Silver(I)-mediated isomerisation of trans-[RuCl2(P-P) 2] (P-P =four-membered ring chelate diphosphine ligand) to cis-[{Ru(P-P)2-(μ-Cl)2}nAg]+ species, and applications in stereoselective syntheses of cis- and trans-[RuCl(L)(P-P)2]+ (L = neutral ligand)

Article abstract of DOI:10.1039/b009645h

When complexes trans-[RuCl₂(diphos)₂] (diphos = (Ph₂P)₂C=CH₂, Ph₂PCH ₂PPh₂ or Ph₂PNHPPh₂) were treated with silver salts of poorly coordinating anions (e.g. AgBF₄, AgO ₃SCF₃) in 1,2-dichloroethane halide abstraction was slow, even at reflux, and the reaction proceeded by an interesting mechanism involving a trans to cis isomerisation. At room temperature, a complex, characterized as [RuCl(diphos)₂(μ-Cl)Ag]⁺, forms. When heated to reflux for 5-10 min this reacts to give a mixture, which includes cis-[RuCl ₂(diphos)₂], and species of the type [{Ru(diphos) ₂(μ-Cl)₂}_n-Ag]⁺ (n = 1 or 2) in which the exact coordination geometry at AgI is unknown. The reactions have been followed by ³¹P-{¹H} NMR spectroscopy, and electronic spectroscopy results also support the isomerisation. The ruthenium and silver K-edge EXAFS spectra of the solid isolated from the reaction of trans-[RuCl ₂(dppm)₂] and AgO₃SCF₃ lend support to this formulation; in particular, the best fit for the silver(I) environment includes coordination to one oxygen (from the ^-O₃SCF ₃; Ru-O 2.43 A), two Cl (2.79 A) and between one and two Ru (3.94 A). Over 30-40 min at reflux, AgCl precipitates, the signals due to cis-[RuCl₂(diphos)₂] disappear from the ³¹P-{¹H} NMR spectra, those due to the [{Ru(diphos) ²(μ-Cl)₂}_nAg]⁺ mixture diminish, and new resonances, attributed to five-coordinate [RuCl(diphos) ₂]⁺, appear. This reaction has been investigated for stereoselective ligand substitution. Treatment of trans-[RuCl ₂(diphos)₂] with AgBF₄ or AgO ₃SCF₃ in DCE in the presence of CH₃CN or CO yields exclusively trans-[RuCl(L)(diphos)₂]⁺ (L = CH ₃CN or CO), but treatment of trans-[RuCl₂(diphos) ₂] with AgBF₄ or AgO₃SCF₃ in DCE, followed by treatment with L, usually gave exclusively cis-[RuCl(L)(diphos) ₂]⁺ (L = CH₃CN or CO). The Royal Society of Chemistry 2001.

Full text of DOI:10.1039/b009645h

View Article Online / Journal Homepage / Table of Contents for this issue
Silver(I)-mediated isomerisation of trans-[RuCl2(P–P)2] (P–P ؍
four-membered ring chelate diphosphine ligand) to cis-[{Ru(P–P)2-
(ꢀ-Cl)2}nAg]؉ species, and applications in stereoselective syntheses
of cis- and trans-[RuCl(L)(P–P)2]؉ (L ؍ neutral ligand)†
Simon J. Higgins*,a Annemarie La Pensée,a Clare A. Stuarta and John M. Charnockb
a Department of Chemistry, University of Liverpool, Chemistry Buildings, Crown Street,
Liverpool, UK L69 7ZD. E-mail: shiggins@liv.ac.uk
b Daresbury Laboratory, Daresbury, Warrington, UK WA4 4AD
Received 1st December 2000, Accepted 31st January 2001
First published as an Advance Article on the web 19th February 2001
When complexes trans-[RuCl (diphos) ] (diphos = (Ph P) C᎐CH , Ph PCH PPh or Ph PNHPPh ) were treated
᎐
2
2
2
2
2
2
2
2
2
2
with silver salts of poorly coordinating anions (e.g. AgBF4, AgO3SCF3) in 1,2-dichloroethane halide abstraction
was slow, even at reflux, and the reaction proceeded by an interesting mechanism involving a trans to cis isomerisation.
At room temperature, a complex, characterized as [RuCl(diphos)2(µ-Cl)Ag]ϩ, forms. When heated to reflux for 5–10
min this reacts to give a mixture, which includes cis-[RuCl2(diphos)2], and species of the type [{Ru(diphos)2(µ-Cl)2}n-
Ag]ϩ (n = 1 or 2) in which the exact coordination geometry at AgI is unknown. The reactions have been followed by
31P-{1H} NMR spectroscopy, and electronic spectroscopy results also support the isomerisation. The ruthenium and
silver K-edge EXAFS spectra of the solid isolated from the reaction of trans-[RuCl2(dppm)2] and AgO3SCF3 lend
support to this formulation; in particular, the best fit for the silver() environment includes coordination to one
oxygen (from the ϪO3SCF3; Ru–O 2.43 Å), two Cl (2.79 Å) and between one and two Ru (3.94 Å). Over 30–40 min
at reflux, AgCl precipitates, the signals due to cis-[RuCl2(diphos)2] disappear from the 31P-{1H} NMR spectra,
those due to the [{Ru(diphos)2(µ-Cl)2}nAg]ϩ mixture diminish, and new resonances, attributed to five-coordinate
[RuCl(diphos)2]ϩ, appear. This reaction has been investigated for stereoselective ligand substitution. Treatment of
trans-[RuCl2(diphos)2] with AgBF4 or AgO3SCF3 in DCE in the presence of CH3CN or CO yields exclusively
trans-[RuCl(L)(diphos)2]ϩ (L = CH3CN or CO), but treatment of trans-[RuCl2(diphos)2] with AgBF4 or AgO3SCF3
in DCE, followed by treatment with L, usually gave exclusively cis-[RuCl(L)(diphos)2]ϩ (L = CH3CN or CO).
Ϫ
CF3SO3 ) and one of dppen in 1,2-dichloroethane (DCE),
Introduction
eventually gave a complex mixture (31P-{1H} NMR spectro-
1
The double bond of the diphosphine (Ph P) C᎐CH (dppen)
᎐
2
2
2
scopic evidence). However, it was clear that some unusual
chemistry was involved. Initially, no AgCl precipitate formed,
although a reaction clearly occurred (colour change; several
cis-[RuX2(dppen)2] and other, broad, resonances seen in the
31P-{1H} NMR spectrum). Furthermore, other unusual
products have been characterised from the reactions of [RuCl2-
(dppm)2] with silver salts in the presence of CO.13
We have therefore investigated the reaction of complex trans-
1a and related diphosphine complexes trans-[RuCl2(diphos)2]
(diphos = Ph2PCH2PPh2, dppm, trans-1b; Ph2PNHPPh2, dppa,
trans-1c; or dppe, trans-1d) with silver() salts in more detail.
Preliminary accounts of this work,14 and of the formation of
related [Ru(diphos)3]2ϩ,15 have appeared. While our studies were
in progress, Wolf and co-workers reported a similar case of
trans to cis isomerisation of [RuCl2(dppm)2], catalysed by
CuCl.16
becomes activated towards nucleophilic addition when this
ligand chelates to a metal centre, to an extent dependent upon
the metal, oxidation state and co-ligands.2–9 We have recently
employed this chemistry with some success to prepare deriv-
atives of the redox-active complex trans-[RuCl2(dppen)2]
trans-1a for anchoring to oxide supports,10,11 and for electro-
polymerisation to afford a polypyrrole derivative.12 However, in
comparison with complexes [MX2(dppen)] (M = Pd, X = OAc;
M = Pt or Pd, X = Cl or I),5,7–9 trans-1a is rather unreactive. For
example, although it will undergo reaction with an excess of
primary amine to give trans-[RuCl2{(Ph2P)2CHCH2NHR}2],
with one exception it proved unreactive towards carbanions,
whereas complexes [MX {(Ph P) C᎐CH }] readily add carb-
᎐
2
2
2
2
anions.7–9 The exception is lithium acetylides, but these also
᎐
displaced the chloride ligands to give trans-[Ru(C᎐CR) -
᎐
2
10
᎐
{(Ph P) CHCH C᎐CR} ] after work-up. Earlier, we showed
᎐
2
2
2
2
that the cationic complex [Pt(dppen)2]2ϩ was significantly more
reactive to nucleophiles than neutral [PtCl2(dppen)].9 We were
therefore interested in the preparation and reactivity of cationic
complexes such as [RuCl(L)(dppen)2]ϩ, [Ru(dppen)2(L)2]2ϩ (L =
neutral ligand) and [Ru(dppen)3]2ϩ.
Experimental
Physical methods
Details of spectrometers used in this study were previously
described.10,17 Except where otherwise noted, the following
standard conditions were employed for spectroscopic meas-
urements: FAB mass spectra were recorded in 3-nitrobenzyl
alcohol matrices using xenon bombardment, 31P-{1H} NMR
spectra in 1,2-dichloroethane solution in a 10 mm diameter
Our early efforts to prepare [Ru(dppen)3]2ϩ, by treatment of
trans-1a with two equivalents of AgOTf (TfOϪ is triflate,
† Electronic supplementary information (ESI) available: FAB-MS
spectra of cis-1a/2a and cis-1b/2b mixtures. See http://www.rsc.org/
suppdata/dt/b0/b009645h/
1
tube, using CDCl3 in a coaxial 5 mm tube for locking, and H
NMR spectra using CDCl3 solutions in 5 mm diameter tubes.
902
J. Chem. Soc., Dalton Trans., 2001, 902–910
DOI: 10.1039/b009645h
This journal is © The Royal Society of Chemistry 2001

View Article Online
Samples for EXAFS spectroscopy were lightly ground with
BN in a pestle and mortar, and packed in aluminium sample
holders between Sellotape windows. Data were collected at the
ruthenium and silver K-edges using a liquid nitrogen-cooled
cryostat, in transmission mode on Station 9.2 of the Daresbury
Synchrotron Radiation Source. A Si(220) crystal monochrom-
ator, detuned to reject 50% of the incident signal to minimise
harmonic contamination, was employed. Data were processed
using the Daresbury program EXCALIB, the background was
subtracted using EXBACK and the isolated EXAFS data were
analysed by fitting a theoretical curve to the experimental
k3-weighted EXAFS spectra using EXCURV 98.18 Theoretical
fits were obtained by adding shells of backscattering atoms
around the central absorber atom, and refining the Fermi
energy Ef, the absorber–scatterer distances r, and the Debye–
Waller factors, to minimise the R-factor. The number of
scatterers, N, was taken as an integer value giving the best fit.
Only shells giving a significant improvement in the R factor
have been included in the final fits.
spectroscopic properties, was similarly prepared in 70% yield
using AgBF4.
cis-[{Ru(dppen)2(µ-Cl)2}nAg]BF4 (mixture of n ؍ 1 or
2
species), cis-2a. To complex trans-1a (0.30 g, 0.31 mmol) in
DCE (50 cm3) was added AgBF4 (0.06 g, 0.31 mmol). The mix-
ture was brought to reflux and stirred for 10 min. The reaction
mixture was then examined by 31P-{1H} NMR spectroscopy at
25 ЊC; see Results and Discussion section. A yellow-green solid
could be isolated by reducing the volume to 20 cm3 in vacuo,
and precipitating with hexanes (10 cm3). Yield 60–70% based
on Ru (assuming 1 : 1 Ru:Ag). Found: C, 55.13; H, 4.12%.
C52H44AgBCl2F4P4Ru requires C, 53.87; H, 3.82% (some vari-
ability; a typical example is quoted). FAB-MS data: m/z 2037
(4), [{Ru(dppen)2(µ-Cl)2}2Ag]ϩ; 964 (22), [RuCl2(dppen)2]; and
929 (65%), [RuCl(dppen)2]. 31P-{1H} NMR: pairs of singlets
at δ ϩ16.89, Ϫ5.92; ϩ15.23, Ϫ4.62; ϩ13.80, Ϫ3.61 respectively.
cis-[{Ru(dppm)2(µ-Cl)2}nAg]BF4 (mixture of n ؍ 1 or
2
species), cis-2b. Prepared similarly from complex trans-1b and
AgBF4 in DCE solution. Characterised in situ by 31P-{1H}
NMR spectroscopy. Yellow-green solid isolated as for cis-2a.
Yield 70%. FAB-MS data: m/z 1990 (10), [{Ru(dppm)2-
(µ-Cl)2}2Ag]ϩ; 940 (14), [RuCl2(dppm)2]; and 905 (100%),
[RuCl(dppm)2]. 31P-{1H} NMR: pairs of triplets at δ ca. ϩ0.4,
Ϫ27.48; ϩ0.4, Ϫ26.02; ϩ0.4, Ϫ24.76 (|JAX ϩ JAXЈ| all 73 Hz).
The corresponding triflate salt, with essentially identical
spectroscopic properties, was similarly prepared using AgOTf.
Synthetic chemistry
All reactions were carried out under nitrogen or argon
atmospheres using Schlenk techniques. The ligands dppm
and dppe were purchased from Aldrich and used as received,
and dppen1 and dppa19 were prepared by literature methods.
Ruthenium() chloride hydrate (41–43% Ru) was a loan from
Johnson-Matthey. Silver() and thallium() salts were purchased
from Aldrich and used as received. Complexes trans-[RuCl2-
(diphos)2] (diphos = dppen, trans-1a;10 dppm, trans-1b;20 dppa,
trans-1c;21 or dppe, trans-1d22) were conveniently prepared
in 80–95% yield from [RuCl2(PPh3)3]23 using the method
published for trans-1a.10 Complexes cis-[RuCl2(diphos)2]
(diphos = dppen, cis-1a;10 or dppm, cis-1b20) were prepared by
thermal isomerisation of the trans isomers. [OsCl2(dppm)2]
(cis and trans isomers) were prepared by a literature method.20
Although methods are given mainly for tetrafluoroborate salts,
in all cases except where noted the corresponding triflate salt
could be prepared similarly.
cis-[{Ru(dppa)2(µ-Cl)2}nAg]BF4 (mixture of n ؍ 1 or
2
species), cis-2c. Yellow-green solid, prepared similarly, from
trans-1c. Yield 73%. FAB-MS data: m/z 1992 (3), [{Ru(dppa)2-
(µ-Cl)2}2Ag]ϩ; 942 (100), [RuCl2(dppa)2]; and 907 (23%),
[RuCl(dppa)2]. 31P-{1H} NMR: pairs of triplets at δ ca. ϩ63.0,
ϩ34.50; ϩ63.0, ϩ36.57; ϩ63.0, ϩ38.50.
Attempted synthesis of [RuCl(dppen)2]BF4 3a. To complex
trans-1a (0.23 g, 0.24 mmol) in DCE (50 cm3) was added AgBF4
(0.047 g, 0.24 mmol) at reflux. After 40 min, the solution was
allowed to cool to room temperature, and then filtered through
Keiselguhr. The volume was reduced to 20 cm3 in vacuo. The
product was precipitated by addition of hexanes (10 cm3). The
dark red solid was filtered off and dried in vacuo. Yield 0.18 g.
FAB-MS data: m/z 964 (10), residual [RuCl2(dppen)2] species;
cis-[RuCl2(dppa)2], cis-1c. Complex trans-1c (2.0 g, 2.12
mmol) was dissolved in the minimum volume of DCE. The
solution was refluxed for 20 h, then allowed to cool to room
temperature. Hexane (100 cm3) was added slowly, with scratch-
ing, and the bright yellow product filtered off and dried
in vacuo. Yield 1.95 g, 98%. Found: C, 59.73; H, 4.37; N, 2.81%.
C48H42Cl2N2P4Ruؒ0.5C2H4Cl2 requires C, 59.31; H, 4.47; N,
2.82%. FAB-MS data: m/z 942 (100), Mϩ; 907 (42), [M Ϫ Cl]ϩ;
and 893 (27%), [M Ϫ Cl Ϫ HCl]ϩ. 31P-{1H} NMR: δ ϩ62.8 and
and 929 (100%), [RuCl(dppen)2]. Electronic spectrum (cmϪ1
)
(DCE, 1 mM Ru): 14,080 and 21,280 (sh). 31P-{1H} NMR:
δ ϩ21.64, Ϫ0.13 (br, s); weak peaks due to cis-2a species.
Attempted syntheses of [RuCl(dppm)2]BF4 3b and of
[RuCl(dppa)2]BF4 3c. The same procedure was followed, using
complexes trans-1b and trans-1c respectively. Neither 3b nor 3c
could be isolated free of cis-2b or cis-2c respectively. Spectro-
scopic data for 3b: electronic spectrum (cmϪ1) (DCE, 1 mM
Ru) 19,300; 31P-{1H} NMR δ: ϩ6.21 and Ϫ20.58 (br, t’s,
|JAX ϩ JAXЈ| 70 Hz). Spectroscopic data for 3c: electronic
spectrum (cmϪ1) (DCE, 1 mM Ru) 19,100; 31P-{1H} NMR:
δ ϩ65.6 and ϩ39.5 (br, t’s, |JAX ϩJAXЈ| 70 Hz).
1
ϩ37.2 (‘virtual’ t’s, AAЈXXЈ, |JAX ϩ JAXЈ| 76 Hz). Selected H
NMR: δ 8.1–6.6 (40 H, m’s, aromatic) and ϩ3.65 (2H, s, br;
NH).
trans-[RuCl(dppen)2(µ-Cl)Ag]BF4, trans-2a. To complex
trans-1a (0.15 g, 0.16 mmol) dissolved in 1,2-dichloroethane
(DCE; 40 cm3) was added AgBF4 (0.034 g, 0.175 mmol) at room
temperature. The mixture was stirred for 1 h. The volume was
then reduced to 20 cm3 in vacuo and the product precipitated by
addition of hexanes (20 cm3). The yellow-green solid was
filtered off and dried in vacuo. Yield 0.135 g, 75%. Found: C,
54.77; H, 4.18%. C52H44AgBCl2F4P4Ru requires C, 53.87; H,
3.82%. FAB-MS data: m/z 964 (100), [RuCl2(dppen)2]; and 929
(95%), [RuCl(dppen)2]. 31P-{1H} NMR: δ ϩ9.52 (s).
[RuCl(dppe)2]BF4 3d. To a solution of complex trans-1d (0.26
g, 0.27 mmol) in CH2Cl2 (50 cm3) was added AgBF4 (0.057 g,
0.29 mmol). The mixture was then heated at reflux for 40 min,
and allowed to cool to room temperature. The AgCl was filtered
from the dark red solution, and the volume reduced to 20 cm3
in vacuo. The product was precipitated with hexanes, filtered off
and dried in vacuo. Yield 0.14 g, 56%. Analytical and spectro-
scopic data were as previously published.24
trans-[RuCl(dppm)2(µ-Cl)Ag]OTf, trans-2b. Prepared simi-
larly from complex trans-1b and AgOTf in 57% yield. Found:
C, 52.17; H, 3.86%. C51H44AgCl2F3O3P4RuS requires C, 51.14;
H, 3.70%. FAB-MS data: m/z 940 (28), [RuCl2(dppm)2]; and
905 (100%), [RuCl(dppm)2]. 31P-{1H} NMR: δ ϩ 13.97 (s). The
corresponding tetrafluoroborate salt, with essentially identical
cis-[RuCl(CH3CN)(dppen)2]CF3SO3 cis-4a. Complex trans-1a
(0.22 g, 0.23 mmol) in DCE (50 cm3) was treated with AgOTf
(0.0643 g, 0.25 mmol), and the solution refluxed for 40 min,
becoming dark red. Upon addition of CH3CN (10 cm3), the
J. Chem. Soc., Dalton Trans., 2001, 902–910
903

View Article Online
colour changed to pale yellow, and AgCl precipitated. The
reaction mixture was refluxed for 10 min and then cooled and
filtered. It was evaporated to dryness, and the residue examined
by 31P-{1H} NMR spectroscopy (see Results). It was then
recrystallised from CH2Cl2–hexanes. Yield 0.17 g, 70%. Found:
C, 58.14; H, 4.21; N, 1.19%. C55H47ClF3NO3P4RuSؒ0.5CH2Cl2
requires C, 58.83; H, 4.27; N, 1.24%. FAB-MS data: m/z 970
(54), Mϩ; 929 (100), [M Ϫ CH3CN]ϩ; and 893 (80%), [M Ϫ CH3-
CN Ϫ HCl]ϩ. 31P-{1H} NMR: δ ϩ0.8 (1P, m, PA, JAB 328, JAM
25, JAX 12), Ϫ1.5 (1P, m, PB, JBM 8, JBX 21 Hz), ϩ12.6 (1P, m,
PX) and ϩ19.3 (1P, m, PM).
was then cooled and filtered. It was evaporated to dryness, and
the residue examined by 31P-{1H} NMR spectroscopy (see
Results). It was then recrystallised from CH2Cl2–hexanes.
Yield 0.256 g, 85%. Found: C, 56.49; H, 3.90%. C54H44-
ClF3O4P4RuSؒ0.5CH2Cl2 requires C, 56.98; H, 3.95%. FAB-MS
data: m/z 957 (100), Mϩ, 929 (32), [M Ϫ CO]ϩ; and 893
(47%), [M Ϫ CO Ϫ HCl]ϩ. IR (νCO, cmϪ1) 1997. 31P-{1H}
NMR: δ ϩ9.7 (1P, m, PC), ϩ4.9 (1P, m, PA, JAB 277, JAM 25,
JAX 18), Ϫ1.5 (1P, m, PB, JBM 18, JBX 28 Hz) and Ϫ9.0 (1P,
m, PX).
cis-[RuCl(CO)(dppm)2]CF3SO3 cis-5b. Synthesized from
complex trans-1b as for trans-5a. Yield 79%. Found: C, 56.06;
H, 4.06%. C52H44ClF3O4P4RuS requires C, 56.06; H, 4.03%.
FAB-MS data: m/z 933 (100), Mϩ; 905 (19), [M Ϫ CO]ϩ; and
869 (27%), [M Ϫ CO Ϫ HCl]ϩ. IR (νCO, cmϪ1) 2001. 31P-{1H}
NMR: δ Ϫ9.3 (1P, m, PC), Ϫ13.5 (1P, m, PA, JAB 278, JAM 43,
JAX 19), Ϫ30.0 (1P, m, PB, JBM 31, JBX 43 Hz) and Ϫ31.5 (1P, m,
PX).
cis-[RuCl(CH3CN)(dppm)2]BF4 cis-4b. Prepared from com-
plex trans-1b as for cis-4a. Yield 87%. Found: C, 59.09; H, 4.50;
N, 1.22%. C52H47BClF4NP4Ruؒ0.5CH2Cl2 requires C, 58.62;
H, 4.50; N, 1.30%. FAB-MS data: m/z 946 (44), Mϩ; 905
(100), [M Ϫ CH3CN]ϩ, and 869 (59%), [M Ϫ CH3CN Ϫ HCl]ϩ.
31P-{1H} NMR: δ ϩ0.1 (1P, m, PM), Ϫ6.1 (1P, m, PX), Ϫ19.6
(1P, m, PA, JAB 322, JAM 43, JAX 24), and Ϫ25.2 (1P, m, PB, JBM
32, JBX 44 Hz).
trans-[RuCl(CO)(dppen)2]CF3SO3 trans-5a. Complex trans-
1a (0.34 g, 0.35 mmol) in DCE (50 cm3) was heated to reflux.
Then CO was passed through the solution, and AgOTf (0.090
g, 0.35 mmol) added. On addition of the silver salt the solution
rapidly became colourless and AgCl precipitated. The solution
was refluxed for 30 min, then allowed to cool to room tem-
perature. It was filtered and then evaporated to dryness. The
residue was examined by 31P-{1H} NMR spectroscopy, then
recrystallised from CH2Cl2–hexanes. Yield 0.27 g, 82%. Found:
C, 55.61; H, 3.83%. C54H44ClF3O4P4RuSؒCH2Cl2 requires C,
55.45; H, 3.89%. FAB-MS data: m/z 957 (100), Mϩ; 929 (5),
cis-[RuCl(CH3CN)(dppa)2]BF4 cis-4c. Prepared from com-
plex trans-1c as for cis-4a. Yield 80%. Found: C, 56.18; H, 4.22;
N, 3.50%. C50H45BClF4N3P4Ruؒ0.5CH2Cl2 requires C, 56.29; H,
4.30; N, 3.90%. FAB-MS data: m/z 948 (34), Mϩ; 907 (100),
[M Ϫ CH3CN]ϩ, 871 (45), [M Ϫ CH3CN Ϫ HCl]ϩ. 31P-{1H}
NMR: δ ϩ59.5 (2P, overlapping m, PX, PM), ϩ43.0 (1P, m, PA,
JAB 324, JAM 43, JAX 24) and ϩ36.2 (1P, m, PB, JBM 32, JBX 44
Hz).
trans-[RuCl(CH3CN)(dppen)2]PF6 trans-4a. Complex trans-
1a (0.30 g, 0.31 mmol) was dissolved in DCE (60 cm3) and the
solution brought to reflux. To this was added [Cu(CH3CN)4]-
PF6 (0.12 g, 0.32 mmol), and reflux continued for 30 min. The
mixture was cooled to room temperature, the CuCl filtered off
and the yellow solution evaporated to dryness. Although the
product was pure by 31P-{1H} NMR spectroscopy, it was
recrystallised from CH2Cl2–hexane. Yield 0.271 g, 78%. Found:
C, 56.44; H, 4.22; N, 1.20%. C54H47ClF6NP5Ruؒ0.5CH2Cl2
requires C, 56.54; H, 4.18; N, 1.21%. FAB-MS data: m/z 970
(81), Mϩ; 929 (100), [M Ϫ CH3CN]ϩ; and 893 (96%), [M Ϫ
[M Ϫ CO]ϩ; and 893 (16%), [M Ϫ CO Ϫ HCl]ϩ. IR (νCO, cmϪ1
)
1972. 31P-{1H} NMR: δ ϩ9.4 (s). Selected 1H NMR: δ 6.53 (m,
4H, C᎐CH ).
᎐
2
trans-[RuCl(CO)(dppm)2]CF3SO3 trans-5b. Synthesized from
complex trans-1b as for trans-5a. Yield 78%. Found: C, 54.99;
H, 3.85%. C52H44ClF3O4P4RuSؒCH2Cl2 requires C, 54.53; H,
3.97%. FAB-MS data: m/z 933 (100), Mϩ; 905 (7), [M Ϫ CO]ϩ;
and 869 (17%), [M Ϫ CO Ϫ HCl]ϩ. IR (νCO, cmϪ1) 1974. 31P-
1
{1H} NMR: δ Ϫ13.7 (s). Selected H NMR: δ 5.20 (br t, 4H,
1
CH3CN Ϫ HCl]ϩ. 31P-{1H} NMR: δ ϩ13.92 (s). Selected H
PCH2P).
NMR: δ 6.28 (‘virtual’ quintet, C᎐CH , app. JPH 6 Hz) and 1.25
᎐
2
(s, 3H, CH3CN).
Results
trans-[RuCl(CH3CN)(dppm)2]BF4 trans-4b. Complex trans-1b
(0.22 g, 0.23 mmol) was dissolved in 5 : 1 DCE–CH3CN (60
cm3) at reflux. AgBF4 (0.048 g, 0.25 mmol) was added, and the
solution refluxed for 30 min. After cooling to room temperature
the AgCl precipitate was filtered off and the solution evapor-
ated to dryness. The pale yellow residue was recrystallised from
CH2Cl2–hexanes. Yield 0.174 g, 72%. Found: C, 59.00; H, 4.37;
N, 1.31%. C52H47BClF4NP4Ruؒ0.5CH2Cl2 requires C, 58.62; H,
4.50; N, 1.30%. FAB-MS data: m/z 946 (54), Mϩ; 905 (100),
[M Ϫ CH3CN]ϩ; and 869 (59%), [M Ϫ CH3CN Ϫ HCl]ϩ.
Reactions of trans-[RuCl2(diphos)2] (diphos ؍ dppen, dppm or
dppa) to form Ag؉ adducts
Solutions of trans-[RuCl2(diphos)2] (diphos = dppen, trans-1a;
dppm, trans-1b; or dppa, trans-1c) in DCE were treated with 1
equivalent of fresh AgBF4 at room temperature and the
progress of the reactions was monitored by 31P-{1H} NMR
spectroscopy. Initially, the singlet due to trans-1a–1c (at δ 14.9,
Ϫ7.2 and 60.7 respectively) was replaced by another singlet
(at δ 9.5, Ϫ14.9 and 54.6 respectively). These are assigned
to chloride-bridged adducts, trans-[Cl(diphos)2Ru(µ-Cl)Ag-
(µ-F–BF3)] trans-2a–2c). The preference of AgI for linear
coordination suggests the latter, at least in the solid state.
Moreover, solids with approximately the correct microanalyses
for this formulation were obtained by precipitation with hexane
for trans-2a and trans-2b, and the IR spectrum of these showed
distinct splitting of the B–F stretching bands. Use of AgOTf
gave identical results. The 31P chemical shifts of the trans-2a–2c
complexes in DCE solution were identical for both anions,
which may suggest that the AgI–BF4 or AgI–OTf interactions
evident in the solid state are weak in solution and may be
replaced by solvent coordination; there is precedent for
coordination of α,ω–dichloroalkanes to AgI.25
1
31P-{1H} NMR: δ Ϫ10.2 (s). Selected H NMR: δ 5.14 (br t,
4H, PCH2P) and 1.04 (s, 3H, CH3CN).
trans-[RuCl(CH3CN)(dppa)2]BF4 trans-4c. Made from com-
plex trans-1c as for trans-4b. Yield 89%. Found: C, 55.37; H,
4.23; N, 3.70%. C50H45BClF4N3P4RuؒCH2Cl2 requires C, 54.69;
H, 4.23; N, 3.75%. FAB-MS data: m/z 948 (32), Mϩ; 907 (100),
[M Ϫ CH3CN]ϩ; and 871 (51%), [M Ϫ CH3CN Ϫ HCl]ϩ.
1
31P-{1H} NMR: δ ϩ58.73 (s). Selected H NMR: δ 3.71 (s,
broad, 2H, PNHP) and 1.31 (s, 3H, CH3CN).
cis-[RuCl(CO)(dppen)2]CF3SO3 cis-5a. Complex trans-1a
(0.26 g, 0.27 mmol) in DCE (50 cm3) was treated with AgOTf
(0.072 g, 0.28 mmol), and the solution refluxed for 30 min,
becoming dark red. Upon exposure to CO gas the solution
became colourless, and AgCl precipitated. The reaction mixture
On warming a solution of complex trans-1a and AgBF4
to 50 ЊC for 5 minutes there was initially no precipitation of
AgCl, but several new peaks appeared in the 31P-{1H} NMR
904
J. Chem. Soc., Dalton Trans., 2001, 902–910

View Article Online
Fig. 1 31P-{1H} NMR spectra showing the progress of a reaction between trans-[RuCl2(dppen)2] trans-1a and AgBF4 (1 equivalent) in 1,2-
dichloroethane. (a) Reaction mixture after addition of AgBF4 at room temperature, and bringing to reflux for ten minutes. (b) Reaction after 40
minutes at reflux. Peaks labeled ‘A’ are assigned to cis-1a, ‘B’ to five-coordinate [RuCl(dppen)2]ϩ 3a.
spectrum, the relative intensities of which suggested that they
were pairs of related singlets (Fig. 1a). The weakest of these
pairs, at δ Ϫ3.6 and 13.9, are very close to the resonances previ-
ously observed for cis-[RuCl2(dppen)2] cis-1a which we made
previously by thermal isomerisation.‡ Two other pairs
appeared at δ Ϫ6.0 and 16.9 and at Ϫ4.7 and 15.2. We ascribe
these as due to a mixture of cis-1a moieties in which the chlor-
ide ligands are interacting with AgI, and in which the complexes
differ only in their coordination at silver. Examples might
include 1 : 1 complexes such as [(dppen)2Ru(µ-Cl)2Ag(µ-F)BF3]
and/or [(dppen)2Ru(µ-Cl)2Ag(DCE)]BF4, and 2 : 1 complexes
such as [{Ru(dppen)2(µ-Cl)2}2Ag]BF4 (Scheme 1). The presence
of a mixture of 1 : 1 and 2 : 1 complexes would account for the
presence of some free cis-1a in the mixture. We designate this
mixture of products cis-2a. We have been unable to isolate its
individual components. However, precipitation with hexanes
gave a yellow-green solid. Microanalyses (C and H) of this
gave values which were high for a formulation [RuCl2-
(dppen)2]ؒAgBF4, and somewhat variable. Interestingly, the
FAB mass spectra of different samples consistently showed a
cluster of peaks centred at m/z 2038, correct for [{Ru-
(dppen)2(µ-Cl)2}2Ag]ϩ, and at m/z 1169, almost the correct
value (1170) required for a complex [(dppen)2Ru(µ-Cl)2-
Ag(DCE)]ϩ. We attempted to optimise the synthesis of indi-
vidual 1 : 1 and 2 : 1 Ru : Ag species by varying the reaction
stoichiometry, but an equilibrium mixture was always obtained.
The electronic spectra of trans-[RuX2(diphosphine)2] usually
show two d–d bands in the visible region, a lower energy band
assigned as 1A1g → 1Eg and a higher energy band assigned as
1A1g → 1A2g, assuming approximate D4h symmetry.26,27 Com-
plexes cis-[RuX2(diphosphine)2] usually have d–d bands at
somewhat higher energy, which accounts for their generally
paler colour. Consistent with this, the spectrum of cis-2a in
CH2Cl2 showed two shoulders in the visible region, a weak one
at 23,300 cmϪ1 and a stronger one at 27,000 cmϪ1, on the low-
energy tail of a charge transfer band. The spectrum was similar
in profile to that of cis-1a, which shows a shoulder at 22,730
and a peak at 27,170 cmϪ1. In contrast, trans-1a shows a single,
It showed three pairs of second order triplets. The three down-
field triplets of each pair all overlap at δ ca. 0.4, and the upfield
triplets occur at δ Ϫ24.8, Ϫ26.0 and Ϫ27.5 (|JAX ϩ JAXЈ| 73 Hz).
It is clear that isomerisation to a mixture, cis-2b, similar to
cis-2a, has occurred. Similarly, on warming solutions of
trans-1c with AgBF4 the singlet at δ 60.7 was replaced by
three pairs of second order triplets, the least intense of which
(at δ 36.6 and 62.7) are very close to the values found for
cis-[RuCl2(dppa)2] cis-1c, which we prepared independently by
thermal isomerisation of trans-1c for comparison. The other
signals are due to a mixture of Agϩ adducts, cis-2c.
The FAB mass spectra of the cis-2b mixture, isolated as a
solid by evaporation of the solvent, showed molecular
ion peaks for [{Ru(diphosphine)2(µ-Cl)2}2Ag]ϩ at m/z 1990,
together with peaks characteristic of the [Ru(dppm)2Cl2]
moiety. There was, however, no sign of a cluster of peaks
corresponding to a complex [(dppm)2Ru(µ-Cl)2Ag(DCE)]ϩ.
This mixture was further characterised using EXAFS; see
below.
Use of AgOTf gave similar results on reaction with all three
complexes 1a–1c, with only minor differences in product
distribution. Interestingly, we have since found that these
reactions can also be performed in more polar solvent systems
(thf/EtOH) with no significant change in products.
Formation of [RuCl(diphos)2]؉
We next examined the use of catalytic quantities (5–10 mol%)
of AgOTf or AgBF4 to effect trans–cis conversions with com-
plexes 1a–1c, but were unsuccessful because further reactions
took place when heating was prolonged, and AgCl precipitated.
We therefore investigated these reactions. When trans–1a was
refluxed for 40 mins with one equivalent of AgBF4 in DCE
two new, broad singlet resonances in the 31P-{1H} NMR
spectrum, at δ 0.5 and 20.9, grew at the expense of the sig-
nals due to cis-1a/2a (Fig. 1b). In particular, the signals due
to cis-1a disappeared completely. The solution changed to
intense red, and AgCl precipitated. The electronic spectrum
of the red solution showed a new band at 14,100 cmϪ1, con-
sistent with the formation of a five–coordinate complex
[RuCl(dppen)2]ϩ 3a.28,29 When the reaction was prolonged the
new resonances continued to grow at the expense of those for
cis-2a, but additional resonances due to unidentified decom-
position products were also apparent, and it was not possible
to isolate 3a pure.
rather broad, peak at 24,150 cmϪ1
.
A DCE solution of complex trans-1b was refluxed with
AgBF4 for 10 min, and the 31P-{1H} NMR spectrum recorded.
‡
The reason why these complexes show two apparent singlets in their
31P-{1H} NMR spectra, instead of the more usual second-order triplets,
has been discussed in an earlier paper.10
When complex trans-1b was refluxed for 30 min with AgBF4
J. Chem. Soc., Dalton Trans., 2001, 902–910
905

View Article Online
Table 1 Comparison of EXAFS and crystallography parameters for
cis-[RuCl2(dppm)2] cis-1b
Crystallography
EXAFS
Shell type C.N.
rcr/Å (av.)
rex/Å
2σ2/Å2
Residual
26.3
1
2
3
4
P
4
2
2
8
2.324
2.447
3.122
3.655
2.32
2.42
3.11
3.70
0.010
0.013
0.016
0.029
Cl
C
C
Fig. 2 31P-{1H} NMR spectrum of the species obtained upon reaction
of trans-[RuCl2(dppm)2] trans-1b and AgBF4 (1 equivalent) in 1,2-
dichloroethane at reflux for 40 min. Peaks labeled ‘C’ are assigned to
[RuCl(dppm)2]ϩ 3b, those labelled ‘i’ to unidentified impurities.
Complex cis-1d, or adduct(s) of cis-1d with AgI, could not be
detected in the reaction mixture under any conditions, although
the reaction was slow enough to follow by 31P-{1H} NMR
spectroscopy.
EXAFS studies on cis-1b/cis-2b mixture
Since we were unable to obtain crystals of any of the Ru–Cl–Ag
complexes we have employed EXAFS in an effort to obtain
some structural information. A sample of the cis-1b/cis-2b mix-
ture was prepared using AgOTf in DCE. cis-1b was used as a
model compound since its crystal structure has been deter-
mined.30 The analysis of the ruthenium K-edge EXAFS spec-
trum of cis-1b is compared in Table 1 with the crystal structure
data. The final model included four P atoms at 2.32 Å and two
Cl atoms at 2.42 Å (the major contribution to the fit), and eight
C atoms at 3.70 Å. Two more C atoms at 3.11 Å improved the fit
only marginally. The data agree well with those from the crystal
structure, as expected.
The Fourier transform of the silver K-edge EXAFS spec-
trum of the cis-1b/cis-2b mixture shows two clear peaks (Fig.
3). The best fit for the first shell was with one O atom at 2.43
Å, and two Cl atoms at 2.79 Å (Table 2). The second shell was
fitted with one, or two, Ru atoms at 3.94 Å, with the latter
giving slightly the better fit. The ruthenium K-edge spectrum
(Fig. 3) did not show a clear outer shell in the Fourier trans-
form. The inner sphere was fitted with four P atoms at 2.33 Å
and two Cl atoms at 2.41 Å, although the contribution of the Cl
atoms made less of an improvement to the fit than for cis-1b.
This may indicate a change in coordination about Ru, the most
likely cause being an asymmetry in the two Ru–Cl distances. A
shell of eight C atoms at 3.69 Å improved the fit significantly.
The addition of a contribution from a silver atom at 4.05 Å did
result in further improvement of the fit, but not sufficiently to
demonstrate conclusively the presence of this shell with the
ruthenium data alone.
Scheme 1
in DCE, AgCl precipitated, the colour became intense red, and
an additional pair of rather broad second order triplets,
assigned as due to [RuCl(dppm)2]ϩ 3b, appeared in the 31P-{1H}
NMR spectrum at δ Ϫ20.6 and 6.2 (Fig. 2). At the same time
the weakest pair of triplets present in the spectrum of the
cis-1b/cis-2b mixture had disappeared (some small addi-
tional resonances due to unidentified impurities are also pres-
ent, e.g. the multiplet at δ ca. Ϫ6). The electronic spectrum
showed a new band at 19,600 cmϪ1. Similarly, treatment of
trans-1c with AgBF4 in refluxing DCE for 40 min resulted in the
growth of two new second order triplets, at δ 39.5 and 65.6, in
the 31P-{1H} NMR spectrum, and the appearance of a new
band at 14,500 cmϪ1 in the electronic spectrum, due to 3c.
Although treatment of cis-1a–1c with silver salts in DCE
resulted in the formation of somewhat larger proportions of
3a–3c (31P-{1H} NMR evidence) this route also failed to yield
pure samples.
Treatment of trans-[RuCl2(dppe)2] trans-1d with AgOTf in
DCE resulted only in halide abstraction to give dark red
[RuCl(dppe)2]OTf, previously isolated (as a PF6Ϫ salt) by treat-
ment of a CH2Cl2 solution of cis-1d with NaPF6 suspension.24
906
J. Chem. Soc., Dalton Trans., 2001, 902–910

View Article Online
Table 2 Comparison of various fits to EXAFS spectra for the cis-1b/cis-2b mixture (triflate). The final fits are in bold characters
Edge
Fit no.
Shell
Atom
No.
rex/Å
2σ2/Å2
Residual
Ag K-edge
Ag K-edge
Ag K-edge
1
2
3
1
1
1
2
1
2
3
1
2
3
O
Cl
O
Cl
O
Cl
Ru
O
4
4
1
2
1
2
1
1
2
2
2.60
2.72
2.42
2.77
2.43
2.78
3.94
2.43
2.79
3.94
0.025
0.033
0.005
0.017
0.001
0.017
0.008
0.001
0.016
0.014
56.3
44.0
42.6
Ag K-edge
4
39.4
Ag K-edge
5
Cl
Ru
38.8
22.3
21.8
Ru K-edge
Ru K-edge
1
2
1
1
2
1
2
3
1
2
3
4
P
P
Cl
P
Cl
C
P
Cl
C
4
4
2
4
2
8
4
2
8
1
2.35
2.34
2.42
2.34
2.41
3.69
2.33
2.41
3.69
4.05
0.011
0.012
0.035
0.012
0.037
0.022
0.012
0.032
0.022
0.022
Ru K-edge
3
20.5
Ru K-edge
4
Ag
20.1
Fig. 3 Ruthenium (left) and silver (right) K-edge EXAFS spectra (top) and Fourier transforms (bottom) for the cis-1b/cis-2b mixture (triflate salt).
The solid lines are experimental data and the dotted lines are for the best fits (those shown in bold in Table 2).
under conditions used successfully for conversion of trans-1b
Treatment of complexes trans-1a–1c with other Lewis acids
into cis-1b.16
It was of interest to see whether the isomerisation of trans-
1a–1c could be effected by other Lewis acid catalysts. Recently,
Zhu et al. reported that the isomerisation of trans-1b to cis-1b
was catalysed by CuCl.16 On treatment of trans-1a with an
excess of CuCl in CH2Cl2 or DCE at room temperature its
31P-{1H} resonance broadened, and shifted slightly to δ ϩ15.9.
On heating at reflux partial conversion into a mixture of cis
species occurred. These are probably analogous to [{Ru(dppm)2-
(µ-Cl)2}2Cu][CuCl2], characterised crystallographically, and
related species differing in their coordination at CuI.16 However,
CuCl did not catalyse the conversion of trans-1a into cis-1a
On treatment of complex trans-1a with the more powerful
Lewis acid [Cu(CH3CN)4]BF4 in DCE, trans-[RuCl(CH3CN)-
(dppen)2]BF4 4a and a precipitate of CuCl were the only
products (see below).
Thallium(I) salts are sometimes used as non–oxidising alter-
natives to silver() salts in halide abstraction reactions. Interest-
ingly, when trans-1a was treated with TlOTf in DCE at room
temperature the singlet at δ ϩ14.9 was replaced by two reson-
ances in the 31P-{1H} NMR spectrum of the reaction mixture, a
broad singlet at δ ϩ13.5 and a sharp singlet at δ ϩ9.5, the latter
similar to that of trans-[RuCl(dppen)2(µ-Cl)Ag]ϩ. Similarly,
J. Chem. Soc., Dalton Trans., 2001, 902–910
907

View Article Online
when trans-1b was treated with TlOTf in DCE, the singlet at
δ Ϫ7.2 was replaced by a broad singlet at δ Ϫ8.1 and a sharp
singlet at δ Ϫ13.9, the latter similar to that of trans-
[RuCl(dppm)2(µ-Cl)Ag]ϩ. FAB mass spectrometry of the solids
precipitated upon addition of hexanes to these reaction
mixtures revealed a cluster of peaks centred at near the
correct values (m/z 1167 and 1145 respectively) for trans-
[RuCl(diphos)2(µ-Cl)Ag]ϩ. However, when these reaction mix-
tures were brought to reflux no further changes were observed,
In particular, no evidence for trans into cis conversion was seen,
and no halide abstraction occurred.
Reactions of [OsCl2(diphos)2] with silver(I) salts
In an attempt to obtain Os–Cl–Ag adducts we investigated the
reaction of trans-[OsCl2(dppm)2] with silver() salts in DCE.
However, we obtained only silver metal, and [OsCl2(dppm)2]ϩ;
it is known that MII–MIII redox potentials for such osmium()
complexes are significantly less positive than for analogous
ruthenium complexes.20,31,32 Since the potentials are also con-
siderably more positive for cis complexes than trans,20 we also
investigated the reaction of cis-[OsCl2(dppm)2] with silver()
salts in DCE. Interestingly, this appeared to result in halide
abstraction initially, with the formation of a red solution and a
AgCl precipitate. However, even in the absence of oxygen, the
red solution rapidly turned green and the isolated product was
paramagnetic.
Fig. 4 31P-{1H} NMR spectrum of the crude product from treatment
of complex trans-1a with AgBF4 in DCE at 50 ЊC for 40 min, followed
by bubbling CO through the solution. It can be seen that the major
product, with an ABMX spin system, is cis-[RuCl(CO)(dppen)2]BF4
cis-5a.
dependent upon whether or not Agϩ-mediated isomerisation is
allowed to occur before introduction of the neutral ligand.
Treatment of complex trans-1b with AgBF4 in DCE at 50 ЊC,
followed by removal of the solvent and treatment of the residue
with CH3CN, gave mainly cis-[RuCl(CH3CN)(dppm)2]BF4
cis-4b, although two minor species were also detected in the
crude product by 31P-{1H} NMR spectroscopy. A pair of
‘virtual’ triplets at δ Ϫ18.0 and Ϫ3.1 was due to cis-
[Ru(CH3CN)2(dppm-P,PЈ)2]2ϩ, and a singlet at δ Ϫ10.15 to
trans-[RuCl(CH3CN)(dppm)2]ϩ trans-4b. Pure cis-4b was
isolated upon recrystallisation. Complex trans-4b was the only
product when trans-1b was refluxed in 5 : 1 CH3CN–DCE with
one equivalent of AgBF4. It too was isolated and fully charac-
terised. Treatment of trans-1b with AgBF4 in DCE at 50 ЊC,
followed by bubbling with CO, gave cis-[RuCl(CO)(dppm)2]BF4
cis-5b in high yield after recrystallisation, whereas treatment of
trans-1b with CO, then AgBF4, in DCE, gave exclusively trans-
[RuCl(CO)(dppm)2]BF4 trans-5b. The carbonyl complexes were
isolated and fully characterised. Both carbonyl cations have
been described previously.33
With the ligand dppa, treatment of complex trans-1c with
one equivalent of AgBF4 in 5 : 1 CH3CN–DCE did not give
exclusively trans-[RuCl(CH3CN)(dppa)2]BF4 trans-4c, but 70%
trans-4c, 25% cis-4c and 5% of an unidentified complex
(31P-{1H} NMR spectroscopy). Treatment of trans-1c with
AgBF4 in DCE for 40 min, followed by treatment with CH3CN,
gave exclusively cis-4c, as did treatment of cis-1c with AgBF4 in
CH3CN, or with [Cu(CH3CN)4]PF6 in DCE (PF6Ϫ salt).
Interestingly, treatment of trans-[RuCl2(dppe)2] trans-1d with
one equivalent of [Cu(CH3CN)4]PF6 in DCE gave, after
recrystallisation, trans-[RuCl(CH3CN)(dppe)2]PF6 trans-4d.
Treatment of trans-1d with AgBF4 or AgOTf in the presence of
CO, however, gave a mixture of cis- and trans-[RuCl(CO)-
(dppe)2]ϩ 5d. The formation of 3d in situ in DCE, followed by
the addition of CH3CN or CO, was stereoselective, giving cis-4d
or cis-[RuCl(CO)(dppm)2]ϩ cis-5d respectively, in agreement
with an earlier publication.24
Syntheses of [RuCl(diphosphine)2(L)]؉ from
[RuCl2(diphosphine)2] by halide abstraction
A preliminary investigation of the use of AgI-mediated isomer-
isation of complexes trans-1a–1c in ligand substitution reac-
tions was undertaken. Treatment of trans-1a with AgBF4 in
DCE at 50 ЊC for 40 min, followed by treatment with CH3CN,
gave a precipitate of AgCl. Treatment of the remaining yellow
solution with hexanes gave, exclusively, cis-[RuCl(CH3CN)-
(dppen)2]BF4 cis-4a. This was identified by the fact that its
31P-{1H} NMR spectrum was almost identical with cis-[RuCl-
(CH3CN)(dppen)2]Cl, which forms when cis-1a is dissolved in
CH3CN,10 and by its FAB mass spectrum, which showed a
cluster of peaks at m/z 970 due to Mϩ, and at 929 due to
[M Ϫ CH3CN]ϩ. Alternatively, cis-4a could be made (as the
Ϫ
PF6 salt) by treatment of cis-1a with one equivalent of
[Cu(CH3CN)4]PF6 in DCE; CuCl precipitated from the
reaction mixture in this case.
Interestingly, in contrast, when complex trans-1a was
brought to reflux in 5 : 1 CH3CN : DCE in the presence of 1 mol
of AgBF4, AgCl precipitated immediately. Pale yellow trans-
[RuCl(CH3CN)(dppen)2]BF4 trans-4a was the main product
(31P-{1H} NMR evidence), but was difficult to separate from an
unidentified impurity (δ ϩ23.3). A better method was treatment
of trans-1a with one equivalent of [Cu(CH3CN)4]PF6 in DCE,
which afforded exclusively trans-4a (as the PF6Ϫ salt) and CuCl.
The trans-4a had a very similar FAB mass spectrum to that of
cis-4a, but its 31P-{1H} NMR spectrum showed a singlet at
δ ϩ13.9.
Treatment of complex trans-1a with AgBF4 in DCE at 50 ЊC
for 40 min, followed by bubbling CO through the solution, gave
> 95% cis-[RuCl(CO)(dppen)2]BF4 cis-5a (31P-{1H} NMR
spectroscopic evidence), from which pure cis-5a was isolated on
recrystallisation. This had ν(CO) at 1997 cmϪ1 in its IR spec-
trum, and its 31P-{1H} NMR spectrum showed an ABMX
spin system (Fig. 4). When trans-1a was treated with CO in
DCE, and one equivalent of AgBF4 was then added, trans-
[RuCl(CO)(dppen)2]BF4 trans-5a was obtained. This had a
similar FAB mass spectrum to that of cis-5a, but ν(CO) at 1972
cmϪ1 in its IR spectrum and a singlet at δ ϩ9.3 in its 31P-{1H}
NMR spectrum.
The complexes cis-5a and -5b were not stable in CH2Cl2
solution, and slowly isomerised to trans-5a and -5b. It is likely
that the CO ligand prefers to be trans to the π-donor ClϪ rather
than the rival π-acceptor phosphine ligand.
Discussion
Ag؉–mediated isomerisations
The first products formed when complexes trans-1a–1c were
treated with AgBF4 were the adducts trans-2a–2c. There is
plenty of precedent for interactions of this type between substi-
Therefore, the stereochemical outcome of the reactions of
complex trans-1a with silver salts and CH3CN, or CO, is
908
J. Chem. Soc., Dalton Trans., 2001, 902–910

View Article Online
tutionally inert metal halide complexes and AgI. For example,
reaction of [Ru2Cl4(PEtPh2)5], [Ru2Cl3(PEtPh2)6]Cl or [RuCl2-
(PEtPh2)3] with AgCl resulted in phosphine migration from
Ru to Ag, and the formation of [Ru2Cl2(µ-Cl)3(PEtPh2)4-
Ag(PEtPh2)], in which the Ag(PEtPh2) unit is coordinated to
one of the terminal and two of the three bridging chloride
ligands.34 Reaction of [M(C5H5)2(NO)I] (M = Mo or W) with
AgBF4 or AgSbF6 in CH2Cl2 gave precipitates of the corre-
sponding adducts, [M(C5H5)2(NO)IؒAgY].35 Moreover, when
trans-1b was treated with CuCl its 31P-{1H} NMR resonance
broadened and shifted slightly; this was attributed to the
formation of trans-[Cl(dppm)2Ru(µ-Cl)CuCl].16
have been suggested as intermediates in ligand substitution
reactions,13,38 and trans-[RuCl2(2,2-dppp)2] (2,2-dppp = 2,2-
bis[diphenylphosphino]propane) dissociates to [RuCl(2,2-
dppp)2]ϩ in CHCl3.39 Although, coincidentally, ruthenium()
complexes [RuCl2(diphosphine)2]BF4 have visible spectra simi-
lar to those of [RuCl(diphos)2]ϩ,31 and silver() salts have been
used to oxidise other [RuX2(diphosphine)2] to RuIII,31 the
ruthenium() formulation can be excluded since (i) for 3a we
confirmed that the precipitate was AgCl (not Ag) by dissolving
it in aqueous ammonia, and (ii) normal 31P-{1H} NMR spectra
are observed, which would not be the case for paramagnetic low
spin d5 RuIII.
Subsequently, and more rapidly (20–30 minutes) on warm-
ing, complexes trans-2a–2c isomerise cleanly to mixtures, cis-
1a–1c/cis-2a–2c. In the absence of Agϩ, refluxing trans-1a in
chlorobenzene for 16 h is necessary to effect 90% conversion
into cis-1a.10 Similarly, refluxing trans-1b in DCE for 16 h is
necessary to effect conversion into cis-1b.20 Recently, other
workers reported that the isomerisation of trans-1b to cis-1b
was catalysed by CuCl and CuI, and they were able to isolate,
and characterise crystallographically, the heterotrimetallic
chloride–bridged complex [{Ru(dppm)2(µ-Cl)2}2Cu][CuCl2].16
On dissolution in CD2Cl2 the 31P-{1H} NMR spectrum showed
that this complex was in equilibrium with a species with similar
chemical shifts and coupling constants, assigned as [(dppm)2-
Ru(µ-Cl)2CuCl] in which CuI has trigonal planar coordination
to all three chlorines, and another species with broad reson-
ances for which the formulation cis-[RuCl(dppm)2(µ-Cl)}CuCl]
was suggested, the broadening being ascribed to exchange of
CuCl between the two chloride ligands.16 Our observation of a
cluster of peaks in the FAB mass spectra of cis-2a–2c corre-
sponding to [{Ru(dppen)2(µ-Cl)2}2Ag]ϩ, the 31P-{1H} NMR
data, and the EXAFS data all suggest that cis-2a–2c consist of
a similar mixture of species differing only in their coordination
at silver. In particular, although the ruthenium K-edge EXAFS
spectrum of cis-2b showed only one distinct shell, and inclusion
of a silver atom only resulted in a marginal fit improvement, the
silver K-edge data did show two distinct shells. The fact that the
inner shell could best be fitted with two Cl atoms at 2.77 Å (a
reasonable distance for a Ru–Cl–Ag-bridge species34) and one
O atom (for this triflate salt) suggests the possibility that a 1 : 1
species such as [(dppen)2Ru(µ-Cl)2Ag(η1-OTf)] may be present,
while the fact that the outer shell could be fitted well by between
one and two ruthenium atoms lends support to the form-
ulation of the mixture as a combination of cis-Ru(µ-Cl)2Ag and
cis-{Ru(µ-Cl)2}2Ag species.
Although our experience is that the chemistry of complexes
cis-1a and cis-1b differs in some respects,10 and other workers
have found differences between the behaviour of dppm and
dppa as ligands,40,41 we did not observe any significant differ-
ence in the reactions of trans-1a–1c with silver salts.
Other Lewis acids
There are several examples of adducts between ruthenium()
halide complexes and TlI, and their chemistry suggests that the
interaction can be remarkably strong. For example, treatment
of a CH2Cl2 solution of [Ru(H)Cl(PP3)] (PP3 = P[CH2CH2-
PPh2]3) with aqueous TlPF6 resulted in extraction of Tlϩ from
the aqueous layer and formation of [(H)(PP3)Ru(µ-Cl)Tl]PF6,42
and treatment of [RuCl2(PPh3)([9]aneS3)] ([9]aneS3 = 1,4,7-
trithiacyclononane) with TlPF6 gave a chloride-bridged Ru2Tl2
cluster dication.43 Both of these complexes were characterised
by X-ray crystallography. However, although 31P-{1H} NMR
spectroscopy indicates that trans-1c and -1b interact in an as yet
undefined way with TlOTf in DCE, this Lewis acid does not
induce trans to cis conversion.
CuCl does not act as a catalyst for conversion of complex
trans-1a into cis-1a, although it does do so for trans-1b. We
previously observed that the thermal trans–cis conversion of 1a
was more difficult than for 1b. Stoichiometric amounts of CuCl
do cause the formation of some cis Ru–Cl–Cu species from
trans-1a, but the reaction was not quantitative. We therefore
tried [Cu(CH3CN)4]PF6, which we anticipated might be a more
powerful Lewis acid than CuCl, but this simply abstracted
halide from trans-1a–1c, giving CuCl and trans–[RuCl-
(CH3CN)(dppen)2]PF6. This is, in fact, the most convenient
route to these complexes.
31P-{1H} NMR spectra of cis-[RuCl(diphosphine)(L)]؉
The mechanism of the Ag–mediated isomerisation to cis
stereochemistry is likely to be related to that suggested for the
CuI-mediated isomerisation of complex trans-1b, namely that
coordination of one of the chlorides to AgI weakens the Ru–Cl
bond, allowing facile conversion into the thermodynamically
preferred cis isomer.16 It is possible that some additional driving
force may come from the chelate effect, on formation of the cis-
Ru(µ-Cl)2Ag bridges. These complexes, from the point of view
of silver() coordination chemistry, make an interesting com-
parison with halogenocarbon complexes [Ag(I{CH2}nI)2]PF6,
described recently.36
The 31P-{1H} NMR spectrum of complex cis-5a (Fig. 4)
showed that all four phosphorus atoms are inequivalent. The
two mutually trans phosphorus resonances, at δ ca. ϩ4.9 and
Ϫ1.5, form an AB system (JAB is 277 Hz, a value typical of P
trans P coupling in ruthenium() complexes44–46), with further
coupling to the two mutually cis phosphorus atoms. The
values of |JPP| for the latter are all very similar (24–26 Hz).
Although it is unusual for the two mutually trans phosphorus
atoms in complexes like cis-5a to show a marked chemical
shift difference, there is precedent for this type of spectrum
with a complex involving a cis-[Ru(dppmЈP,PЈ)2] moiety.
Crystallographically characterised [(dppm-P,PЈ)2Ru(µ-Cl)-
(µ-H)Rh(cod)]BF4 (cod = cycloocta-1,5-diene) showed two
resonances for the trans phosphorus atoms at δ Ϫ2.0 and
Ϫ18.9 (JPP 294 Hz).47
Formation of [RuCl(diphos)2]؉
Like these silver()–iodocarbon complexes, cis-2a–2c slowly
react to form silver halide. The colour changes and the appear-
ance of new peaks in the 31P-{1H} NMR spectra both indicate
that five-coordinate [RuCl(diphosphine)2]ϩ 3a–3d are formed.
The tendency for six–coordinate [RuCl2(diphosphine)2] to dis-
sociate in polar solvents to [RuCl(diphosphine)2]Cl increases
with increasing chelate ring size and with increasing size
of substituents at phosphorus,24,28,29,37 and no examples of
[RuCl(diphosphine)2]ϩ have been structurally characterised
for four-membered chelate ring ligands. However, such species
The 31P-{1H} NMR spectrum of complex cis-4a was then
re-recorded with improved signal/noise ratio. The ‘multiplet’
near δ 010 was found to be, in fact, the two inner lines of a
closely spaced AB system (δ PA 1.0; δ PB Ϫ0.4; JAB 320 Hz)
with further coupling to the other cis phosphorus atoms.
Moreover, all four phosphorus atoms in the 31P-{1H} NMR
spectrum of cis-4b are also inequivalent. The two mutually
trans phosphines of cis-4b resonate at δ Ϫ20.13 and Ϫ24.89
(JAB 319 Hz).
J. Chem. Soc., Dalton Trans., 2001, 902–910
909

View Article Online
19 H. Noeth and L. Meinel, Z. Anorg. Allg. Chem., 1967, 349, 225.
In conclusion, we have discovered a silver(I)-mediated isom-
erisation of trans-[RuCl2(diphos)2] (diphos = four-membered
chelate diphosphine) to cis-oriented Ru–Cl–Ag species, and
have shown that, as a consequence, the stereochemistry of
complexes [RuCl(L)(diphos)2]ϩ, synthesized from trans-[RuCl2-
(diphos)2] by AgI-mediated halide abstraction, can be con-
trolled by the order of addition of the reagents.
20 B. P. Sullivan and T. J. Meyer, Inorg. Chem., 1982, 21, 1037.
21 J. Ellermann, C. Schelle, M. Moll, R. A. Knoch and W. Bauer,
Z. Anorg. Allg. Chem., 1995, 621, 789.
22 J. Chatt and R. G. Hayter, J. Chem. Soc., 1961, 896.
23 P. S. Hallman, T. A. Stevenson and G. Wilkinson, Inorg. Synth.,
1970, 12, 237.
24 B. Chin, A. J. Lough, R. H. Morris, C. T. Schweitzer and C.
D’Agostino, Inorg. Chem., 1994, 33, 6278.
25 D. M. Van Seggen, P. K. Hurlbright, O. P. Anderson and S. H.
Strauss, Inorg. Chem., 1995, 34, 3453.
Acknowledgements
26 D. M. Klassen and G. A. Crosby, J. Mol. Spectrosc., 1968, 25, 398.
27 A. B. P. Lever, Inorganic Electronic Spectroscopy, Elsevier,
Amsterdam, 1984.
28 J. C. Briggs, C. A. McAuliffe and G. Dyer, J. Chem. Soc., Dalton
Trans., 1984, 423.
We thank Mr Allan Mills for FAB mass spectra, the EPSRC for
studentships (C.A.S. and A.L.) and for funding the beamtime at
Daresbury through the academic DARTS scheme, and
Johnson–Matthey for generous loans of RuCl3ؒnH2O.
29 A. A. Batista, L. A. C. Cordeiro and G. Oliva, Inorg. Chim. Acta,
1993, 203, 185.
30 A. R. Chakravarty, F. A. Cotton and W. Schwotzer, Inorg. Chim.
Acta, 1984, 84, 179.
31 N. R. Champness, W. Levason, D. Pletcher and M. Webster,
J. Chem. Soc., Dalton Trans., 1992, 3243.
32 N. R. Champness, W. Levason, D. Pletcher, M. D. Spicer and
M. Webster, J. Chem. Soc., Dalton Trans., 1992, 2201.
33 A. B. Nikol’skii, A. M. Popov, M. B. Egorova and V. V. Khorunzhii,
Coord. Chem. (USSR), 1990, 15, 800.
34 T. Easton, R. O. Gould, G. A. Heath and T. A. Stevenson, J. Chem.
Soc., Chem. Commun., 1985, 1741.
35 P. Legzdins and C. R. Nurse, Inorg. Chem., 1982, 21, 3110.
36 J. Powell, M. J. Horvath and A. Lough, J. Chem. Soc., Dalton Trans.,
1996, 1669.
References
1 I. J. Colquhoun and W. McFarlane, J. Chem. Soc., Dalton Trans.,
1982, 1915.
2 G. R. Cooper, F. Hassan, B. L. Shaw and M. Thornton-Pett,
J. Chem. Soc., Chem. Commun., 1985, 614.
3 G. R. Cooper, D. M. McEwan and B. L. Shaw, Inorg. Chim. Acta,
1986, 122, 207.
4 F. S. M. Hassan, B. L. Shaw and M. Thornton-Pett, J. Chem. Soc.,
Dalton Trans., 1988, 89.
5 F. S. M. Hassan, S. J. Higgins, G. B. Jacobsen, B. L. Shaw and
M. Thornton-Pett, J. Chem. Soc., Dalton Trans., 1988, 3011.
6 H.-J. Haupt, E. Kleineberg and U. Flörke, Z. Anorg. Allg. Chem.,
1993, 619, 869.
7 A. M. Herring, S. J. Higgins, G. B. Jacobsen and B. L. Shaw, J. Chem.
Soc., Chem. Commun., 1986, 882.
8 S. J. Higgins and B. L. Shaw, J. Chem. Soc., Dalton Trans., 1989,
1527.
9 G. King, S. J. Higgins and A. Hopton, J. Chem. Soc., Dalton Trans.,
1992, 3403.
37 M. Bressan and P. Rigo, Inorg. Chem., 1975, 14, 2286.
38 J. Ruiz, M. E. G. Mosquera and V. Riera, J. Organomet. Chem.,
1997, 527, 35.
39 J. V. Barkley, J. C. Grimshaw, S. J. Higgins, P. B. Hoare, M. K.
McCart and A. K. Smith, J. Chem. Soc., Dalton Trans., 1995,
2901.
10 J. V. Barkley, S. J. Higgins, M. K. McCart and T. J. Pounds,
Inorg. Chem., 1997, 36, 6188.
11 S. J. Higgins, M. K. McCart, M. McElhinney, D. C. Nugent and
T. J. Pounds, J. Chem. Soc., Chem. Commun., 1995, 2129.
12 S. J. Higgins, H. L. Jones, M. K. McCart and T. J. Pounds, Chem.
Commun., 1997, 1907.
40 M. Knorr, E. Hallauer, V. Huch, M. Vieth and P. Braunstein,
Organometallics, 1996, 15, 3868.
41 S. C. Browning and D. H. Farrar, J. Chem. Soc., Dalton Trans., 1995,
521.
42 C. Bianchini, D. Masi, K. Linn, C. Mealli, M. Peruzzini and
F. Zanobini, Inorg. Chem., 1992, 31, 4036.
13 G. Smith, D. J. Cole-Hamilton, A. C. Gregory and N. C. Gooden,
Polyhedron, 1982, 1, 97.
14 S. J. Higgins, C. A. Stuart and A. Mills, Inorg. Chem. Commun.,
2000, 3, 208.
15 A. A. LaPensée, S. J. Higgins, C. A. Stuart and J. F. Bickley, Inorg.
Chem. Commun., 1999, 2, 524.
16 Y. Zhu, M. O. Wolf and G. P. A. Yap, Inorg. Chem., 1997, 36, 5483.
17 C. L. Jones and S. J. Higgins, J. Mater. Chem., 1999, 9, 865.
18 N. Binstead, EXCURV 98, CCLRC Daresbury Laboratory,
Warrington, 1998.
43 A. J. Blake, R. M. Christie, Y. V. Roberts, M. J. Sullivan,
M. Schröder and L. J. Yellowlees, J. Chem. Soc., Chem. Commun.,
1992, 848.
44 B. Chaudret, G. Commenges and R. Poilblanc, J. Chem. Soc.,
Dalton Trans., 1984, 1635.
45 M. V. Baker and L. D. Field, Inorg. Chem., 1987, 26, 2010.
46 E. Lindner, A. Möckel, H. A. Mayer, H. Kühbauch, R. Fawzi and
M. Steimann, Inorg. Chem., 1993, 32, 1266.
47 B. Delavaux, B. Chaudret, F. Dahan and R. Poilblanc, J. Organomet.
Chem., 1986, 317, 69.
910
J. Chem. Soc., Dalton Trans., 2001, 902–910