Reactions of Tp–Os nitrido complexes with the nucleophiles hydroxide and thiosulfate

Inorganica Chimica Acta 359 (2006) 2842–2849 www.elsevier.com/locate/ica Reactions of Tp–Os nitrido complexes with the nucleophiles hydroxide and thiosulfate Adam Wu, Ahmad Dehestani, Erik Saganic, Thomas J. Crevier, Werner Kaminsky 1, Dawn E. Cohen 1, James M. Mayer * Department of Chemistry, University of Washington, Box 351700, Seattle, WA, 98195-1700, USA Received 14 October 2005; accepted 20 November 2005 Available online 5 January 2006 Dedicated to Professor Brian R. James on the occasion of his 70th birthday and in recognition of his many contributions to inorganic chemistry. Abstract The reaction between TpOs(N)Cl2 (1) [Tp = hydrotris(1-pyrazolyl)borate] and aqueous (nBu4N)(OH) in THF-d8 forms the nitrosyl complex TpOs(NO)Cl2 (5) among other products, suggesting an initial hydroxide attack at the nitrido ligand. In contrast, the reaction of the acetate complex TpOs(N)(OAc)2 (2) with NaOH in Me2CO/H2O yields the osmium bis-hydroxide complex TpOs(N)(OH)2 (3), which has been structurally characterized by single-crystal X-ray diffraction. Acetate for hydroxide exchange could occur by ligand substitution or by nucleophilic attack at the carbonyl carbon of the acetate ligands (saponification). Reacting 2 with Na18OH in H218O/ CD3CN yields predominantly doubly 18O-labeled TpOs(N)(18OH)2 (3-18O2) and unlabeled acetate, by ESI/MS and 13C{1H} NMR. This indicates that hydroxide reacts by substitution rather than by attack at the ligand. The reaction of 2 with the softer nucleophile thiosulfate occurs at the nitrido ligand, giving the thionitrosyl complex TpOs(NS)(OAc)2 (4). Reacting 4 with NaOH in (CD3)2CO/D2O also generates the bis-hydroxide complex 3.  2005 Elsevier B.V. All rights reserved. Keywords: Osmium; Tp; Hydroxide; Nitrido; Acetate; Thionitrosyl; Nitrosyl 1. Introduction Metal complexes typically involve electron-rich ligands binding to Lewis-acidic metal centers. Therefore, the addition of nucleophiles usually proceeds by attack at the metal center, while electrophiles typically add to ligands. In some complexes, however, these roles can be reversed. The nitrido ligand is interesting in this regard because it can react with both nucleophiles and electrophiles. Shapley and coworkers have reported that anionic OsVI-nitrido complexes with electron-donating alkyl ligands react with electrophilic methylating agents (Me3O)(BF4) or CF3SO3Me at the nitr* 1 Corresponding author. Tel.: +1 206 543 2083; fax: +1 206 685 8665. E-mail address: mayer@chem.washington.edu (J.M. Mayer). UW Chemistry crystallographic facility. 0020-1693/$ - see front matter  2005 Elsevier B.V. All rights reserved. doi:10.1016/j.ica.2005.11.033 ido ligand to produce imido complexes Os(NMe)R4 (R = Me, CH2SiMe3, CH2CMe3, or CH2Ph) [1]. Recently, protonation of (g5-C5H5)Os(N)(CH2SiMe3)2 by HBF4 or CF3SO3H has been found to occur at the nitrido ligand rather than at an alkyl group [2]. In contrast, the nitrido ligand in TpOs(N)Cl2 (1) [Tp = hydrotris(1-pyrazolyl)borate] is electrophilic. For example, PPh3 reacts rapidly with the nitride to form the phosphinidine complex TpOs(NPPh3)Cl2 [3], and PhMgCl and BPh3 effect Ph addition to the nitrido, yielding TpOs(NHPh)Cl2 after hydrolytic workup [4]. An unusual bicyclic Os-amido complex TpOs(NC6H8)Cl2, is prepared from 1 and 1,3-cyclohexadiene [5]. In these latter cases, there is a two-electron reduction of the metal center, OsVI ! OsIV, as is common when a nucleophile adds to a ligand. A formal fourelectron reduction to OsII occurs when a chalcogen atom 2843 A. Wu et al. / Inorganica Chimica Acta 359 (2006) 2842–2849 Scheme 1. is transferred to the nitrido ligand in 1, from Me3NO or elemental sulfur or selenium, to give the chalconitrosyl complexes TpOs(NE)Cl2 (E = O, S, or Se) [6]. Most of the reactions of 1 that we have examined occur at the nitrido ligand, while the chloride ligands remain inert. With the addition of AgOAc, however, the chloride ligands are replaced by acetate to form TpOs(N)(OAc)2 (2) (Scheme 1) [7]. Under acidic conditions, the acetate ligands in 2 are labilized by protonation, allowing preparation of a series of TpOs(N)X2 complexes using acidic HX reagents (X = Br, NO3, O2CCF3, O2CCCl3, O2CCBr3, or X2 = oxalate) [7]. Reported here are reactions of 2 and related compounds with the non-acidic nucleophilic reagents hydroxide and thiosulfate, leading to the syntheses of TpOs(N)(OH)2 (3) and TpOs(NS)(OAc)2 (4). The reactivity of 1, 2, and 3 with hydroxide, at the metal or at a ligand, is also described. 2. Results and discussion 2.1. Synthesis and spectroscopic characterization of TpOs(N)(OH)2 (3) Stirring TpOs(N)(OAc)2 (2) and 4 equiv. of NaOH in Me2CO/H2O at room temperature for 30 min yielded the orange bis-hydroxide complex TpOs(N)(OH)2 (3) in 60% yield, isolated via silica gel chromatography (Scheme 2). The 1H NMR spectrum of 3 displays a diamagnetic 2:2:2:1:1:1 integration pattern of six pyrazole resonances, and six pyrazole peaks are seen in the 13C{1H} NMR spectrum as well. Thus, 3 has Cs symmetry similar to 1, 2, and other TpOs(N)X2 compounds [7]. A resonance at d 5.09 observed for 3 in dry CD2Cl2 is assigned to the hydroxide protons. The addition of a drop of D2O results in an upfield shift of this peak, which indicates rapid exchange between TpOs(N)(OH)2 and water and confirms their assignment. The IR spectrum of 3 in KBr shows m(O–H) at 3420 cm1 and m(Os„N) at 1080 cm1, typical of TpOs(N)X2 complexes [7]. The labeled compound TpOs(15N)(OH)2 (3-15N) exhibits m(Os„15N) at 1044 cm1, very close to the value of 1046 cm1 calculated from a simple diatomic harmonic oscillator model [8]. The ESI/MS spectrum of 3 shows the most abundant peaks at 454 [M+H]+, 476 [M+Na]+, and 492 m/z [M+K]+; 3-15N shows similar isotopic clusters one m/z unit greater. Electrochemical reduction of 3 in MeCN is irreversible at 1.53 V (Ep,c) versus FeCp2 þ=0 , analogous to other TpOs(N)X2 complexes [7]. High valent metal–hydroxide complexes are not so common because they often readily condense to l-oxo derivatives. Other OsVI examples include trans-[OsO2(OH)4]2 [9,10], trans-[Os(N)(OH)(CN)4]2 [11], and trans–cisOsO2(OH)2(phen), which can be converted to the bis-loxo complex [OsO2(phen)(l-O)]2 in boiling water (phen = 1,10-phenanthroline) [12,13]. Related OsVI and RuVI bis-lhydroxo dimeric complexes [M(N)(CH2SiMe3)2(l-OH)]2 (M = Ru or Os) [14], ½OsO2 Lðl-OHÞ2 2 , and [Os(N)L(lOH)]2 have also been described (H2L = 1,2-bis(p-toluenesulfonylamido)benzene) [15]. The OsVIII complexes [OsO4(OH)] and cis-[OsO4(OH)2]2 are hydroxide adducts of OsO4 [9]. A ReV analog of 3, TpMe2ReO(OH)2, has been prepared by reduction of TpMe2ReO3 by PPh3 in H2O/ THF [TpMe2 = hydrotris(3,5-dimethylpyrazolyl)borate] [16]. 2.2. X-ray structure of TpOs(N)(OH)2 (3) Crystals of 3 were grown from CH2Cl2/hexanes solutions, and the structure was solved by direct methods. There are two independent molecules in the unit cell, one of which is disordered about the quasi-threefold axis of the TpOs unit. The discussion that follows uses the metrical parameters (Table 1) of the non-disordered molecule, which is drawn in Fig. 1 (see Section 4.7 for selected crystallographic data). The ORTEP of 3 shows a distorted octahedral molecule with all of the ligands bent away from Table 1 Selected bond lengths and angles for TpOs(N)(OH)2 (3)a Bond length (Å) Os(1)–N(1) Os(1)–N(3) Os(1)–N(5) Os(1)–N(7) Os(1)–O(1) Os(1)–O(2) B(1)–N(2) B(1)–N(4) B(1)–N(6) N(1)–N(2) N(3)–N(4) N(5)–N(6) Scheme 2. a Bond angle () 2.084(10) 2.099(10) 2.328(9) 1.651(10) 1.941(8) 1.956(7) 1.527(18) 1.544(18) 1.544(17) 1.384(13) 1.374(14) 1.356(12) N(1)–Os(1)–N(3) N(3)–Os(1)–N(5) N(1)–Os(1)–N(7) N(3)–Os(1)–N(7) N(5)–Os(1)–N(7) O(1)–Os(1)–N(7) O(2)–Os(1)–N(7) O(1)–Os(1)–O(2) O(1)–Os(1)–N(3) O(2)–Os(1)–N(1) O(2)–Os(1)–N(3) O(2)–Os(1)–N(5) 89.2(4) 78.5(4) 93.2(4) 90.7(4) 166.9(4) 106.0(4) 102.6(4) 87.7(3) 163.3(4) 164.1(3) 88.3(4) 84.6(3) Metrical data for the non-disordered molecule in the structure. 2844 A. Wu et al. / Inorganica Chimica Acta 359 (2006) 2842–2849 Scheme 3. Fig. 1. ORTEP of the non-disordered molecule in the structure of TpOs(N)(OH)2 (3). Hydrogen atoms are omitted for clarity, except for the hydroxides. the nitrido N(7), typical of complexes with a single nitrido or oxo group [17]. The Os„N distance of 1.651(10) Å is within the 1.602(20)–1.70(2) Å range of such bond lengths in TpOs(N)X2 complexes (X = O2CCF3, NO3, Cl, Me, Ph) [4a,7]. The Os–OH distances of 3 at 1.956(7) and 1.941(8) Å are shorter than those of trans–cis-OsO2(OH)2(phen) (1.982(4) and 1.984(5) Å) [12], trans-[Os(N)(OH)(CN)4]2 (2.123(5) Å) [11], trans-[OsO2(OH)4]2 (2.03 Å) [9,10], and cis-[OsO4(OH)2]2 (2.10 and 2.17 Å) [9], and those of bridging hydroxide complexes [15]. patterns centered at 456/458 [3-16O18O+H/3-18O2+H]+, 480 [3-18O2+Na]+, and 496 m/z [3-18O2+K]+ (Fig. 2a). The patterns for the Na+ and K+ clusters are roughly four m/z units greater than those of unlabeled 3 (Fig. 2b). Analysis of these clusters indicates an average isotopic composition of 84% 3-18O2, 16% 3-18O16O, and <3% 3-16O2 (each ±10%) (calculated isotopic patterns for 3-18O2 are shown in the insets of Fig. 2a). The protonated cluster at 456/458 m/z shows a higher abundance of the mono-labeled species [3-18O16O+H]+ (456 m/z), because protonation occurs at a hydroxide ligand and facilitates 16O exchange between [3-18O2+H]+ and trace H216O in the electrospray solvent. In a control experiment, unlabeled 3 was reacted with Na18OH in H218O/CH3CN. The ESI/MS spectrum of the reaction solution showed predominately peaks for unlabeled 3, with some 18O exchange observed for the 2.3. 18O-Labeled study: ligand substitution of TpOs(N)(OAc)2 (2) and Na18OH The formation of 3 from 2 and NaOH is a formal substitution of hydroxide for acetate. This contrasts with reactions of other nucleophiles such as alkoxides, which do not substitute for acetate. We therefore considered two possible mechanisms for this reaction: saponification via an initial hydroxide attack at the carbonyl carbon of the acetate ligands, or direct substitution of hydroxide for acetate. The mechanisms can be distinguished by reacting 2 with Na18OH. Ligand substitution would involve cleavage of the Os–O bonds to form TpOs(N)(18OH)2 (3-18O2) and unlabeled acetate, while saponification would retain the Os–O bonds and yield unlabeled 3 and singly 18O-labeled 18 OOCMe (Scheme 3). A saponification mechanism has previously been indicated for hydroxide reactions with [Co(NH3)5(O2CCX3)]2+ (X = H, Cl, or F) [18]. The reaction of 2 with Na18OH was conducted in CD3CN instead of acetone, in order to eliminate the possible exchange of 18O label with the solvent. Addition of 95%-enriched Na18OH to 2 in H218O/CD3CN produced isotopically enriched 3. The ESI/MS spectrum showed isotopic Fig. 2. Partial ESI/MS spectra of (a) labeled TpOs(N)(*OH)2 (mostly 3-18O2) from TpOs(N)(OAc)2 (2) + Na18OH and (b) TpOs(N)(OH)2 (3) from 2 + NaOH, with calculated isotopic patterns shown in the insets. A. Wu et al. / Inorganica Chimica Acta 359 (2006) 2842–2849 protonated cluster [3+H]+ (454 m/z). This shows that 3 does not exchange with basic H218O/CH3CN but that some exchange occurs under the acidic conditions of the electrospray experiment. The isotopic enrichment of the acetate product was probed by 13C{1H} NMR, using the known upfield shift of the carboxylate carbons on substitution of 16O for 18O [19]. For aqueous acetic acid, this shift is reported to be 25 ppb at pH 2.0 and 27 ppb at pH 8.0 [20]. 18O-Enriched sodium acetate was prepared from H218O hydrolysis of triethylorthoacetate (MeC(OEt)3) [21], and the 13C{1H} NMR spectrum (Fig. 3a) showed two carboxylate 13C resonances, for singly and doubly 18O-labeled acetate (40% 18 O2, 50% 18O16O, and <10% 16O2, based on integration and Lorentzian line fitting [22]). The 13C{1H} NMR spectrum of the 2 + Na18OH reaction mixture in H218O/ CD3CN showed a single peak in the carboxylate region (Fig. 3b). Spiking this solution with unlabeled NaOAc gave only a single carboxylate carbon resonance (Fig. 3c). Spiking with the Na18O2CMe/Na18O16OCMe prepared from triethylorthoacetate showed three carboxylate 13C peaks (Fig. 3d, Dd @ 27 ppb), assigned to unlabeled, singly, and doubly 18O-labeled acetate. Thus, we conclude that pre- 2845 dominantly unlabeled acetate (<20% 18O16OCMe, Fig. 3b) is formed in the reaction of 2 with Na18OH. These experiments must be done by spiking with authentic samples because the absolute chemical shift of the carboxylate carbon varies from sample to sample (d 180.2 ± 0.2), apparently due to changes in effective pH. In a control experiment, a reaction of 2 + Na16OH in 16 H2 O/CD3CN was spiked with Na18O2CMe/ 18 16 13 1 Na O OCMe. The C{ H} NMR spectrum showed three carboxylate 13C resonances, for all three isotopomers. Thus, under the reaction and spectroscopic conditions (ca. 12 h), exchange of the acetate oxygen atoms with the H216O solvent is too slow to significantly deplete the 18O label from the acetate. If exchange with the excess water had been fast, only 16O2CMe would have been observed. In sum, the ESI/MS and NMR data both indicate that 18 OH substitution for acetate occurs predominately with Os–O bond cleavage, yielding 3-18O2 and unlabeled acetate. The data rule out a saponification pathway, which would have yielded unlabeled 3 and 18OOCMe. The observation that hydroxide and not alkoxides substitute for acetate may be due to the small size of hydroxide. 2.4. Synthesis and spectroscopic characterization of TpOs(NS)(OAc)2 (4) Fig. 3. Partial 13C{1H} NMR spectra of the carboxylate carbon resonances of free acetate prepared from (a) H218O + MeC(OEt)3, (b) TpOs(N)(OAc)2 (2) + Na18OH in H218O/CD3CN, (c) 2 + Na18OH in H218O/CD3CN then + Na16O2CMe, and (d) 2 + Na18OH in H218O/ CD3CN then + Na18O2CMe/Na18O16OCMe. All peaks appear at d 180.2 ± 0.2, varying by a few tenths of ppm from sample to sample presumably depending on the pH. Refluxing 2 and 10 equiv. of sodium thiosulfate (Na2S2O3 Æ 5H2O) in MeCN for 2 h yielded the blue OsIIthionitrosyl complex TpOs(NS)(OAc)2 (4) in 41% yield, isolated via silica gel chromatography (Scheme 2). As with 3, the 1H and 13C{1H} NMR spectra of 4 indicate Cs symmetry. The EI/MS spectrum shows an isotopic cluster at 569 m/z (M+ for 4), which appears at 570 m/z for TpOs(15NS)(OAc)2 (4-15N). The IR spectrum of 4 shows a band at 1278 cm1, which appears at 1257 cm1 in 4-15N, consistent with a thionitrosyl ligand. This vibration likely has some of the character of a diatomic N–S stretch [m(N„S)] and some of the asymmetric stretch of a triatomic Os–N–S unit (see [23]). This is why a simple diatomic oscillator approximation overestimates the isotopic shift: calculated m(15N„S) = 1248 cm1. The cyclic voltammogram of 4 in MeCN shows reversible oxidation and reduction waves centered at 0.96 and 1.32 V versus FeCp2 þ=0 , respectively. These are 0.2 V more negative than those of the chloride analog, TpOs(NS)Cl2 (E1/2 = 1.20, 1.15 V) [6], indicating stronger electron donation of acetate versus chloride ligands. The thionitrosyl complex 4 is likely formed by direct sulfur atom transfer from thiosulfate to 2. This is a formal fourelectron reduction of the osmium, two-electron reduction of the central sulfur in thiosulfate, and a six-electron oxidation of the nitrido nitrogen atom. The use of thiosulfate as a sulfur atom donor to a nitrido ligand has previously been described in the formation of mer-L3Tc(NS)Cl2 (L = py, 4-Me-py, or 3,5-Me2-py) from (Bu4N)[Tc(N)Cl4], Na2S2O3, and L [24]. We find that thiosulfate can also be used to convert 1 to TpOs(NS)Cl2, which has previously been prepared 2846 A. Wu et al. / Inorganica Chimica Acta 359 (2006) 2842–2849 from 1 and propylene sulfide [6], S8 [6], CS2 þ N3  [25], or Mo(N)(S2CNEt2)3 [26]. Closely related thionitrosyl Os complexes [(tpm)Os(NS)Cl2]+ (tpm = tris(1-pyrazolyl)methane) [25] and trans-[(tpy)Os(NS)Cl2]+ (tpy = 2,2 0 :6 0 ,200 -terpyridine) [27] have also been reported. Dilworth et al. first synthesized thionitrosyl metal complexes Mo(NS)(S2CNR2)3 (R = Me, Et, or R2 = (CH2)5), from Mo(N)(S2CNR2)3 and propylene sulfide or S8 [28]. Crystals of 4 were grown from Me2CO/Et2O solutions, and the structure was examined by direct methods. The structure is disordered about a mirror plane containing C(1), C(2), Os(1), and C(13), with all of the other atoms being 50/50 disordered about this plane. The ORTEP diagram in Fig. 4 shows one set of atoms; a drawing showing all of the duplicated atoms is given in the Supporting information. Due to the extensive disorder, a large number of restraints were required to achieve a reasonable refinement. Thus, the bond lengths and angles are not well defined, and only evidence of connectivity is obtained (see Section 4.7 for selected crystallographic data). 2.5. Reactions of hydroxide with TpOs(N)Cl2 (1), TpOs(N)(OH)2 (3), and TpOs(NS)(OAc)2 (4) The reaction between 1 and 1.2 equiv. of aqueous (nBu4N)(OH) in THF-d8 immediately produces a number of products, including the known nitrosyl complex TpOs(NO)Cl2 (5) in 24% yield by 1H NMR integration. Complex 5 was identified by 1H NMR and IR spectra (e.g., m(N„O) = 1832 cm1). It was previously prepared from 1 by oxygen transfer from Me3NO [6], similar to the synthesis of [(tpy)Os(NO)Cl2]+ [29]. The reaction of 1 with NO also generates 5, in a multi-step reaction [30]. Column chromatography separates a number of the species of this reaction. One of the paramagnetic products was Fig. 4. ORTEP of one of the disordered positions of TpOs(NS)(OAc)2 (4), with hydrogen atoms omitted for clarity. Scheme 4. identified by ESI/MS (976 m/z) and IR (m(N„N) = 2011 cm1) as the l-N2 OsII–OsIII dimer [TpCl2Os–(lN2)–OsCl2Tp], previously reported by Meyer et al. [31]. The reaction of hydroxide to 1 likely proceeds by an initial hydroxide attack at the nitrido ligand, parallel to the addition of other nucleophiles such as phosphines and carbanions [3,4]. This would give an anionic OsIV(NOH) intermediate, which must undergo subsequent loss of two electrons and a proton to form the nitrosyl product 5 (Scheme 4). The reducing equivalents are in part consumed in the formation of [TpCl2Os–(l-N2)–OsCl2Tp], which can also be formed by reduction of 1 with an one-electron reductant cobaltocene [31]. We were not able to determine the fate of the proton (there does not appear to be any gas evolution in the reaction). While no OsNOH complexes have been reported, OsCl2(NHOH)(NO)(PPh3)2 is formed on protonation of Os(NO)2(PPh3)2 with 2 equiv. of HCl [32]. Complex 3 does not react with 4 equiv. of NaOH in CD3CN/D2O over half an hour at room temperature. After 16 h, 1H NMR indicated 19% decomposition of 3 and 11% of free Tp, which was confirmed by spiking with KTp. The addition of hydroxide to 3 is much slower than the analogous reaction of 1. This is most likely because reduction of 3 is less favorable, as indicated by the irreversible electrochemical reduction potentials: Ep,c = 1.53 V for 3, 1.34 V for 1 [7] (both versus FeCp2 þ=0 in MeCN). The more facile reduction of 1 favors hydroxide attack at the nitrido ligand to form the reduced OsIV(NOH) intermediate. Similarly, hydroxide does not rapidly react with the nitride in 2, due to an even more negative Ep,c (1.83 V [7]). In contrast, hydroxide reacts rapidly with 2 at the osmium, likely because acetate is a better leaving group than the chloride ligands in 1. The thionitrosyl complex 4 also reacts with hydroxide. Reacting 4 with 4 equiv. of NaOH in (CD3)2CO/D2O gave a 26% yield of TpOs(N)(OH)2 (3) in 30 min (by 1H NMR). This conversion involves both desulfurization of the thionitrosyl to a nitrido ligand and substitution of the acetate ligands of 4. The other osmium products were apparently paramagnetic, as 3 was the only Tp complex observed by 1 H NMR. ESI/MS confirmed 3 as a product, and showed a small amount of the nitrido-bis(acetate) complex 2 as well as other isotopic clusters for as yet unidentified osmium species. The fate of the sulfur is not evident. A possible pathway for this reaction would involve initial addition of hydroxide to the NS ligand in 4, resulting in desulfurization and formation of 2, which is then converted to 3. Desulfurization of thionitrosyl metal complexes with phosphines as reductants has been accomplished in a number of A. Wu et al. / Inorganica Chimica Acta 359 (2006) 2842–2849 cases, including Mo(NS)(S2CNR2)3 + PnBu3 ! Mo(N)(S2CNR2)3 (R = Me, Et, or R2 = (CH2)5) [28]. Treatment of TpOs(NS)Cl2 with PPh3 yields TpOs(NPPh3)Cl2 via the formation of SPPh3 and 1, which is rapidly trapped by a second equivalent of PPh3 [6]. [(tpm)Os(NS)Cl2]+ reacts with PPh3 analogously [25]. The reaction of [(tpm)OsII(NS)Cl2]+ with Me3NO gives [(tpm)OsIII(NSO)Cl2] by oxygen atom transfer to the sulfur [25]. Hydroxide is also known to add to the nitrogen of nitrosyl ligands in ruthenium complexes [33]. 3. Conclusions Hydroxide reacts with each of TpOs(N)(OAc)2 (2) and TpOs(NS)(OAc)2 (4) to form TpOs(N)(OH)2 (3). An 18Olabeling study indicates that the mechanism for the formation of 3 from 2 + hydroxide is direct substitution rather than saponification of the acetate ligands. In contrast, hydroxide does not substitute for the chloride ligands in TpOs(N)Cl2 (1), but instead reacts rapidly with the nitride to form the nitrosyl complex TpOs(NO)Cl2 (5) among other products. That hydroxide attacks the nitride ligand rapidly in 1 but not in 2 or 3 is likely due to the facility of reduction of the osmium centers in the three compounds, as indicated by the irreversible reduction potentials: Ep,c = 1.34 (1), 1.83 (2), and 1.53 V (3) (versus FeCp2 þ=0 in MeCN). 4. Experimental All reactions were conducted under air, unless stated otherwise. 1H and 13C{1H} NMR spectra were recorded on Bruker spectrometers (300 and 500 MHz) at ambient temperatures and referenced to a residual solvent peak: d (multiplicity, number of protons, assignment). All pyrazole resonances display 3JHH = 2 Hz. Electrospray ionization mass spectra (ESI/MS) were obtained on a Bruker Esquire-LC ion trap mass spectrometer, and reported as m/z for the most isotopically abundant peak in an Os isotopic pattern. Samples were infused as a MeCN solution and acquired in positive or negative ionization mode. Electron impact mass spectra (EI/MS) were obtained on a Kratos Profile HV-3 direct probe instrument. UV–Vis spectra were acquired with a Hewlett–Packard 8453 diode array spectrophotometer in anhydrous CH2Cl2, and reported as kmax/nm (e/M1 cm1). IR spectra were obtained as KBr pellets using Perkin–Elmer 1720 and Bruker Vector 33 FT-IR spectrometers, and are reported in cm1 at 4 cm1 resolution. Cyclic voltammetry (CV) measurements in 0.1 M (nBu4N)(PF6)/MeCN were performed using a Pt disc working electrode, a Pt wire auxiliary electrode, and an Ag wire/AgNO3 reference electrode with FeCp2 as an internal standard, and potentials are reported versus FeCp2 þ=0 . Elemental analyses were performed by Atlantic Microlab. All reagent grade solvents were purchased from Fisher Scientific or EMD Chemicals. Deuterated solvents and H218O (95% 18O-enrichment) were obtained from Cam- 2847 bridge Isotope Laboratories. CD2Cl2 was dried over CaH2 and vacuum transferred, and CD3CN over CaH2 then P2O5. (nBu4N)(OH) (40% wt. in H2O), NaH, NaOH, and NaOAc (Aldrich) were used without purification. KTp [34], TpOs(N)Cl2 (1) [4], TpOs(N)(OAc)2 (2) [7], and TpOs(15N)(OAc)2 (2-15N) [7] were prepared according to the literature procedures. 18 O-Enriched sodium acetate was prepared from H218O hydrolysis of triethylorthoacetate (MeC(OEt)3) following the literature procedure [21]. 1H NMR: 1.78 (s, CH3). 13 C{1H} NMR: 24.55 (CH3); 180.042 (C18O2), 180.065 (C18O16O). Integration of the 13C{1H} NMR spectrum indicated an isotopic composition of 40% 18O2, 50% 18 16 O O, and <10% 16O2. The NMR spectra were obtained in 350 lL CD3CN + 100 lL H216O containing 1 mg NaH, in order to be comparable to spectra obtained as part of the 18O-labeling study described in Section 4.3. 4.1. Synthesis of TpOs(N)(OH)2 (3) A solution of 2 (240 mg, 0.45 mmol) and NaOH (71 mg, 1.78 mmol) in Me2CO/H2O (20 mL/2 mL) was stirred for 30 min. The solvent was removed in vacuo, and the residue was chromatographed on silica gel with EtOAc/MeOH (90:10) and dried in vacuo at room temperature to yield 121 mg (0.27 mmol, 60%) of orange powder. 1H NMR (CD2Cl2): 6.00 (t), 7.32 (d), 7.45 (d) (1H each, pz); 6.49 (t), 7.91 (d), 8.17 (d) (2H each, pz 0 ); 5.09 (br s, 2H, OH). 13 C{1H} NMR (CD2Cl2): 105.69, 134.19, 141.51 (pz); 108.48, 138.54, 145.92 (pz 0 ). ESI/MS: 454 [M+H]+, 476 [M+Na]+, 492 [M+K]+. UV–Vis: 409 (210). IR: 1080 m(Os„N), 3420 m(O–H). CV: Ep,c = 1.53 V (OsVI/V). Anal. Calc. for C9H12N7O2BOs: C, 23.95; H, 2.68; N, 21.73. Found: C, 23.65; H, 2.57; N, 21.62%. TpOs(15N)(OH)2 (3-15N) was synthesized analogously from 2-15N. ESI/MS: 455 [M+H]+, 477 [M+Na]+, 493 [M+K]+. IR: 1044 m(Os„15N). 4.2. Synthesis of TpOs(NS)(OAc)2 (4) A suspension of 2 (100 mg, 0.19 mmol) and Na2S2O3 Æ 5H2O (464 mg, 1.87 mmol) in MeCN (30 mL) was refluxed for 2 h. The mixture was cooled to room temperature and filtered. The filtrate was evaporated to dryness, and the residue was chromatographed on silica gel with CH2Cl2/MeOH (90:10). The blue fraction was stripped to dryness, re-precipitated with CH2Cl2/hexanes, filtered off, and dried in vacuo at 78 C to yield 44 mg (0.08 mmol, 41%) of light blue powder. 1H NMR (CD2Cl2): 6.27 (t), 7.65 (d), 7.78 (d) (1H each, pz); 6.42 (t), 7.77 (d), 8.01 (d) (2H each, pz 0 ); 2.16 (s, 6H, CH3). 13 C{1H} NMR (CD2Cl2): 106.77, 136.50, 143.02 (pz); 108.30, 137.82, 145.79 (pz 0 ); 23.55 (CH3); 177.18 (CO2). EI/MS: 569 [M]+, 510 [MOAc]+, 464 [MOAcNS]+. UV–Vis: 413 (260), 642 (190). IR: 1278 m(N„S). CV: E1/2 = 0.96 (OsIII/II), 1.32 V (OsII/I). Anal. Calc. for C13H16N7O4BSOs: C, 27.52; H, 2.84; N, 17.28. Found: 2848 A. Wu et al. / Inorganica Chimica Acta 359 (2006) 2842–2849 C, 27.71; H, 2.81; N, 17.01%. TpOs(15NS)(OAc)2 (4-15N) was synthesized analogously from 2-15N. EI/MS: 570 [M]+, 511 [MOAc]+, 464 [MOAc15NS]+. IR: 1257 m(15N„S). 4.3. 18O-Labeled study: reaction of TpOs(N)(OAc)2 (2) and Na18OH NaH (1 mg, 0.04 mmol) and dry CD3CN (350 lL) were charged into a J-Young NMR tube, and H218O (100 lL) was added to generate a solution of Na18OH in H218O/ CD3CN. Complex 2 (5.6 mg, 0.01 mmol) was then added, and the tube was shaken until a solution was formed, to generate TpOs(N)(*OH)2 (mostly 3-18O2) and 16O2CMe. 13 C{1H} NMR: 180.209 (C16O2). The tube was then spiked with Na16O2CMe (3 mg), and 13C{1H} NMR was obtained: 180.281 (C16O2). The above experiment was duplicated except spiking with Na18O2CMe/Na18O16OCMe (3 mg). 13 C{1H} NMR: 180.044 (C18O2), 180.068 (C18O16O), 180.097 (C16O2). ESI/MS of 3-18O2: 458 [M+H]+, 480 [M+Na]+, 496 [M+K]+. In a control experiment, NaH (1 mg), CD3CN (350 lL), H216O (100 lL), and 2 (5.6 mg) were added into a J-Young NMR tube to generate unlabeled 3 and 16O2CMe, and the tube was spiked with Na18O2CMe/Na18O16OCMe (3 mg). 13C{1H} NMR: 180.186 (C18O2), 180.214 (C18O16O), 180.234 (C16O2). 4.4. Reaction of TpOs(N)Cl2 (1) with (nBu4N)(OH) In an NMR tube, (nBu4N)(OH) (40% wt. in H2O, 3.2 lL, 4.9 lmol) was added to a solution of 1 (2 mg, 4.1 lmol) in THF-d8 (400 lL), containing a small amount of C6H6 as an internal standard. The tube was shaken, and a 1H NMR spectrum showed TpOs(NO)Cl2 (5) in 24% yield, based on integration relative to the standard. The product was confirmed by comparing its 1H NMR spectrum with that of 5 prepared by the literature method [6]. 1H NMR: 6.27 (t), 7.78 (d), 7.96 (d) (1H each, pz); 6.49 (t), 7.94 (d), 8.00 (d) (2H each, pz 0 ). IR: 1832 m(N„O). Another product [TpCl2Os–(l-N2)–OsCl2Tp] [31] was identified by ESI/MS and IR; the observed ESI/MS isotopic pattern matched the calculated one. ESI/MS: 976 [M]. IR: 2011 m(N„N). 4.5. Reaction of TpOs(N)(OH)2 (3) with NaOH An NMR tube was charged with a solution of 3 (2 mg, 4.4 lmol) in CD3CN/D2O (350/100 lL) and a capillary containing (Me3Si)2O in CD3CN as a standard. After taking an 1H NMR, 18 lL of 1 M NaOH/D2O (18 lmol) was added. No change by NMR was detected after 30 min. 81% of 3 remained after 16 h, at which time 11% of free Tp was observed, based on integration relative to the standard. The identity of Tp was confirmed by spiking with authentic KTp. 1H NMR of 3: 6.02 (t), 7.29 (d), 7.55 (d) (1H each, pz); 6.52 (t), 8.03 (d), 8.19 (d) (2H each, pz 0 ); free Tp: 6.13 (t), 7.31 (d), 7.46 (d) (1H each, pz). 4.6. Reaction of TpOs(NS)(OAc)2 (4) with NaOH An NMR tube was charged with a solution of 4 (2 mg, 3.5 lmol) in (CD3)2CO/D2O (350/100 lL) and a capillary containing (Me3Si)2O in CD3CN as a standard. After taking an 1H NMR, 14 lL of 1 M NaOH/H2O (14 lmol) was added, and the tube was allowed to stand for 30 min. 1H NMR showed the formation of 3 in 26% yield, based on integration relative to the standard. The tube was spiked with authentic 3 to confirm the identity of the product. 1 H NMR: 5.97 (t), 7.33 (d), 7.55 (d) (1H each, pz); 6.53 (t), 8.07 (d), 8.23 (d) (2H each, pz 0 ). ESI/MS: 454 [M+H]+, 476 [M+Na]+, 492 [M+K]+. 4.7. X-ray structural determination of TpOs(N)(OH)2 (3) and TpOs(NS)(OAc)2 (4) Crystals of 3 were obtained from slow evaporation of CH2Cl2/hexanes solutions and were mounted onto a glass capillary with oil. The data were collected on a Nonius Kappa CCD diffractometer. Selected crystallographic data for 3: C9H12BN7O2Os, formula weight = 451.27, 0.14 · 0.14 · 0.10 mm, monoclinic, space group P21/c (No. 14), a = 13.3050(9) Å, b = 14.6520(13) Å, c = 15.0850(8) Å, b = 115.984(4), V = 2643.5(4) Å3, qcalc = 2.268 Mg m3, Z = 8, 2hmax = 4.14–24.71, Mo Ka radiation (k = 0.71070 Å), F(0 0 0) = 1696, T = 130(2) K, total/independent reflections = 21 816/4287 (Rint = 6.18%), observed data = 7165 (I > 2 r(I)), restraints/parameters = 3/377, absorption correction: semi-empirical (hklSCALEPACK), maximum (minimum) transmission: 0.4450 (0.1037), R1 (wR2) = 5.21 (10.4)% for I > 2r(I), R1 (wR2) = 9.27 (11.5)% for all data, GOF = 0.958, largest difference in peak (hole) = 1.307 (1.581) e Å3. Solution by direct methods (SIR-92) produced a complete heavy-atom phasing model consistent with the proposed structure. The heavy atoms were refined anisotropically by full-matrix least-squares, and the hydrogen atoms were placed using a riding model. However, one of the two chemically identical molecules was found to be disordered about the TpOs pseudo-threefold axis. O3b, O4b, and N14b of the minor form, 19.07(3)% of the molecules were refined isotropically with fixed thermal parameters (0.05). Crystals of 4 were grown from slow evaporation of Me2CO/Et2O solutions. Selected crystallographic data for 4: C13H16BN7O4SOs, formula weight = 567.44, 0.17 · 0.17 · 0.08 mm, orthorhombic, space group Pbcm (No. 57), a = 9.2380(6) Å, b = 15.1010(10) Å, c = 13.3830(10) Å, a = b = c = 90, V = 1867.0(2) Å3, qcalc = 2.019 Mg m3, Z = 4, 2hmax = 2.20–28.31, Mo Ka radiation (k = 0.71073 Å), F(0 0 0) = 1088.0, T = 130(2) K, total/independent reflections = 29 644/2393 (Rint = 7.44%), observed data = 4254 (I > 2r(I)), restraints/parameters = 64/228, absorption correction: semi-empirical from equivalents, maximum (minimum) transmission: 0.58 (0.32), R1 (wR2) = 5.19 (11.9)% for I > 2 r(I), R1 (wR2) = 12.7 (15.1)% for all data, GOF = 1.011, largest difference in A. Wu et al. / Inorganica Chimica Acta 359 (2006) 2842–2849 peak (hole) = 0.880 (1.557) e Å3. The structure of 4 was solved by direct methods (SIR-97), and the heavy and hydrogen atoms were refined similar to those of 3. The structure is disordered, and one set of the molecule is reflected into the other by application of a crystallographic mirror plane through C(1), C(2), Os(1), and C(13). Additional crystallographic information is given in the Supporting information. Acknowledgements We are grateful to Dr. Martin Sadı́lek and Mr. Loren Kruse for assistance with mass spectrometry, and Dr. Eric Watson and Dr. Ian Rhile for helpful discussions. We thank the US National Science Foundation for financial support to J.M.M. (CHE0204697), and for funds toward the purchase of the Esquire-LC mass spectrometer (CHE9807748). Appendix A. Supporting information CCDC 243292 and 286567 contain the crystallographic data for 3 and 4. The data can be obtained free of charge via www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cambridge Crystallographic Data Center, 12, Union Road, Cambridge CB2 1EZ, UK; fax: +44 1223 336 033; or deposit@ccdc.cam.ac.uk). Supplementary data associated with this article, including additional discussion of the structure of 4, with ORTEP and the unit cell diagrams showing all the disordered atoms can be found in the online version, at doi:10.1016/j.ica.2005.11.033. References [1] P.A. Shapley, Z.-Y. Own, J.C. Huffman, Organometallics 5 (1986) 1269. [2] C.M. Lutz, S.R. Wilson, P.A. Shapley, Organometallics 24 (2005) 3350. [3] B.K. Bennett, E. Saganic, S. Lovell, W. Kaminsky, A. Samuel, J.M. Mayer, Inorg. Chem. 42 (2003) 4127. [4] (a) T.J. Crevier, B.K. Bennett, J.D. Soper, J.A. Bowman, A. Dehestani, D.A. Hrovat, S. Lovell, W. Kaminsky, J.M. Mayer, J. Am. Chem. Soc. 123 (2001) 1059; (b) T.J. Crevier, J.M. Mayer, J. Am. Chem. Soc. 120 (1998) 5595; (c) T.J. Crevier, J.M. Mayer, Angew. Chem. Int. Ed. 37 (1998) 1891. [5] A.G. Maestri, K.S. Cherry, J.J. Toboni, S.N. Brown, J. Am. Chem. Soc. 123 (2001) 7459. [6] T.J. Crevier, S. Lovell, J.M. Mayer, A.L. Rheingold, I.A. Guzei, J. Am. Chem. Soc. 120 (1998) 6607. [7] A. Dehestani, W. Kaminsky, J.M. Mayer, Inorg. Chem. 42 (2003) 605. 2849 [8] R.M. Silverstein, F.X. Webster, in: Spectrometric Identification of Organic Compounds, sixth ed., Wiley, New York, 1998, p. 71. [9] W.P. Griffith, in: G. Wilkinson (Ed.), Comprehensive Coordination Chemistry, vol. 4, Pergamon Press, Oxford, 1987, p. 519. [10] (a) L.O. Atovmyan, V.G. Andrianov, M.A. Porai-Koshits, J. Struct. Chem. 3 (1962) 660 (Engl. Transl.); (b) M.A. Porai-Koshits, L.O. Atovmyan, V.G. Andrianov, J. Struct. Chem. 2 (1961) 686 (Engl. Transl.). [11] H.J. van der Westhuizen, S.S. Basson, J.G. Leipoldt, W. Purcell, Transition Met. Chem. 19 (1994) 582. [12] A.M.R. Galas, M.B. Hursthouse, E.J. Behrman, W.R. Midden, G. Green, W.P. Griffith, Transition Met. Chem. 6 (1981) 194. [13] C.-H. Chang, W.R. Midden, J.S. Deetz, E.J. Behrman, Inorg. Chem. 18 (1979) 1364. [14] (a) P.A. Shapley, J.J. Schwab, S.R. Wilson, J. Coord. Chem. 32 (1994) 213; (b) P.A. Shapley, W.A. Reinerth, Organometallics 15 (1996) 5090. [15] W.-H. Leung, E.K.-F. Chow, S.-M. Peng, Polyhedron 12 (1993) 1635. [16] K.P. Gable, E.C. Brown, J. Am. Chem. Soc. 125 (2003) 11018. [17] W.A. Nugent, J.M. Mayer, in: Metal–Ligand Multiple Bonds, Wiley, New York, 1988, p. 157. [18] F. Basolo, R.G. Pearson, in: Mechanisms of Inorganic Reactions: A Study of Metal Complexes in Solution, second ed., Wiley, New York, 1967, p. 229. [19] P.E. Hansen, in: G.A. Webb (Ed.), Annual Reports on NMR Spectroscopy, vol. 15, Academic Press, London, 1983, p. 105. [20] J.M. Risley, R.L. Van Etten, J. Am. Chem. Soc. 103 (1981) 4389. [21] C.R. Hutchinson, C.T. Mabuni, J. Labelled Compd. Radiopharm. 13 (1977) 571. [22] NUTS – NMR Utility Transform Software for Windows, 1D version; Acorn NMR Inc., 2003. [23] (a) A.M. English, K.R. Plowman, I.S. Butler, Inorg. Chem. 20 (1981) 2553; (b) A.M. English, K.R. Plowman, I.S. Butler, Inorg. Chem. 21 (1982) 338. [24] J. Lu, M.J. Clarke, Inorg. Chem. 29 (1990) 4123. [25] E.-S. El-Samanody, K.D. Demadis, L.A. Gallagher, T.J. Meyer, P.S. White, Inorg. Chem. 38 (1999) 3329. [26] S.B. Seymore, S.N. Brown, Inorg. Chem. 41 (2002) 462. [27] K.D. Demadis, T.J. Meyer, P.S. White, Inorg. Chem. 37 (1998) 3610. [28] (a) J. Chatt, J.R. Dilworth, J. Chem. Soc., Chem. Commun. (1974) 508; (b) M.W. Bishop, J. Chatt, J.R. Dilworth, J. Chem. Soc., Dalton Trans. (1979) 1. [29] D.S. Williams, T.J. Meyer, P.S. White, J. Am. Chem. Soc. 117 (1995) 823. [30] M.R. McCarthy, T.J. Crevier, B. Bennett, A. Dehestani, J.M. Mayer, J. Am. Chem. Soc. 122 (2000) 12391. [31] K.D. Demadis, E.-S. El-Samanody, G.M. Coia, T.J. Meyer, J. Am. Chem. Soc. 121 (1999) 535. [32] K.R. Grundy, C.A. Reed, W.R. Roper, J. Chem. Soc., Chem. Commun. (1970) 1501. [33] F. Roncaroli, M.E. Ruggiero, D.W. Franco, G.L. Estiú, J.A. Olabe, Inorg. Chem. 41 (2002) 5760. [34] S. Trofimenko, J. Am. Chem. Soc. 89 (1967) 3170.