Water-soluble hydroxyalkylated phosphines: Examples of their differing behaviour toward ruthenium and rhodium

Article abstract of DOI:10.1039/b411701h

The reaction of P(CH₂OH)₃ (I) and P(C ₆H₅)(CH₂OH)₂ (II) with RuCl ₃ in methanol eliminates two equivalents of formaldehyde to yield the mixed tertiary and secondary phosphine com

Full text of DOI:10.1039/b411701h

View Article Online / Journal Homepage / Table of Contents for this issue
F U L L P A P E R
Water-soluble hydroxyalkylated phosphines: examples of their
differing behaviour toward ruthenium and rhodium
Lee J. Higham,*a Michael K. Whittleseyb and Paul T. Woodc
a Chemistry Department, The Centre for Synthesis and Chemical Biology, Conway Institute of
Biomolecular and Biomedical Sciences, University College Dublin, Belfield, Dublin 4, Ireland
b Department of Chemistry, University of Bath, Claverton Down, Bath, UK BA2 7AY
c University Chemical Laboratory, Lensfield Road, Cambridge, UK CB2 1EW
Received 30th July 2004, Accepted 18th October 2004
First published as an Advance Article on the web 16th November 2004
The reaction of P(CH2OH)3 (I) and P(C6H5)(CH2OH)2 (II) with RuCl3 in methanol eliminates two equivalents of
formaldehyde to yield the mixed tertiary and secondary phosphine complexes all-trans-[RuCl2(P(CH2OH)3)2
(P(CH2OH)2H)2] (1) and [RuCl2(P(C6H5)(CH2OH)2)2(P(C6H5)(CH2OH)H)2] (2), respectively. There is a high degree
of hydrogen-bonding interactions between the hydroxymethyl groups in 1 and 2, although the phenyl groups of the
latter reduce the extent of the network compared to 1. The generation of these mixed secondary and tertiary
phosphine complexes is unprecedented. Under the same reaction conditions, the tris(hydroxypropyl)phosphine III
formed no ruthenium complex. The reaction of P(CH2OH)3, P(C6H5)(CH2OH)2 and P{(CH2)3OH}3 with
[RhCl(1,5-cod)]2 in an aqueous/dichloromethane biphasic medium yielded [RhH2(P(CH2OH)3)4]+ (3),
[RhH2(P(C6H5)(CH2OH)2)4]+ (4) and [Rh(P(C6H5)(CH2OH)2)4]+ (5) and [Rh(P{(CH2)3OH}3)4]+ (6), respectively.
Treating 5 with dihydrogen rapidly gave 4. The hydroxypropyl compound 6 formed the corresponding dihydride
much more slowly in aqueous solution, although [RhH2(P{(CH2)3OH}3)4]+ (7) was readily formed by reaction with
dihydrogen. Two separate reaction pathways are therefore involved; for P(CH2OH)3 and to a lesser extent
P(C6H5)(CH2OH)2, the hydride source in the product is likely to be the aqueous solvent or the hydroxyl protons,
whilst for P{(CH2)3OH}3 an oxidative addition of H2 is favoured. The protic nature of 3 and 4 was illustrated by the
H–D exchange observed in d2-water. Dihydrides 3 and 4 reacted with carbon monoxide to yield the dicarbonyl
cations [Rh(CO)2(P(CH2OH)3)3]+ (8) and [Rh(CO)2(P(C6H5)(CH2OH)2)3]+ (9). The analogous experiment with
[RhH2(P{(CH2)3OH}3)4]+ resulted in phosphine exchange, although our experimental evidence points to the
possibility of more than one fluxional process in solution.
be expected to alter the solubility/hydrogen bonding properties
Introduction
of any complex. The compound [Mo(CO)5(P(C6H5)(CH2OH)2)]
Despite continued interest in the development of water-
is a rare report of its coordination chemistry.9 We also in-
soluble phosphines for use in biphasic catalysis and medicinal
cluded the commercially available phosphine P{(CH2)3OH}3
III in our study, to determine the effects of extending the
number of CH2 spacer groups between the alcohol and the
phosphorus. In addition, aside from the use of this ligand in
metallodendrimers,10 telomerisations11 and some hydrogenation
reactions5,7,12 its coordination chemistry is seldom encountered
in the literature.
The P(C6H5)(CH2OH)2 reaction was carried out in an analo-
gous fashion to that of I (Scheme 1), with the solution changing
from dark red to a green colour and the subsequent precipitation
chemistry,1 there is still much to be done on studying the funda-
mental reactivity of the hydroxyalkyl derivatives. Rhenium and
technetium complexes of this ligand family have been shown
to possess desirable attributes for use in radiopharmaceuticals.2
RhIII,3 PdII, PtII and ReV complexes have been shown to be
1c
kinetically stabilised by such groups, and both rhodium4 and
ruthenium5 compounds are active hydroformylation and hydro-
genation catalysts, respectively. We have previously commu-
nicated6 that the reaction of tris(hydroxymethyl)phosphine and
ruthenium trichloride forms the all-trans complex [RuCl2-
(P(CH2OH)3)2(P(CH2OH)2H)2] 1 with the highly unusual gener-
ation of two secondary phosphine groups on the metal. Ikariya
et al. have since shown this compound to be a superior catalyst
than P(CH3)3 or water-soluble aryl phosphine complexes for
the hydrogenation of supercritical carbon dioxide.7 In light of
the growing number of applications of these ligands, this novel
reactivity prompted us to extend our study of water-soluble
monodentate phosphines to both ruthenium and rhodium
precursors. We report here our findings that the behaviour of
these ligands differs widely between the two metals.
Results and discussion
Ruthenium reactions
The striking formation of all-trans-[RuCl2(P(CH2OH)3)2
(P(CH2OH)2H)2] (1) prompted us to prepare the ligand
P(C6H5)(CH2OH)2 II for its simple synthesis,8 the presence of
two a-carbons bearing hydroxyl groups (to confer water solu-
bility) and the fact that the hydrophobic phenyl moiety would
Scheme 1 Formation of the all-trans-complexes 1 and 2.
4 2 0 2
D a l t o n T r a n s . , 2 0 0 4 , 4 2 0 2 – 4 2 0 8
T h i s j o u r n a l i s
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4
©

View Article Online
Table 1 A comparison of relevant bond lengths and angles for 2
of small yellow crystals of 2 over 24 h. These were suitable for
an X-ray crystallographic study (Fig. 1).
Ru–P(1)
Ru–P(2)
Ru–Cl(1)
2.352(2)
2.383(2)
2.443(2)
1.55(8)
1.853(8)
1.871(9)
1.390(11)
1.834(8)
1.849(8)
1.863(9)
1.426(11)
1.395(12)
1.34(2)
P(1)–Ru–P(1A)
P(2)–Ru–P(2A)
Cl(1)–Ru–Cl(1A)
P(1)–Ru–P(2A)
P(1)–Ru–P(2)
P(1)–Ru–Cl(1A)
P(1)–Ru–Cl(1)
P(2)–Ru–Cl(1)
P(2)–Ru–Cl(1A)
C(1)–P(1)–C(7)
C(1)–P(1)–Ru
180
180
180
P(1)–H(1)
P(1)–C(1)
P(1)–C(7)
C(7)–O(1)
P(2)–C(11)
P(2)–C(18)
P(2)–C(17)
C(17)–O(2)
C(18)–O(3)
C(18)–O(13)
91.45(7)
88.55(7)
87.09(7)
92.91(7)
94.26(7)
85.74(7)
104.3(4)
124.1(2)
114.1(3)
102.5(4)
98.9(4)
100.3(4)
118.8(3)
118.8(3)
114.2(3)
C(7)–P(1)–Ru
C(11)–P(2)–C(18)
C(11)–P(2)–C(17)
C(18)–P(2)–C(17)
C(11)–P(2)–Ru
C(18)–P(2)–Ru
C(17)–P(2)–Ru
a
˚
O · · · O distances (A)
O(1) · · · O(2B)
O(1) · · · O(3B)
O(2) · · · O(1B)
O(2) · · · O(13B)
O(3) · · · O(1D)
O(13) · · · O(2D)
2.698(13)
2.775(14)
2.698(13)
2.67(2)
2.775(14)
2.67(2)
a Selected possible H-bonds based on O · · · O distances (the concerned
OH hydrogen atoms could not be localised).
(2.323 and 2.331 A)13 and trans-[RuCl2(P(C6H5)2H)4] (2.3665(8)
˚
14
˚
and 2.3505(8) A).
The 31P NMR spectra of the reaction mixtures of both 1 and
=
2 share common features; the formation of O PR(CH2OH)2
and [PR(CH2OH)3]+ is observed in both cases (see later). For
1 the resonances are observed as an A2B2 pattern at d 9.7 ppm
2
2
(t, JPP = 37.5 Hz) and d 13.5 ppm (t, JPP = 37.5 Hz), with
the higher field triplet assigned to the secondary phosphine on
account of the large P–H coupling (1JPH = 253.0 Hz) in the 31P–
1H coupled spectrum. The IR spectrum (KBr) shows a band at
2374 cm−1 assigned to m(P–H).
Fig. 1 View of 2 showing the atom numbering scheme and the
possible hydrogen-bonding (dashed lines). The bold dashed lines indicate
disorder in one of the hydroxymethyl groups (see text). The inset displays
the [110] direction showing the hydrogen bonding (dashed line).
The formation of 1 and 2 is not instantaneous but occurs
slowly over time out of a complicated spectral region (reaction
time <1 h) which precluded identification of the initial ruthe-
nium phosphine species present. There is no NMR evidence
for uncoordinated P(CH2OH)2H or P(C6H5)(CH2OH)H, and
for 1 we see no scrambling of the phosphine ligands when
the solutions are kept in the dark at 0 ◦C (light appears to
promote isomerisation/scrambling). Only dimethyl sulfoxide
and dimethylformamide dissolved 2. Upon dissolution of 2
in d6-dimethyl sulfoxide, the 31P NMR spectrum was recorded
immediately and showed signals similar to 1; the 2JP–P coupling
is comparable (35.0 Hz compared to 37.5 Hz for 1) and the
presence of the secondary phosphine was confirmed by the
1JP–H coupling on the upfield resonance (263.9 Hz compared to
253.0 Hz for 1). Other complicated areas of the NMR spectrum
are probably attributable to the ‘non-innocent’ behaviour of the
solvent. The solvent reactivity may also explain why neither 1
nor 2 can be synthesised from the commonly used precursor
[RuCl2(DMSO)4], which in both cases yield a complex mixture
of products. Solutions of 2 in deuterated dimethylformamide
showed essentially the same behaviour.
The compound crystallises in a centrosymmetric space group
with the metal on an inversion centre which imposes an all-
trans geometry and also dictates that both isomers of the chiral
secondary phosphine ligand are present in each molecule. The
position of one of the oxygen atoms of the tertiary phosphine
is disordered over two sites. The site with one-third occupancy
(O(13)) is eclipsed with respect to the other alcohol function on
that ligand whilst that with two-thirds occupancy (O(3)) points
away. In common with the other oxygen atoms in the molecule
both of these positions have a number of intermolecular
˚
hydrogen bonds shorter than 2.8 A. These give rise to sheets that
lie in the ab plane but the presence of the hydrophobic phenyl
ring interactions between these sheets prevents the formation
of a three-dimensional hydrogen-bonded network, and helps to
explain the apparent insolubility of 2. There are similarities with
the P(CH2OH)3 complex 1; both the tertiary and secondary
phosphines are involved in intermolecular hydrogen bonding,
but with the P–H group playing no part in this.
Important bond lengths and angles for 2 are summarised in
Table 1. In the case of 2 the metal to phosphine bond lengths
are similar in both the tertiary and secondary cases, unlike 1
Grosselin et al.15 noted a number of pertinent points in
their reactions of ruthenium chloride with TPPTS (TPPTS =
tris(m-sulfophenyl)phosphine trisodium salt) when conducted
in a polar solvent (in this case water). They postulate that the
reduction of RuIII to RuII and phosphine oxidation occurs prior
to coordination (Scheme 2) and invoke a similar mechanism to
where there is a difference of about 0.1 A (for 1: secP–Ru 2.414(2)
˚
ter
˚
cf. P 2.318(2) A). The increased steric bulk of the ligand and
lowered basicity of II may be expected to increase the Ru–P
bond length with respect to 1, and this is observed when the
tertiary phosphine bond lengths are compared (for 2: terP–Ru
˚
2.383(2) A). However the reverse trend is found when comparing
the respective secondary phosphines (for 1: secP–Ru 2.414(2) cf.
sec
˚
P 2.352(2) A for 2). These differences may be attributable
to the extensive hydrogen bonding. The P(C6H5)(CH2OH)H
bond length in 2 is in close agreement with that found in the
analogous organo-soluble complexes trans-[RuCl2(P(CH3)2H)4]
Scheme 2 Grosselin’s proposed reduction of RuIII and concomitant
phosphine oxidation.
D a l t o n T r a n s . , 2 0 0 4 , 4 2 0 2 – 4 2 0 8
4 2 0 3

View Article Online
Scheme 3 Proposed formation for the secondary phosphine groups in 1 and 2.
that which accounts for the aqueous oxidation of TPPTS by
rhodium chloride.16
An alternate precursor was employed, and in a bipha-
sic reaction of P(CH2OH)3 and [RuCl2(P(C6H5)3)3] (wa-
ter/dichloromethane) ligand exchange took place to yield 1,
with less phosphine oxide and phosphonium salt present, but
other ruthenium–phosphine species evident by 31P NMR. How-
ever, when P{(CH2)3OH}3 was used in this manner no re-
action took place (31P NMR showed only uncoordinated
P{(CH2)3OH}3). The synthesis of [Ru4H4(CO)8(P(CH2OH)3)4]
A similar mechanism would explain many of the observations
made during the formation of 1 and 2, including the presence
=
of O PR(CH2OH)2 and the generation of HCl – necessary for
the formation of the observed [PR(CH2OH)3]+ (together with a
mole of eliminated formaldehyde from the secondary phosphine
forming step).
20
5
The synthesis of a bidentate phosphine–phosphinite complex
[Pt{(HOCH2)2PCH2OP(CH2OH)2}2]Cl2 prepared by Pringle
and the preparation5 of the catalytic precursors [(g -
C5H5)RuCl(CO)L] where L is I or III are evidence that both
ligands can coordinate to ruthenium in a ‘regular’ way depend-
ing on the other ligands present. As a comparison, we reacted
17
et al. requires the elimination of two units of formaldehyde,
and was accompanied by the formation of [P(CH2OH)4]Cl. The
elimination proceeds via nucleophilic attack by uncoordinated
P(CH2OH)3 on the a-carbon of a cyclic adduct intermediate,
which then eliminates as [P(CH2OH)4]+. Upon ring closure the
formation of the bidentate ligand is complete. We propose that
the formation of the secondary phosphine in 1 and 2 occurs by
attack of free PR(CH2OH)2 on the a-carbon of a coordinated
PR(CH2OH)2 ligand. The formation of [PR(CH2OH)3]+ can
then be accounted for by elimination of formaldehyde and the
presence of HCl, generated as a by-product of the metal reduc-
tion by phosphine (Scheme 3). This supports our findings that a
ratio of 6 : 1 phosphine : ruthenium is required, and that additio-
nal phosphine (up to 10 equivalents) has no effect on the yield.
These results clearly demonstrate that the hydroxymethyl
group is significantly activated by coordination to the ruthenium
centre even when only two such groups are present on the
ligand. It is unclear at present why an all-trans arrangement
prevails for 1 and 2 and we found no evidence of further
5
2 equivalents of P(CH2OH)3 and [(g -C5H5)RuCl(P(C6H5)3)2]
in a biphasic reaction but found no exchange took place. In
an attempt to generate ruthenium hydride complexes, biphasic
ligand exchange reactions were attempted using P(CH2OH)3
21
22
and RuH2(PPh3)4
and RuH(CH3CO2)(PPh3)3,
but no
compounds could be isolated. No hydride was generated either
by reaction of 1 with sodium borohydride.
Rhodium reactions
Chatt et al.23 reported the synthesis of mer-[RhCl3
(P(CH2OH)3)3], by reaction of P(CH2OH)3 and RhCl3 in reflux-
ing alcohol, and trans-[RhCl(CO)(P(CH2OH)3)2] was prepared
by a ligand exchange reaction with [RhCl(CO)2]2 in methanol. Of
significance is the more recent report describing the formation
of fac-[RhCl3(P(CH2OH)3)3] from the reaction of RhCl3 and
I in aqueous solution.3 To the best of our knowledge, no
rhodium complexes of P(C6H5)(CH2OH)2 and only one of
P{(CH2)3OH}3 have been reported.10 We have found that
reactions of I with [RhCl(1,5-cod)]2 in a biphasic reaction
(water/dichloromethane) led to the instantaneous formation
of the dihydride cis-[RhH2(P(CH2OH)3)4]+ 3; water-soluble
phosphine hydrides are uncommon in the literature but they
are increasingly attracting attention.24–26
18
formaldehyde elimination when sodium hydrogen sulfite or
19
sodium metabisulfite were added to solutions of 1; instead
they only promoted phosphine oxidation. The addition of
formaldehyde solution to 1 and 2 did not lead to any detectable
insertion into the P–H bonds, whilst ligand solutions give
hemiacetal mixtures with the same reagent i.e., a reaction which
is in effect the reversal of the purification step of P(CH2OH)3.
It is plausible that the formation of 1 and 2 occurs by a
secondary process of ligand rearrangement after formaldehyde
elimination, driven perhaps by favourable hydrogen bonding.
This is supported by the stability of 1 to ligand dissocia-
tion/exchange. The reactions were repeated using aqueous
methanol (75 : 25 and 50 : 50 methanol : water) and we
found increased amounts of phosphonium salt and oxide with
increased water content. When water was employed as sole
solvent these were the only products observed in the 31P NMR
spectrum, together with colloidal ruthenium. The proportion
of phosphine oxide present is also substantially increased if the
methanolic reaction is carried out under reflux.
The common factor to both phosphines is the hydroxymethyl
groups and thus we carried out the analogous reaction with
P{(CH2)3OH}3 III to see if the functionality at the end of
the propyl chain affects the reaction in any way. Surprisingly
there was no complex formation (given that III has been used
as a ligand for ruthenium based catalysis – see earlier); only
a precipitate of what appeared to be ruthenium metal and a
green solution containing phosphine oxide was observed. This
contrasts to 1 and 2, where the immediate presence of substantial
amounts of [PR(CH2OH)3]+ was observed.
The reaction is easily monitored by the transfer of coloration
from the dichloromethane layer to the aqueous phase (the
1
31P NMR spectrum of 3: d 33.7 ppm, dt, JRh–P = 86.3 Hz,
1
2
2
d 21.0 ppm, dt, JRh–P = 98.8 Hz, JP–P = 22.0 Hz). A JP–H
coupling constant of 120.3 Hz is observed for the phosphorus
signal at d 21.0 ppm, indicating these phosphines are trans
to the hydride ligands which appear at d −11.10 ppm in
the 1H NMR spectrum. Broadband 31P decoupling of the
spectrum shows a doublet splitting of 1JH–Rh = 12.1 Hz. Selective
decoupling of the phosphines trans to the hydrides produced a
broadened pseudo-quartet, and signifies that the magnitude of
2JP
must be similar in magnitude to 1JRh–H – approximately
12 aHxiazl–.H Selective decoupling of the axial phosphines produces
a second-order spectrum. There are strong similarities with
cis-[RhH2(P(CH3)3)4]+,27 where the authors calculated a best-fit
spectrum and established that the hydride multiplet consisted of
a seven-spin AAꢀMXXꢀY2 system, rather than the more simple
A2X2MY2 with the equatorial cis-hydrides and equatorial phos-
phines assigned as AAꢀ and XXꢀ, respectively. This is because
of the chemical equivalence but magnetic non-equivalence of
these groups in the equatorial plane, where both the larger trans-
and smaller cis-2JP–H coupling must be considered. The axial
4 2 0 4
D a l t o n T r a n s . , 2 0 0 4 , 4 2 0 2 – 4 2 0 8

View Article Online
phosphines, i.e. the phosphines trans to one another are denoted
as Y2. Accordingly a AAꢀMXXꢀY2 system is also assigned for 3.
During this time Komiya et al. noted the unexpected for-
mation of the same compound in tetrahydrofuran (along with
other uncharacterised species) and speculated that traces of
adventitious water may be the source of the hydride.4 We find
that the reaction of [RhI(1,5-cod)]2, [RhCl(1,5-cod)(P(C6H5)3)],
and [Rh(1,5-cod)(P(C6H5)3)2]B(C6H5)4 with excess P(CH2OH)3
(ratio = 1 Rh : 4.5 P) in dichloromethane/water all gave 3 (Fig. 2)
as the main Rh containing product. The complexes [RhI(1,5-
cod)]2 and [RhCl(1,5-cod)(P(C6H5)3)] gave the cleanest routes
with 3 as the only Rh complex. In our hands the reaction of
dihydride 4. This correlates with the report by Osborn et al.
who synthesised [Rh(P(C6H5)(CH3)2)4]+ (1JRh–P = 132 Hz) and
converted it into the dihydride, an organo-soluble analogue of
4, in a similar fashion.27
Further experiments showed 4 can also be formed in the
absence of water, using alcoholic solvents, excess ligand and
dihydrogen, although the stability of both 5 and 4 in such
solvents appeared to be greatly reduced to that found in aqueous
solution.
Analogous to the reaction with P(CH2OH)3, the dideuteride
[RhD2(P(C6H5)(CH2OH)2)4]+ was prepared and the spectra
2
showed essentially the same features; notably JP–D coupling
RhCl3 with P(CH2OH)3 in aqueous solution led to the formation
which again broadened the upfield equatorial phosphine signal
in the 31P NMR spectrum, and no hydride signal in the proton
spectrum. Addition of water regenerates the dihydride 4, as
discussed for 3. The doublet species [Rh(P(C6H5)(CH2OH)2)4]+
(5) was present throughout these experiments, indicating the
conversion of 5 to [RhH2(P(C6H5)(CH2OH)2)4]+ was slower than
the D–H exchange process and serves to demonstrate the protic
character of 4.
+
=
of O P(CH2OH)3, [P(CH2OH)4] (as seen for 1 and 2) and
two independent doublets (d 37.2 and 15.2 ppm) as part of
a mixture of rhodium-phosphine species (one of which may
be the fac-[RhCl3(P(CH2OH)3)3] compound communicated by
Katti3). Complex 3 was not formed under these conditions,
but when RhCl3, tetrakis(hydroxymethyl)phosphonium chloride
and triethylamine were combined in aqueous solution it was
synthesised relatively cleanly (by 31P NMR spectroscopy).
Next we reacted P{(CH2)3OH}3 with [RhCl(1,5-cod)]2 and
found that after 5 min stirring, complete colour transfer to the
aqueous layer had taken place.
There was however, no evidence of two doublets of triplets
even after a 24 h period, only a doublet species 6 which
showed no JP–H coupling at d 4.4 ppm (136.3 Hz). This echoes
the formation of 5, and the highly hydrophilic compound is
likely to be the analogous RhI cation [Rh(P{(CH2)3OH}3)4]+.
The notable difference between 5 and 6 however, is the time
required for the latter to show even partial conversion to the
dihydride (several months cf. 60 min). Further support for this
assignment came from treating solutions of 6 with dihydrogen;
a colour change was observed from orange to light yellow over a
15 min period and the familiar AA’MXX’Y2 spin system became
evident as 6 was quantitatively converted to the dihydride
Fig. 2 The rhodium dihydride and dicarbonyl compounds.
The reaction of [RhCl(1,5-cod)]2 with P(CH2OH)3 in d-
chloroform/d2-water gave a yellow aqueous layer after 30 min. A
1
[RhH2(P{(CH2)3OH}3)4]+ 7 (dP 13.7 ppm, JRh–P = 99.6 Hz,
1
1
31P{ H} NMR spectrum of the d2-water layer was recorded and
2JP–P = 22.3 Hz; −0.4 ppm, JRh–P = 87.4 Hz,2JP–P = 22.3 Hz).
showed that the upfield resonance observed for 3 i.e. that of the
phosphorus trans to the hydride ligands had changed to a broad
signal, resulting from trans JP–D coupling. No hydride signal
was seen in the proton spectrum. Unfortunately no M–D band
was identified in the IR spectrum as the region (∼1410 cm−1)
was obscured, although the M–H band at m = 1966 cm−1
was no longer evident. Addition of water to these solutions
sees the reappearance of the dihydride 3. We carried out some
selective decoupling experiments on 3 and found that when the
water resonance was irradiated and a 31P–1H coupled spectrum
The upfield resonance broadens in the 31P–1H coupled NMR,
but the signal for the hydrides is very similar to that of 3 and 4
and appears at d −11.85 ppm in the proton spectrum.
Solutions of 7 are stable in d6-dimethyl sulfoxide and d2-water,
and show no sign of H–D exchange. The former behaviour
mirrors that of 4, but contrasts with that observed with the
P(CH2OH)3 complex 3, where dimethyl sulfoxide promotes
decomposition via phosphine dissociation. The absence of
exchange suggests the protic nature of 7 is considerably less than
either [RhH2(P(CH2OH)3)4]+ or [RhH2(P(C6H5)(CH2OH)2)4]+,
an unsurprising observation given the lack of reactivity of
[Rh(P{(CH2)3OH}3)4]+ in aqueous solution.
2
recorded, JP–H coupling was observed, which indicates that
exchange must be slow on the NMR timescale.
The reaction with [RhCl(1,5-cod)]2 was repeated with
P(C6H5)(CH2OH)2 and the 31P NMR spectrum after 10 min
showed two doublets of triplets (d 28.5 ppm, 1JRh–P = 99.0 Hz,
Of the three dihydride complexes, the phenyl derivative 4 is the
least soluble on account of the lower number of hydroxyl groups
present. Solutions of 3, 4 and 7 show similar stability at room
temperature, of up to two weeks, when stored under an inert
atmosphere, and none of the complexes demonstrated reversible
or irreversible loss of H2 under reduced pressure. Exposure
to oxygen and/or light promoted decomposition, whilst the
alcoholic solutions show a marked decrease in stability, leading
to decomposition to the corresponding oxides.
1
2JP–P = 21.1 Hz; 15.8 ppm, JRh–P = 87.0 Hz,2JP–P = 21.1 Hz),
consistent with the formation of the analogous complex cis-
[RhH2(P(C6H5)(CH2OH)2)4]+ 4. A measure of the splitting on
the upfield resonance gives an approximate value of ∼113 Hz,
2
indicative of trans JP–H coupling for 4 (compared to 120.7 Hz
for 3). It was not possible to record a more accurate value owing
to the broad nature of the signal. The hydride chemical shift (d
−11.20 ppm) of the multiplet is similar to that observed for the
P(CH2OH)3 analogue 3, and also to that reported for the closely
related [RhH2(P(C6H5)(CH3)2)4]+.27
The addition of excess phosphine to aqueous solutions of the
1
dihydrides showed no effect on the 31P{ H} NMR spectrum
for 3, suggesting there is no P(CH2OH)3 exchange. With
P(C6H5)(CH2OH)2 however, there was a broadening of the res-
onances of 4 and uncoordinated phosphine, although the 1JRh–P
and2JP–P couplings were retained. In the case of P{(CH2)3OH}3,
the addition of excess phosphine to aqueous solutions of 7
resulted in a broad phosphine signal and the re-emergence of
the doublet resonance assigned to 6 i.e. [Rh(P{(CH2)3OH}3)4]+.
This appears to suggest that the phosphine promotes a reduc-
tive elimination reaction not observed with P(CH2OH)3 and
P(C6H5)(CH2OH)2.
Also evident in the 31P NMR spectrum after 10 min re-
action time was a large doublet resonance at d −17.5 ppm
(1JRh–P = 134.7 Hz), which we attribute to the four-coordinate
rhodium(I) complex [Rh(P(C6H5)(CH2OH)2)4]+ 5 on account
1
of the magnitude of JRh–P and the absence of significant P–
H coupling. Furthermore 4 is formed gradually over 60 min
with a concomitant decrease in the signal for 5. Treating 5
with dihydrogen resulted in the immediate conversion to the
D a l t o n T r a n s . , 2 0 0 4 , 4 2 0 2 – 4 2 0 8
4 2 0 5

View Article Online
The formation of 3 proceeds without any evidence for the
tetrakisphosphine complex [Rh(P(CH2OH)3)4]+. NMR analysis
of the reaction analysed immediately showed only signals for 3,
phosphine oxide and a large signal for P(CH2OH)3. Even with
low phosphine ratios of 1 : 1 [RhCl(1,5-cod)]2 : P(CH2OH)3, the
only metal complex observed in the 31P NMR spectrum of the
aqueous layer was 3.
showing the expected isotopic shift (1976 cm−1). On this basis
9 is assigned as trans-[Rh(13CO)2(P(C6H5)(CH2OH)2)3]+. No
evidence for phosphine dissociation was observed by 31P NMR,
suggesting the fluxionality of 9 is not due to phosphine ligand
exchange.
The solution behaviour of 9 resembles is some ways that
of [Rh(CO)2(P(C6H5)3)3]+, which exhibits significant dynamic
◦
Conversely, interconversion of 5 into the dihydride 4 can be
monitored over time, and unlike the reaction with P(CH2OH)3,
reactant ratios of P : Rh lower than 4 : 1 for P(C6H5)(CH2OH)2
often show only phosphine oxide in the aqueous layer. This
slower rate of formation may be due to the electron-withdrawing
properties of the phenyl substituent. The rate is dramatically
increased however upon the addition of H2 to the aqueous
solution of 5, in an oxidative addition to a d8 square-planar
complex. Such reactions are common, but there are notable
exceptions; for instance [Rh(P(C6H5)2CH3)4]+ does not react
with H2,27 and [Rh(P(CH3)3)4]+ is more reactive but tends to
dissociate P(CH3)3 in the absence of excess phosphine.23
The tetrakis-cation [Rh(P{(CH2)3OH}3)4]+ 6 reacted very
behaviour in solution; only on cooling solutions to ≤ −60 C
does the broad singlet observed for the carbonyl ligands become
a well defined doublet of quartets (d 198.6 ppm, 1JC–Rh = 50 Hz,
2JC–P = 14 Hz).28 However, in contrast to 8 and 9, there is consid-
erable phosphine dissociation at ro◦om temperature, with 1JRh–P
coupling only resolved below −40 C (113.0 Hz), and two CO
bands in the room-temperature solution IR spectrum, suggestive
of a cis arrangement of these ligands. These results demonstrate
that the P(CH2OH)3 and P(C6H5)(CH2OH)2 complexes 8 and 9
differ from many [Rh(CO)2(PR3)3]+ complexes reported in the
literature which are commonly unstable and readily dissociate
PR3 and/or CO.
The reaction with [RhH2(P{(CH2)3OH}3)4]+ 7 and 13CO saw
slowly toward water to give
7
after several months.
the emergence of a doublet species (10) at d 11.0 ppm (1JRh–P
=
Hence the formation of the dihydrides [RhH2(P(CH2OH)3)4]+
and [RhH2(P{(CH2)3OH}3)4]+ proceed primarily via two
independent pathways, whilst for the phenyl derivative
[RhH2(P(C6H5)(CH2OH)2)4]+ both protonation and oxidative
addition of H2 appear accessible. These observations are sup-
ported by the H–D exchange reactions discussed earlier. The
non-conversion of 6 to 7 in aqueous solution (in the absence of
hydrogen) is surprising if water is the protonating agent during
the formation of 3 and 4, in that one may expect the most basic
phosphine to promote protonation. One possible explanation
of this trend is that the source of the hydride is not water but
the hydroxyl protons of the ligands, a hypothesis which also
supports the reported synthesis of 3 under anhydrous conditions
(but which is currently attributed to trace amounts of water 4).
This may also account for the slow formation of 7 from 6 in
aqueous solution, where the hydroxyl groups could be repelled
from the metal centre by the longer, hydrophobic propyl chains.
The synthesis of the dideuteride analogues of 3 and 4 could then
be explained by the hydroxyl ligand protons undergoing prior
exchange in the deuterated solvent, followed by reaction at the
rhodium centre.
95.5 Hz), accompanied by a broad resonance for uncoordinated
P{(CH2)3OH}3. On warming d4-methanolic solutions of 10 to +
55 ◦C, loss of 1JRh–P coupling was observed as both resonances
broadened; the chemical shift for the uncoordinated phosphine
flattened over an area of ∼15 ppm, consistent with a phosphine
exchange process. On lowering the temperature the doublet
resolved again until at −60 ◦C the signal broadened and the
coupling was lost. The 13C NMR spectra showed a broadened
doublet resonance at 198.5 ppm (1JRh–C = 53.7 Hz) at room
temperature, which lost the 1JRh–C coupling at +50 ◦C. At −60 ◦C
the resonance was also significantly broadened. 13CO labelled 10
kept under a 12CO atmosphere fully incorporated that isotope,
indicating facile CO exchange. In contrast to 8 and 9 neither the
13C nor the 31P variable-temperature NMR spectra showed JPC
coupling. The IR spectrum of 12CO labelled 10 gave a strong
intensity band at m(CO) = 1966 cm−1 (another medium band
at 1989 cm−1 was also present), with a corresponding value of
1916 cm−1 for the 13CO analogue. The spectroscopic evidence of
rapid dissociation and exchange precludes definitive assignment
of 10 other than as a [Rh(CO)n(P{(CH2)3OH}3)m]+ type species
and serves to illustrate further the difference in reactivity of
P{(CH2)3OH}3 to the hydroxymethyl phosphines.
Reactivity of 3, 4 and 7 towards carbon monoxide
Aqueous and methanolic solutions of all the dicarbonyl
cationic complexes are prone to oxidation on standing, and
prolonged exposure to a reduced pressure also leads to decom-
position. Repeated attempts at recrystallising the chloride salts
of 8 and 9 from aqueous solution, methanol, and methanol–
diethyl ether solvent systems were unsuccessful.
Bubbling aqueous or methanolic solutions of 3 with carbon
monoxide for 15 min gave the phosphine oxide and a new species
8, which appeared as a doublet at d 24.4 (1JRh–P = 107.2 Hz in d4-
methanol) in the 31P NMR spectrum (there was no evidence for
any hydride ligands). The analogous spectrum of 13CO labelled 8,
showed coupling between the P(CH2OH)3 ligands and 13CO, and
the original doublet appeared as a doublet of triplets. The 1JRh–P
Attempts were also made to insert carbon dioxide into the
Rh–H bond/s of the dihydrides, but when aqueous solutions of
3, 4 and 7 were stirred under an atmosphere of the gas for 24 h,
no reactivity was observed.
2
coupling remained at 107.2 Hz, with the JPC triplet splitting
equal to 13.4 Hz. The 13C NMR spectrum showed the methylene
resonances at d 58.2 ppm, and a clear, well-defined doublet of
1
Experimental
quartets at d 197.1 ppm, with a JRh–C coupling constant of
48.8 Hz and 2JP–C of 13.4 Hz. The IR spectrum of 12CO labelled
8 showed a strong, single band at 2020 cm−1 (1958 cm−1 for
13CO labelled 8) indicative of a trans-disposition of the carbonyl
ligands. The data is consistent with a trigonal bipyramidal
structure for trans-[Rh(CO)2(P(CH2OH)3)3]+.
General methods
All manipulations of oxygen sensitive compounds were car-
ried out using vacuum-line techniques. All organic solvents
were distilled and stored under nitrogen; from sodium (hex-
ane and toluene), sodium/potassium (diethyl ether), sodium
and benzophenone (tetrahydrofuran), phosphorus pentox-
ide (dichloromethane) and magnesium/iodine (methanol and
ethanol). The solvent d4-methanol (Cambridge Isotope Lab-
For [RhH2(P(C6H5)(CH2OH)2)4]+ (4), stirred under a 12CO
atmosphere over 17 h, a broad doublet species 9 was observed
at 19.0 ppm (1JRh–P ∼ 107.8 Hz). As in the corresponding
P(CH2OH)3 reaction, no other rhodium–phosphine species were
observed. When 13CO was used, the signal resolves at 0 ◦C
˚
oratories) was dried over 4 A molecular sieves, and degassed
1
2
into a doublet of triplets, with JRh–P = 106.6 Hz and JPC
=
with argon, as was the d6-dimethyl sulfoxide and d-chloroform.
The d2-water was purchased from Apollo Scientific Ltd and was
also purged with argon. Triethylamine (Aldrich) was dried and
distilled from phosphorus pentoxide and stored under argon.
The gases carbon monoxide (both isotopes), carbon dioxide
11.3 Hz. The 13C NMR spectrum shows a corresponding
1
doublet of quartets at the same temperature, with JRh–C
=
53.2 Hz. An IR spectrum of 12CO labelled 9 gave a single
strong band at m(CO) = 2018 cm−1, with the 13CO analogue
4 2 0 6
D a l t o n T r a n s . , 2 0 0 4 , 4 2 0 2 – 4 2 0 8

View Article Online
and dihydrogen were used as supplied (Aldrich). Lancaster
Synthesis Ltd supplied NaB(C6H5)4, NaPF6, NaBF4, AgBF4
and 1,5-cyclooctadiene, which was used without further pu-
rification. The phosphines P(C6H5)H2 and P{(CH2)3OH}3 were
purchased from Strem Chemicals Inc. The phosphonium salt
[P(CH2OH)4]Cl was donated as an 80% w/w aqueous solution
by Rhodia.
precipitate. This was filtered off, dissolved in the minimum
amount of warm methanol, layered with hexane and left for
2 weeks at 0 ◦C. Small yellow crystals were deposited and
examined by X-ray crystallography which gave the structure
as all-trans-[RuCl2(P(CH2OH)3)2(P(CH2OH)2H)2] (1). Yield =
35 mg, 24.2% (Found: C, 20.64; H, 5.38. C10H32O10P4RuCl2
requires C, 19.75; H, 5.30%). IR (KBr, cm−1) 2374 m, m(P–H).
NMR (d4-methanol, ppm): dH 4.48 (8 H, br s, CH2), 4.53 (12
The metal salts RuCl3·H2O and RhCl3 were loaned
by Johnson Matthey plc. The ruthenium compound
[RuCl2(P(C6H5)3)3] was prepared by following the literature
1
3
H, br s, CH2); dC 57.4 (vt, | JP–C + JP–C| 27.2 Hz), 57.9 (vt,
1
3
2
| JP–C + JP–C| 26.8 Hz); dP 13.5 (t, JP–P = 37.5 Hz), 9.7 (t,
2JP–P = 37.5 Hz). Crystallographic data for 1 have been reported
previously (CCDC 182/861).
method.29 The rhodium starting materials [RhCl(1,5-cod)]2,
30
[RhI(1,5-cod)]2 and [RhCl(1,5-cod)P(C6H5)3]
and [Rh(1,5-
cod)(P(C6H5)3)2][B(C6H5)4] 27 were all prepared according to the
literature methods.
trans-[RuCl2 (P(C6H5 )(CH2OH)2 )2 (P(C6H5 )(CH2OH)H)2 ]
(2). P(C6H5)(CH2OH)2 (281 mg, 1.7 mmol) was dissolved
in methanol (10 cm3) and added to a solution (20 cm3) of
RuCl3 in the same solvent (76 mg, 0.4 mmol). The red solution
became green in colour on stirring overnight, after which
time the product precipitated as small yellow crystals. The
crystals were filtered off, washed with methanol and dried in
vacuo. The structure by X-ray crystallography was found to be
trans-[RuCl2(P(C6H5)(CH2OH)2)2(P(C6H5)(CH2OH)H)2] (2).
Yield = 30 mg, 10.5% (Found: C, 45.71; H, 5.04; Cl, 8.88.
C30H40O6P4RuCl2 requires C, 45.47; H, 5.09; Cl, 8.95%).
Physical and analytical measurements
NMR spectra were recorded on Jeol JNM-Ex 270 MHz
FT-NMR and Varian Mercury 400 MHz NMR spectrome-
ters (d relative to tetramethylsilane for 1H and 13C spectra
and 85% phosphoric acid for 31P spectra). IR spectra were
recorded on a Nicolet Impact 410 FT-IR System spectrometer
in aqueous solutions between calcium fluoride plates or as
Nujol mulls/potassium bromide disks. Elemental analyses were
carried out at the University of East Anglia (UEA). Electrospray
mass spectra were recorded on a Fison VG Autospec mass
spectrometer in 1 : 1 water–methanol with 1% acetic acid.
cis-[RhH2(P(CH2OH)3)4]+ (3) and cis-[RhH2(P(C6H5)-
(CH2OH)2)4]+ (4). The relevant phosphine (9.0 mmol)
was dissolved in a degassed aqueous solution (20 cm3) and
slowly added to [RhCl(1,5-cod)]2 (0.51 g, 1.0 mmol) in
dichloromethane (10 cm3). Over a 10 min period complete
colour transfer took place; the organic layer became colourless
Syntheses
P(CH2OH)3 (I). This was based on the literature pro-
cedure.31
and the aqueous layer turned yellow. For P(CH2OH)3 (I), the 31
P
An 80% w/w solution of [P(CH2OH)4]Cl was evaporated to
dryness on a rotary evaporator. The crystalline product (24 g,
0.19 mol) was further dried by azeotropic toluene distillation
NMR spectrum indicated that dihydride 3 was the only rhodium
phosphine species in solution, but for P(C6H5)(CH2OH)2 (II)
longer reaction times are required (typically 1 h) for complete
conversion of 5 to 4. The organic layer was then discarded and
the aqueous phase washed with dichloromethane (20 cm3). The
water was removed under vacuum (at 40 ◦C) and the orange oil
dissolved in methanol (5 cm3). This was then washed repeatedly
with diethyl ether to remove phosphine oxide that precipitated
from the solution. The resultant oil was redissolved in methanol
and the process repeated three times to give 3 as a fine yellow
powder and 4 as a sticky solid. Attempts to characterise 4
by elemental analysis were thwarted by the presence of small
quantities of remaining oxide. Separating the aqueous layer
after 5 min reaction time and slowly bubbling dihydrogen
through the solution for 10 min can also be used to prepare 4.
3: (Found: C, 23.14; H, 5.25. C12H38O12P4RhCl requires C,
22.60; H, 6.02%). NMR (d4-methanol, ppm}: dH 3.48 (12 H, s,
CH2), 3.39 (12 H, s, CH2), −11.10 (2 H, m, RhH); dC 61.0 (vt,
◦
(110 C, 1 atm). The anhydrous salt was then transferred to a
250 cm3 round-bottomed flask, to which was added 150 cm3
of dry triethylamine. This was stirred on an oil-bath at 60 ◦C
for 1 h. After cooling, the resultant amine salt was removed
by filtration. The excess triethylamine and formaldehyde by-
product were then removed by distillation (88 ◦C, 1 atm), the
formaldehyde subliming out into the condenser. The remaining
product fraction consisted of P(CH2OH)3 and a quantity of
hemiacetal impurities. The product was purified by vacuum
distillation (90 ◦C, 0.1 mmHg) using a nitrogen-bleed inlet
to prevent decomposition of P(CH2OH)3 to PH3. The yield
is quantitative (Found: C, 29.23; H, 7.12. Calc. for C3H9O3P:
C, 29.04; H, 7.31%). NMR (d2-water, ppm): dH 4.15 (6 H, d,
2JP–H = 5.5 Hz, CH2); dP −23.7 (s). m/z 124 (M+). Other physical
properties of P(CH2OH)3 can be found in the literature.32
1
3
1
3
| JP–C + JP–C|17.7 Hz), 59.9 (vt, | JP–C + JP–C| 35.4 Hz); dP
P(C6H5)(CH2OH)2 (II). Based on a private communication
by Prof. Katti.8
1
2
1
33.7 (dt, JRh–P = 86.3 Hz, JP–P = 22.0 Hz), 21.0 (dt, JRh–P
=
98.8 Hz, JP–P = 22.0 Hz); IR (KBr, cm−1) 1966 w, m(Rh–H).
When biphasic reaction conditions were again employed and
the Rh : P ratio was 1 : 4.5 then 3 was similarly prepared
from [RhCl(1,5-cod)P(C6H5)3], [RhI(1,5-cod)]2, and [Rh(1,5-
cod)(P(C6H5)3)2][B(C6H5)4], with [Cl]−, [I]−, and [B(C6H5)4]−
acting as the respective counter-ions.
2
Ethanol (15 cm3) was added to an aqueous solution of
formaldehyde (40% w/v, 2.2 g, 29.5 mmol) and degassed with
argon. P(C6H5)H2 (1.6 g, 14.5 mmol) was added dropwise to the
solution and stirred for 5 h. The solvent was removed in vacuo
to yield P(C6H5)(CH2OH)2 as a viscous, colourless oil in near
quantitative yield (2.37 g, 96%). NMR (d2-water, ppm): dH 7.34
(5 H, m, ArH), 4.11 (4 H, d, 2JP–H = 5.9 Hz, CH2); dP −20.4 (s).
4: IR (KBr, cm−1) 1974 w, m(Rh–H). NMR (d4-methanol,
ppm): dH 7.35 (20 H, m, ArH), 4.07 (16 H, m, CH2), −11.20
trans-[RuCl2(P(CH2OH)3)2(P(CH2OH)2H)2]
(1). RuCl3·
1
3
(2 H, m, RhH); dC 62.6 (vt, | JP–C + JP–C| 13.4, 19.5, 18.3 Hz),
xH2O (50 mg, 0.24 mmol) was dissolved in dry methanol
(5 cm3), and to this 6 equivalents of P(CH2OH)3 (178 mg,
1.44 mmol) as a methanol solution (8 cm3) was added. After
10 min a green colour resulted, and a precipitate appeared
to form, which quickly redissolved. After 1 h the solvent
was removed under vacuum, and the resultant oil solidified.
This was dissolved in the minimum amount of methanol
and hexane was added (20 cm3). The hexane layer became
opaque as impurities precipitated out. This layer was filtered
off and the process repeated until a yellow solid started to
1
3
1
61.0 (vt, | JP–C + JP–C| 13.4, 17.0 Hz); dP 28.5 (dt, JRh–P
=
=
2
1
2
99.0 Hz, JP–P = 21.1 Hz), 15.8 (dt, JRh–P = 87.0 Hz, JP–P
21.1 Hz); ESMS m/z 785 [RhH2(P(C6H5)(CH2OH)2)4]+.
cis-[RhH2(P{(CH2)3OH3}4]+ (7). P{(CH2)3OH}3 (110 mg,
0.53 mmol) was dissolved in deoxygenated water (10 cm3) and
added by cannula transfer to a stirred dichloromethane solution
(10 cm3) of [RhCl(1,5-cod)]2 (29 mg, 0.06 mmol). After stirring
for 5 min the aqueous layer became orange in colour, and the
colourless dichloromethane layer was discarded. The solution
D a l t o n T r a n s . , 2 0 0 4 , 4 2 0 2 – 4 2 0 8
4 2 0 7

View Article Online
at this stage contained [Rh(P{(CH2)3OH}3)4]+ (6) as the major
product. Dihydrogen was bubbled through the stirred solution
for 5 min, the resultant light yellow solution indicative of the full
conversion of 6 to 7. As with 4, removal of the solvent yielded 7
as a yellow oil with a small amount of phosphine oxide present
as an impurity. Most of the oxide can be removed by diethyl
ether extraction of methanolic solutions, but our attempts to
remove all oxide impurity for elemental analysis purposes were
unsuccessful.
References
1 (a) Aqueous Phase Organometallic Chemistry, ed. B. Cornils and
W. A. Herrmann, VCH, Weinheim, 1998; (b) F. Joo´ and A. Katho´,
J. Mol. Catal., 1997, 116, 3; (c) K. V. Katti, H. Gali, C. J. Smith and
D. E. Berning, Acc. Chem. Res., 1999, 32, 9; (d) N. Pinault and D. W.
Bruce, Coord. Chem. Rev., 2003, 241, 1.
2 (a) R. Schibli, K. V. Katti, W. A. Volkert and C. L. Barnes, Inorg.
Chem., 2001, 40, 2358–2362, and references therein; (b) C. J. Smith,
G. L. Sieckman, N. K. Owen, D. L. Hayes, D. G. Mazuru, R. Kannan,
W. A. Volkert and T. J. Hoffman, Cancer Res., 2003, 63, 4082–4088.
3 K. Raghuraman, N. Pillarsetty, W. A. Volkert, C. Barnes, S. Jurisson
and K. V. Katti, J. Am. Chem. Soc., 2002, 124, 7276.
4 A. Fukuoka, W. Kosugi, F. Morishita, M. Hirano, L. McCaffrey, W.
Henderson and S. Komiya, Chem. Commun., 1999, 489.
5 B. Drießen-Ho¨lscher and J. Heinen, J. Organomet. Chem., 1998, 570,
141.
IR (KBr, cm−1) 2012 w, m(Rh–H). NMR (d6-dimethyl sulfox-
ide, ppm): dH 3.63 (24 H, m, CH2), 1.90 (48 H, m, CH2), −11.85
(2 H, m, RhH); dP 13.7 (dt, 1JRh–P = 99.6 Hz, 2JP–P = 22.3 Hz),
−0.4 (dt, 1JRh–P = 87.4 Hz, 2JP–P = 22.3 Hz).
Synthesis of the rhodium dicarbonyl cations (8) and (9) and
data for species Rh(13CO)n(P{(CH2)3OH}3)m]+ 10. These com-
pounds were all prepared by stirring the aqueous or methanolic
solutions of the corresponding dihydride under an atmosphere
of CO for 48 h. The 13CO derivatives were prepared in an entirely
analogous fashion.
6 L. J. Higham, A. K. Powell, M. K. Whittlesey, S. Wocadlo and P. T.
Wood, Chem. Commun., 1998, 1107.
7 Y. Kayaki, T. Suzuki and T. Ikariya, Chem. Lett., 2001, 1016.
8 See the experimental section for a simple procedure based upon
a private communication by Professor Katti. The ligand may also
be prepared using the literature procedure: J. B. Rampal, G. D.
Macdonell, J. P. Edasery, K. D. Berlin, A. Rahman, D. van der Helm
and K. M. Pietrusiewicz, J. Org. Chem., 1981, 46, 1156.
9 J. C. Kane, G. P. A. Yap, A. L. Rheingold and E. H. Wong, Inorg.
Chim. Acta, 1999, 287, 209–211.
10 M. Petrucci-Samija, V. Guillemette, M. Dasgupta and A. K. Kakkar,
J. Am. Chem. Soc., 1999, 121, 1968.
11 T. Prinz and B. Driessen-Ho¨lscher, Chem. Eur. J., 1999, 5, 2069–2076.
12 C-C. Tai, J. Pitts, J. C. Linehan, A. D. Main, P. Munshi and P. G.
Jessop, Inorg. Chem., 2002, 41, 1606–1614.
13 F. A. Cotton, B. A. Frenz and D. L. Hunter, Inorg. Chim. Acta, 1976,
16, 203.
14 E. B. McAslan, A. J. Blake and T. A. Stephenson, Acta Crystallogr.,
Sect. C, 1989, 45, 1811.
15 J. M. Grosselin, C. Mercier, G. Allmang and F. Grass,
Organometallics, 1991, 10, 2126–2133.
16 C. Larpent, R. Dabard and H. Patin, Inorg. Chem., 1987, 26, 2922.
17 P. A. T. Hoye, P. G. Pringle, M. B. Smith and K. Worboys, J. Chem.
Soc., Dalton Trans., 1993, (2), 269.
Spectroscopic data for [Rh(13CO)2(P(CH2OH)3)3]+ (8): NMR
1
(d4-methanol, 298 K, ppm): dC 197.1 (dq, JRh–C = 48.8 Hz,
1
2
2JP–C = 13.4 Hz, RhC); dP 24.4 (dt, JRh–P = 107.2 Hz, JP–C
2020 s, m(CO) (1958 for 13CO 8).
=
13.4 Hz). For [Rh(12CO)2(P(CH2OH)3)3]+ IR: (water, cm−1)
Spectroscopic data for [Rh(13CO)2(P(C6H5)(CH2OH)2)3]+ 9:
1
NMR (d4-methanol, 273 K, ppm): dC 197.6 (dq, JRh–C
=
53.2 Hz, 2JP–C = 11.3 Hz, RhC): dP 19.0 (dt, 1JRh–P = 106.6 Hz,
2JP–C = 11.3 Hz). For [Rh(12CO)2(P(C6H5)(CH2OH)2)3]+ IR:
(water, cm−1) 2018 s, m(CO) (1976 for 13CO 9).
Spectroscopic data for (10): NMR (d4-methanol, 273 K, ppm):
1
1
dC 198.5 (d, JRh–C = 53.7 Hz, RhC); dP 11. 0 (d, JRh–P
=
95.5 Hz). For [Rh(12CO)2(P{(CH2)3OH}3)3]+ IR: (water, cm−1)
1966 s, m(CO) (1916 for 13CO 10).
Attempts to isolate pure samples of the dicarbonyl products
were hindered due to contamination by small quantities of the
phosphine oxides.
18 For an example, see: J. H. Downing, V. Gee and P. G. Pringle, Chem.
Commun., 1997, 1527–1528.
19 For an example, see: N. J. Goodwin, W. Henderson, B. K. Nicholson,
J. Fawcett and D. R. Russell, J. Chem. Soc., Dalton Trans., 1999, (11),
1785–1793.
20 A. Salvini, P. Frediani, M. Bianchi, F. Piacenti, L. Pistolesi and L.
Rosi, J. Organomet. Chem., 1999, 582, 218–228.
21 R. Young and G. Wilkinson, Inorg. Synth., 1977, 17, 75.
22 R. Young and G. Wilkinson, Inorg. Synth., 1977, 17, 79.
23 J. Chatt, G. J. Leigh and R. M. Slade, J. Chem. Soc., Dalton Trans.,
1973, (19), 2021.
Crystal structure determination for complex 2
¯
C30H40Cl2O6P4Ru, M = 792.48, triclinic, space group P1, a =
˚
9.283(2), b = 10.008(6), c = 10.568(1) A, a = 99.87(2), b =
◦
3
˚
105.12(1), c = 109.15(1) , U = 859.1(6) A , T = 293 K, Z = 1,
−1
−3
˚
l(Mo-Ka) = 0.838 mm , Dc= 1.516 Mg m , k = 0.71073 A.
5570 data were collected on a Rigaku RAXIS IIc image plate of
which 2962 were unique (Rint = 0.0825), 2003 had Fo > 4r(Fo),
4.94 < 2h < 51.18◦, no absorption correction was applied.
Structure solved by direct methods using SHELXS and all non-
hydrogen atoms refined anisotropically using full-matrix least
squares on F2 (SHELXL-93).33 R1 = 0.0641 (for 4r data), wR2 =
0.2084, S = 1.010 (for all data).
24 J. D. Gilbertson, N. K. Szymczak and D. R. Tyler, Inorg. Chem.,
2004, 43, 3341.
25 B. J. Frost and C. A. Mebi, Organometallics, 2004, 23, 5317.
26 D. P. Paterniti, P. J. Roman, Jr. and J. D. Atwood, Organometallics,
1997, 16, 3371.
27 R. R. Schrock and J. A. Osborn, J. Am. Chem. Soc., 1971, 93, 2397.
28 A. R. Siedle, R. A. Newmark and R. D. Howells, Inorg. Chem., 1988,
27, 2473.
29 P. S. Hallman, T. A. Stephenson and G. Wilkinson, Inorg. Synth.,
1970, 12, 237.
30 J. Chatt and L. M. Venanzi, J. Chem. Soc. A, 1957, 4735.
31 K. A. Petrov, V. A. Parshina and M. B. Luzanova, Zh. Obshch. Khim.,
1962, 32(2), 553 (English transl., p. 542).
32 A. W. Frank, D. J. Daigle and S. L. Vail, Textile Res. J., 1982, 738;
M. Reuter and L. Orthner [to Farbwerke Hoechst A.-G.], Ger. Pat.,
1,035,135, 1958.
CCDC reference number 228753.
See http://www.rsc.org/suppdata/dt/b4/b411701h/ for cry-
stallographic data in CIF or other electronic format.
Acknowledgements
We wish to thank the EPSRC and UEA for financial sup-
port, Johnson Matthey for precious metals and Rhodia for
tetrakis(hydroxymethyl)phosphonium chloride.
33 G. M. Sheldrick, University of Go¨ttingen, 1993.
4 2 0 8
D a l t o n T r a n s . , 2 0 0 4 , 4 2 0 2 – 4 2 0 8