. Akad. Nauk SSSR 1976 226 853. 30 R. Massart R. Contant J.-M. Fruchart J.-P. Ciabrini and M. Fournier Inorg. Chem. 1977 16 2916. 31 L. C. Francesconi G. Graczyk S. Wehrli S.N. Shaikh D. McClinton S. Liu J. Zubieta and H. F. Kung Inorg. Chem. 1993 32 3114; C. S. John L. C. Francesconi S. Wehrli G. Grazyck P. A. Carroll and H. F. Kung Polyhedron 1992 11 1145; S. Sukal J. E. Bradshaw J. He G. P. A. Yap A. Rheingold H. F. Kung and L. C. Francesconi Polyhedron 1998 in the press. 32 M. J. Abrams C. E. Costello S. N. Shaikh and J. Zubieta Inorg. Chim. Acta. 1991 180 9. 33 Electrospray Ionization Mass Spectrometry Fundamentals Instrumentation and Applications ed. Richard B. Cole John Wiley and Sons New York 1997. 34 (a) T.-C. Lau J. Wang R. Guevremont and K. W. M. Siu J. Chem. Soc. Chem. Commun. 1995 877; (b) M. J. Deery O. W. Howarth and K. R. Jennings J. Chem Soc. Dalton Trans. 1997 4783. 35 K. S. Rothenberger M. J. Nilges T. E. Altman K. Glab W. Froncisz and J.S. Hyde Chem. Phys. Lett. 1986 124 295. 36 M. J. Nilges and R. L. Belford J. Magn. Reson. 1979 35 259. 37 G. D. Simpson PhD Thesis University of Illinois Urbana- Champaign 1978. 38 D. P. Smith H. So J. Bender and M. T. Pope Inorg. Chem. 1973 12 685. 39 J. Bendix M. Brorson and C. E. SchäVer Inorg. Chem. 1993 32 2838. 40 B. R. McGarvey Transition Met. Chem. 1966 3 89. 41 J. R. Morton and K. F. Preston J. Magn. Reson. 1978 30 577. 42 S. P. Harmalker M. A. Leparulo and M. T. Pope J. Am. Chem. Soc. 1983 105 4286. 43 W. Froncisz and J. S. Hyde J. Chem. Phys. 1980 73 3123. 44 For [RePW11O40]52 and [ReSiW11O40]62 markedly rhombic hyper- fine tensors were reported.10 We have found that this results when the non-coincidence between the g and A tensors is neglected. These systems like the molybdenum and rhenium systems reported here should have marked non-coincidence between g and A in the xz plane (yz plane according to our coordinate definition) and nearly axial A-tensors. Paper 8/05286C