Chelate ring-opening aminophosphine complexes of ruthenium(II)

Article abstract of DOI:10.1016/S0020-1693(02)01133-7

The preparations and X-ray crystal structures of the Ru(II) aminophosphine complexes trans, cis-[RuCl₂(H(Bz)NCH₂CH₂PPh₂-N,P) ₂] (1), trans, cis-[RuCl₂(H₂NCH₂CH₂PPh ₂-N,P)₂] (3) and [(η⁶-C₆H₆)Ru(PPh₂CH ₂CH₂NMe₂-N,P)Cl][PF₆] (5) are described. Complexes 1 and 3 are octahedral and contain a marked distortion of the Cl-Ru-Cl axis from linearity. Complex 3 undergoes facile chelate ring-opening in DMSO and CH₃CN solutions to give complexes containing one chelated aminophosphine ligand and one pendant ligand coordinated only through P. The X-ray structure of trans, cis-[RuCl₂(DMSO-S)₂(H₂NCH₂CH ₂P(O)Ph₂-N,O)] (4) containing an oxidised aminophosphine ligand is also reported.

Full text of DOI:10.1016/S0020-1693(02)01133-7

Inorganica Chimica Acta 339 (2002) 551ꢀ559
/
www.elsevier.com/locate/ica
Chelate ring-opening aminophosphine complexes of ruthenium(II)
Robert Morris, Abraha Habtemariam, Zijian Guo, Simon Parsons, Peter J. Sadler *
School of Chemistry, The University of Edinburgh, King’s Building, West Mains Road, Edinburgh EH9 3JJ, UK
Received 22 February 2002; accepted 17 May 2002
Dedicated in honor of Professor Helmut Sigel
Abstract
The preparations and X-ray crystal structures of the Ru(II) aminophosphine complexes trans, cis-[RuCl2(H(Bz)NCH2CH2PPh2-
N,P)2] (1), trans, cis-[RuCl2(H2NCH2CH2PPh2-N,P)2] (3) and [(h6-C6H6)Ru(PPh2CH2CH2NMe2-N,P)Cl][PF6] (5) are described.
Complexes 1 and 3 are octahedral and contain a marked distortion of the ClÃ
/
RuÃ/Cl axis from linearity. Complex 3 undergoes facile
chelate ring-opening in DMSO and CH3CN solutions to give complexes containing one chelated aminophosphine ligand and one
pendant ligand coordinated only through P. The X-ray structure of trans, cis-[RuCl2(DMSO-S)2(H2NCH2CH2P(O)Ph2-N,O)] (4)
containing an oxidised aminophosphine ligand is also reported.
# 2002 Elsevier Science B.V. All rights reserved.
Keywords: Ruthenium(II); Aminophosphine ligands; X-ray structures; Chelate ring-opening
1. Introduction
ruthenium(II) [7], trans-dichloro-N,N?-bis[(o-(di-phe-
nylphosphino)benzylidene]ethylenediamine-rutheniu-
m(II) where the aminophosphine ligand is tetradentate
There are many previous reports of the potential use
of aminophosphine complexes, mainly in relation to
catalysis [1]. However, very few studies [2] have inves-
tigated the biological activity of such complexes even
though they contain cis amine ligands, a feature found
in many active platinum anti-cancer agents [3]. Chelate
ring-opened aminophosphine complexes also possess
lipophilic-cationic character and have the potential to
exhibit cytotoxicity by disrupting mitochondrial func-
tion [4]. Octahedral d6 Ru(II) centres can provide more
reaction sites than square-planar d8 Pt(II) and Pd(II)
and in addition, nitrogen ligands would be expected to
bind more weakly than phosphorus, thus offering the
possibility of activation via chelate ring-opening [5,6].
[9], and the dimeric structure of [Ru2Cl2(PÃ
/
N)4]2ꢁ
,
where PÃN is 1-(diphenyl-phosphino)-2-(2-pyridy-
/
l)ethane [10]. We have previously reported [11] the
synthesis and characterisation of the Ru(II) complex
trans, cis-[RuCl2(Me2NCH2CH2PPh2-P,N)2]. This
complex underwent a two step reaction in methanol
involving chloride dissociation followed by PÃ
/N chelate
ring-opening.
In this paper, we report the synthesis and crystal
structures of Ru(II) aminophosphine complexes of the
ligands shown in Fig. 1. We have studied chelate ring-
opening reactions and considered possible biological
applications.
Although some Ruꢀaminophosphine complexes have
/
been synthesised previously [7,8], only a few X-ray
crystal structures have been reported. These include
the five coordinate complex dichloro(o-diphenylphos-
phino-N,N-dimethylaniline)-[tris(p-tolyl)phosphine]-
2. Experimental
2.1. Chemicals
The following were used as received: RuCl3×
/3H2O
(Johnson Matthey), 1,4-cyclohexadiene (Acros). Aceto-
nitrile used for cyclic voltammetric studies was pre-dried
with K2CO3 followed by distillation from CaH2.
* Corresponding author. Tel.: ꢁ
6452
E-mail address: p.j.sadler@ed.ac.uk (P.J. Sadler).
/
44-131-650 4729; fax: ꢁ44-131-650
/
0020-1693/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 0 2 0 - 1 6 9 3 ( 0 2 ) 0 1 1 3 3 - 7

552
R. Morris et al. / Inorganica Chimica Acta 339 (2002) 551ꢀ559
/
Me2NCH2CH2PPh2 (DMAP), H(Bz)NCH2CH2PPh2×
/
2HCl (DPEBA) [5] and the complexes cis- and trans-
[RuCl2(DMSO)4] [12], [(COD)RuCl2]x [13], and [(h6-
C6H6)Ru(CH3CN)2Cl][PF6] [14].
2.4. Preparation of complexes
2.4.1. Synthesis of trans, cis-
[RuCl2(H(Bz)NCH2CH2PPh2-N,P)2] (1)
cis-[RuCl2(DMSO)4] (0.24 g, 0.5 mmol) was stirred in
acetone (10 ml) and the ligand DPEBA (0.36 g, 1 mmol)
added in one portion. This was sealed and the reaction
stirred overnight at ambient temperature. Ether was
added to induce precipitation of a pink solid, which was
Fig. 1. Structures of the ligands used. Abbreviations: DHAP, 1-
amino-2-diphenylphosphinoethane; DMAP, 1-dimethylamino-2-di-
phenylphosphinoethane; DPEBA, 1-benzylamino-2-diphenylphosphi-
noethane.
collected and recrystallised from acetoneꢀether.
/
Yield: 4 mg, 1%. Calc. for C42H44Cl2N2P2Ru: C,
62.22; H, 5.47; N, 3.46. Found: C, 61.63; H, 5.80; N,
3.71%. Single crystals of complex 1 suitable for X-ray
2.2. Measurements
analysis were grown from an acetoneꢀether mixture.
/
Ultraviolet and visible spectra were obtained on a
Perkinꢀ
/
Elmer Lambda 16 UVꢀVis recording spectro-
/
2.4.2. Synthesis of trans, cis-
[RuCl2(Me2NCH2CH2PPh2-N,P)2] (2)
photometer using 1 cm path length quartz cuvettes. The
temperature was controlled using a PTP1 Peltier Tem-
perature Programmer. Spectra were normally referenced
to solvent alone. Spectra were processed with UVWIN-
LAB for WINDOWS ’95.
1H and 31P{1H} NMR spectra were recorded at 298 K
on Bruker DMX500 (1H 500 MHz; 31P 202 MHz) and
Varian-Inova 600 spectrometers (1H 600 MHz), using 5
mm NMR tubes. The chemical shift references were as
follows: H (internal, TSP), 31P (external, 30% H3PO4).
The 2D gradient-selected 1H,1H-COSY and TOCSY
(mixing time of 0.12 s) experiments were carried out using
The ligand DMAP (0.36 g, 1.4 mmol) dissolved in
benzene (5 ml) was added to a stirred suspension of
[Ru(COD)Cl2]x (0.39 g, 1.4 mmol) in benzene (10 ml).
The reaction was refluxed for 24 h and filtered while hot
to remove unreacted ruthenium starting material. The
red filtrate was evaporated to dryness and recrystallised
from toluene.
Yield: 0.33 g, 34.6%. Calc for C32H40Cl2N2P2Ru: C,
55.98; H, 5.87; N, 4.08. Found: C, 55.53; H, 5.63; N,
3.92%. 31P{1H} NMR (d8-toluene): d 57.0.
1
standard sequences. Data sets with 2048ꢂ512 points
/
2.4.3. Synthesis of trans, cis-
[RuCl2(H2NCH2CH2PPh2-N,P)2] (3)
were acquired with frequency widths of 8000 Hz in both
dimensions and 16 scans per t1 increment. The t1
dimension was zero-filled to 2048 data points, and the
spectrum was processed with a combination of expo-
nential and Gaussian weighting functions. The 2D
[1H,31P] HSQC NMR spectra were recorded using a
standard sequence. The 31P-spins were decoupled by
irradiationwiththeGARP-1sequenceduringacquisition.
Cyclic voltammetry experiments were carried out in
The ligand DHAP (1.15 g, 5 mmol) was added in one
portion to a stirred suspension of [Ru(COD)Cl2]x (0.7 g,
2.5 mmol) in benzene (40 ml) and the mixture refluxed
under argon. After 4 h, a yellow precipitate had formed.
This was collected after cooling and recrystallised from
CH2Cl2ꢀether.
/
Yield: 0.20 g, 12.8%. Calc. for C28H32Cl2N2P2Ru: C,
53.34; H, 5.12; N, 4.44. Found: C, 53.95; H, 5.16; N,
3.07%. 31P{1H} NMR (d6-DMSO): d 60.7 (d, Jꢃ
CH3CNꢀ[n-Bu4N][BF4] (0.4 M), with an Autolab
/
/24.6
PGSTAT20 potentiostat equipped with GPES 4.2 soft-
ware. The working electrode was a platinum microdisc
of diameter 0.5 mm and the potential reported is quoted
Hz), 45.7 (d, Jꢃ
/
24.6 Hz). Crystals of complex 3 suitable
for X-ray analysis were grown by the slow evaporation
of an ethanolic solution.
relative to a Ag/AgCl reference electrode via FeCp2ꢀ
/
ꢁ
FeCp2 (0.375 V).
2.4.4. Synthesis of trans, cis-[RuCl2(DMSO-
S)2(H2NCH2CH2P(O)Ph2-N,O)] (4)
2.3. Preparation of ligands and starting complexes
trans-[RuCl2(DMSO)4] (0.06 g, 0.13 mmol) was set
stirring in freshly dried ethanol (10 ml). DHAP (0.03 g,
0.13 mmol) was added in one portion and the yellow
solution stirred overnight. An orange precipitate formed
The following were synthesised according to pub-
procedures:
lished
H2NCH2CH2PPh2
(DHAP),

R. Morris et al. / Inorganica Chimica Acta 339 (2002) 551ꢀ
/559
553
during this time. This was collected and recrystallised
ether.
Yield: 6 mg, 7.3%. 31P{1H} NMR (CDCl3): d 51.6 (s).
Complex 4 was obtained as single crystals suitable for
X-ray analysis by diffusion of ether into an ethanol
solution.
and all 5534 unique data to 2uꢃ
/
1208). The final
ꢄ3
˚
0.81 e A
from ethanolꢀ
/
difference map extremes were ꢁ
/
0.70 and ꢄ
/
.
2.5.2. Crystal data for 3
C28H32Cl2N2P2Ru, Mꢃ
/
630.47, triclinic, space group
˚
¯
P
/
1; aꢃ
aꢃ77.502(2), bꢃ
1375.5(2) A , Tꢃ
Orange block 0.20ꢂ
rather weakly diffracting, and so data were collected
/
9.9837(8), bꢃ
75.694(2), gꢃ
150 K, Zꢃ2, rcalc
0.06ꢂ0.06 mm. The crystal was
/
10.4950(9), cꢃ
63.886(2)8, Vꢃ
1.522 mg mꢄ3
.
/
15.2046(13) A,
2.4.5. Synthesis of [(h6-C6H6)Ru(Me2NCH2CH2PPh2-
N,P)Cl][PF6] (5)
/
/
/
/
3
˚
/
/
ꢃ
/
DMAP (0.077 g, 0.23 mmol) dissolved in dry aceto-
nitrile (2 ml) was added in one portion to a solution of
[(h6-C6H6)Ru(CH3CN)2Cl][PF6] (0.102 g, 0.23 mmol) in
dry acetonitrile (8 ml). The mixture was stirred at
ambient temperature for 10 h in a sealed flask with no
special precautions to exclude air. The orange solution
was evaporated to an orange oil which was taken up in 2
ml CHCl3. Ether was added to induce precipitation and
the orange solid was collected, washed with ether and
recrystallised as the 0.5 CH3CN solvate by slow diffu-
sion of ether into an acetonitrile solution.
/
/
˚
using synchrotron radiation (lꢃ
tion correction was applied using Sadabs [17] (mꢃ
mmꢄ1, range of transmission: 0.449ꢀ
0.928) [18]. R1ꢃ
3.78% (based on F and 6496 data with F ꢀ4s(F)) and
wR2ꢃ
9.55% (based on F2 and all 7321 unique data to
2uꢃ598). The final difference map extremes were ꢁ1.63
and ꢄ
0.93 e Aꢄ3. Dispersion terms were calculated
/
0.6878 A); an absorp-
/0.901
/
/
/
/
/
/
˚
/
using FPRIME [18].
Yield: 0.085 g, 57.8%. Calc. for C23H27.5Cl-
F6N1.5P2Ru: C, 43.30; H, 4.35; N, 3.29. Found: C,
43.05; H, 4.40; N, 3.03%. 1H NMR (CD3CN): d 5.68 (d,
6H), 3.19 (s, 3H), 3.11 (s, 3H). 31P{1H} NMR (CD3CN):
d 57.9 (s). Crystals of complex 5 suitable for X-ray
analysis were grown by diffusion of ether into an
acetonitrile solution.
2.5.3. Crystal data for 4
C18H28Cl2NO3PRuS2, Mꢃ
c, aꢃ11.459(9), bꢃ11.938(8), cꢃ
106.63(6)8, Vꢃ2344.4(8) A , Tꢃ
1.625 mg mꢄ3. Orange block, 0.45ꢂ
Mo Ka radiation mꢃ
1.162 mmꢄ1. The peak profiles
/
573.47, monoclinic, P21/
˚
/
/
/
17.886(16) A, bꢃ
220 K, Zꢃ4, rcalc
0.27ꢂ0.19 mm,
/
3
˚
/
/
/
ꢃ
/
/
/
/
were rather asymmetric and misshapen, and possibly for
this reason an attempt to perform an absorption
correction based on c-scan measurements was unsuc-
cessful; a correction using SHELXA [16] was applied after
2.5. X-ray crystallographic analysis
Data for 1, 4 and 5 were collected on Stoe Stadi-4
diffractometers using conventional sealed-tube sources,
the data set for 3 was collected using synchrotron
radiation with a Bruker SMART diffractometer on
Station 9.8 at the Synchrotron Radiation Source at
Daresbury, UK. Both instruments were equipped with
Oxford Cryosystems low-temperature devices. The
structures were solved by Patterson methods (DIRDIF
or SHELXTL) [15,16] and refined by full-matrix least-
squares against F2 (SHELXTL). In all cases H-atoms were
placed in calculated positions.
isotropic refinement (range of transmission: 0.05ꢀ
Only the Ru, Cl, P, S and O atoms were refined
anisotropically. R1ꢃ12.0% (based on F and 1370 data
with F ꢀ4s(F)) and wR2ꢃ
35.710% (based on F2 and
all 3061 unique data to 2uꢃ458). The final difference
map extremes were ꢁ1.25 and ꢄ
/
0.46).
/
/
/
/
/
/
1.67 e Aꢄ3. Clearly
˚
this is a low-precision determination, and the results
were only used to establish chemical connectivity.
2.5.4. Crystal data for 5
2.5.1. Crystal data for 1
C42H44Cl2N2P2Ru, Mꢃ
C23H27.5ClF6N1.5P2Ru, Mꢃ
c, aꢃ10.954(3), bꢃ16.076(5), cꢃ
110.45(2)8, Vꢃ2501.2(12) A , Tꢃ
rcalc
1.693 mg mꢄ3. Orange block, 0.45ꢂ
0.45 mm, Mo Ka radiation mꢃ
0.921 mmꢄ1. An
absorption correction was applied using c-scan data
(range of transmission: 0.478ꢀ0.548). One phenyl group
/
637.43, monoclinic, P21/
˚
/
810.70, triclinic, space group
/
/
/
15.159(4) A, bꢃ
150 K, Zꢃ4,
0.45ꢂ
/
3
˚
˚
¯
P
/
1; aꢃ
/
10.920(6), bꢃ
106.74(5), gꢃ
220 K, Zꢃ2, rcalc
block 0.19ꢂ0.16ꢂ0.12 mm. The crystal was rather
/
12.061(7), cꢃ
111.50(4)8, Vꢃ
1.424 mg mꢄ3. Orange
/
16.696(9) A, aꢃ
/
/
/
/
97.35(5), bꢃ
/
/
/1890.8(18)
ꢃ
/
/
/
3
˚
A , Tꢃ
/
/
ꢃ
/
/
/
/
weakly diffracting, and so data were collected with Cu
Ka radiation; an absorption correction was applied
/
ꢄ
and the PF6 anion are disordered over two orienta-
tions; the acetonitrile of solvation is disordered about a
using c-scan data (mꢃ
/
5.700 mmꢄ1, range of transmis-
0.704) [16]. One carbon atom (labelled C3 in
sion: 0.474ꢀ
/
crystallographic inversion centre. R1ꢃ
F and 4092 data with F ꢀ4s(F)) and wR2ꢃ
(based on F2 and all 4414 unique data to 2uꢃ
508). The
final difference map extremes were ꢁ0.86 and ꢄ0.65 e
/3.69% (based on
the tables) is disordered over two positions; similarity
restraints were applied to the geometries of the two part-
weight components. R1ꢃ
/
/9.27%
/
/
7.64% (based on F and 2833
17.56% (based on F2
/
/
ꢄ3
˚
A
data with F ꢀ4s(F)) and wR2ꢃ
/
/
.

554
R. Morris et al. / Inorganica Chimica Acta 339 (2002) 551ꢀ559
/
3. Results and discussion
ligand in a 4:1 ratio. When this was reacted with trans-
[RuCl2(DMSO)4] in a 1:1 molar ratio we obtained the
aminophosphine oxide complex trans, cis-[RuCl2-
(DMSO-S)2(H2NCH2CH2P(O)Ph2-N,O)] (4), in very
low yield. Ruthenium phosphine oxide complexes can
be difficult to isolate due to the liability of the phosphine
oxide ligands in solution [22], for example a Ru(II)terpy
complex with (O)Ph2PCH2CH2PPh2(O) has been re-
ported to dissociate rapidly in solution [21]. The 31P
NMR spectrum of complex 4 in CDCl3 showed a singlet
at 51.6 ppm.
3.1. Preparation and properties of ligands
The ligands, prepared by literature methods [5], were
obtained as viscous oils (DMAP and DHAP) or as a
hydrochloride salt (DPEBA). The oily ligands were
readily purified by passing ether solutions through a
short column of alumina or by reduced-pressure dis-
tillation. The DPEBA ligand was recrystallised from
methanolꢀether.
/
The reaction of [(h6-C6H6)Ru(CH3CN)2Cl][PF6] with
DMAP in acetonitrile followed by work-up produced
red crystals of the complex [(h6-C6H6)Ru(DMAP-
N,P)Cl][PF6] (5) in moderate yield. Complex 5 is soluble
in alcohols and coordinating solvents such as DMSO
and acetonitrile, but insoluble in water or mixed solvents
The mixed donor ligands DMAP, DHAP and
DPEBA are soluble in almost all organic solvents. The
1H NMR spectra of the ligands in CDCl3 showed the
expected peaks with the backbone Ã
plets and the ÃNH2 a broad peak at d 1.5 (DHAP). The
31P{1H} NMR spectra showed singlet resonances in the
region of ꢄ19 to ꢄ22 ppm.
/
CH2Ã
/
giving multi-
/
/
/
containing an appreciable amount (ꢀ20%) of water.
/
1
The H NMR spectrum in CD3CN shows a doublet for
the p-coordinated benzene ring due to coupling with 31P,
implying that all the protons on the benzene ring are
equivalent and hence there is free rotation of the p-
coordinated benzene ligand. The NMe2 groups of the
chelated ligand gave rise to two singlets at d 3.11 and
3.19, indicating that the two methyl groups are none-
quivalent, as expected since the Ru(II) centre is chiral.
The 31P NMR spectrum showed a singlet at d 58,
appropriate for a ring-closed structure. The 13C{1H}
3.2. Preparation and properties of ruthenium(II)
complexes
cis-[RuCl2(DMSO)4] is a very versatile reagent known
to provide a very good entry point to Ru(II) chemistry.
A large number of Ru(II) complexes have been synthe-
sised from it [19], including the chelate-ring opening
aminophosphine complex [RuCl2(DMAP)2] [11]. The
easily displaced DMSO ligands make possible the
formation of Ru(II) chloride complexes with a variety
of ligands.
The reaction of cis-[RuCl2(DMSO)4] with DPEBA as
its hydrochloride salt in a 1:2 molar ratio, in acetone,
gave rise to trans, cis-[RuCl2(H(Bz)NCH2CH2PPh2-
N,P)2] (1). Attempts to prepare 1 in CH2Cl2 resulted
in an air-sensitive, uncharacterisable reddish solid.
The 1:1 reaction of the ligand DMAP and
[Ru(COD)Cl2]x in refluxing benzene produced the
complex trans, cis-[RuCl2(Me2NCH2CH2PPh2-N,P)2]
(2). It was hoped that this product would contain one
coordinated ligand and an h4-coordinated cyclocta-
diene. However, under the conditions used, the COD
ligand proved to be too labile and was displaced by two
aminophosphine ligands. Complex 2 is a known com-
plex [11,20].
NMR spectrum contained resonances at d 125ꢀ
the PÃ
Ph2 carbons, and a doublet at d 89.3 for the h6-
coordinated benzene ring, due to coupling to31P, (JCP
2.95 Hz). The NÃCH2Ã signal appeared as doublet at d
60.5 (JCP CH2Ã resonance as
2.98 Hz), and the PÃ
doublet at d 25.6 (JCP CH3
26.66 Hz). The two NÃ
/135 for
/
ꢃ
/
/
/
ꢃ
/
/
/
ꢃ
/
/
signals were singlets at d 58.3 and 56.5. There was no
change in the NMR spectrum after 3 days incubation at
300 K, showing that chelate-ring opening does not occur
in CD3CN, a coordinating solvent.
3.3. NMR analysis of chelate ring-opening of complex 3
Complex
3 trans, cis-[RuCl2(H2NCH2CH2PPh2-
N,P)2] underwent a rapid chelate ring-opening reaction
in DMSO. A fresh solution of 3 in DMSO-d6 was
monitored by NMR. The ring-opening reaction ap-
peared complete by the time the first spectrum was
acquired. The presence of a one ring-open, one ring-
closed conformation was strongly suggested by the
31P{1H} NMR spectrum. Two doublets were present,
The reaction of [Ru(COD)Cl2]x with DHAP in
refluxing benzene in a 1:2 molar ratio resulted in the
displacement of the COD ligand and the coordination of
two
DHAP
ligands
to
form
trans,
cis-
[RuCl2(H2NCH2CH2PPh2-N,P)2] (3). The complex
was characterised by elemental analysis and X-ray
crystallography. Complex 3 is reasonably soluble only
in coordinating solvents such as DMSO and acetonitrile,
and sparingly soluble in alcohols.
one at d 60.7 and the other at d 45.7. The J(PÃ
/
P)
coupling constant of 24.6 Hz is indicative of two cis
phosphorus nuclei [23]. The signal at d 60.7 corresponds
to the expected value for a chelated aminophosphine
ligand bound through both P and N [11,24]. If ring-
One batch of our aminophosphine ligand DHAP
contained both reduced (31P d ꢄ
/21.0) and oxidised
(PÄ
/
O d 32.6, c.f. Ph3PÄ
/
O dꢃ/31.0 [21]) forms of the

R. Morris et al. / Inorganica Chimica Acta 339 (2002) 551ꢀ
/559
555
opening of a chelated ligand occurs (RuÃ
/N bond
Table 1
Selected bond lengths (A) and angles (8) for complex 1, 3, 4 and 5
breaks), the 31P signal for the phosphorus of that ligand
would be expected to shift to higher field. This was
observed for trans, cis-[RuCl2(DMAP)] [11] and also
the chemical shift is similar to that observed for the ring-
opened Pt(II) complexes of aminophosphine ligands
coordinated via P only [5,6]. This phenomenon has also
been observed for the complex [Cp*Ru(DMAP)Cl] [31],
for which the 31P chemical shift of the ring-closed form
of the complex is d 60.9 and the ring-opened form
appears at higher field at d 38.9. After 1 week it was
noted that a third signal had appeared in the 31P
spectrum of 3 at d 51.4. This was a singlet, indicating
equivalent phosphorus nuclei in the complex. However,
even after 2 months this species never became dominant
and a definite assignment was not possible. This peak
may arise from a Ru-complex with 2 equiv. ring-opened
ligands. No free ligand was detected and therefore,
complexes with only one ligand, due to loss of a DHAP
can be ruled out. The spectrum in CD3CN is very
similar. Thus 3 reacts with DMSO and CH3CN to give a
complex with one chelated aminophosphine ligand and
one pendant ligand coordinated only through phos-
˚
1
3
4
5
Bond lengths
RuÃP(1)
RuÃP(2)
RuÃN(1)
RuÃN(2)
RuÃCl(1)
RuÃCl(2)
RuÃO
2.247(4)
2.252(4)
2.212(11)
2.230(10)
2.401(4)
2.410(4)
2.2552(6)
2.2565(6)
2.3122(11)
2.212(3)
2.1642(19) 2.20(2)
2.1796(18)
2.4053(6)
2.4192(6)
2.400(7) 2.3986(11)
2.401(7)
2.18(2)
RuÃS(1)
RuÃS(2)
2.200(7)
2.250(5)
Bond angles
P(1)ÃRuÃP(2)
N(1)ÃRuÃN(2)
Cl(1)ÃRuÃCl(2)
P(1)ÃRuÃN(2)
P(2)ÃRuÃN(2)
N(1)ÃRuÃO(1)
101.48(14) 102.80(2)
93.1(4) 90.73(7)
166.98(11) 164.29(2)
170.6(2)
81.68(9)
82.5(3)
83.0(3)
82.99(5)
83.51(5)
94.4(6)
Table 2
1H NMR chemical shifts (d) for the chelated and pendant (ring-
openend) ligands of complex 3 in DMSO-d6
1
phorus. The H NMR spectrum showed a variety of
peaks due to the 2 inequiv. ligands, and from COSY,
TOCSY and [31PÃ
/
1H] HSQC experiments it was possi-
ble to distinguish between the pendant and chelated
ligands. The [31PÃ1H] HSQC spectrum is shown in Fig.
/
2 and the COSY assignments are given in Table 2. The
combined COSY and HSQC spectra allowed assign-
ments of the ring-opened and ring-closed aminopho-
sphine complexes to be made, and the TOCSY
experiment allowed the NÃ/H protons to be assigned.
d (chelated ligand)
d (pendant ligand)
The ligand coordinated at the vacated site after ring-
opening, is believed to be DMSO, the solvent used for
the NMR experiments.
NÃHx
NÃHy
H1
4.97
4.71
3.13
2.57
3.07
2.95
4.23
4.23
2.89
1.86
2.57
1.97
7.89
H1?
H2
H2?
a
Ha, Ha?
Hb, Hb?
Hc
7.78 (7.36)
b
7.4 (7.27)
b
7.4 (6.75)
b
7.4
7.4
b
a
Figures in brackets refer to the second phenyl group.
Overlapped.
b
3.4. Cyclic voltammetry of complex 3 in CH3CN
The Ru(III)/(II) reduction potential of ruthenium
complexes may play an important role in their biological
activity. It has been hypothesised that Ru(III) complexes
may be prodrugs that are activated in vivo by reduction
to Ru(II) [25]. Therefore, we investigated the electro-
chemistry of complex 3. Acetonitrile was used to provide
good solubility and allow ease of handling at various
temperatures. With CH3CN as solvent, a reversible
Fig. 2. [1H, 31P] 2D HSQC NMR spectrum of complex 3 after
standing for 1 week in d6-DMSO. The dotted boxes contain peaks for
the major species, the mono chelate ring-opened complex: 31P doublet
for ring-opened ligand at low field, and for the ring-closed ligand at
1
high field. H,1H 2D COSY was used to confirm the assignments (see
Table 2).

556
R. Morris et al. / Inorganica Chimica Acta 339 (2002) 551ꢀ559
/
oxidation at 0.82 V (vs. Ag/AgCl) was observed at 298
K. The oxidised species appeared to be unstable and no
peak was detected using a slow sweep rate. When the
sample was cooled to 227 K, this couple was more
readily observed. The redox process can tentatively be
assigned to the metal centred Ru(II)ꢀRu(III) couple.
/
The value of 0.82 V is much higher than the correspond-
ing value for [RuCl2(DMAP)2] [11] which was observed
at 0.326 V, and is likely to be too high to be of biological
relevance.
3.5. UVꢀ
/
Vis spectroscopy of complex 5
The UVꢀ/Vis spectrum of 5 in DMSO changed with
time (Fig. 3). The band at 326 nm gradually decreased in
intensity and that at 300 nm increased concomitantly,
with an isosbestic point at 320 nm. By following the
absorbance decrease at 326 nm with time, assuming a
first order reaction, and treating the data by the
Guggenheim method [26], a first order rate-constant of
Fig. 4. X-ray crystal structure of complex 1. Probability surfaces
enclose 50% ellipsoids. The minor disorder component has been
omitted.
2.03ꢂ
10ꢄ4 sꢄ1 was obtained. This reaction can be
attributed to substitution of a chloride ligand by DMSO
and was inhibited in the presence of 100 mM LiCl.
/
3.6. Crystallography
The X-ray crystal structures of 1, 3, 4 and 5 are shown
in Figs. 4ꢀ7. Data collection and refinement parameters
are summarised in Section 2. Bond distances and angles
are listed in Table 1.
/
Fig. 5. X-ray crystal structure of complex 3.
The Ru(II) centres in these complexes have distorted
octahedral geometries. In complex 1 the two Cl ligands
are trans and the two chelated PÃ
/
N ligands are in cis
positions (Fig. 4). This arrangement appears to be
sterically crowded, but is apparently preferred to that
in which the two P atoms are trans because of their
strong trans effects [27]. This is in agreement with the
previously reported structure of trans, cis-[RuCl2(D-
MAP)2] [11]. There are no strong intermolecular inter-
actions involved in the crystal packing. The RuÃ
/
P
˚
Fig. 6. X-ray crystal structure of complex 4.
(2.247(4) and 2.252(4) A) and RuÃ
/
Cl (2.401(4) and
˚
2.410(4) A) bond lengths are within the range of values
typical for Ru(II) complexes [11,28]. In trans, cis-
[RuCl2(DMAP)2] [11] it was found that the RuÃ
/
N
Fig. 3. The electronic absorption spectrum of complex 5 in DMSO at
298 K recorded every hour for 12 h. A pseudo first order kinetic plot
using the absorbance change 326 nm gave a rate constant of 2.03ꢂ
/
Fig. 7. X-ray crystal structure of complex 5. The PF6 anion is not
shown.
10ꢄ4 sꢄ1
.

R. Morris et al. / Inorganica Chimica Acta 339 (2002) 551ꢀ
/559
557
˚
2.391 A) were considerably longer
bond lengths (meanꢃ
/
expect two p-accepting ligands to be located orthogonal
/
˚
˚
O bond length is 2.18(2) A,
similar to those in the chelated phosphine-oxide com-
(0.2 A) than values for PÃ
/
N chelated Ru(II) complexes
previously reported, for example, N trans to P in
to each other [32]. The RuÃ
complexes such as trans-dichloro-bis(diphenylphosphi-
plex trans-[RuCl2{o-C6H4(PMePh)2}{o-C6H4(PMePh)-
˚
no)pyridineruthenium(II) (2.13 and 2.06 A) [29], trans-
dichloro-N,N-bis-[o-(diphenylphosphino)benzylidene]-
˚
(ethylenediamine)ruthenium(II) (2.17 and 2.16 A) [9],
and dichlorotetra(1-(diphenylphosphino)-2-(2-pyridyl)-
(P(O)MePh)}] which has a RuÃ
/
O bond length of
6
˚
2.166(5) A [33], the arene complex [(h -p-cymene)RuCl-
/ /
(h2-Ph2PCH2P(O)Ph2)][SbF6] (RuÃ
˚
Oꢃ2.154(3) A), [34]
and the mixed-valence Ru(II/III) binuclear compound
[Ru2(m-O2CC4H3S)4(OPPh3)2]ꢁ (RuÃ
[35]).
The RuÃ
complex 4 are significantly different from each other,
/
/
˚
ethane)diruthenium(II) (2.152 and 2.157 A) [30]. How-
˚
O(P)ꢃ2.216(7) A
ever, this effect is not seen for complex 1, despite the
similar geometry and despite having the potentially
more bulky benzyl group on the nitrogens. It is possible
that the presence of H as the second substituent on N
reduces the steric effects. The benzyl groups point away
from each other, thereby greatly reducing the steric
hindrance. The crystal structures of Pd(II) and Pt(II)
complexes with the same ligand (DPEBA) do not
/
S bonds for the two DMSO ligands of
˚
the longest is RuÃ
/
S trans to N (2.250(5) A). The RuÃ
/
S
˚
bond trans to oxygen, however, is quite short, 2.200(7) A,
similar to that for [Ru(DMSO)(NH3)5]2ꢁ, 2.188(3) A
˚
[36]. The shortness of the RuÃ
complex has been attributed [36] to an Mꢁ ꢃ
component, a conclusion supported by the relatively long
/
S bond in the latter amine
/
SÃ
/
Oꢄ
contain any metalÃ
/
N bonds probably because the
˚
benzyl groups are too sterically demanding and prevent
chelation in these square-planar structures [5]. The
SÃ
/
O bond length of 1.527(7) A. For uncoordinated
˚
O bond distance is 1.4918(9) A.
˚
This value is lengthened to 1.528(1) A upon oxygen
coordination to metal ions, while it is reduced to
sulfoxides, the average SÃ
/
bending of the ClÃ
/
RuÃ
/
Cl axis (166.98(11)8) probably
results from crowding caused by the phenyl substituents
on phosphorus and is greater than was found for
˚
1.4731(6) A upon sulfur coordination [37]. Complex 4,
˚
O bond length (1.48(2) A), a feature
[RuCl2(DMAP)2] (1728) [11]. ClÃ
/
RuÃ
/
Cl bond angles
has a typical SÃ
/
as low as 165.48 have been reported in the literature, e.g.
for trans-dichloro-N,N-bis-[o-(diphenylphosphino)ben-
zylidene](ethylenediamine)ruthenium(II) [9].
found in S-bound DMSO complexes such as [Ru(II)Cl3-
2ꢁ
(DMSO)3]ꢄ (mean 1.48 A) [38] and [Ru(DMSO)6]
˚
˚
(mean 1.482 A) [39]. Both RuÃ
/
S bonds in complex 4 are
shorter than those in cis-[RuCl2(DMSO)4] (mean 2.267
The structure of complex 3 is shown in Fig. 5. The two
chloride ligands are trans and the aminophosphine
ligands are cis. The lack of steric hindrance from the
˚
˚
A), its isomer trans-[RuCl2(DMSO)4] (mean 2.352 A)
[12], and [Ru(DMSO)6]2ꢁ (mean 2.259 A). The ClÃ
/
RuÃ
/
˚
NÃ
/
H groups is apparent from the RuÃ
/N bond lengths of
Cl angle of 170.6(2)8 is again distorted from linearity. The
˚
˚
N bond length (2.20(2) A) is in the expected range
2.1642(19) and 2.1796(18) A which are the shortest RuÃ
/
RuÃ
/
N bond lengths of the complexes 1, 4, 5, [RuCl2(DM-
AP)2], [11] and [Cp*Ru(DMAP)Cl] [31]. The hydrogens
allow the nitrogens to move closer to the metal as
compared with bulkier substituents such as methyl or
benzyl groups. The bite angle of the chelate rings is
approximately 838, smaller than for the other complexes.
The closer approach of the N atoms in 3 compared with
the other bis-aminophosphine complexes discussed here
[9,29,30].
[(h6-C6H6)Ru(Me2NCH2CH2PPh2-N,P)Cl][PF6] (5)
(Fig. 7) has a piano-stool type geometry typical of
Ru(II) arene complexes, with an h6-arene ring forming
the seat and the three ligands forming the legs of the
˚
P bond length in 5 (2.3122(11) A) is
stool The RuÃ
/
˚
similar to those of [Cp*Ru(DMAP)Cl] (2.289(1) A [31]).
˚
N bond length for 5, 2.212(3) A, is
However the RuÃ
/
˚
can be seen from the NÃ
compared with 93.1(4)8 for 1 and 94.6(3)8 for trans, cis-
[RuCl2(DMAP)2] [11]. The RuÃCl bond lengths
(2.4053(6) and 2.4192(6) A) are typical for octahedral
Ru(II) complexes. The ClÃRuÃCl bond angle of
/
RuÃN angle of 90.73(7)8,
/
significantly shorter than the value of 2.260(2) A for the
˚
Cp* complex, but longer than the values of 2.11ꢀ2.13 A
/
/
found for [(h6-arene)Ru(en)Cl][PF6] where arene is
˚
benzene or a substituted benzene, and en is ethylene-
˚
Cl bond (2.3986(11) A) is shorter
/
/
diamine [40]. The RuÃ
/
˚
164.29(2)8 is very low, highly distorted from linearity.
The phenyl rings of adjacent PPh2 groups are almost
perpendicular, allowing them to pack closely together.
In complex 4 (Fig. 6), the aminophosphine oxide
ligand forms a six-membered chelate ring, binding to Ru
through O and N in cis equatorial positions. The two
chloride ligands are trans in the axial positions and the
two DMSO ligands fill the other two equatorial sites.
Both DMSO molecules are coordinated through sulfur,
and one is trans to N and the other is trans to O. Sulfur
is a mildly p-accepting ligand, and it is reasonable to
than for [Cp*Ru(DMAP)Cl] (2.441(1) A), but within the
range for RuÃCl bonds of half-sandwich complexes of
/
Ru(II) [40,41]. The PÃ
/
RuÃN angle of 81.68(9)8 is close
/
to those found for [Cp*Ru(DMAP)Cl] and [(h6-
C6H6)Ru(en)Cl][PF6] [31,40].
4. Conclusions
We have previously reported that square-planar Pt(II)
aminophosphine complexes can exist in ring-closed

558
R. Morris et al. / Inorganica Chimica Acta 339 (2002) 551ꢀ559
/
(chelated) and ring-opened (with the pendant arm
coordinated through phosphorus only) forms in aqu-
eous solution, and that the extent of ring opening can be
controlled by the substituents on N, the size of the
chelate ring, the pH and chloride concentration [5,6,42].
Some of these complexes were also cytotoxic to cancer
cells. In this work we have investigated the structures
and solution properties of Ru(II) analogues, since there
is also significant current interest in the anticancer
activity of ruthenium complexes [25].
References
[1] (a) M. Basset, D.L. Davies, J. Neild, L.J.S. Prouse, D.R. Russell,
Polyhedron 10 (1991) 501;
(b) K. von Issleib, A. Kipke, V. Hanfeld, Z. Anorg. Allg. Chem.
444 (1978) 5;
(c) K. von Issleib, A. Kipke, Z. Anorg. Allg. Chem. 464 (1980)
176;
(d) G.K. Anderson, R. Kumar, Inorg. Chem. 23 (1984) 4064;
(e) M.M.T. Khan, E.R. Rao, Polyhedron 6 (1987) 1727;
(f) M.M.T. Khan, V.V.S. Reddy, H.C. Bajat, Inorg. Chim. Acta
130 (1987) 163;
The X-ray crystal structures of the Ru(II) aminopho-
sphine complexes trans, cis-[RuCl2(H(Bz)NCH2-
CH2PPh2-N,P)2] (1) and trans, cis-[RuCl2(H2NCH2-
CH2PPh2-N,P)2] (3) showed that analogues of the Pt(II)
complexes can be prepared, but now as octahedral
complexes containing additional axial chloride ligands
as expected for Ru(II). Steric crowding causes a marked
(g) V.V.S. Reddy, S. Vijay, J. Mol. Catal. 45 (1988) 73.
[2] S. Chatterjee, D.C.R. Hockless, G. Salem, P. Waring, J. Chem.
Soc., Dalton Trans. (1997) 3889.
[3] J. Reedijk, J. Chem. Soc., Chem. Commun. (1996) 801.
[4] S.J. Berners-Price, R.J. Bowen, M.J. McKeage, P. Galettis, L.
Ding, B.C. Baguley, W. Brouwer, J. Inorg. Biochem. 67 (1997)
154.
[5] A. Habtemariam, B. Watchman, B.S. Potter, R. Palmer, S.
Parsons, A. Parkin, P.J. Sadler, J. Chem. Soc., Dalton Trans.
(2001) 1306.
distortion of the ClÃ
/
RuÃ
/
Cl axis from linearity. Com-
plex 3 underwent facile chelate ring-opening in DMSO
and CH3CN solutions which might be promising for
anticancer activity involving attack on DNA bases.
However, in general we obtained only low yields of
these complexes and, unlike the Pt(II) species which are
charged, the aqueous solubility of the neutral Ru(II)
complexes was very low. This problem will need to be
addressed if the biological properties of these complexes
are to be further explored.
[6] A. Habtemariam, P.J. Sadler, J. Chem. Soc., Chem. Commun.
(1996) 1785.
[7] C. Muddalige, S.J. Rettig, W.R. Cullen, R.B. James, J. Chem.
Soc., Chem. Commun. (1993) 830.
[8] M.M.T. Khan, A.P. Reddy, Polyhedron 6 (1987) 2009.
[9] J.-X. Gao, H.-L. Wan, W.-K. Wong, M.-C. Tse, W.-T. Wong,
Polyhedron 15 (1996) 1241.
[10] L. Costella, A. del Zotta, A. Mezzetti, E. Zangrando, P. Rigo, J.
Chem. Soc., Dalton Trans. (1993) 3001.
The aminophosphine complex [(h6-C6H6)Ru(Me2N-
CH2CH2PPh2-N,P)Cl][PF6] (5) is an analogue of active
ethylenediamine complexes which bind strongly to
DNA [40], and complexes in this arene Ru(II) series
may be useful anticancer agents.
[11] Z. Guo, A. Habtemariam, P.J. Sadler, B.R. James, Inorg. Chim.
Acta 273 (1998) 1.
[12] E. Alessio, G. Mestroni, G. Nardin, W.M. Attia, M. Calligaris, G.
Sava, S. Zorzet, Inorg. Chem. 27 (1988) 4099.
[13] E.W. Abel, J. Chem. Soc. (1959) 3178.
[14] F.B. McCormick, D.D. Cox, W.B. Gleason, Organometallics 12
(1993) 610.
[15] P.T. Beurskens, G. Beurskens, W.P. Bosman, R. de Gelder, S.
Garcia-Granda, R.O. Gould, R. Israel, J.M.M. Smits, Crystal-
lography Laboratory, University of Nijmegen, Nijmegen, The
Netherlands, 1996.
5. Supplementary material
[16] G.M. Sheldrick, SHELXTL, version 5, Bruker Analytical X Ray,
Madison, WI, 1995.
CCDC deposition numbers: complex 1, 178906;
complex 3, 178903, complex 4 178905; complex 5,
178904. Copies of this information may be obtained
free of charge from The Director, CCDC, 12 Union
[17] G.M. Sheldrick, SADABS: Area-Detector Absorption Correction,
Siemens Industrial Automation, Madison, WI, 1996.
[18] D. Cromer, R. von Dreele, in: L.J. Farrugia (Ed.), FPRIME
Incorporated in the WINGX Suite of Programs, J. Appl. Crystal-
logr. 32 (1999) 837.
Road, Cambridge, CB2 1EZ, UK (fax: ꢁ44-1223-336-
/
033; e-mail: deposit@ccdc.cam.ac.uk or http://
www.ccdc.cam.ac.uk).
[19] (a) E.B. Boyer, P.A. Harding, S.D. Robinson, J. Chem. Soc.,
Dalton Trans. (1986) 1771;
(b) G. Sia, A.L. Rheingold, B.J. Haggerty, D.W. Meek, Inorg.
Chem. 31 (1992) 900;
(c) A. Albinati, Q. Jiang, H. Ruegger, L.M. Venanzi, Inorg.
Chem. 32 (1993) 4940.
Acknowledgements
[20] J.Y. Shen, C. Slugovc, P. Wiede, K. Mereiter, R. Schmid, K.
Kirchner, Inorg. Chim. Acta 268 (1998) 69.
We thank the EPSRC, BBSRC, Wolfson Foundation
and Royal Society for their support for this work, Dr.
Lesley Yellowlees for assistance with electrochemistry,
Dr. John Parkinson with NMR spectroscopy, members
of EC COST D8 and D20 for stimulating discussions,
CCLRC and Dr. Simon Teat (Daresbury Laboratory)
for Synchrotron beam time, and Johnson Matthey for
the loan of some of the ruthenium.
[21] A. Dovletoglou, S.A. Adeyemi, M.H. Lynn, D.J. Hodgson, T.J.
Meyer, J. Am. Chem. Soc. 112 (1990) 8989.
[22] B.A. Mayer, B.K. Sipe, T.J. Mayer, Inorg. Chem. 20 (1981) 1475.
[23] (a) M.M. Taqui-Khan, V. Vijay, V.V.S. Reddy, Inorg. Chem. 25
(1986) 208;
(b) J.C. Briggs, C.A. McAuliffe, G. Dyer, J. Chem. Soc., Dalton
Trans. (1984) 423;
(c) D.F. Gill, B.E. Mann, B.L. Shaw, J. Chem. Soc., Dalton

R. Morris et al. / Inorganica Chimica Acta 339 (2002) 551ꢀ
/559
559
Trans. (1973) 311;
(d) G.K. Anderson, R. Turner, Inorg. Chem. 24 (1984) 4064.
[24] P.E. Garrou, Chem. Rev. 81 (1981) 229.
[33] S.R. Hall, B.W. Shelton, A.H. White, Aust. J. Chem. 36 (1983)
267.
[34] J.W. Faller, P.B. Patel, M.A. Albrizzio, M. Curtis, Organome-
tallics 18 (1999) 3096.
[25] M.J. Clarke, F. Zhu, D.R. Frasca, Chem. Rev. 99 (1999) 2511.
[26] D.P. Shoemaker, C.W. Garland, Experiments in Physical Chem-
istry, McGraw-Hill, New York, 1962, p. 222.
[35] M.C. Barral, R. Jime´nez, J.L. Priego, E.C. Royer, M.J. Saucedo,
F.A. Urbanos, Polyhedron 14 (1995) 2419.
[27] F.A. Cotton, G. Wilkinson, Advanced Inorganic Chemistry, 4th
ed., Wiley, New York, 1980, p. 1200.
[36] F.C. March, G. Ferguson, Can. J. Chem. 49 (1971) 3590.
[37] (a) M. Calligaris, Croat. Chem. Acta 72 (1999) 147;
(b) M. Calligaris, O. Carugo, Coord. Chem. Rev. 153 (1996) 83.
[38] R.S. McMillan, A. Mercer, B.R. James, J. Trotter, J. Chem. Soc.,
Dalton Trans. (1995) 1006.
[28] (a) A.G. Orpen, L. Brammer, F.H. Allen, O. Kennard, D.G.
Watson, R. Taylor, J. Chem. Soc., Dalton Trans. (1989) 83;
(b) A.M. Joshi, I.S. Thorburn, S.J. Rettig, B.R. James, Inorg.
Chim. Acta 283 (1992) 198.
[39] A.R. Davies, F.W.B. Einstein, N.P. Farrell, B.R. James, R.S.
McMillan, Inorg. Chem. 17 (1978) 1965.
[29] D. Drommi, F. Nicolo, C.G. Arena, G. Bruno, F. Faraone, R.
Gobetto, Inorg. Chim. Acta 221 (1994) 109.
[40] R.E. Morris, R.E. Aird, P.del S. Murdoch, H. Chen, J.
Cummings, N.D. Hughes, S. Parsons, A. Parkin, G. Boyd, D.I.
Jodrell, P.J. Sadler, J. Med. Chem. 44 (2001) 3616.
[41] I. los de Rios, M.J. Tenorio, J. Padilla, M.L. Puerta, P. Valerga, J.
Chem. Soc., Dalton Trans. (1996) 377.
[30] L. Costella, A. Del Zotto, A. Mezzetti, E. Zangrando, P. Rigo, J.
Chem. Soc., Dalton Trans. (1993) 3001.
[31] K. Mauthner, C. Slugovc, K. Mereiter, R. Schmid, K. Kirchner,
Organometallics 16 (1997) 1956.
[32] M.M. Olmstead, Y. Maisonnat, J.P. Farr, A.L. Balch, Inorg.
Chem. (1981) 4061.
[42] A. Habtemariam, J.A. Parkinson, N. Margiotta, T.W. Hambley,
S. Parsons, P.J. Sadler, J. Chem. Soc., Dalton Trans. (2001) 362.