View PDF VersionPrevious ArticleNext Article
 
DOI: 10.1039/B411487F (Paper) New J. Chem., 2005, 29, 193-204

New d4 dihydrides of Ru(IV) and Os(IV) with π-donor ligands: M(H)2(chelate)(PiPr3)2 with chelate = ortho-XYC6H4 with X, Y = O, NR; R = H or CH3

German Ferrando-Miguel , Peng Wu , John C. Huffman and Kenneth G. Caulton *
Department of Chemistry and Molecular Structure Center, Indiana University, Bloomington, IN 47405-7102. E-mail: caulton@indiana.edu

Received (in Durham, UK) 27th July 2004 , Accepted 22nd November 2004

First published on 15th December 2004

Abstract

The synthesis and characterization of Ru(H)2(ortho-OC6H4E)L2 (L = PiPr3, E = NH, O) show these to be dihydrides with a nonoctahedral structure. The former compound reacts with H2 to give Ru(H)3(OC6H4NH2)L2, which has the ability to hydrogenate tBuHC[double bond, length as m-dash]CH2. Osmium analogs are available from Os(H)3ClL2, and the mono-N-methyl example Os(H)2[N(Me)C6H4NH]L2 shows inequivalent hydrides by 1H NMR at 20 °C. It exchanges with D2 faster into the NH site than into the OsH sites. Triflic acid protonates Os(H)2[N(Me)C6H4NH]L2 at the metal to give a trihydride. Catechol protonates Os(H)2[N(Me)C6H4NH]L2 to displace the ortho-diamine to give Os(H)2(OC6H4O)L2. The collective evidence is consistent with a dianionic, not a quinoid oxidation state for the chelate ligands and a d4, six-coordinate potential energy surface, often with low barriers between alternative non-octahedral structures.

Introduction

Six-coordinate species of ruthenium and osmium in oxidation state +4 represent an interesting development, especially when they are unsaturated. Synthetic access to M(H)2Cl2L2 (L = PiPr3) in recent years1,2 was followed by an exploration of their reactivity towards alkynes (yielding carbenes and carbynes), as well as binding H2 and catalytic alkene and alkyne hydrogenation.3–11 The molecules M(H)2Cl2L2 are exceptional in not being octahedral, but also in having the two bulky phosphines cisoid (∠P–M–P ∼112°). The actual structure has been deduced2,12,13 to be a compromise between the conflicting dictates of the d4 electronic configuration and the minimization of steric pressures and this feature of the d4 configuration has recently been generalized.14,15 In part because these are crowded molecules, reactivity can sometimes require heating (atypical for 16-electron species). Finally, because the +2 oxidation state is preferred (or at least very common) for Ru and Os, reactivity by reductive elimination can occur, most often by loss of HCl to a Brønsted basic reagent, if it is available.12,16

We report here the outcome of efforts to broaden the class of molecules M(H)2X2L2 by incorporation of chelating ortho catecholates or diamides A. These permit further exploration of the unusual nonoctahedral structure of this class of compounds and, when the two donor groups E are not identical, provide further insight into the fluxionality of these molecules.

ugraphic, filename = b411487f-u1.gif

Experimental

General

All manipulations were performed using standard Schlenk techniques or in an argon-filled glovebox unless otherwise noted. Solvents were distilled from Na, Na/benzophenone, P2O5, or CaH2, degassed prior to use, and stored over 4 Å molecular sieves in airtight vessels. Lithium salts o-C6H4(OH)OLi and o-C6H4(NH2)OLi were prepared by reacting nBuLi with catechol (see below) and 2-aminophenol (1∶1 molar ratio) in benzene, respectively, then isolated as white solids by removal of the volatiles in vacuo. [RuHCl(PiPr3)2]2,17 written in this section as a monomer to clarify reaction stoichiometries, and Os(H)3Cl(PiPr3)218 were prepared as previously reported; all other reagents were used as received from commercial vendors. 1H NMR chemical shifts are reported in ppm relative to protio impurities in the deutero solvents. 31P NMR spectra are referenced to an external standard of 85% H3PO4 (0 ppm). NMR spectra were recorded with either a Varian Gemini 2000 (300 MHz 1H; 121 MHz 31P; 75 MHz 13C), a Varian Unity INOVA instrument (400 MHz 1H; 162 MHz 31P; 101 MHz 13C), or a Varian Unity INOVA instrument (500 MHz 1H, 126 MHz 13C). The following abbreviations are used: s = singlet, d = doublet, dd = doublet of doublets, dt = doublet of triplets, t = triplet, td = triplet of doublets, q = quartet, vt = virtual triplet, dvt = doublet of virtual triplets, m = multiplet, br = broad, ap = apparent. Infrared spectra were recorded using a Nicolet 510P FT-IR spectrometer.
o-C6H4(OH)OLi. In a 100 ml Schlenk flask, 5.11 ml (8.18 mmol) of BuLi (1.6 M in hexane) was added via syringe to a solution of 0.9 g (8.18 mmol) catechol in 40 ml of benzene. White precipitate formed immediately. After stirring for another 0.5 h, the solvent was removed under reduced pressure. The white product was washed by hexane and dried in vacuo. IR: vOH 3487 cm−1.
Ru(H)2(PiPr3)2(OC6H4O). In a 100 ml Schlenk flask, 116 mg C6H4(OH)OLi (1 mmol) was added to a solution of RuHCl(PiPr3)2 (458 mg, 1 mmol) in 40 ml benzene. C6H4(OH)OLi dissolved gradually and the resulting red brown solution was stirred for 14 h. The reaction mixture was then passed through a 1 cm plug of neutral alumina with benzene and taken to dryness. Yield: 480 mg (90%). 1H NMR (300 MHz, C6D6, 20 °C): δ − 13.63 (t, 2JP–H = 35 Hz, 2H Ru–H), 1.10 (dd, JP–H=14 Hz, 3JH–H = 6.9 Hz, 36H, P(CHMe2)3), 2.18 (m, 6H, PCH), 7.06 (m, 2H, Ar–H), 7.53 (m, 2H, Ar–H). 31P{1H} NMR (121 MHz, C6D6, 20 °C): δ 90.5 (s).
Isomerization of 1-pentene catalyzed by Ru(H)2(PiPr3)2(OC6H4O). A 5.0 mg (9.4 μmol) amount of Ru(H)2(PiPr3)2(OC6H4O) was dissolved in 0.5 ml of C6D6 in an NMR tube. Via syringe, 50 μl (0.46 mmol) of 1-pentene was added. The tube was fitted with a Teflon stopcock, shaken and heated to 55 °C. The sample was monitored periodically by 1H and 31P{1H} NMR and showed only the signals of Ru(H)2(PiPr3)2(OC6H4O). 1H NMR showed 94% conversion to 2-pentene in 80 h (trans∶cis = 2.19∶1).
Ru(CO)2(PiPr3)2(OC6H4O). A solution of Ru(H)2(PiPr3)2(OC6H4O) (15 mg, 0.028 mmol) in C6D6 (0.5 ml) was placed in an NMR tube fitted with a Teflon stopcock. The solution was frozen in liquid N2, the headspace was evacuated, and excess CO (1 atm.) was introduced into the tube. When the solution warmed to room temperature and the tube was shaken, the solution color immediately changed from deep red brown to pale yellow. 1H and 31P{1H} NMR and IR spectra showed complete conversion to Ru(CO)2(PiPr3)2(OC6H4O). 1H NMR (300 MHz, C6D6, 20 °C): δ 1.17 (dvt, JP–H = 6.3 Hz, 3JH–H = 7.2 Hz, 36H, P(CHMe2)3), 2.16 (m, 6H, P(CHMe2)3), 6.81 (m, 2H, Ar–H), 6.97 (m, 2H, Ar–H). The peak of free H2 is observed at 4.50 ppm. 31P{1H} (121 MHz, C6D6, 20 °C): δ 43.4 (s). 13C{1H} NMR (75 MHz, C6D6, 20 oC): δ 19.9 (s, P(CHMe2)3), 24.5 (vt, JP–C = 9.4 Hz, P(CHMe2)3), 115.4 (s, Ar–C), 116.1 (s, Ar–C), 162.1 (s, Ar–C), 184.4 (CO). IR: vCO (benzene) 2014, 1945 cm−1.
Synthesis of Ru(H)2(PiPr3)2(O(3-CH3)C6H3O). A solution of RuHCl(PiPr3)2(20 mg, 0.044 mmol) in benzene-d6 (0.5 ml) was placed in an NMR tube with a Teflon stopcock. Lithium 3-methyl-catecholate (5.6 mg, 0.044 mmol) was added and the tube was shaken. The lithium salt dissolved gradually. 1H and 31P{1H} NMR showed complete conversion to Ru(H)2(PiPr3)2(O2C7H6) in 14 h. 1H NMR (300 MHz, C6D6, 20 °C): δ −13.74 (t, 2JP–H = 35.7 Hz, 2H Ru–H), 1.15 (dd, JP–H = 13.5 Hz, 3JH–H = 7.2 Hz, 36H, P(CHMe2)), 2.22 (m, 6H, P(CHMe2)), 2.71 (s, 3H, CH3–Ar), 6.97 (d, 1H, Ar–H), 7.03 (dd, 1H, Ar–H), 7.44 (d, 1H, Ar–H). 31P{1H} NMR (121 MHz, C6D6, 20 °C): 89.5 (s).
Synthesis of Ru(H)2(PiPr3)2(OC6H4NH). In a 100 ml Schlenk flask, 115 mg (o)-C6H4(NH2)OLi (1 mmol) was added to a solution of RuHCl(PiPr3)2 (458 mg, 1 mmol) in 40 ml toluene. (o)-C6H4(NH2)OLi dissolved gradually and the resulting red brown solution was stirred for 6 h. After separation of LiCl by filtration, the solution was taken to dryness. Yield: 488 mg (92%). 1H NMR (300 MHz, C7D8, 20 °C): δ −11.46 (t, 2JP–H = 35 Hz, 2H, Ru–H), 1.05 (dd, 3JP–H = 13 Hz, 3JH–H = 6.8 Hz, 36H, P(CHMe2)3), 1.99 (m, 6H, P(CHMe2)3), 6.92 (t, 1H, Ar–H), 7.06 (t, 1H, Ar–H), 7.14 (d, 1H, Ar–H), 7.45 (d, 1H, Ar–H), 7.74 (s, 1H, N–H), 31P{1H} NMR (121 MHz, C7D8, 20 °C): δ 95.1 (s). 13C{1H} NMR (75 MHz, C7D8, 20 °C): 19.7 (s, P(CHMe2)3), 28.6 (vt, JP–C = 11.4 Hz, P(CHMe2)3), 29.0 (vt, JP–C = 11.5 Hz, P(CHMe2)3), 115.3 (s, Ar–C), 116.5 (s, Ar–C), 116.7 (s, Ar–C), 119.0 (s, Ar–C), 150.6 (s, Ar–C), 164.2 (s, Ar–C).
RuH3(PiPr3)2(OC6H4NH2). A solution of Ru(H)2(PiPr3)2(OC6H4NH) (20 mg, 0.038 mmol) in C7D8 (0.5 ml) was placed in an NMR tube fitted with a Teflon stopcock. The solution was frozen in liquid N2, the headspace was evacuated, and excess H2 (1 atm.) was introduced into the tube. When the solution warmed to room temperature and the tube was shaken, the solution color changed from deep red brown to pale yellow in less than 1 h. 1H and 31P{1H} NMR spectra showed complete conversion to RuH3(PiPr3)2(OC6H4NH2). 1H NMR (300 MHz, C7D8, 20 °C): −12.39 (t, 2JP–H = 14.7 Hz, 3H, Ru–H), 1.05 (dd, 3JP–H = 11.7 Hz, 3JH–H = 6.3 Hz, 18H, P(CHMe2)3), 1.13 (dd, 3JP–H = 12.9 Hz, 3JH–H = 6.3 Hz, 18H, P(CHMe2)3), 1.76 (m, 6H, P(CHMe2)3), 4.34 (s, 2H, NH2), 6.35 (m, 1H, Ar–H), 6.81 (d, 1H, Ar–H), 6.99 (m, 2H, Ar–H). 31P{1H} NMR (121 MHz, C7D8, 20 °C): δ 60.4 (s).
Synthesis of (η5-(o)-CH3-C6H4O)RuH(PiPr3)2. A solution of RuHCl(PiPr3)2 (20 mg, 0.044 mmol) in benzene-d6 (0.5 ml) was placed in an NMR tube with a Teflon stopcock. Lithium 2-methylphenolate (5 mg, 0.044 mmol) was added and the tube was shaken. The lithium salt dissolved gradually and the solution color became lighter. 1H and 31P{1H} NMR showed complete conversion to (η5-(o)-CH3-C6H4O)RuH(PiPr3)2 in 14 h. 1H NMR (300 MHz, C6D6, 20 °C): δ −14.15 (ap t, 2JP–H = 37.5 Hz, 1H, Ru–H), 0.96 (dd, 3JP–H = 11.7 Hz, 3JH–H = 7.2 Hz, 9H, P(CHMe2), 1.02 (dd, 3JP–H = 12 Hz, 3JH–H = 6.6 Hz, 9H, P(CHMe2), 1.24 (m, 18H, P(CHMe2), 1.679 (m, 3H, P(CHMe2), 2.18 (s, 3H, Ar–CH3), 2.27 (m, 3H, P(CHMe2), 3.85 (br, 1H, Ar–H), 4.66 (br, 1H, Ar–H), 5.20 (br, 1H, Ar–H), 5.58 (br, 1H, Ar–H). 31P{1H} NMR (121 MHz, C6D6, 20 °C): δ 62.8 (s), 65.9 (s). 13C{1H} NMR (75 MHz, C6D6, 20 °C): δ 16.7 (s, Ar–CH3), 20.4 (s, P(CHMe2), 20.7 (s, P(CHMe2)), 21.0 (s, P(CHMe2)), 21.7 (s, P(CHMe2)), 28.1 (d, 1JP–C = 17.5 Hz, P(CHMe2)), 29.0 (d, 1JP–C = 17.2 Hz, P(CHMe2)), 67.8 (s, Ar–C), 75.8 (s, Ar–C), 91.3 (s, Ar–C), 97.8 (s, Ar–C), 101.4 (s, Ar–C), 165.5 (s, Ar–C).
Synthesis of Os(H)2(N(H)C6H4NH)(PiPr3)2. In an NMR tube, Os(H)3Cl(PiPr3)2 (0.01g, 0.018 mmol) and 1,2-phenylenediamine (0.004 g, 0.036 mmol) were dissolved in 0.5 ml of benzene-d6. The reaction proceeds for 48 h and a white precipitate was observed (the corresponding hydrochloride salt of the amine reagent). After 48 h the solution had turned yellow-orange and the volatiles were removed under vacuo. The mixture consists of ca. equimolar amounts of the title compound and the amine hydrochloride. 1H NMR (300 MHz, C6D6, 20 °C): −13.51 (t, J(H–P) = 38 Hz, Os–H, 2H), 1.00 (dvt, N = 19.2 Hz, Os–P(CH(CH3)2)), 1.97 (m, Os–P(CH(CH3)2)), 7.11 and 7.44 (AA′MM′ system, 4H), 8.34 (br, Os–[N(H)C6H4NH], 2H). 31P{1H} NMR (121.4 MHz, C6D6, 20 °C): 43.0.
Synthesis of Os(H)2(N(CH3)C6H4NH)(PiPr3)2. The compound Os(H)3Cl(PiPr3)2 (0.1 g, 0.18 mmol) and N-methyl-1,2-phenylenediamine (44.4 μl, 0.36 mmol) were dissolved in 10 ml of benzene. After 30 min at room temperature tert-butylethylene (25 μl, 0.18 mmol) was added to the reaction mixture via syringe. The reaction was allowed to proceed for 24 h at room temperature. A white precipitate was observed to form in solution. The volatiles were removed under vacuo and the product was extracted with pentane (2 × 5 ml). The volatiles were removed under vacuo to produce a red solid. The 1H NMR spectrum shows that this solid still contains some free amine. Several attempts to further purify the compound were unsuccessful. 1H NMR (300 MHz, C6D6, 20 °C): −12.02 (t, J(H–P) = 35 Hz, Os–H, 1H), −13.67 (td, J(H–P) = 37 Hz, J(H–H) = 3 Hz, Os–H, 1H), 0.88 (dvt, N = 18.3 Hz, Os–P(CH(CH3)2)), 0.90 (dvt, N = 19.2 Hz, Os–P(CH(CH3)2)), 1.98 (m, Os–P(CH(CH3)2)), 4.41 (t, J(H–P) = 1.8 Hz, Os–[N(CH3)C6H4NH], 3H), 7.10 (vt, J(H–H) = 7.4 Hz, aryl H, 1H), 7.21 (vt, J(H–H) = 7.2 Hz, aryl H, 1H), 7.41 (d, J(H–H) = 8.4 Hz, aryl H, 1H), 7.55 (d, J(H–H) = 8.4 Hz, aryl H, 1H), 8.85 (br, Os–[N(CH3)C6H4NH], 1H). 31P {1H} NMR (121.4 MHz, C6D6, 20 °C): 54.8.
Synthesis of [OsH3(N(Me)C6H4NH)(PiPr3)2]+ O3SCF3−. In an NMR tube, Os(H)2(N(CH3)C6H4NH)(PiPr3)2 (0.01g, 0.016 mmol) was dissolved in 0.5 ml of benzene-d6 and triflic acid (0.7 μl, 0.016 mmol) was added to the solution via syringe. The reaction is instantaneous and the solution turned green-brown. The volatiles were removed under vacuo and a greenish residue was obtained. 1H NMR (300 MHz, C6D6, 20 °C): −8.15 (td, J(H–P) = 16 Hz, J(H–H) = 3 Hz, Os–H, 3H), 0.70 (dvt, N = 21.9 Hz, Os–P(CH(CH3)2)), 0.75 (dvt, N = 21.9 Hz, Os–P(CH(CH3)2)), 1.72 (m, Os–P(CH(CH3)2)), 4.43 (s, OsN(CH3)C6H4NH, 3H), 6.79 (m, aryl H, 2H), 7.02 (d, J(H–H) = 9 Hz, aryl H, 1H), 8.46 (d, J(H–H) = 9 Hz, aryl H, 1H), 14.65 (br, OsN(CH3)C6H4NH, 1H). 31P {1H} NMR (121.4 MHz, C6D6, 20 °C): 41.8. 13C {1H} NMR (100 MHz, C6D6, 20 °C): 19.4 (s, Os–P(CH(CH3)2)), 29.5 (s, Os–P(CH(CH3)2)), 54.3 (s, OsN(CH3)C6H4NH2), 113.9, 123.0, 124.6, 125.9 (s, aryl C), 156.3, 158.9, (s, aryl ipso-C). 1H NMR (300 MHz, d8-THF, −85 °C): −8.03 (t, J(H–P) = 16 Hz, Os–H, 3H, T1 min (−84 °C) = 85 ms).
Reaction of Os(H)2(N(CH3)C6H4NH)(PiPr3)2 with catechol. In an NMR tube, Os(H)2(N(CH3)C6H4NH)(PiPr3)2 (0.01 g, 0.016 mmol) and catechol (0.007 g, 0.064 mmol) were dissolved in 0.5 ml of toluene-d8. After 5 min of mixing the reagents at room temperature, 5 mol% population of the compound [OsH3(N(Me)C6H4NH)(PiPr3)2]+ is detected in the NMR tube, along with unreacted diamido compound and 1 mol% of Os(H)2(OC6H4O)(PiPr3)2, the substitution product. When the reaction is allowed to proceed for 8 h, only traces of the cationic trihydrido compound remain and the major product is Os(H)2(OC6H4O)(PiPr3)2. Vacuum removal of the volatiles under vacuo with constant heating does not remove the excess catechol or the released free amine. The substitution product was therefore extracted with portions of pentane, in which it is slightly soluble. Data for Os(H)2(OC6H4O)(PiPr3)2: 1H NMR (300 MHz, C7D8, 20 °C): −16.61 (t, J(H–P) = 36 Hz, Os–H, 2H), 1.01 (dd, J(H–H) = 13 Hz, J(H–P) = 8 Hz, Os–P(CH(CH3)2)), 2.14 (m, Os–P(CH(CH3)2)), 6.85 and 7.46 (AA′MM′ system, 4H). 31P {1H} NMR (121.4 MHz, C7D8, 20 °C): 40.6. Only diagnostic data are given for the cation.

[OsH3(η2-N(CH3)C6H4NH)(PiPr3)2]+. 1H NMR (300 MHz, C7D8, 20 °C): −8.26 (br.t, J(H–P) = 16 Hz, Os–H, 3H), 0.63 (br.m, Os–P(CH(CH3)2)), 1.60 (br.m, Os–P(CH(CH3)2)), 3.66 (s, OsN(CH3)C6H4NH, 3H), 14.34 (s, OsN(CH3)C6H4NH, 1H). 31P {1H} NMR (121.4 MHz, C7D8, 20 °C): 42.1.

Synthesis of Os(H)2(N(H)C6H4O)(PiPr3)2. In an NMR tube, Os(H)3Cl(PiPr3)2 (0.01 g, 0.018 mmol) and lithium o-aminophenolate (0.0011 g, 0.018 mmol) were dissolved in 0.5 ml benzene-d6. The reaction proceeds for 4 d at 70 °C and a white precipitate was observed. The volatiles are removed under vacuo and the product extracted with ether. 1H NMR (300 MHz, C6D6, 20 °C): −14.72 (t, J(H–P) = 38 Hz, Os–H, 2H), 1.04 (dd, J(H–H) = 7.2 Hz, J(H–P) = 12.9 Hz, Os–P(CH(CH3)2)), 2.08 (m, Os–P(CH(CH3)2)), 6.97, 7.09, 7.33, 7.65 (t, t, d, d, respectively, J(H–H) = 6.3 Hz, aryl protons), 7.82 (br, Os–[N(H)C6H4O], 1H). 31P{1H} NMR (121.4 MHz, C6D6, 20 °C): 41.4. The intermediate OsH3(NH2C6H4O)(PiPr3)2 is observed all through the reaction: 1H NMR (300 MHz, C6D6, 20 °C): −10.17 (br, Os–H, 1H),−13.54 (br, Os–H, 2H), 1.12 (dvt, N = 19.2 Hz, Os–P(CH(CH3)2)), 1.14 (dvt, N = 19.2 Hz, Os–P(CH(CH3)2)), 1.85 (m, Os–P(CH(CH3)2)), 5.06 (br, Os–[NH2C6H4O], 2H), 6.39, 6.76, 7.00, 7.12 (t, d, t, d, respectively, J(H–H) = 6.3 Hz, aryl protons). 31P{1H} NMR (121.4 MHz, C6D6, 20 °C): 30.7.
Synthesis of Os(H)2(O(3-Me)C6H3O)(PiPr3)2. In an NMR tube, OsH3Cl(PiPr3)2 (0.01 g, 0.018 mmol) and lithium 3-methylcatecholate (0.0024 g, 0.018 mmol) were dissolved in 0.5 ml of benzene-d6. The reaction proceeded for 18 h at 80 °C and a white precipitate was observed. The volatiles were removed under vacuo. 1H NMR (300 MHz, C6D6, 20 °C): −17.17 (t, J(H–P) = 38 Hz, Os–H, 2H), 1.09 (dd, J(H–H) = 6 Hz, J(H–P) = 13 Hz, Os–P(CH(CH3)2), 2.21 (m, Os–P(CH(CH3)2)), 2.69 (s, aryl-CH3, 3H), 6.90, 6.97, 7.01 (m, aryl protons, 3H). 31P {1H} NMR (121.4 MHz, C6D6, 20 °C): 37.8.
Synthesis of Os(H)2(HN(3-Me)C6H3NH)(PiPr3)2. In an NMR tube, OsH3Cl(PiPr3)2 (0.01 g, 0.018 mmol) and 2,3-diaminotoluene (0.0045 g, 0.036 mmol) were dissolved in 0.5 ml of benzene-d6. The reaction was allowed to proceed at room temperature for 48 h. The mother liquor was decanted and separated from the white precipitate (hydrochloride salt of the amine reagent) formed during the reaction. 1H NMR (300 MHz, C6D6, 20 °C): −13.52 (t, J(H–P) = 38 Hz, Os–H, 2H), 1.03 (dvt, N = 20 Hz, Os–P(CH(CH3)2)), 2.02 (m, Os–P(CH(CH3)2)), 2.55 (s, aryl-CH3, 3H), 6.93, 7.07, 7.37 (d,t,d, J(H–H) = 8 Hz, aryl protons, 3H), 8.40 and 8.54 (each a brs, Os(NH(Me)C6H3NH, 1+1H). 31P {1H} NMR (121.4 MHz, C6D6, 20 °C): 42.9.

X-ray structure determinations

(1) Ru(H)2(OC6H4O)(PiPr3)2·C6H6. The selected crystal was affixed to the side of a glass fiber using silicone grease and cooled for data collection. The data were collected using 5 s frames with an omega scan of 0.30° on a Bruker SMART 6000 CCD. Examination of the lattice points using RLATT revealed that the crystal was twinned. An initial solution was obtained by using locally written software to integrate one of the two twin components. The Bruker Gemini software was then used to separate the two data sets. The crystal was twinned 180° about [001]. Data were corrected for Lorentz and polarization effects and equivalent reflections averaged using the Bruker SAINT software as well as utility programs from the XTEL library. The structure was solved using SHELXTL and Fourier techniques. All hydrogen atoms were located and refined, including the hydrides. A final difference Fourier was essentially featureless with a maximum peak height of 0.87 e Å−3.
(2) Os(H)2(HNC6H4NMe)(PiPr3)2. The selected crystal fragment was affixed to a glass fiber using silicone grease and transferred to the Bruker SMART 6000 CCD system where it was cooled to 158 °C using a gas-flow cooling system of local design. The data were collected using 30 s frames with an omega scan of 0.30°. Data were corrected for Lorentz and polarization effects and equivalent reflections averaged using the Bruker SAINT software as well as utility programs from the XTEL library. The structure was solved using SHELXTL and Fourier techniques. The structure consists of two independent molecules in the non-centrosymmetric space group P21; there was some concern about this space group assignment. The BMFIT program of Nyburg was used to compare the two independent molecules, and it was obvious that, while very similar, they are different due to a 60° twist of one P(iPr)3 ligand. During refinement, it became apparent that the crystal chosen was also a racemic twin, so the refinement was completed using the SHELX series of programs, and the BASF value converged to 0.265(6). Hydrogen atoms were placed in fixed idealized positions for the refinement, and the hydrides were not located. A final difference Fourier map was essentially featureless, although there were several peaks over 1.00 e Å−3 near the two metal atoms.

Results

Ru(H)2(OC6H4O)L2. Reaction of [RuHClL2]2 with singly-deprotonated catechol (1∶1 Ru∶catecholate) occurs over 14 h at 20 °C in benzene to produce a single product (Scheme 1, eqn. (a); the reaction rate is apparently limited by the low solubility of the lithium catecholate in benzene. NMR spectra in C6D6 serve to define the product stoichiometry. The hydride ligands appear as a sharp triplet (intensity 2H integrated vs. aryl ortho H), the aryl region shows only two chemical shifts (AA′XX′ pattern), and the iPr methyls appear as a doublet of doublets; the implied lack of virtual coupling in the latter suggests that the phosphines are not trans (i.e., 2JPP′ is not large). The 31P NMR spectrum is a singlet near 91 ppm, which compares well to the value19 103.2 ppm in Ru(H)2Cl2(PiPr3)2.20
scheme, filename = b411487f-s1.gif
Scheme 1

The presence of equimolar added catechol detectably alters the spectral parameters of Ru(H)2(OC6H4O)L2 in benzene at 20 °C. The hydride chemical shift moves 0.3 ppm downfield, and δ (31P) moves 2 ppm downfield. Both aryl proton δ values of added catechol move downfield (0.2 and 0.4 ppm compared to catechol in benzene), but they are not coalesced with those of coordinated catecholate. Since the hydroxyl proton δ values of added catechol move downfield 2.9 ppm, the collective spectral data are consistent with hydrogen bonding between Ru(H)2(OC6H4O)2L2 and catechol.

Structure of Ru(H)2(PiPr3)2(OC6H4O). This structure (Tables 1 and 2)† is remarkable for being non-octahedral. As seen in Figs. 1 and 2, the planes of Ru and any two identical ligands are neither orthogonal, nor coplanar, as would be true in an octahedron. Even considering only the heavier atoms, which are more accurately determined than the hydrides, the RuP2 and the RuO2 planes are neither orthogonal nor coplanar. This structure thus bears a remarkable similarity to that of M(H)2Cl2(PiPr3)2 (M = Ru19 and Os1), and it was reasoned there2,12,13 that the observed structure was a compromise between the nonoctahedral structure dictated electronically for a d4 six-coordinate species and the steric dictates of PiPr3 in conflict with chloride. Those arguments thus also pertain here, even if they force the two bulky phosphines into a small ∠P–Ru–P (114.4°) and even though the catecholate ligand has a small bite angle (∠O–Ru–O = 80.5°). Note, however, that the Cl–Ru–Cl angle in Ru(H)2Cl2(PiPr3)2 is also only 84.3° indicating that the chelate ring constraint naturally meets the angle preferred even by monodentate ligands. The molecule has an idealized C2 axis bisecting the PRuP angle and the ORuO angle, and there is no evidence for an ortho quinone structure in the aryl ring C/C distances. The catecholate chelate ring involving Ru is essentially planar (interior angles sum to 539° vs. 540° for planar), and the two Ru–O distances are indistinguishable (2.03 Å), and considerably shorter than to those in phenoxide oxygen in cis-RuH(O-tolyl)(PMe3)4,21 where Ru–O is 2.152(3) Å and ∠Ru–O–C is 133.3°.
ORTEP drawing (50% probability ellipsoids in all figures) of Ru(H)2(OC6H4O)(PiPr3)2 with H on C deleted, showing selected atom labeling. The idealized C2 axis is vertical.
Fig. 1 ORTEP drawing (50% probability ellipsoids in all figures) of Ru(H)2(OC6H4O)(PiPr3)2 with H on C deleted, showing selected atom labeling. The idealized C2 axis is vertical.

ORTEP drawing of Ru(H)2(OC6H4O)(PiPr3)2 viewed down the idealized C2 axis.
Fig. 2 ORTEP drawing of Ru(H)2(OC6H4O)(PiPr3)2 viewed down the idealized C2 axis.
Table 1 Crystal data
Empirical Formula C30H54O2P2Ru C25H52N2OsP2
a b
Color of crystal orange orange
Space group P[1 with combining macron] P21
a 11.224(2) 8.9882(3)
b 11.907(2) 21.8475(9)
c 12.510(2) 14.1438(6)
α 77.529(5) 90
β 89.453(5) 97.4220(10)
γ 76.548(5) 90
Z 2 4
V/Å3 1586.60 2754.14(19)
ρ calcd/g cm−3 1.276 1.526
T/K −118 113
λ/Å 0.71073 0.71073
R(Fo)a 0.0351 0.0322
Rw(Fo)b 0.0346 0.0698

Table 2 Selected bond distances (Å) and angles (°) for Ru(H)2(OC6H4O)(PiPr3)2
RU(1) H(1) 1.56(5)  
RU(1) H(2) 1.39(6)  
RU(1) P(10) 2.2833(15)  
RU(1) P(20) 2.2876(14)  
RU(1) O(2) 2.027(3)  
RU(1) O(9) 2.031(3)  
O(2) C(3) 1.348(6)  
O(9) C(8) 1.350(6)  
C(3) C(4) 1.389(7)  
C(3) C(8) 1.406(6)  
C(4) C(5) 1.382(8)  
C(5) C(6) 1.376(8)  
C(6) C(7) 1.362(8)  
C(7) C(8) 1.392(7)  
       
P(10) RU(1) P(20) 114.39(5)
P(10) RU(1) O(2) 152.75(10)
P(10) RU(1) O(9) 87.22(10)
P(20) RU(1) O(2) 86.21(10)
P(20) RU(1) O(9) 151.58(10)
O(2) RU(1) O(9) 80.51(13)
RU(1) O(2) C(3) 112.9(3)
RU(1) O(9) C(8) 112.9(3)
P(10) RU(1) H(1) 66.0(17)
P(10) RU(1) H(2) 78.6(23)
P(20) RU(1) H(1) 78.2(16)
P(20) RU(1) H(2) 66.4(24)
O(2) RU(1) H(1) 139.1(17)
O(2) RU(1) H(2) 95.2(24)
O(9) RU(1) H(1) 95.7(17)
O(9) RU(1) H(2) 139.5(24)
H(1) RU(1) H(2) 112.(3)

Comparison of this structure to that22 of Os(H)2(OCH2CF3)2(PiPr3)2 is instructive (Os and Ru have negligibly different covalent radii). Bond lengths and angles in this molecule are remarkably similar to those in Ru(H)2(OC6H4O)L2. The angle O–Os–O, at 81.0°, suggests that there is no major chelate constraint in the catecholate, and the similarity in the “twisted” orientation of the M(H)2, M(O)2 and MP2 planes suggests that the competition between steric and electronic factors reaches the same equilibrium even with different metals and different alkoxides. This structural similarity also extends to Os(H)2X2(PiPr3)2 where X = OSO2CF322 or Cl.1

Ru(H)2[OC6H4(NH)]L2. Reaction of [RuHClL2]2 with the monolithium salt of ortho-aminophenol in benzene at 20 °C proceeds to completion within 6 h to give only one product (Scheme 1, eqn. (b)). This shows four aryl protons and one NH proton (7.7 ppm, broad), together with a hydride triplet; the latter integrates for two protons. The second hydride (cf. [RuHClL2]2 reagent) thus results from migration of one H from N to Ru; the reaction is oxidative addition. The iPr groups show one diastereotopic 1H NMR methyl chemical shift. These do not show virtual coupling, which suggests ∠P–Ru–P is significantly less than 180°. The 31P NMR spectrum shows equivalent phosphines, and these become a triplet when only the hydrides are permitted to couple to P; this further supports the dihydride formulation. This is confirmed by the T1 min (300 MHz, −75 °C) value of 226 (±4) ms. These data are consistent with structure B, provided some fluxional process effects site exchange of Ha and Hb, which are inequivalent in this structure(syn vs. anti to O). These hydrides remain a single 1H NMR triplet at −85 °C in toluene-d8.
ugraphic, filename = b411487f-u2.gif
C–H Cleavage?23. Given the occurrence of N–H oxidative addition above, we considered the possibility of an analogous reaction for a benzylic C–H bond. This reaction (eqn. (1)) proceeds to completion within 14 h at 20 °C in benzene, and the only observed product has the expected elemental composition, but four resonances of the aryl protons are sufficiently far upfield (3.8–5.6 ppm) to be diagnostic of a π-arene structure, and the benzylic protons integrate for 3 H.
 
ugraphic, filename = b411487f-u3.gif(1)
The single hydride ligand is an apparent triplet, but the individual broad components are equally consistent with a doublet of doublets structure. The two phosphine ligands are inequivalent, but separated by only ∼3 ppm. This inequivalence arises from the methyl substituent, which destroys mirror symmetry for the molecule, leaving it chiral, and thus with inequivalent phosphines. The iPr methyls appear as four chemical shifts, and without virtual coupling.

These spectral parameters are numerically very similar to those already reported24 for (η5-C6H5O)RuH(PPh3)2. In sum, forming a saturated η5-cyclohexadienoyl product is preferred to forming D. Heating C for 2 h in benzene at 60 °C gives only partial decomposition, but no D. Carrying out the reaction of eqn. (1) at 50 °C in benzene for 1 h also gives only C.


ugraphic, filename = b411487f-u4.gif
Reactivity: (a) CO and 1-pentene. Reaction of Ru(H)2(OC6H4O)(PiPr3)2 with CO (1 atm.) is complete in time of mixing at 22 °C in benzene to give only cis, trans-Ru(CO)2(OC6H4O)(PiPr3)2, identified by 1H and 31P{1H} NMR spectra (C2V symmetry) and its two CO stretching frequencies. The speed of this reaction is consistent with “operational unsaturation”25–27 of the metal center, and the loss of H2 consistent with the π-basicity of Ru being dominated by the arrival of the π-acidic carbonyl ligands; that is depletion of metal reducing power by two CO triggers conversion of 2 H− to H2.

While Ru(H)2(OC6H4O)(PiPr3)2 is unchanged by 1-pentene, the terminal olefin is isomerized to trans- and cis-2-pentene over 80 h at 55 °C. This olefin thus can bind to the complex, but only to a kinetically significant, rather than a detectable (NMR) extent.

(b) Hydrogenation and catalysis. Molecular hydrogen (1 atm.) fails to react with Ru(H)2(OC6H4O)(PiPr3)2 in C6D6 during 10 h at 20 °C. In contrast, Ru(H)2(OC6H4NH)(PiPr3)2 reacts with H2 (1 atm.) readily in toluene-d8 to generate a pale yellow solution in less than 1 h (eqn. (2)).
 
ugraphic, filename = b411487f-u5.gif(2)
The complete conversion to product is confirmed by the appearance of a new 1H NMR triplet at −12.39 ppm (intensity 3H integrated vs. aryl ortho H) as well as the appearance of a broad singlet at 4.34 ppm (intensity 2H integrated vs. aryl ortho H), the latter assigned as NH2. Diastereotopic iPr methyl hydrogens are also observed for the molecule, at 1.05 and 1.13 ppm. 31P{1H} NMR reveals a singlet at 60.4 ppm. All these are consistent with the formation of an 18-electron Ru(IV) complex with the structure E.28 The hydride signal becomes broad at 50 °C in toluene-d8, but there is only incomplete decoalescence by 90 °C. The value of T1 goes through a minimum at 70 °C and 300 MHz, and the value 4529,30 (±1) ms is too long to involve an H2 ligand; a trihydride formulation with a small ∠H–Ru–H is thus indicated.

This reactivity towards H2 enables catalytic olefin hydrogenation. If tBuHC[double bond, length as m-dash]CH2 in C6D6 at 20 °C is treated with H2 (1 atm.) in the presence of 5 mol% Ru(H)2(OC6H4NH)L2, all hydrogen (the limiting reagent) is consumed within 2 h, with conversion of olefin to neohexane. The olefin is clearly capable of “dehydrogenation” of RuH3(OC6H4NH2)L2 because, under the hydrogen-poor conditions at the end of this catalytic reaction, the only metal complex observed to be present is Ru(H)2(OC6H4NH)L2; 50% of the charged olefin remains unreacted due to the limited H2 available.

Osmium analogs. In an effort to develop greater electron richness (and thus reducing power) at the metal, we moved from Ru to Os, and from a catecholate to a diamide ligand. Since “OsHClL2” is not synthetically available, we employed, as our OsL2 source, the more hydrogen-rich Os(H)3ClL2. Since this Os source has a higher oxidation state than that in our Ru reagent, this has impact on the synthetic method, as well as the primary product formed.
(a) Phenylenediamine. We investigated the reaction (Scheme 2, eqn. (a)) of this osmium reagent with 1,2-phenylenediamine (1∶2 stoichiometry) in benzene-d6. The primary product is an adduct formed by these two reagents (eqn. (3)), characterized in the 1H NMR by a very broad peak in the hydride region at −15.69 ppm, a very broad peak at 3.98 ppm (due to ArNH2) and a very broad peak in the 31P {1H} NMR at 36.1 ppm.
 
ugraphic, filename = b411487f-u6.gif(3)
No decoalesced 1H NMRsignals for bound and free amine (i.e. excess) are observed at 22 °C in this solution, and the phosphine methyl proton region displays only one apparent quartet, indicating a dissociative equilibrium (eqn. (3)). Nevertheless, some of the final product Os(H)2(NHC6H4NH)(PiPr3)2 is already present in the reaction mixture within 1 h after mixing. After 24 h the reaction is complete, including precipitation of a colorless salt, the amine hydrochloride. The dihydride Os(H)2(NHC6H4NH)(PiPr3)2 is characterized in the 1H NMR by a single triplet of intensity two at −13.51 ppm in the hydride region, two intensity two multiplets in the aryl region and a broad peak at 8.33 ppm due to bound NH groups. The iPr methyl region displays one apparent quartet, indicating equivalent methyls. The single hydride peak and the absence of diastereotopic methyl groups indicate that this molecule is (time averaged) mirror-symmetric. The 31P {1H} NMR spectrum is a singlet at 43 ppm. We assign these signals to species F which shows that the second mole of amine serves to remove HCl.
scheme, filename = b411487f-s2.gif
Scheme 2
(b) N-Methylphenylenediamine. Synthesis and characterization. The reaction of Os(H)3Cl(PiPr3)2 with N-methyl-1,2-phenylenediamine is additionally revealing (Scheme 2, eqn. (b) and also eqn. (3)). When the reaction is carried out in 1∶2 stoichiometry (Os∶amine) in benzene-d6 an adduct is instantaneously and essentially completely formed (by 1H and 31P {1H} NMR). This adduct is characterized by a broad singlet in the 1H NMR spectrum at −14.44 ppm. Free and coordinated amine show coalesced 1H NMR signals. The chemical shift of the methyl group and the amino protons shift downfield by 0.1 ppm and 1 ppm, respectively, the latter a singlet at 3.80 ppm. The relative magnitude of these shifts indicate that the donor group interacting with the metal center is the primary amine, NH2. The aromatic protons of the amine have also shifted from their values for the pure arene in benzene. The phosphine methyl region 1H NMR displays one apparent quartet, indicating a fast equilibrium. The 31P {1H} NMR displays a broad singlet at 27.0 ppm. Under these reaction conditions, no further change is observed over 24 h, and heating the reaction mixture to 55 °C for 17 h does not produce a significant amount of any other product. However, upon addition of 1 equivalent of tert-butylethylene, the reaction proceeds further at 25 °C and the conversion is complete after 24 h, producing the new compound Os(H)2(N(Me)C6H4NH)(PiPr3)2. Production of neo-hexane (hydrogenated olefin) is stoichiometric and a white precipitate is observed, presumably the corresponding chloride salt of the amine reagent. Two signals at −12.02 and −13.67 ppm, the former a triplet and the latter a triplet of doublets (due to coupling to two phosphines and coupling to NH, confirmed by selective NH decoupling) characterize the final product in the hydride region of the 1H NMR spectrum. The NMe group and NH proton resonate as a broad triplet at 4.41 ppm and as a broad singlet at 8.85 ppm, respectively. Four signals are observed in the aryl region with the expected multiplicities and relative intensities for an ortho-C6H4XY ring, and diastereotopic apparent quartets are observed in the methyl phosphine region. The 31P {1H} NMR displays a singlet at 54.8 ppm. This is consistent with structure F, but, in contrast to the Ru analogs, the barrier to any mutual site exchange between inequivalent hydrides (Ha and Hb in eqn. (3)) is higher.

Intrigued by the large (11 ppm) difference in the 31P{1H} chemical shift between these two seemingly similar analogs, Os(H)2(N(Me)C6H4NH)(PiPr3)2 (FMe) and Os(H)2(N(H)C6H4NH)(PiPr3)2 (FH), we investigated the 1H NMR spectra of compound FMe at low temperatures to detect any possible C–H agostic interaction involving the N-methyl group in FMe. The 1H NMR signals of compound FMe suffer no significant transformation in the range from 20 °C to −70 °C. Moreover, the value of the C–H coupling constant of the methyl substituent remains constant over this range of temperatures (J(C−H) = 132 Hz as measured from the 13C satellites in the 1H NMR), indicating a typical non-distorted sp3 hybridized carbon. We also sought to detect a dynamic process (hydride site exchange) for compound FMe. The two distinct hydride signals show broadening at 60 °C, the signal at −13 ppm no longer showing resolvable coupling to the NH proton. Broadening is also observed for the N-methyl signal and the N–H signal, but no coalescence of the hydrides occurs. Compound FMe does not decompose at 60 °C over 14 h in toluene-d8.

Operationally unsaturated Os(H)2(N(Me)C6H4NH)(PiPr3)2 does not react with 1 atm. CO (in contrast to Ru(H)2(OC6H4)(PiPr3)2) or C2H4 at 20 °C in benzene, or with equimolar PhC≡CH (20 °C, benzene).

Molecular structure. Given (eqn. (4)) that the chelate ligand G has a redox alternative31–34 form (H) that would leave Os in oxidation state +2, we determined the structure of FMe.
 
ugraphic, filename = b411487f-u7.gif(4)
We were also interested in how structurally different are the NMe and NH groups in their bonds to Os and to the phenyl rings. Figs. 3 and 4 are chosen to allow comparison to Figs. 1 and 2. These reveal (Table 3) a similar overall structure, but the planar diamidobenzene ligand is twisted more from the OsP2 plane, perhaps dictated by NMe-to-PiPr3 repulsions. This twist, if it originates from the methyl group, could be the origin of the 11 ppm different 31P NMR chemical shifts for these two Os complexes. Probably as a consequence of this twist, the P–Os–P angle is 3° larger than the P–Ru–P angle. The N–Os–N angle, at 76.7°, is nearly 4° smaller than the O–Ru–O angle (80.5°). The Os–NMe distance, 2.041 Å, is longer than the Os–NH distance, 2.003 Å. The three angles around the MeN9 sum to 360°, indicative of planarity. Carbon–carbon distances in the aryl ring show some weak evidence for the diimino form (H), especially in contrast to the ruthenium catecholate (Scheme 3). However, the best criterion to choose between forms G and H is the O–C or N–C distances.35,36 By this criterion both ligands here (i.e. the Ru and Os compounds in Scheme 3) are in their fully reduced, dianionic form.
ORTEP drawing of the non-hydrogen atoms of one molecule of Os(H)2(HNC6H4NMe)(PiPr3)2, oriented as in Fig. 1. Hydrides were not located in this X-ray study.
Fig. 3 ORTEP drawing of the non-hydrogen atoms of one molecule of Os(H)2(HNC6H4NMe)(PiPr3)2, oriented as in Fig. 1. Hydrides were not located in this X-ray study.

ORTEP drawing of Os(H)2(HNC6H4NMe)(PiPr3)2, oriented as in Fig. 2. This molecule is approximately the enantiomer of the Ru complex in Fig. 2.
Fig. 4 ORTEP drawing of Os(H)2(HNC6H4NMe)(PiPr3)2, oriented as in Fig. 2. This molecule is approximately the enantiomer of the Ru complex in Fig. 2.
Table 3 Selected bond distances (Å) and angles (°) for two independent molecules of Os(H)2(NMeC6H4NH)(PiPr3)2
Os(1) N(2) 2.004(5)  
Os(1) N(9) 2.041(6)  
Os(1) P(21) 2.2843(16)  
Os(1) P(11) 2.3035(15)  
N(2) C(3) 1.385(9)  
C(3) C(4) 1.394(10)  
C(3) C(8) 1.415(10)  
C(4) C(5) 1.379(10)  
C(5) C(6) 1.407(12)  
C(6) C(7) 1.381(11)  
C(7) C(8) 1.403(9)  
C(8) N(9) 1.376(9)  
N(9) C(10) 1.460(9)  
Os(31) N(32) 1.998(6)  
Os(31) N(39) 2.045(6)  
Os(31) P(51) 2.2892(16)  
Os(31) P(41) 2.3079(16)  
N(32) C(33) 1.370(10)  
C(33) C(34) 1.391(10)  
C(33) C(38) 1.412(11)  
C(34) C(35) 1.390(11)  
C(35) C(36) 1.366(13)  
C(36) C(37) 1.382(12)  
C(37) C(38) 1.421(10)  
C(38) N(39) 1.371(10)  
N(39) C(40) 1.455(9)  
       
N(2) Os(1) N(9) 76.7(2)
N(2) Os(1) P(21) 96.45(17)
N(9) Os(1) P(21) 131.58(15)
N(2) Os(1) P(11) 132.04(16)
N(9) Os(1) P(11) 100.74(16)
P(21) Os(1) P(11) 117.22(6)
C(3) N(2) Os(1) 119.4(5)
C(8) N(9) Os(1) 117.2(5)
C(10) N(9) Os(1) 127.3(5)
N(32) Os(31) N(39) 76.5(2)
N(32) Os(31) P(51) 96.58(17)
N(39) Os(31) P(51) 130.26(16)
N(32) Os(31) P(41) 133.35(16)
N(39) Os(31) P(41) 100.87(16)
P(51) Os(31) P(41) 117.18(6)
C(33) N(32) Os(31) 120.1(5)
C(38) N(39) Os(31) 116.6(5)
C(40) N(39) Os(31) 127.7(5)


Comparison of arene ring C–C distances.
Scheme 3 Comparison of arene ring C–C distances.

A recent study35 of Os(H)2(OC6H4NH)(PPh3)2, which also shows this same distortion away from octahedral geometry, also concluded that this contains Os(IV) and a dianionic chelate ligand. In contrast, the RuCl2(OC6H4NH)(PPh3)2 analog has an octahedral geometry; a ground state with Ru(III) strongly spin-coupled to a radical monoanion chelate was deduced. Thus, the Os(H)2 moiety is more reducing than the RuCl2 unit.

(c) Reactivity of Os(H)2(NMeC6H4NH)(PiPr3)2 with protons. Triflic acid protonates compound FMe in arene solvents at room temperature, to form the triflate salt of [Os(H)3(N(CH3)C6H4NH)(PiPr3)2]+ (I). The metal in Os(H)2(NMeC6H4NH)(PiPr3)2 is thus more basic than either nitrogen. A triplet of doublets (coupling to 2P and one NH) in the hydride region characterizes cation I in the 1H NMR at −8.15 ppm. The methyl protons and the amide proton resonate as a singlet at 4.43 and a broad singlet at 14.65 ppm, respectively. Selective decoupling of the amido signal induces a multiplicity change in the hydride signal to a triplet, confirming that the doublet structure in the hydride peak is due to coupling to the amido proton. Complicated multiplets are observed in the aryl region, which are completely resolved into two triplets and two doublets when the solvent is d8-THF, and the diastereotopic P-methyl groups appear as overlapping apparent quartets. The 19F NMR chemical shift indicates no interaction of the triflate with the cation, and the 31P {1H} NMR spectrum is a singlet.

Due to the symmetry of the diamido ligand, the three hydride atoms cannot be chemically equivalent, even though we observe one sharp signal at 20 °C. In an attempt to decoalesce the hydrides, we investigated the low temperature properties of compound I in d8-THF. Neither the hydride nor the NH proton signals in the 1H NMR spectra undergo a significant change from 20 °C to −110 °C. The 31P {1H} NMR signal remains unchanged in the same range of temperatures. Relaxation rate measurements of the hydride ligands give a T1 min. of 85 ms at −84 °C, confirming a trihydride formulation with ∠H–Os–H < 90°. This protonation with a strong acid helps to understand the following reaction with catechol.

(d) Substitution reactions on Os(H)2(N(CH3)C6H4NH)(PiPr3)2. The above evidence for hydrogen bonding between Ru(H)2(OC6H4O)L2 and added catechol suggests that weak acids might replace the chelated diamide from Os. We therefore investigated a substitution reaction (Scheme 2, eqn. (c)) in toluene-d8 at room temperature with excess catechol (1∶4, Os∶catechol). Within (5 min) the time of acquiring the first NMR spectrum, a broadening of the NH proton in compound FMe was observed. This we attribute to hydrogen bonding between catechol and the NH proton of FMe. The only product detected at this stage is the catecholate salt of [Os(H)3(N(CH3)C6H4NH)(PiPr3)2]+ (I). The spectroscopic features of this compound are similar to the triflate salt of I, though not identical, thus indicating minor differences due to counterion identity. After 2 h, the substitution product Os(H)2(OC6H4O)(PiPr3)2 (IV) is observed (1H, 31P {1H} NMR). The hydride region of the 1H NMR spectrum displays a triplet at −16.61 ppm. The aryl region displays two multiplets and the P-methyl region displays one doublet of doublets, confirming the presence of two fold symmetry in the molecule. The 31P {1H} NMR spectrum of compound IV is a singlet at 40.6 ppm.

To prove that the most exchangeable site is the amide nitrogen, MeOD was added in excess to a solution of Os(H)2(NMeC6H4NH)(PiPr3)2. The NH proton signal immediately disappeared from the 1H NMR spectrum and there was no intensity decrease in the OsH signal after 10 min at 22 °C. Given the hydrogen bonding evident between Ru(H)2(OC6H4O)L2 and catechol,37–39 we suggest that this selective exchange of MeOD with the NH proton occurs via a structure such as J.


ugraphic, filename = b411487f-u8.gif
Reactivity towards H2 and D2. NMR spectra are unchanged over 24 h when compound I is subjected to 1 atm. of H2 in C6D6 at room temperature. Under 1 atm. of D2, deuterium is first incorporated in the NH signal after 10 min, and incorporation into the hydride positions is complete after 4 h. No further reactivity is observed in this solution at room temperature or at 70 °C for 24 h by 2H NMR (e.g., no free amine or OsD6(PiPr3)2 is observed). This result proves that compound I does bind H2 (D2), but in an unfavorable equilibrium at 1 atm. H2 (solubility ∼10−3 M).
(e) Synthesis of Os(H)2(NHC6H4O)(PiPr3)2. To continue the study of six-coordinate d4 osmium complexes with chelating ligands, we decided to study a more strongly asymmetric ligand, one bearing ortho alkoxide and amido donor substituents. The compound Os(H)3Cl(PiPr3)2 reacts very slowly with the lithium salt of ortho-aminophenolate at room temperature in benzene-d6 (Scheme 2, eqn. (d)). At 70 °C, the reaction proceeds faster, and after 4 d the product obtained is Os(H)2(NHC6H4O)(PiPr3)2. The 1H NMR spectrum of this molecule shows a triplet in the hydride region at −14.72 ppm, two doublets and two triplets in the aryl region, and a broad singlet at 7.82 ppm corresponding to the NH amido proton. The 31P{1H} NMR spectrum shows a singlet. The 1H NMR in the methyl phosphine region displays only one doublet of doublets, in agreement with the fact that we only observe one chemical shift for the hydride ligands. Nevertheless, any compound with such ligand distribution has chemically inequivalent hydrides and diastereotopic methyl groups (as observed for Os(H)2(NHC6H4NMe)(PiPr3)2). These results indicate there must exist a dynamic process that renders the two hydrides equivalent, and that the barrier for this process is lower when the chelate NMe is replaced by O.

During the course of the reaction at or above 20 °C, the intermediate OsH3(OC6H4NH2)(PiPr3)2 is observed by 1H and 31P{1H} NMR. This intermediate is characterized in the 1H NMR spectrum by two unstructured signals at −10.7 and −13.54 ppm in relative intensities 1∶2, respectively. This behavior is analogous to that of compounds of general formula OsH3Cl(NR3)(PiPr3)2 at −20 °C or below.40 The aryl region of the 1H NMR spectrum displays two doublets and two triplets; the NH2 protons resonate at 5.06 ppm as a broad singlet and the phosphine region shows diastereotopic methyl signals. The 31P{1H} NMR spectrum shows a singlet. Free H2 is also observed during the reaction at room temperature (at higher temperature, it is lost rapidly to the head space).

The hydride region of OsH3(OC6H4NH2)(PiPr3)2 persists as two signals, 1∶2 intensity, down to −70 °C, but each shows triplet structure between −40 °C and −70 °C. There is additional broadening of the intensity 2 hydride signal from −70 °C to −90 °C, and the minimum T1 values (300 MHz, and generally at −80 ± 5 °C) are shown in K, together with a proposed assignment. None of these is short enough (<10 ms) to indicate an H2 ligand, but the 73 ms value indicates an acute H–Os–H angle. Note also the short value for the NH2 protons. The 108 and 120 ms values are suggestive of mutual approach, and mutual dipole relaxation; for this reason, we suggest a hydrogen bond between them. These hydride values are all short compared to that measured(235 ms) here for the classical hydrides in Os(H)2(OC6H4NH)(PiPr3)2. Intramolecular hydrogen bonding illustrated in K can explain the lack of rapid site exchange of one of the three hydrides.

ugraphic, filename = b411487f-u9.gif

Interestingly, the unsaturated Os(H)2(NHC6H4O)(PiPr3)2 does not produce OsH3(NH2C6H4O)(PiPr3)2 under equimolar amounts of H2 in toluene-d8 after 4 h at 20 °C. The added free H2 is observed as a singlet at 4.5 ppm in the 1H NMR spectrum. The absence of reactivity indicates that the two species are not yet at equilibrium during the course of the synthesis, but that the trihydride is the primary product, and it only slowly (20 °C) loses H2, with loss of one amine H.

Fluxionality (Scheme 4). The equivalent hydrides seen even at low temperature for M(H)2(OC6H4NH)(PiPr3)2 could be accounted for by the typically facile site exchange of hydrides (e.g., facile bond angle distortion and possible conversion to a rapidly rotating H2 ligand). We were, however, puzzled by the lack of inequivalent (i.e., diastereotopic) iPr methyls in M(H)2(OC6H4NH)L2 (M = Ru and Os). Fluxional process aalone (involving heavy atom motion of the E/E′ chelate) cannot account for the observations; it makes the two hydrides equivalent, as well as Me and Me* equivalent, but it leaves the two phosphines inequivalent. Process balone is unable to create time-average equivalence of all the sites observed to be equivalent; it leaves the hydrides inequivalent. However, both processes, the equivalent of full 180° rotation of the plane of the chelate ligand, must occur rapidly for M(H)2(OC6H4NH)L2.

To test whether the above observations of a single iPr methyl chemical shift might be due to two cases of accidental degeneracy (Ru and Os), we synthesized a more subtly asymmetric example for both metals L (Scheme 1, eqn. (c) and Scheme 2, eqn. (e)). For M = Ru, the resulting molecule

ugraphic, filename = b411487f-u10.gif
shows equivalent hydrides and one iPr methyl 1H NMR chemical shift, and one 31P{1H} NMR chemical shift. For M = Os, the heterogeneous reaction of LiOC6H3(CH3)OH with Os(H)3ClL2 proceeds in benzene, without any detectable intermediate, to give the Os example of L which shows equivalent hydrides, one iPr methyl 1H NMR chemical shift, and one 31P{1H} NMR chemical shift. In sum, the likelihood of accidental degeneracy of three sites in four different molecules is so low that we conclude the two fluxional processes in Scheme 4 occur rapidly at 20 °C for the 3-methyl catecholates of both Ru and Os.

To test whether the frozen fluxionality of Os(H)2(HNC6H4NMe)(PiPr3)2 is due to sterics (Me preventing rotation) or electronics (E = O vs. NR), we investigated the molecule containing the 2,3-diamidotoluene ligand. Reaction of Os(H)3Cl(PiPr3)2 with excess 2,3-diaminotoluene (1∶2 stoichiometry) in benzene-d6 (Scheme 2, eqn. (a)) proceeds spontaneously at room temperature to form an amino adduct, analogously to all the amines previously described (eqn. (3)). This adduct is characterized by a broad hydride signal at −14.29 ppm in the 1H NMR spectrum and a broad peak at 25.9 ppm in the 31P {1H} NMR spectrum. No decoalesced signals due to coordinated and free amine are observed, and only one chemical shift is present in the methyl phosphine region of the 1H NMR spectrum. After 10 min some diamido product is already observed. After 48 h the reaction is complete and a substantial amount of white precipitate is observed in the reaction mixture (hydrochloride salt of the amine reagent). The final dihydrido diamido product is characterized in the 1H NMR spectrum by a triplet in the hydride region at −13.51, two amido proton signals at 8.54 and 8.40 ppm, a singlet at 2.55 corresponding to the benzylic methyl, one apparent quartet in the iPr methyl phosphine region and the corresponding aryl signals. The 31P {1H} NMR spectrum displays a sharp singlet at 42.9 ppm. Thus, we can conclude that this molecule is also fluxional. In conclusion, only for Os(H)2(NMeC6H4NH)L2 is the barrier higher and we suggest that this is due to steric conflicts between NCH3 and PiPr3 in process a of Scheme 4.


scheme, filename = b411487f-s4.gif
Scheme 4

Discussion

The synthetic methods employed here show similarities, but also fundamental differences. The fact that both involve oxidative addition of O–H or N–H bonds testifies to the considerable reducing power of Ru and Os (even if they are unsaturated) in an environment devoid of π-acid ligands (e.g., CO). The difference is that the Ru examples introduce the chelate as an anion, by chloride metathesis. Certain of the Os cases in Scheme 2 employ neutral amine, with excess amine effecting reductive elimination of HCl; variable amine basicity can influence the success of this dehydrochlorination. Because the osmium source, Os(H)3ClL2 contains two more hydrides per metal than does “RuHClL2”, only Os involves loss of H2 (eqn. (3)). The osmium reactions begin because unsaturated Os(H)3ClL2 has enough Lewis acidity to bind aryl amines.
 
Os(H)2(chelate)L2 + D2 ⇌ OsH2D2(chelate)L2(5)
We suggest that intramolecular hydrogen bonding41 (M) can be a more favorable source of evolved H2 than would be reductive elimination from an Os(H)2 substructure. The observed deuteration of NH as the most rapid isotope exchange in eqn. (5) supports the importance of coordinated D2 in hydrogen bonding to the NHAr ligand (Scheme 5).
ugraphic, filename = b411487f-u11.gif
scheme, filename = b411487f-s5.gif
Scheme 5

The selectivity of this H/D exchange with D2, being faster with N–H than with Os–H protons, demands that the effective binding site for D2 to Os (Scheme 6) is at c, not a or b (these latter would effect deuteration of Os–H in a pentagonal bipyramidal intermediate). In addition, D2 at site c must reversibly form deuterido/amine complex N (where the notation (H,H) leaves open the question of whether these are two hydrides, or H2) at a rate faster than Os–D in N exchanges with (H,H).


scheme, filename = b411487f-s6.gif
Scheme 6

These experiments permit some comparisons of Ru vs. their Os analogs.

(1) Chloride replacement by a mono lithium catecholate succeeds not only for the chloride bridged Ru2(H)2(μ-Cl)2(PiPr3)4 but also for the more hydride rich Os(H)3Cl(PiPr3)2, where H2 loss occurs.

(2) The analogous reactions occur for ortho-amine phenoxide. The intermediates O and P are suggested, and the next step must be H–E bond cleavage.

ugraphic, filename = b411487f-u12.gif
For Ru, this directly furnishes the observed product, but the suggested RuH3 intermediate is not detected. For Os, the intermediate OsH3(OC6H4NH2)(PiPr3)2is detected, and it only slowly loses H2 at 70 °C. The Ru analog of this molecule, RuH3(OC6H4NH2)(PiPr3)2, can be synthesized by addition of H2 (1 atm.) to unsaturated Ru(H)2(OC6H4NH)(PiPr3)2, a net heterolytic splitting of H2. The H2adducts of Os(H)2(EC6H4E′)L2 are thermodynamically unfavorable even under 1 atm. H2, and the lack of observation of these in the amine reactions (Scheme 2, eqn. (a)) originates in their different mechanism vs. the lithium phenoxide reactions.

(3) The species MH3(OC6H4NH2)(PiPr3)2 show a dramatic difference: for M = Os, the hydrides are not fully coalesced into a single 1H NMR signal (being instead a 2∶1 intensity pattern), while for M = Ru the hydrides show site exchange averaging even at −70 °C. Other examples are known where rearrangement barriers are higher for 5d than for 4d analogs.42,43 In addition, if the mechanism of site exchange involves a η2-H2 intermediate (eqn. (6)),

 
ugraphic, filename = b411487f-u13.gif(6)
then the energetics of reduction of M would typically have a larger ΔG° for M = Os than for Ru.

(4) The H2 loss process from MH3(OC6H4NH2)(PiPr3)2 is likely to involve one of the NH2 protons (eqn. (7)).

 
ugraphic, filename = b411487f-u14.gif(7)
This may be the lowest energy path because it does not require a divalent MIIH(H2) intermediate (especially costly for a 5d metal), and because that d6 intermediate might bind H2 especially strongly. In addition, there is abundant evidence37–39 for hydrogen bonding between a hydride ligand and a nearby acidic proton, so Ha and Hb in Q may already have some bonded character. Concrete evidence for this is the observed higher barrier for scrambling one hydride with the other two in OsH3(OC6H4NH2)(PiPr3)2vs. the Ru analog. The more electropositive character of a 5d metal (vs. 4d) should make hydride on Os more Brønsted basic, hence forming a stronger hydrogen bond.

(5) For osmium only, we have studied ortho-phenylenediamine reagents for their potential to introduce an ortho-diamido chelate. Here, a second mole of diamine was anticipated to serve the role of base, perhaps making a deprotonated (lithium) amide unnecessary. This reaction proceeds satisfactorily for ortho-phenylenediamine and its ring methyl derivative, with loss of two amine protons, to give the dihydride of OsIV. That is, H2 is released during the reaction to give an OsIV diamide.

(6) When one ortho-phenylenediamine nitrogen carries one methyl group, this secondary amine center changes the reaction progress considerably. Although an adduct is formed between Os(H)3Cl(PiPr3)2 and NHMe(C6H4)NH2, there is no further reaction in the presence of additional (free) NHMe(C6H4)NH2. The adduct persists, with no loss of H2 or HCl. The addition of tBuHC[double bond, length as m-dash]CH2, as a hydrogen acceptor, delivers Os(H)2(NMeC6H4NH)(PiPr3)2, together with amine hydrochloride and hydrogenated olefin. Steric factors at the secondary amine center may be the source of the trouble, until a lower coordination number can be achieved by olefin hydrogenation.

(7) The molecule Os(H)2(NMeC6H6NH)(PiPr3)2 is unique among all those 6 coordinate species studied here in showing inequivalent hydrides (but still time average equivalent phosphorus nuclei). This is attributed to steric conflicts between the NMe group and a PiPr3 group in the transition state for equivalencing the hydrides, which is deduced to involve a planar P2OsN2 substructure.

(8) The mechanism of the ligand exchange reaction of Os(H)2(NMeC6H4NH)(PiPr3)2 with catechol involves proton transfer, and that catechol proton goes to Os, not N, indicating that each amide lone pair is strongly involved in bonding to its attached π-acids, OsIV and the aryl π* orbital.

(9) With regard to the question of intramolecular redox between M and N, the fact that the quinoid structure involves divalent metal leads to some predictions. First, this should adopt an octahedral structure, which is not observed in the two structures reported here. In addition, such an octahedral structure is less likely to be fluxional, given the general rigidity that characterizes d6, octahedral molecules. Taken together, the structures and fluxionality observed here establish a picture of the d4, six-coordinate potential energy surface as having low barriers between nearly degenerate alternative structures. Not even catecholate ligand, a weak π-donor, brings the RuH2(PiPr3)2 moiety to oxidation state II.

Acknowledgements

This work was supported by the National Science Foundation.

References

  1. M. Aracama, M. A. Esteruelas, F. J. Lahoz, J. A. Lopez, U. Meyer, A. Oro and H. Werner, Inorg. Chem., 1991, 30, 288 CrossRef.
  2. D. F. Gusev, R. Kuhlman, J. Rambo, H. Berke, O. Eisenstein and K. G. Caulton, J. Am. Chem. Soc., 1995, 117, 281 CrossRef CAS.
  3. J. Espuelas, M. A. Esteruelas, F. J. Lahoz, L. A. Oro and N. Ruiz, J. Am. Chem. Soc., 1993, 115, 4683 CrossRef CAS.
  4. R. A. Sanchez-Delgado, M. Rosales, M. A. Esteruelas and L. A. Oro, J. Mol. Catal. A: Chem., 1995, 96, 231 CrossRef CAS.
  5. A. J. Edwards, M. A. Esteruelas, F. J. Lahoz, A. M. Lopez, E. Onate, L. A. Oro and J. I. Tolosa, Organometallics, 1997, 16, 1316 CrossRef CAS.
  6. M. A. Esteruelas, A. M. Lopez, N. Ruiz and J. I. Tolosa, Organometallics, 1997, 16, 4657 CrossRef CAS.
  7. G. Barea, M. A. Esteruelas, A. Lledos, A. M. Lopez and J. I. Tolosa, Inorg. Chem., 1998, 37, 5033 CrossRef CAS.
  8. P. Crochet, M. A. Esteruelas, A. M. Lopez, M.-P. Martinez, M. Olivan, E. Onate and N. Ruiz, Organometallics, 1998, 17, 4500 CrossRef CAS.
  9. M. L. Buil, O. Eisenstein, M. A. Esteruelas, C. Garcia-Yebra, E. Gutierrez-Puebla, M. Olivan, E. Onate, N. Ruiz and M. A. Tajada, Organometallics, 1999, 18, 4949 CrossRef CAS.
  10. R. Castarlenas, M. A. Esteruelas, E. Gutierrez-Puebla, Y. Jean, A. Lledos, M. Martin and J. Tomas, Organometallics, 1999, 18, 4296 CrossRef CAS.
  11. R. Castarlenas, M. A. Esteruelas, E. Gutierrez-Puebla, Y. Jean, A. Lledos, M. Martin, E. Onate and J. Tomas, Organometallics, 2000, 19, 3100 CrossRef CAS.
  12. N. Dolker and G. J. Frenking, Organomet. Chem., 2001, 617–618, 225 CrossRef CAS.
  13. F. Maseras and O. Eisenstein, New J. Chem., 1998, 22, 5 RSC.
  14. S.-Y. Yang, W.-H. Leung and Z. Lin, Organometallics, 2001, 20, 3198 CrossRef CAS.
  15. G. B. Richter-Addo, R. A. Wheeler, C. A. Hixon, L. Chen, M. A. Khan, M. K. Ellison, C. E. Schulz and W. R. Scheidt, J. Am. Chem. Soc., 2001, 123, 6314 CrossRef CAS.
  16. W. Stuer, J. Wolf, H. Werner, P. Schwab and M. Schulz, Angew. Chem. Int. Ed., 1998, 37, 3421 CrossRef CAS.
  17. J. N. Coalter, J. C. Bollinger, J. C. Huffman, U. Werner-Zwanziger, K. G. Caulton, E. R. Davidson, H. Gerard, E. Clot and O. Eisenstein, New J. Chem., 2000, 24, 9 RSC.
  18. G. Ferrando, H. Gerard, G. J. Spivak, J. N. Coalter, J. C. Huffman, O. Eisenstein and K. G. Caulton, Inorg. Chem., 2001, 40, 6610 CrossRef CAS.
  19. J. Wolf, W. Stuer, C. Grunwald, O. Gevert, M. Laubender and H. Werner, Eur. J. Inorg. Chem., 1998, 1827 CrossRef CAS.
  20. We have independently prepared Ru(H)2Cl2(PiPr3)2 and find that the literature value for 2JPH, 19 Hz, is incorrect. We measure it as 30 Hz, which is much closer to the value in Ru(H)2(catecholate)(PiPr3)2, 35 Hz.
  21. J. F. Hartwig, R. A. Andersen and R. G. Bergman, Organometallics, 1991, 10, 1875 CrossRef CAS.
  22. R. Kuhlman, W. E. Streib, J. C. Huffman and K. G. Caulton, J. Am. Chem. Soc., 1996, 118, 6934 CrossRef CAS.
  23. D.-H. Lee, J. Chen, J. W. Faller and R. H. Crabtree, Chem. Commun., 2001, 213 RSC.
  24. D. J. Cole-Hamilton, R. J. Young and G. Wilkinson, J. Chem. Soc., Dalton Trans., 1976, 1995 RSC.
  25. T. J. Johnson, P. S. Coan and K. G. Caulton, Inorg. Chem., 1993, 32, 4594 CrossRef CAS.
  26. C. C. Bickford, T. J. Johnson, E. R. Davidson and K. G. Caulton, Inorg. Chem., 1994, 33, 1080 CrossRef CAS.
  27. T. J. Johnson, K. Folting, W. E. Streib, J. D. Martin, J. C. Huffman, S. A. Jackson, O. Eisenstein and K. G. Caulton, Inorg. Chem., 1995, 34, 488 CrossRef CAS.
  28. R. Kuhlman, D. G. Gusev, I. L. Eremenko, H. Berke, J. C. Huffman and K. G. J. Caulton, Organomet. Chem., 1997, 536/537, 139 CrossRef.
  29. D. V. Yandulov, D. Huang, J. C. Huffman and K. G. Caulton, Inorg. Chem., 2000, 39, 1919 CrossRef CAS.
  30. D. G. Gusev, R. Kuhlman, G. Sini, O. Eisenstein and K. G. Caulton, J. Am. Chem. Soc., 1994, 116, 2685 CrossRef CAS.
  31. H. Chun, C. N. Verani, P. Chaudhuri, E. Bothe, E. Bill, T. Weyhermueller and K. Wieghardt, Inorg. Chem., 2001, 40, 157 CrossRef CAS.
  32. C. G. Pierpont and C. W. Lange, Prog. Inorg. Chem., 1994, 41, 331 CAS.
  33. C. G. Pierpont, Coord. Chem. Rev., 2001, 216–217, 99 CrossRef CAS.
  34. E. J. Miller and T. B. Brill, Inorg. Chem., 1983, 22, 2392 CrossRef CAS.
  35. S. Bhattacharya, P. Gupta, F. Basuli and C. G. Pierpont, Inorg. Chem., 2002, 41, 5810 CrossRef CAS.
  36. D. Herebian, E. Bothe, E. Bill, T. Weyhermueller and K. Wieghardt, J. Am. Chem. Soc., 2001, 123, 10012 CrossRef CAS.
  37. J. A. Ayllon, S. Sabo-Etienne, B. Chaudret, S. Ulrich and H.-H. Limbach, Inorg. Chim. Acta, 1997, 259, 1 CrossRef CAS.
  38. R. H. Crabtree, J. Organomet. Chem., 1998, 557, 111 CrossRef CAS.
  39. D. G. Gusev, A. J. Lough and R. H. Morris, J. Am. Chem. Soc., 1998, 120, 13138 CrossRef CAS.
  40. R. Kuhlman, E. Clot, C. Leforestier, W. E. Streib, O. Eisenstein and K. G. Caulton, J. Am. Chem. Soc., 1997, 119, 10153 CrossRef CAS.
  41. R. Custelcean and J. E. Jackson, Chem. Rev., 2001, 101, 1963 CrossRef CAS.
  42. J. P. Jesson and P. Meakin, J. Am. Chem. Soc., 1974, 96, 5760 CrossRef CAS.
  43. P. Meakin and J. P. Jesson, J. Am. Chem. Soc., 1974, 96, 5751 CrossRef CAS.

Footnote

† CCDC reference numbers 245545 and 245546. See http://www.rsc.org/suppdata/nj/b4/b411487f/ for crystallographic data in .cif or other electronic format.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2005
Click here to see how this site uses Cookies. View our privacy policy here.