Ruthenium complexes with tridentate PNX (X = O, S

Download PDF
0 downloads 0 Views 314KB Size Report
Comment
Silver triflate can act as an activator and improve enantioselectivity23 and was used in some reactions. The results are shown in Table 2. Complexes with activity ...
PAPER

www.rsc.org/dalton | Dalton Transactions

Ruthenium complexes with tridentate PNX (X = O, S) donor ligands† Simon R. Bayly, Andrew R. Cowley, Jonathan R. Dilworth* and Caroline V. Ward Received 2nd November 2007, Accepted 22nd January 2008 First published as an Advance Article on the web 6th March 2008 DOI: 10.1039/b717025d The ligands, PhPNXMe (1), PhPNXPh (2), and PhPNSMe (3), (PhPNX = 2-Ph2 P-C6 H4 CH= NC6 H4 X-2; X = O, S) have been prepared. A range of new ruthenium complexes were synthesised using these and related ligands, namely: [{RuCl(PhPNO)}2 Cl] (4), [Ru(PhPNO)2 ] (5), [RuCl(PhPNXR)(PPh3 )]BPh4 [X = O, R = Me (6); X = O, R = Ph (7); X = S, R = Me (8)], [{RuCl(PhPNX R)}2 Cl]X [X = O, R = Me, X = Cl− (9); X = S, R = Me, X = BPh4 − or PF6 − (10)], and [RuCl(PhPNO-g6 C6 H5 )]BPh4 (11). The catalytic activity of these complexes with respect to the hydrosilyation of acetophenone and the hydrogenation of styrene has been investigated, giving an insight into the requirements for an active complex in these reactions.

Introduction Recently the potential for catalytic activity of platinum group metal complexes incorporating phosphorus–sulfur donor ligands has been realised.1–3 The differing binding capabilities of the two donor atoms gives rise to the possibility of group displacement by substrates and thus hemilability. Recently we synthesised a series of ruthenium aromatic phosphinothioether complexes and investigated their catalytic activity towards the hydrosilyation of acetophenone. Disappointingly such complexes were found to be essentially inactive.4 It was therefore of interest to synthesise tridentate aromatic ligands with three heteroatoms, P, N and X (X = S, O) and investigate the effect of changing denticity and donor atom type upon coordination chemistry and catalytic activity. A review of the literature reveals that little ruthenium chemistry of this type has been carried out. Previous investigations have been mostly concerned with ruthenium–PNOH and –PNO− systems.5–7 Some of these ruthenium–PNO complexes have been applied to the asymmetric transfer hydrogenation of ketones and have been found to be active, with both conversions and enantioselectivities often above 70%.7,8

Scheme 1

A generalised PhPNX ligand.

Scheme 2

Formation of thiazoline and thiazole rings.

Scheme 3

Attempted synthesis of the disulfide imine.

Results and discussion This investigation began with the attempted synthesis of PNS tridentate ligand systems (Scheme 1). The literature shows that aldehydes react with aminothiolates to form benzothiazolines, which can subsequently be oxidised to benzothiazoles (Scheme 2).9–11 While thiazolines can ring open in the presence of ions to give complexes of iminothiolates, the thiazoles cannot. In an attempt to avoid this ring formation the reaction of 2-(diphenylphosphino)benzaldehyde was carried out with 2aminodisulfide (Scheme 3). However, in our hands this reaction yielded PhP(O)NHS-cyclic (Scheme 4). This structure has Chemistry Research Laboratory, University of Oxford, Mansfield Road, Oxford, UK OX1 3TA. E-mail: [email protected]; Tel: +44 (0)1865 285151, +44 (0)1865 285155 † CCDC reference numbers 666144–666152. For crystallographic data in CIF or other electronic format see DOI: 10.1039/b717025d

2190 | Dalton Trans., 2008, 2190–2198

previously been reported.10 It appears that both oxidation of phosphorus and reduction of the disulfide bond occurs during the reaction or subsequent recrystallisation. Attention was therefore This journal is © The Royal Society of Chemistry 2008

˚ ) and bond angles (◦ ) for crystallographTable 1 Selected bond lengths (A ically characterised compounds

Scheme 4 PhP(O)NHS-cyclic ligand.

focused on the neutral thioether ligand, PhPNSMe, which does not undergo cyclisation reactions. Synthesis of PhPNX ligands 2-(Diphenylphosphino)benzaldehyde,12 PhPNOH9 and PhPNOMe13 (1) were synthesised by the literature procedures.9 PhPNOPh (2). The new ligand, 2, was prepared from the reaction of 2-(diphenylphosphino)benzaldehyde with 2phenoxyaniline, in ethanol. Analysis showed the ligand to be contaminated by small amounts of both starting materials but attempts at purification by vacuum distillation resulted in decomposition. However, the ligand was found to be sufficiently pure for subsequent reaction with ruthenium precursors. PhPNSMe (3). Ligand 3 was prepared by the reaction of 2-(methylmercapto)aniline with 2-(diphenylphosphino)benzaldehyde, in chloroform. This ligand has similar solubility to ligands 1 and 2. The structure of 3 was confirmed by single crystal X-ray analysis (Fig. 1). Selected bond lengths and angles for crystallographically characterised compounds are presented in Table 1. The geometry around the C=N bond is essentially co-planar, with the C(1)–C(6) and C(8)–C(13) phenyl groups trans to one another. In order for the ligand to be able to coordinate a metal centre in a tridentate manner, through all three donors, rotation is required around the N(1)–C(8) bond, rendering the two phenyl rings cis. All three donors then have the correct configuration for mer coordination.

3 P(1)–C(1) S(1)–C(14) C(7)–N(1)

1.848(3) 1.789(3) 1.281(3)

P(1)–C(1)–C(6) C(7)–N(1)–C(8) C(1)–C(6)–C(7) C(13)–S(1)–C(14)

120.28(18) 117.9(2) 120.8(2) 102.43(14)

4 Ru(1)–Ru(2) Ru(1)–Cl(1) Ru(1)–O(1) Ru(1)–N(1) Ru(1)–P(1) Ru(1)–O(2)

2.7528(5) 2.412(1) 2.115(3) 2.000(3) 2.242(1) 2.141(3)

Cl(1)–Ru(1)–Cl(2) Ru(2)–Ru(1)–O(1) Cl(1)–Ru(1)–O(1) Ru(2)–Ru(1)–N(1) Ru(2)–Ru(1)–P(1) Ru(2)–Ru(1)–O(2) Ru(1)–Cl(1)–Ru(2) Ru(1)–O(1)–Ru(2) Cl(1)–Ru(2)–Cl(3) Ru(1)–Ru(2)–P(2)

93.23(4) 49.28(7) 87.50(8) 116.5(1) 132.33(3) 48.68(7) 69.71(3) 81.29(9) 93.75(4) 127.14(3)

4a Ru(1)–Cl(1) Ru(1)–P(1) Ru(1)–N(1) Ru(1)–O(1) O(1)–H(1)

2.3428(8) 2.2463(8) 2.066(3) 2.157(2) 0.87(5)

Cl(1)–Ru(1)–Cl(2) Cl(2)–Ru(1)–Cl(3) Cl(1)–Ru(1)–P(1) Cl(1)–Ru(1)–N(1) Cl(2)–Ru(1)–N(1) Cl(1)–Ru(1)–O(1) P(1)–Ru(1)–O(1) N(1)–Ru(1)–O(1) Ru(1)–P(1)–C(1)

96.38(3) 167.60(3) 92.31(3) 171.85(7) 84.07(7) 92.40(7) 174.74(7) 79.5(1) 109.85(11)

5 Ru(1)–P(1) Ru(1)–N(1) Ru(1)–O(1)

2.2499(8) 2.053(2) 2.097(2)

P(1)–Ru(1)–N(1) P(1)–Ru(1)–O(1) N(1)–Ru(1)–O(1) P(1)–Ru(1)–P(2) N(1)–Ru(1)–P(2) O(1)–Ru(1)–P(2) Ru(1)–P(1)–C(1)

94.43(8) 171.38(8) 80.9(1) 97.00(3) 95.03(7) 90.63(8) 113.12(11)

MS contains peaks which correspond to the molecular ion and two further fragments, [RuCl2 (PhPNO)]− and [RuCl(PhPNO)]− . Paramagnetism and lability precluded NMR assignment. X-Ray analysis of a single crystal grown under nitrogen from a solution of the isolated solid in dry methanol revealed the symmetrical dimeric structure of [{Ru2.5/2.5 Cl(PhPNO)}2 Cl] (4) (Fig. 2). On standing in air, the solution changes in colour from purple to brown and after work-up crystals were obtained from methanol/diethyl ether. X-Ray analysis of these

Fig. 1 ORTEP representation of ligand 3 (40% thermal ellipsoids, hydrogen atoms omitted for clarity).

Synthesis of PhPNX complexes [{RuCl(PhPNO)}2 Cl] (4). Complex 4 was prepared from the reaction of diruthenium-tetrachlorobis(4-cymene) with PhPNOH, in methanol. This complex is air sensitive and changes colour from purple to brown over tens of minutes in the presence of air, in both the solid state and solution. It is soluble in tetrahydrofuran and methanol and is insoluble in diethyl ether. Negative ion ES This journal is © The Royal Society of Chemistry 2008

Fig. 2 ORTEP representation of 4 (40% thermal ellipsoids, hydrogen atoms omitted for clarity).

Dalton Trans., 2008, 2190–2198 | 2191

showed them to be [RuIII Cl3 (PhPNOH)] (4a) (Fig. 3). From these structures we propose the following decomposition pathway: [{RuII/II Cl(PhPNO)}2 Cl]− is initially formed in the reaction and oxidises to give 4, which subsequently, on prolonged standing in solution, forms 4a. The analogous reaction using a 1 : 2 ratio of ruthenium precursor to ligand (based on the number of moles of dimer), was carried out in the presence of two equivalents of triethylamine. A brown powder was isolated (40%, assuming the formulation below). After several hours of standing in air, the solid became black. Analysis of the brown solid, including positive ion ES MS, suggests the formation of [RuII (PhPNO)2 ] (5). This complex was also prepared using two equivalents of ligand with [RuCl2 (PPh3 )2 ] as discussed below.

A second complex 5a was isolated from the solution after filtering off 5. ES MS and analytical data were consistent with this being monoprotonated 5. An X-ray crystal structure confirmed the coordination about the metal but was not of sufficient quality to discuss further. The addition of excess sodium tetraphenylborate in methanol solution to a sample of the initial reaction mixture does not produce a precipitate and this supports the formation of a neutral rather than a charged product. Identical complexes are formed whether one or two equivalents of ligand are utilised in the reaction. However, the yield is maximised in the latter. Complex 5 can also be prepared from the reaction of [{RuCl2 (cymene)}2 ] with one equivalent of ligand in the presence of base. X-Ray crystal structure of 5. The structure of 5 was confirmed by single crystal X-ray analysis (Fig. 4). This complex has a mer– mer configuration with the two nitrogens trans to one another and the two phosphorus atoms and two oxygens cis. Ru–N, Ru–O and ˚ , ∼2.1 A ˚ and ∼2.3 A ˚, Ru–P bond lengths in this complex are ∼2.1 A respectively. These are very similar to those measured for the ruthenium dimer 5 and agree with those of other ruthenium(II)– PNO complexes with similar ligands.6,7

Fig. 3 ORTEP representation of 4a (40% thermal ellipsoids, hydrogen atoms omitted for clarity).

X-Ray crystal structures of 4 and 4a. A significant feature of the structure of 4 is the presence of a Ru–Ru bond, indicated by the ˚ . The total valence electron count Ru–Ru separation of 2.7528(5) A for the dimer is 35 electrons. The Ru–N, Ru–O and Ru–P bond ˚ , ∼2.1 A ˚ and ∼2.3 A ˚ , respectively. These are in lengths are ∼2.0 A agreement with those of other ruthenium complexes reported.12–15 ˚ . The chloride All Ru–Cl bond lengths are very similar, at ∼2.4 A shared between the two ruthenium atoms lies equidistant between them. The equivalent angles on each side of the dimer are also very similar. For example Cl(1)–Ru(1)–Cl(2) is 93.23(4)◦ and Cl(1)– Ru(2)–Cl(3) is 93.75(4)◦ . Also Ru(1)–Ru(2)–P(2) is 127.14(3)◦ and Ru(2)–Ru(1)–P(1) is 132.33(3)◦ . All these details are consistent with a delocalised structure with an average oxidation state for each ruthenium of 2.5. Complex 4a exists as two polymorphs. Above temperatures of ∼220 K the material has a smaller unit cell and at 250 K the space group is P21 /n. The low temperature form, for which crystallographic data has been analysed, has a space group of P21 /c. The overall geometry about the Ru centres is pseudooctahedral, with the PhPNO ligand adopting its familiar mer ˚ and are configuration. All Ru–Cl bond lengths are ∼2.3 A consistent with those for other reported ruthenium complexes. The same is true for the Ru–N, Ru–O and Ru–P bond lengths.12–15 [Ru(PhPNO)2 ] (5). Reaction of [RuCl2 (PPh3 )3 ] with two equivalents of PhPNOH, in methanol under reflux, produced complex 5. In air this complex decomposes rapidly in the solid state and in solution. It is soluble in methanol and chlorinated solvents and insoluble in diethyl ether. Positive ion EI MS shows the molecular ion and also a peak corresponding to the fragment [Ru(PhPNO)]+ . 2192 | Dalton Trans., 2008, 2190–2198

Fig. 4 ORTEP representation of 5 (40% thermal ellipsoids, hydrogen atoms omitted for clarity).

Solution spectroscopic studies of the Ru/PhPNO system Although it is possible to isolate them pure in the solid state, 31 P and 1 H NMR (CDCl3 ) spectra show that 4 and 5 form complex mixtures of species in solution. The Ru/PhPNO system can participate in a number of potential equilibria (Scheme 5) and the 31 P NMR spectra in particular show resonances consistent with species of the type shown. [RuCl(PhPNXR)(PPh3 )]BPh4 [XR = OMe (6), OPh (7), SMe (8)]. Complexes 6, 7 and 8 were synthesised from the reaction of [RuCl2 (PPh3 )3 ] with one equivalent of ligand in methanol. The three complexes have similar solubilities. All are soluble in chlorinated solvents, acetonitrile and methanol and are insoluble in diethyl ether and hexane. Complex 7 changes in colour from This journal is © The Royal Society of Chemistry 2008

Scheme 5 Solution chemistry of the Ru/PhPNO system.

red to yellow in acetonitrile, possibly due to coordination of solvent, and in dichloromethane becomes brown over a period of several days. Elemental analysis has been carried out and the results for 6 and 7 agree with the complexes formulated. Satisfactory values for 8 could not be obtained. Positive ion ESMS and FAB support the formation of all three complexes. Each complex also gives an absorption in the IR spectrum due to the C=N stretch. These are observed at 1579 cm−1 and 1638 cm−1 for 6, 1580 cm−1 for 7 and 1579 cm−1 for 8. In solution these species gave rise to complex mixtures due to dissociation and/or decomposition, precluding NMR assignment. The reaction of PhPNOMe with excess [RuCl2 (PPh3 )3 ] in tetrahydrofuran solution has previously been reported.16 The red solid obtained was formulated as [RuCl2 (PhPNOMe)(PPh3 )] (this is similar to 6 but with the chloride counter-ion now bound to the metal centre). However, no crystal structure of this compound has been reported. Its catalytic activity towards the transfer hydrogenation of acetophenone to 1-phenylethanol was measured (conditions: 82 ◦ C, propan-2-ol, acetophenone : ruthenium : NaOH = 500 : 1 : 24) and the yield after four hours was found to be 98%. The TOF was found to be 730 h−1 and the reaction was not enantioselective.16 [{RuCl(PhPNX R)}2 Cl]X [X R = OMe, X = Cl (9), X R = SMe, X = BPh4 − or PF6 − (10)]. Complexes 9 and 10 were prepared by reaction of diruthenium-tetrachlorobis(4-cymene) with ligands 1 and 3, respectively, in methanol solution. Similar chloride bridged ruthenium complexes are reported in the literature.17 Although analogous in formulae and presumably in structure, complexes 9 and 10 differ from each other in their sensitivity to oxygen. Upon exposure to air, complex 9 changes in colour from blue to brown over several hours in the solid state. In contrast complex 10 does not visibly react. A solution of complex 9 in acetonitrile rapidly changes colour to red, presumably due to incorporation of acetonitrile. This is supported by the observation of peaks in the positive ion 



−

This journal is © The Royal Society of Chemistry 2008



ESMS (using a solution of the complex in MeCN) attributable to [RuCl(MeCN)2 (PhPNOMe)]+ and [RuCl(MeCN)(PhPNOMe)]+ . In methanol no colour change was observed and a peak corresponding to [RuCl(MeOH)(PhPNOMe)]+ was observed. Complex 9 is insoluble in hexane and diethyl ether. Complex 10 has a similar range of solubilities but does not show any colour change in solution. Positive ion FAB and ESMS and elemental analysis were in agreement with the calculated formula. However, satisfactory values could not be obtained for 9. Despite this the HPLC of 9 displays a single peak at RT 1.22 min. Also the IR spectrum contains an absorption at 1590 cm−1 , assigned to the C=N stretching frequency. Similarly the IR spectrum of 10 has an absorption at 1579 cm−1 (like 8 this is also shifted to lower wavenumber compared to the free ligand). Solution spectroscopic studies of Ru/PhPNXR (X = O, R = Me, Ph; X = S, R = Me) systems Similarly to 4 and 5, complexes 6–10 form complex mixtures in solution. Possibilities for species formed include the presence of different geometrical isomers, monomer–dimer equilibria, and dissociation of triphenylphosphine where present. [RuCl(PhPNO-g6 C6 H5 )]BPh4 (11). Complex 11 (Scheme 6) was formed from the reaction of diruthenium-tetrachlorobis(4cymene) with PhPNOPh, in methanol. It is soluble in chlorinated solvents and methanol and is insoluble in hexane and diethyl

Scheme 6 Proposed structure of 11.

Dalton Trans., 2008, 2190–2198 | 2193

ether. Unlike many of the ruthenium PNX systems, 11 appears relatively stable in solution and only one species is observed by NMR and a full spectral assigment can be made. Comparison with previously reported g6 -arene chelates confirms the proposed g6 -coordination of the phenoxy group.18 The 31 P NMR spectrum shows a single peak at 42.8 ppm. The 1 H NMR spectrum, measured in d6 -DMSO, shows five g6 -C6 H5 resonances for the inequivalent protons, between 2.4 ppm and 6.0 ppm, each of which has a J H–H coupling of ∼7 Hz. No J P–H couplings are observed. The six g6 -C6 H5 carbons show six broad signals (up to 19 Hz in width at half height) in the 13 C NMR spectrum, in the region 30–100 ppm. The broad signals may conceal J C–P couplings. The remaining phenyl resonances occur in typical regions in both the 1 H and 13 C spectra and the CH signal has a chemical shift of 8.79 ppm in the 1 H spectrum (J H–P 1 Hz) and of 175.4 ppm in the 13 C spectrum (J C–P 7 Hz). Catalysis The newly synthesised complexes were screened for their activity towards the hydrosilylation of acetophenone to 1-phenylethanol. Hydrosilylation is generally taken as an indicator of catalytic activity and when encouraging results were obtained, further catalysis measurements were made. These included hydrosilylation with the addition of silver triflate as an activator and the hydrogenation of styrene to ethylbenzene. It was hoped that the results would permit some discussion of the dependence of activity on ligand denticity, charge, donor atom set and ligand steric requirements. Hydrosilylation. Catalytic activity towards the hydrosilylation of acetophenone, to 1-phenylethanol using diphenylsilane was tested. Four known complexes were used as a standard: diruthenium-tetrachlorobis(4-cymene), [RuCl2 (PPh3 )3 ],19 [RuCl2 (dppe)2 ]20 and [RuCl2 (PhPOMe)2 ].21,22 The method used was adapted from the procedure described by Uemura, Hidai et al.23 Silver triflate can act as an activator and improve enantioselectivity23 and was used in some reactions. The results are shown in Table 2. Complexes with activity higher than [RuCl2 (PPh3 )3 ] and [RuCl2 (PhPOMe)2 ] (TOF of 19 h−1 and 16 h−1 , respectively), in

the absence of silver triflate, include 4, 5, 7 and 9. In particular the activity of 5 is surprising as it was anticipated that chelation by two tridentate ligands would reduce the activity by making reactive sites unavailable. Interestingly all these complexes contain PNO donors and presumably the activity is a consequence of the poor oxygen to metal binding strength. In contrast complexes containing PNS are essentially inactive. This is attributed to the much stronger donation of sulfur to ruthenium. HPLC studies of 8 and 10 also suggest stability of these complexes in solution, with the elution of one major peak in each case. This implies that the strength of the Ru–X bond (where X is S or O) is important, in part, in determining catalytic activity. The addition of silver triflate in these reactions has little effect or in fact decreases the activity in most cases. Exceptions do exist however and in the reactions with 9 and 10 with this additive, higher activities result. We postulate that this is due to abstraction of chloride from the chloride bridge and generation of vacant coordination sites for incoming substrate molecules. The activity is also increased by the addition of silver triflate in the reaction with 6. Again this is presumably due to chloride abstraction but as yet we do not have a rationale as to why this is the case for this complex and not for the analogous species 7 and 8. From these studies it appears that ligand lability is a critical factor in determining catalytic activity and this in turn is dependent on the donor set. Oxygen is a much weaker donor to ruthenium than sulfur. Also ligand denticity seems to play a part. In general, the complexes with the highest activities above have a single tridentate ligand, rather than multiple bidentate coordinating ligands. This is possibly because tridentate ligands can act in a hemilabile mode and maintain complex stability with two donor atoms, whilst the third, is detached creating a vacant coordination site. In these studies the charge of the ligand appears to have little effect upon activity. Hydrogenation. Hydrogenation of styrene was used to test the catalytic activity of complexes towards hydrogenation. This was carried out in the same way for each sample and in the absence of catalyst, no conversion was observed. The activity of [RuCl2 (PPh3 )3 ]19 was measured, so that comparisons could be drawn with the newly synthesised complexes. The results obtained

Table 2 Results of catalytic hydrosilylation using Ru/PNX complexes Catalyst

Conversion (%) (with AgOTf)

TONa (with AgOTf)

[RuCl2 (PPh3 )3 ] [RuCl2 (cymene)]2 Cl2 [RuCl2 (dppe)2 ] [RuCl2 (PhPOMe)2 ] [{RuCl(PhPNO)}2 Cl] [Ru(PhPNO)2 ] [Ru(PhPNO)2 ]b [Ru(PhPNO)2 ]c [RuCl(PhPNOMe)(PPh3 )]BPh4 [RuCl(PhPNOPh)(PPh3 )]BPh4 [RuCl(PhPNSMe)(PPh3 )]BPh4 [{RuCl(PhPNOMe)}2 Cl]Cl [{RuCl(PhPNSMe)}2 Cl]BPh4 [RuCl(PhPNO-g6 C6 H5 )Cl]BPh4

19 1 0.2 16 20 (13) 22 (11) 6 0.5 18 (23) 36 (29) 2 (2) 26 (36) 0 (3) 0.5 (0.6)

19 1 0.2 16 20 (13) 22 (11) 6 0.4 9 (12) 18 (15) 1 (1) 26 (36) 0 (3) 0.3 (0.4)

4 5 6 7 8 9 10 11

TON = TOF (h−1 ). b Product from the reaction of [RuCl2 (PPh3 )3 ], PhPNOH and triethylamine—formula unconfirmed. c Product from the reaction of [{RuCl2 (cymene)}2 ], PhPNOH and triethylamine—formula unconfirmed.

a

2194 | Dalton Trans., 2008, 2190–2198

This journal is © The Royal Society of Chemistry 2008

Table 3 Results of catalytic hydrogenation using Ru/PNX complexes Catalyst [RuCl2 (PPh3 )3 ] [{RuCl(PhPNO)}2 Cl] [Ru(PhPNO)2 ] [RuCl(PhPNOMe)(PPh3 )]BPh4 [RuCl(PhPNOPh)(PPh3 )]BPh4 [{RuCl(PhPNOMe)}2 Cl]Cl a

4 5 6 7 9

Conversiona (%)

TON

TOF/h−1

6 0 6 0 5 0

30 0 30 0 13 0

15 0 15 0 6 0

Estimated error ±0.5%.

are illustrated in Table 3. [RuCl2 (PPh3 )3 ] forms [RuCl(H)(PPh3 )3 ] upon reaction with base and investigations have shown this species to be very effective in the hydrogenation of terminal alkenes.24,25 Using the procedure described above it was found to have a TON of 30 and TOF of 15 h−1 . These values are relatively low because of the mild reaction conditions used, low concentration of metal complex and the absence of base. Two of the new complexes described are active under these conditions, 5 and 7. The activity of 5 is comparable to that of [RuCl2 (PPh3 )3 ], which is significant. However it is not clear why the other ruthenium–PNO complexes are inactive under the same conditions.

Conclusions Ligands of the type Ph2 PC6 H4 CH=NC6 H4 XR (X = O, R = H, Me, Ph; X = S, R = Me) react with [{RuCl2 (cymene)}2 ] to give products with structures highly dependent on the XR group. When XR = OH a phenoxy bridged dimer is formed whereas with XR = OMe a triple chloride bridge results. The reaction with XR = OPh yields a complex with a g6 -bound phenoxy phenyl group. By contrast reaction of these same ligands with [RuCl2 (PPh3 )3 ], generally leads to monomeric cationic complexes with the retention of PPh3 . In general complexes containing one tridentate ligand were found to be most active for hydrosilyation. Higher activities were obtained for alkoxy rather than thioether analogues, suggesting the dissociation of an oxygen donor atom is a crucial part of the catalytic cycle. Two of the new complexes (5 and 7) were found to be reasonably active for hydrogenation, but no structure–activity relationship could be elucidated.

Experimental General procedures Unless otherwise stated, all manipulations were carried out under an atmosphere of dinitrogen with dry solvents and using Schlenk line techniques. Solvents were pre-dried over sodium wire or calcium chloride prior to heating at reflux over the appropriate drying agent/indicator: methanol (magnesium methoxide), tetrahydrofuran (sodium/benzophenone), toluene (sodium). All solvents were degassed prior to use. [RuCl2 (PPh3 )3 ],19 PhPNOH9 and 2-(diphenylphosphino)-benzaldehyde26 were prepared according to literature procedures. All other reagents and solvents were purchased from commercial sources and were used as received. 1 H, 13 C and 31 P NMR spectra were recorded on a Varian Mercury-Vx 300 spectrometer at 300.0, 75.4 and 121.4 MHz, respectively, or a Varian Unit 500 MHz spectrometer at 499.9 MHz, This journal is © The Royal Society of Chemistry 2008

125.7 MHz and 202.4 MHz, respectively, at ambient temperature unless otherwise stated. 1 H and 13 C NMR spectra were referenced internally to residual solvent resonances or tetramethylsilane at zero ppm (for 1 H NMR) and 31 P spectra were externally referenced to 85% H3 PO4 at zero ppm. Electrospray mass spectrometry was performed on a Micromass LCT Time of Flight Mass Spectrometer. Electron impact mass spectrometry was performed on a Micromass GCT Time of Flight Mass Spectrometer, with a heated solid probe. FAB mass spectrometry was performed on a Fisons Autospec, using dithiothreitol and dithioerythritol or m-nitrobenzyl alcohol as the matrix. HPLC was carried out on a GILSON analytical/preparative instrument with a UV/VIS detector (k = 250 nm). A reversephase HAMILTON PRP-1 analytical column (4.1 × 150 mm, 10 lm) was used with an isocratic mobile phase of 95 : 5% acetonitrile : water and a flow rate of 1.5 mL min−1 . Gas chromatography was carried out on a Unicam ProGC gas chromatogram system using helium as the carrier gas and a flame ionisation detector (FID) with a hydrogen/air flame. A 2 metre Carbowax 20M packed column was used. Elemental analyses were performed by the elemental analysis department of the Inorganic Chemistry Laboratory, University of Oxford. Crystallography X-Ray crystallography was performed on a Bruker-Nonius KappaCCD diffractometer, using graphite-monochromated Mo-Ka ˚ ). Single crystals were mounted on a radiation (k = 0.71073 A glass fibre using perfluoropolyether oil and were cooled rapidly to 150 K in a stream of cold nitrogen using an Oxford Cryosystems CRYOSTREAM unit. Intensity data were processed using the DENZO-SMN package.27 Structures were solved using the directmethods programme SIR92,28 which located all non-hydrogen atoms. Subsequent full-matrix least-squares refinement was carried out using the CRYSTALS programme suite.29 Coordinates and anisotropic thermal parameters of all non-hydrogen atoms were refined. Hydrogen atoms were positioned geometrically after each cycle of refinement. A 3-term Chebychev polynomial weighting scheme was applied. Selected bond lengths and angles are presented in Table 1. Crystallographic data is shown in Table 4. Synthetic procedures PhPNOPh (2). 2-(Diphenylphosphino)benzaldehyde (520 mg, 1.79 mmol) was dissolved in non pre-dried ethanol (15 mL) with mild heating. An ethanolic solution of 2phenoxyaniline (332 mg, 1.79 mmol, in 5 mL ethanol) was added Dalton Trans., 2008, 2190–2198 | 2195

Table 4 Crystallographic data for compounds 3, 4, 4a, and 5 Compound

3

4

4a

5

Chemical formula

C26 H22 NPS

C30 H34 Cl3 NO3 PRu

Formula weight T/K ˚ k/A Crystal system Space group ˚ a/A ˚ b/A ˚ c/A a/◦ b/◦ c /◦ ˚3 V /A Z D/Mg m−3 l/mm−1 F 000 Crystal size/mm Reflections measured Unique reflections Rint Parameters refined Goodness of fit R wR

411.50 150 0.71073 Monoclinic P 21 /a 10.4453(4) 15.2797(6) 13.4194(5) 90 97.7604(6) 90 2122.1 4 1.288 0.240 864.912 0.05 × 0.20 × 0.30 19458 4976 0.090 262 1.0530 0.0380 0.0437

C50 H38 Cl3 N2 O2 P2 Ru2 + xCH4 O (x = 0.691) 1123.54 150 0.71073 Monoclinic P 21 /n 10.8467(2) 33.4618(4) 13.0799(2) 90 95.8351(5) 90 4722.8 4 1.580 0.924 2252.705 0.12 × 0.17 × 0.17 41954 10882 0.063 591 1.0563 0.0332 0.0352

C50 H38 N2 O2 P2 Ru + xCH4 O (x ∼ 1.63) 914.03 150 0.71073 Monoclinic P 21 /n 16.4446(2) 12.5025(2) 20.7078(4) 90 93.0930(7) 90 4251.3 4 1.428 0.493 1880.535 0.04 × 0.24 × 0.30 45136 10081 0.051 539 1.0213 0.0532 0.0510

dropwise. The reaction mixture was heated under reflux in air for three hours and after cooling to room temperature, the solvent was removed in vacuo. PhPNOPh was obtained as a yellow oil which contained small amounts of unreacted starting materials. Attempts at purification failed, therfore the crude product was used in complexation reactions. FTIR (Nujol mull), cm−1 : 1697 (w), C=N stretch; Mass spectrum [ES(+)]: m/z 458 [M]+ ; 1 H NMR: (CDCl3 ) d 6.30–8.35 (m, Ph, 23H), 9.35 [d, CH, 1H (J H–P 6 Hz)]; 31 P NMR: (CDCl3 ) d −13.7; 13 C NMR: (CDCl3 ) d 116.4–144.0 (Ph), 160.1 [d, CH (J C–P 25 Hz)] PhPNSMe (3). 2-(Methylmercapto)aniline (0.27 mL, 2.191 mmol) was dissolved in non pre-dried chloroform (50 mL) and 2-(diphenylphosphino)benzaldehyde (636 mg, 2.19 mmol), also dissolved in chloroform (10 mL), was added dropwise to this. The reaction mixture was heated under reflux in air for four hours, using a Dean–Stark trap. After cooling to room temperature the solvent was removed in vacuo to leave a yellow oil. Diethyl ether (10 mL) was added and the reaction mixture stirred overnight. A yellow precipitate formed and was collected by filtration in air, washed with diethyl ether and dried in vacuo. PhPNSMe was obtained as a yellow powder (548 mg, 61%). Elemental analysis, found (required)%: C 76.01 (75.89), H 5.30 (5.39), N 3.44 (3.40); FTIR (Nujol mull), cm−1 : 1618 (m), C=N stretch; Mass spectrum [ES(+)]: m/z 412 [M]+ ; HPLC: RT 8.00 min. (broad peak); 1 H NMR: (CDCl3 ) d 2.41 (s, SMe, 3H), 6.46 [m, HB1 , 1H (J H–H 8, 1 Hz)], 6.93 [m, HA3 , 1H(J H–P 5 Hz, J H–H 8, 1 Hz)], 7.02 [m, HB2 , 1H (J H–H 8, 1 Hz)], 7.14 [m, HB3 , 1H (J H–H 1 Hz)], 7.16 [m, HB4 , 1H (J H–H 1 Hz)], 7.27–7.39 (m, Ph, 10H), 7.33 (m, HA4 , 1H), 7.46 [m, HA2 , 1H (J H–H 1 Hz)], 8.40 [m, HA1 , 1H (J H–P 4 Hz, J H–H 8, 1 Hz)], 9.12 [d, CH, 1H (J H–P 6 Hz)]; 31 P NMR: (CDCl3 ) d −13.7; 13 C NMR: (CDCl3 ) d 14.8 (s, SMe), 117.6 (s, CB1 ), 124.4 (s, CB3 ), 2196 | Dalton Trans., 2008, 2190–2198

695.01 150 0.71073 Monoclinic P 21 /c 16.3162(2) 17.8674(3) 20.7202(3) 90 95.0265(7) 90 6017.3 8 1.534 0.873 2832.962 0.08 × 0.10 × 0.14 62008 14138 0.052 700 1.0300 0.0383 0.0467

125.1 (s, CB2 ), 126.3 (s, CB4 ), 127.7 [d, CA1 (J C–P 4 Hz)], 128.6–139.2 (Ph), 129.0 [d, CA2 (J C–P 5 Hz)], 131.1 (s, CA4 ), 133.2 (s, CA3 ), 158.3 [d, CH (J C–P 25 Hz)]. (4). Diruthenium-tetrachlorobis(4[{RuCl(PhPNO)}2 Cl] cymene) (63 mg, 0.10 mmol) and PhPNOH (78 mg, 0.21 mmol) were suspended in methanol (15 mL) and heated under reflux overnight. The reaction mixture became purple and after cooling to room temperature the solvent was removed in vacuo. Diethyl ether (5 mL) was added and the mixture triturated to yield a precipitate. This was collected by filtration under nitrogen, washed with diethyl ether and dried in vacuo. [{RuCl(PhPNO)}2 Cl] was obtained as an air sensitive, purple solid (48 mg, 44%). Elemental analysis, found (required)%: C 54.46 (56.16), H 3.92 (3.58), N 2.51 (2.62), Cl 10.20 (9.95); Mass spectrum [ES(−)]: m/z 1071 [M]− , 552 [RuCl2 (PhPNO)]− , 516 [RuCl(PhPNO)]− . [Ru(PhPNO)2 ] (5). [RuCl2 (PPh3 )3 ] (131 mg, 0.14 mmol) and PhPNOH (104 mg, 0.27 mmol) were suspended in methanol (15 mL) and the mixture was heated under reflux overnight. The solution became red and after cooling to room temperature the solvent was removed in vacuo. Diethyl ether (10 mL) was added and trituration took place. A red/brown precipitate formed and was collected by filtration under nitrogen, washed with diethyl ether and dried in vacuo. [Ru(PhPNO)2 ] was obtained as an air sensitive, red/brown powder (50 mg, 43%). FTIR (Nujol mull), cm−1 : 1585, (w), C=N stretch; Mass spectrum [EI(+)]: m/z 861.1160 [M]+ , 481.5463 [Ru(PhPNO)]+ ; HPLC: RT 4.13 min. [RuCl(PhPNOMe)(PPh3 )]BPh4 (6). [RuCl2 (PPh3 )3 ] (189 mg, 0.20 mmol) and PhPNOMe (78 mg, 0.20 mmol) were suspended in methanol (15 mL) and the mixture heated under reflux overnight. The solution became red and after cooling to room temperature, This journal is © The Royal Society of Chemistry 2008

was filtered in air and a small amount of a brown solid was removed. A methanolic solution of sodium tetraphenylborate (337 mg, 0.99 mmol, in 3 mL of non pre-dried methanol) was added to the filtrate. A red/orange precipitate formed and was collected by filtration in air, washed with methanol and dried in vacuo. [RuCl(PhPNOMe)(PPh3 )]BPh4 was obtained as an orange solid (128 mg, 58%). Elemental analysis, found (required)%: C 72.62 (73.35), H 5.29 (5.16), N 1.53 (1.26), Cl 3.03 (3.18); FTIR (Nujol mull), cm−1 : 1579 (m), 1638 (m), C=N stretch; Mass spectrum [ES(+)]: m/z 794 [M]+ , 758 [M − Cl]+ ; HPLC: RT 0.82 min. [RuCl(PhPNOPh)(PPh3 )]BPh4 (7). Procedure as for the synthesis of 7, using [RuCl2 (PPh3 )3 ] (685 mg, 0.72 mmol), PhPNOPh (327 mg, 0.72 mmol) and sodium tetraphenylborate (1.223 g, 3.57 mmol). A red precipitate formed and was collected by filtration in air, washed with methanol and dried in vacuo. [RuCl(PhPNOPh)(PPh3 )]BPh4 was obtained as a red solid (494 mg, 59%). Elemental analysis, found (required)%: C 73.90 (74.59), H 4.86 (5.06), N 1.20 (1.19), Cl 2.61 (3.02); FTIR (Nujol mull), cm−1 : 1580 (w), C=N stretch; Mass spectrum [ES(+)]: m/z 856 [M]+ , 822 [M − Cl]+ , 559 [M − (Cl + PPh3 )]+ ; HPLC: RT 1.10 min. [RuCl(PhPNSMe)(PPh3 )]BPh4 (8). Procedure as for the synthesis of 7, using [RuCl2 (PPh3 )3 ] (119 mg, 0.12 mmol), PhPNSMe (51 mg, 0.12 mmol) and sodium tetraphenylborate (212 mg, 0.62 mmol). A red/brown precipitate formed and was collected by filtration in air, washed with methanol and diethyl ether and dried in vacuo. [RuCl(PhPNSMe)(PPh3 )]BPh4 was obtained as a pink solid (76 mg, 54%). Elemental analysis, found (required)%: C 73.08 (72.31), H 5.08 (5.09), N 1.33 (1.24), Cl 2.12 (3.14); FTIR (Nujol mull), cm−1 : 1579 (w), C=N stretch; Mass spectrum [FAB(+)]: m/z 810.3 [M]+ , 775.2 [M − Cl]+ , 548.0 [M − PPh3 ]+ , 512.0 [M − (Cl + PPh3 )]+ ; HPLC: RT 0.99 min. (9). Diruthenium-tetrachloro[RuCl(PhPNOMe)}2 Cl]Cl bis(4-cymene) (49 mg, 0.08 mmol) and PhPNOMe (63 mg, 0.16 mmol) were dissolved in methanol (15 mL) and the mixture was heated under reflux overnight. The solution became purple in colour and after cooling to room temperature the solvent was removed in vacuo. An oily solid remained and was titurated with diethyl ether (10 mL) and a precipitate formed. This was collected by filtration under nitrogen, washed with diethyl ether and dried in vacuo. [{RuCl(PhPNOMe)}2 Cl]Cl was obtained as an air sensitive, blue/purple solid (62 mg, 68%). FTIR (Nujol mull), cm−1 : 1590 (m), C=N stretch; Mass spectrum [ES(+)], in acetonitrile: m/z 614 [RuCl(MeCN)2 (PhPNOMe)]+ , 573 [RuCl(MeCN)(PhPNOMe)]+ , in methanol: m/z 564 [RuCl(MeOH)(PhPNOMe)]+ ; HPLC: RT 1.22 min. [{RuCl(PhPNSMe)}2 Cl]X (X = BPh4 − , PF6 − ) (10). This complex was formed in a reaction analogous to that used in the synthesis of 9, but with slight modification. Dirutheniumtetrachlorobis(4-cymene) (36 mg, 0.06 mmol) and PhPNSMe (48 mg, 0.12 mmol) were reacted using the procedure described above and, after cooling to room temperature, a solution of sodium tetraphenylborate (204 mg, 0.60 mmol) (or ammonium hexafluorophosphate, 0.60 mmol) in non pre-dried methanol (∼5 mL) was added to the resulting red solution. A pink precipitate This journal is © The Royal Society of Chemistry 2008

formed and was collected by filtration in air, washed with diethyl ether and dried in vacuo. [{RuCl(PhPNSMe)}2 Cl]X was obtained as a red powder (84 mg, 98% X = BPh4 − ; 30% X = PF6 − ). Elemental analysis, found (required)%: (X = BPh4 − ) C 63.03 (62.92), H 4.54 (4.45), N 1.98 (1.93), Cl 5.45 (7.33); FTIR (Nujol mull), cm−1 : 1579 (w), C=N stretch; Mass spectrum [ES(+)]: m/z 1131 [M]+ , 1100 [M − Cl]+ , 548 [M − (PhPNSMe + 2Cl)]+ ; HPLC: RT 0.97 min. [RuCl(PhPNO-g6 C6 H5 )]BPh4 (11). Diruthenium-tetrachlorobis(4-cymene) (127 mg, 0.21 mmol) and PhPNOPh (190 mg, 0.42 mmol) were dissolved in methanol (15 mL) and the mixture was heated at reflux for five hours. The solution remained orange. After cooling to room temperature a solution of sodium tetraphenylborate (377 mg, 1.09 mmol) in non pre-dried methanol (∼5 mL) was added. An orange precipitate formed which was collected by filtration in air, washed with methanol and dried in vacuo. [RuCl(PhPNOPh-g6 C6 H5 )]BPh4 was obtained as an orange powder (284 mg, 75%). Elemental analysis, found (required)%: C 72.16 (72.33), H 5.32 (4.86), N 1.21 (1.53), Cl 2.58 (3.88); FTIR (Nujol mull), cm−1 : 1580 (m), C=N stretch; Mass spectrum [FAB(+)]: m/z 594.1 [M]+ , 558.1 [M − Cl]+ , 482.1 [M − (Cl + Ph)]+ ; HPLC: RT 0.87 min.; 1 H NMR: (d 6 -DMSO) d 2.43 [m, g6 -C6 H5 H1 , 1H (J H–H 7 Hz)], 4.46 [d, g6 -C6 H5 H2 , 1H (J H–H 6 Hz)], 4.83 [d, g6 -C6 H5 H3 , 1H (J H–H 6 Hz)], 5.79 [m, g6 -C6 H5 H4 , 1H (J H–H 7 Hz)], 5.95 [m, g6 -C6 H5 H5 , 1H (J H–H 7 Hz)], 6.73–8.01 (m, Ph, 38H), 8.79 [d, CH, 1H (J H–P 1 Hz)]; 31 P NMR: (d 6 -DMSO) d 42.8; 13 C NMR: (d 6 -DMSO) d 30.1 (s, g6 -C6 H5 H1 ), 83.5 (s, g6 -C6 H5 H2 ), 92.0 (s, g6 -C6 H5 H5 ), 92.2 (s, g6 -C6 H5 H4 ), 98.2 (s, g6 -C6 H5 H6 ), 103.5 (s, g6 -C6 H5 H3 ), 118.2–155.8 (Ph), 162.4–164.2 (BPh4 − ), 175.4 [d, CH (J C–P 7 Hz)].

Catalytic hydrosilyation of acetophenone The method utilised to ascertain the activity of the complexes was derived from that of Nishibayashi et al.23 The catalyst (0.01 mmol, 1 mol%) was dissolved in tetrahydrofuran (10 mL), in a Schlenk flask. Acetophenone (117 lL, 1 mmol) was added and the reaction mixture was heated to 65 ◦ C. An excess of diphenylsilane (371 lL, 2 mmol) was added and the flask sealed. Heating continued for a further one hour. After cooling to room temperature the reaction mixture was quenched with non pre-dried methanol (1 mL), to consume the remaining silane. After stirring at room temperature for 30 minutes, 0.2 M hydrochloric acid (2.5 mL) was added to hydrolyse the silyl ether. The reaction mixture was stirred overnight in air and was then analysed by GC. When utilised as an additive silver triflate (0.0026 g, 0.01 mmol) was added to the Schlenk flask with the solid complex, before THF addition. The flask was covered with aluminium foil to prevent photolytic decomposition of any silver salts. Percentage conversions were calculated from the ratio of areas of the eluted peaks of acetophenone and 1-phenylethanol. Calibration using standards showed the two substances to elute in a 1 : 1 ratio when equal amounts of both were injected. No correction factors were therefore applied to the areas. The temperature profile for GC analysis: maintain 60 ◦ C for 2 min, increase 10 ◦ C per min for 14 min, maintain 200 ◦ C for 4 min. Injector temperature: 220 ◦ C, detector temperature: 300 ◦ C. Dalton Trans., 2008, 2190–2198 | 2197

Catalytic hydrogenation of styrene To a solution of the catalyst (0.01 mmol) in toluene (5 mL) in a Schlenk flask was added styrene (571 lL, 5 mmol) and the flask sealed. This was degassed using three freeze–pump–thaw cycles and after warming to room temperature, hydrogen gas was permitted into the flask at a pressure of 2 atmospheres. This pressure was maintained for two hours whilst the reaction was stirred at room temperature. Samples of 0.5 mL were taken, diluted with non-pre-dried THF (2.5 mL) and analysed by GC. A correction factor of 0.91 was applied to the ethylbenzene peak. The temperature profile for GC analysis: maintain 60 ◦ C for 2 min, increase 10 ◦ C per min for 14 min, maintain 200 ◦ C for 19 min. Injector temperature: 350 ◦ C, detector temperature: 350 ◦ C.

Acknowledgements We thank the EPSRC for financial support, Johnson-Matthey for the donation of materials and Dr N. H. Rees of the University of Oxford for his assistance with NMR.

Notes and references 1 J. R. Dilworth, J. R. Miller, N. Wheatley, M. J. Baker and J. G. Sunley, J. Chem. Soc., Chem. Commun., 1995, 1579. 2 E. Hauptman, P. J. Fagan and W. Marshall, Organometallics, 1999, 18, 2016. ´ C. Claver and A. M. Masdeu-Bulto, ´ Coord. Chem. Rev., 3 J. C. Bayon, 1999, 193–195, 73. 4 C. V. Ward, A. R. Cowley and J. R. Dilworth, manuscript in preparation. 5 K. Nakajima, S. Ishibashi and M. Kojima, Chem. Lett., 1998, 10, 997. 6 P. Bhattacharyya, M. L. Loza, J. Parr and A. M. Z. Slawin, J. Chem. Soc., Dalton Trans., 1999, 2917. 7 H.-L. Kwong, W.-S. Lee, T.-S. Lai and W.-T. Wong, Inorg. Chem. Commun., 1999, 2, 66. 8 H. Dai, X. Hu, H. Chen, C. Bai and Z. Zheng, Tetrahedron: Asymmetry, 2003, 14, 1467.

2198 | Dalton Trans., 2008, 2190–2198

9 J. R. Dilworth, S. D. Howe, A. J. Hutson, J. R. Miller, J. Silver, R. M. Thompson, M. Harman and M. B. Hursthouse, J. Chem. Soc., Dalton Trans., 1994, 3553. ˜ ´ 10 A. Castineiras, M. C. Gomez and P. Sevillano, J. Mol. Struct., 2000, 554, 301. 11 R. G. Charles and H. Freiser, J. Org. Chem., 1953, 18, 422. 12 N. R. Champness, R. J. Forder, C. S. Frampton and G. Reid, J. Chem. Soc., Dalton Trans., 1996, 1261. 13 A. A. Batista, E. A. Polato, S. L. Queiroz, O. R. Nascimento, B. R. James and S. J. Rettig, Inorg. Chim. Acta, 1995, 230, 111. 14 B. Serli, E. Zangrando, E. Iengo, G. Mestroni, L. Yellowlees and E. Alessio, Inorg. Chem., 2002, 41, 4033. 15 T. Hirano, M. Kuroda, N. Takeda, M. Hayashi, M. Mukaida, T. Oi and H. Nagao, J. Chem. Soc., Dalton Trans., 2002, 2158. 16 P. Crochet, J. Gimeno, J. Borge and S. Garc´ıa-Granda, New J. Chem., 2003, 27, 414. 17 L. DiMichele, S. A. King and A. W. Douglas, Tetrahedron: Asymmetry, 2003, 14, 3427. 18 J. R. Dilworth, Y. F. Zheng, S. F. Lu and Q. J. Wu, Inorg. Chim. Acta, 1992, 194, 99–103. 19 P. S. Hallman, T. A. Stephenson and G. Wilkinson, Inorg. Synth., 1970, 12, 237. 20 C. W. Jung, P. E. Garrou, P. R. Hoffman and K. G. Caulton, Inorg. Chem., 1984, 23, 726. 21 J. C. Jeffrey and T. B. Rauchfuss, Inorg. Chem., 1979, 18, 2658. 22 T. B. Rauchfuss, F. T. Patino and D. M. Roundhill, Inorg. Chem., 1975, 14, 652. 23 Y. Nishibayashi, I. Takei, S. Uemura and M. Hidai, Organometallics, 1998, 17, 3420. 24 P. S. Hallman, D. Evans, J. A. Osborn and G. Wilkinson, Chem. Commun., 1967, 305. 25 P. S. Hallman, B. R. McGarvey and G. Wilkinson, J. Chem. Soc. A, 1968, 3143. 26 J. E. Hoots, T. B. Rauchfuss and D. A. Wrobleski, Inorg. Synth., 1982, 21, 175. 27 Z. Otwinowski and W. Minor, Processing of X-Ray Diffraction Data Collected in the Oscillation Mode, Methods Enzymol., 1997, 276, 307. 28 A. Altomare, G. Cascarano, G. Giacovazzo, A. Guagliardi, M. C. Burla, G. Polidori and M. Camalli, J. Appl. Crystallogr., 1994, 27, 435. 29 D. J. Watkin, C. K. Prout, J. R. Carruthers, P. W. Betteridge and R. I. Cooper, CRYSTALS issue 11, Chemical Crystallography Laboratory, Oxford, UK, 2001.

This journal is © The Royal Society of Chemistry 2008