Diferrocenyltriphosphines 2. Reversible phosphine deligation in the chemistry of diferrocenyltriphosphine Ru(II) dichloride complexes with nitriles and pyridines: Towards a pH-switchable catalyst?

Article abstract of DOI:10.1016/S0022-328X(01)01135-4

The reaction of the complexes [P₃-(P₃fc₂)RuCl₂], P₃fc₂=[η⁵-(R₂PC₅H₄)Fe-η⁵:η⁵-(C₅H₄P(R′)C₅H

Full text of DOI:10.1016/S0022-328X(01)01135-4

Journal of Organometallic Chemistry 637–639 (2001) 538–548
www.elsevier.com/locate/jorganchem
Diferrocenyltriphosphines 2. Reversible phosphine deligation in the
chemistry of diferrocenyltriphosphine Ru(II) dichloride complexes
with nitriles and pyridines: towards a pH-switchable catalyst?
Ian R. Butler a,*, Simon J. Coles e, Marco Fontani d, Michael B. Hursthouse e,
Eric Lewis a, K.L.M. Abdul Malik c, Marc Meunier b, Piero Zanello d
a Department of Chemistry, Uni6ersity of Wales, Bangor, Gwynedd LL57 2UW, UK
b Centre for Computational Chemistry, Uni6ersity of Wales, Bangor, Gwynedd LL57 2UW, UK
c Department of Chemistry, Uni6ersity of Wales, Cardiff, PO Box 912, Park Place, Cardiff CF10 3TB, UK
d Dipartimento di Chimica dell’Uni6ersita` di Siena, 6ia Aldo Moro, 53100 Siena, Italy
e Department of Chemistry, Uni6ersity of Southampton, Highfield, Southampton SO17 1BJ, UK
Received 2 February 2001; received in revised form 18 June 2001; accepted 20 June 2001
Abstract
The reaction of the complexes [P3-(P3fc2)RuCl2], P3fc2=[h5-(R2PC5H4)Fe-h5:h5-(C5H4P(R%)C5H4)Fe(h5-C5H4PR2)], R, R%=
Ph, iPr with linear nitriles results in the formation of the pseudooctahedral complexes, e.g. [P3-(P3fc2)Ru(NCCH2R)Cl2], R=H,
Ph while the reaction of [(P3fc2)RuCl2] with pyridines results in partial deligation of one of the ferrocenyl phosphine ligand arms
to give the products [P2-(h5-(R2PC5H4)Fe-h5:h5-(C5H4P(R%)C5H4Fe(h5-C5H4PR2))Ru(pyr)2Cl2]. The latter products revert to the
starting complexes in the absence of excess pyridine in solution. The deligated pendant phosphine may be ‘trapped’ by reaction
with Pd(II) or elemental sulfur. A molecular modelling study has been carried out to verify the conformations of the products.
The single crystal structures of mer-[P3-(P3fc)2Ru(CO)Cl2] and mer-[P2-(P3fc)2Ru(pyr)2Cl2] have been determined. Cyclic voltam-
metry supports the role assigned to pyridine in the reaction with the P3Fc2-RuCl2 complexes. © 2001 Published by Elsevier
Science B.V.
Keywords: Ferrocene; Phosphine ligand; Ruthenium complexes; Electrochemistry
1. Introduction
a number of triphosphinodiferrocenyl ligands of the
type 1 (Fig. 1) was described [1]. These ligands may be
obtained in high yields using a relatively simple syn-
thetic methodology. The synthesis of Pd(II), Rh(I) and
Ru(II) complexes of these ligands has also been docu-
mented [1,2].
The ruthenium complexes 2 (Fig. 1) are of special
interest because of their similarity to some well known
five co-ordinate complexes, such as [Ru(triphos)Cl2] [3],
triphos=[(Ph2P(CH2)2)2PPh] which exhibit interesting
co-ordination and reaction chemistry. We were inter-
ested in developing the co-ordination chemistry of the
new compounds, starting with simple addition reaction
to the vacant pseudo octahedral site in the five co-ordi-
nate complexes. One such addition, reported earlier by
this research group, was the addition of carbon monox-
ide to the vacant site, which proceeds smoothly [2]
The design and synthesis of ligands for use in transi-
tion-metal catalysed processes is one of the most impor-
tant areas of organometallic chemistry research. There
is now an abundance of ligands available for use in
such processes, so many in fact that it makes the
selection of an optimum ligand for a particular reaction
difficult for the practising organic chemist. It may be
argued that it is no longer enough to concentrate
resources on the synthesis of more ligands. This re-
search is focussed on the design and synthesis of smart
ligands, ligands that have interesting chemical and com-
plexation behaviour. In a recent paper the synthesis of
* Corresponding author. Fax: +44-1248-370-528.
E-mail address: i.r.butler@bangor.ac.uk, i.r.butler@ferrocene.com
(I.R. Butler).
0022-328X/01/$ - see front matter © 2001 Published by Elsevier Science B.V.
PII: S0022-328X(01)01135-4

I.R. Butler et al. / Journal of Organometallic Chemistry 637–639 (2001) 538–548
539
Fig. 1. Diferrocenyltriphosphines and their ruthenium complexes.
(Scheme 1a). The single crystal X-ray structure of one
such compound is reported herein to clearly define the
geometry of the carbonyl-containing products. The
chemistry of the system in this case mimics that of
[Ru(triphos)Cl2]. It was decided to continue with this
investigation to examine the use of nitriles and pyridi-
nes as potential ligands to complexes 2.
diffusion of a chloroform/petrol layered mixture. The
visible spectra of complexes 2a–2d were obtained by
standard methods: UV: umax (nm) values 2a, 484, 2b,
508, 2c, 488, 2d, 504 (CH2Cl2). Molecular modelling
was carried out using the Molecular Simulations Inc.
Suite of Programs. Dynamics calculations were per-
®
formed using the DISCOVER program.
2.2. X-ray crystallography for [(P3 fc)2Ru(CO)Cl2] (3c)
2. Experimental
The crystal used for X-ray work was obtained by
slow diffusion of a CHCl3–hexane mixture. All mea-
surements were made on a Delft Instruments FAST
area detector diffractometer positioned at the window
of a rotating anode generator with Mo–Ka radiation
by following procedures described earlier [4].
2.1. General experimental details
1H-NMR and 31P-NMR (relative to 85% phosphoric
acid) spectra were recorded at 250 and 101 MHz,
respectively using a Bruker WC-250 instrument. All
other solvents used were predried using conventional
methods and then were distilled prior to use. The
complexes 2a–2b were prepared according to the pub-
lished procedure [2]. Materials and apparatus for the
electrochemistry have been previously described [2].
Potential values are referred to the saturated calomel
electrode (SCE). Under the present experimental condi-
tions the one-electron oxidation of ferrocene occurs at
E°= +0.39 V. Compound 3c was prepared as previ-
ously described [2] and it was crystallised by slow
2.2.1. Crystal data
Formula C48H42Cl2Fe2OP3Ru, 2CHCl3, FW 1250.13,
monoclinic, space group P21/n (no. 14), a=19.818(4),
,
b=11.977(2), c=21.558(4) A, i=98.79(3)°, U=
3
,
5057(2) A (least-squares refinement of diffractometer
angles for 250 reflections within 1.915q525.07°),
Z=4,
D , v(Mo–Ka)=1.417
calc=1.642 Mg m−3
mm−1, F(000)=2516, crystal size=0.26×0.18×0.15
mm3, T=150(2) K. Intensities of 18447 reflections
Scheme 1. The binding of CO to the five-coordinate complex and pyridine coordination which triggers the deligation of a phosphine ligand.

540
I.R. Butler et al. / Journal of Organometallic Chemistry 637–639 (2001) 538–548
Table 1
The structure was solved by direct methods (SHELXS
-
,
Selected bond lengths (A) and bond angles (°) for [(P3fc)2Ru(CO)Cl2]
(3c)
86) [6] and refined by full-matrix least-squares on F2
using all unique data with intensities greater than 0
(SHELXL-93) [7]. The non-hydrogen atoms were all an-
Bond lengths
Ru(1)–P(1)
Ru(1)–P(3)
Ru(1)–Cl(2)
Fe(1)–C(5)
Fe(1)–C(6)
Fe(1)–C(7)
Fe(1)–C(10)
Fe(1)–C(8)
Fe(2)–C(18)
Fe(2)–C(12)
Fe(2)–C(16)
Fe(2)–C(20)
Fe(2)–C(14)
2.438(3)
2.404(3)
2.436(3)
Ru(1)–P(2)
Ru(1)–Cl(1)
Ru(1)–C(1)
Fe(1)–C(3)
Fe(1)–C(4)
Fe(1)–C(11)
Fe(1)–C(9)
Fe(1)–C(2)
Fe(2)–C(19)
Fe(2)–C(17)
Fe(2)–C(15)
Fe(2)–C(21)
Fe(2)–C(13)
2.529(3)
2.427(3)
isotropic. There were two molecules of CHCl3 solvate
(per complex) with some of the chlorine positions par-
tially occupied. The atoms C(1), C(49), C(50) Cl(3)–
Cl(11) were refined with ISOR=0.01 to prevent them
becoming ‘non-positive-definite’. The C–Cl distances in
the solvates were also refined with the constraint 1.76
1.947(14)
2.005(11)
2.022(11)
2.026(10)
2.039(11)
2.082(11)
1.995(11)
2.015(11)
2.029(11)
2.061(11)
2.071(11)
1.997(11)
2.015(11)
2.023(11)
2.029(11)
2.053(12)
1.992(11)
1.996(11)
2.016(11)
2.042(11)
2.063(12)
,
A. The H atoms of CHCl3 were ignored; other H atoms
were included in calculated positions (riding model)
with Uiso=1.2×Ueq of the parent carbon. Final R {on
2067 data with I\2|(I)}=0.0524 and wR2 (on F2, all
7395 data)=0.1168, w=1/[|2(Fo2)]. Selected bond
lengths and angles are given in Table 1.
Bond angles
C(1)–Ru(1)–P(3) 85.7(4)
P(3)–Ru(1)–Cl(1) 82.74(11)
P(3)–Ru(1)–Cl(2) 89.18(11)
C(1)–Ru(1)–P(1) 85.5(4)
Cl(1)–Ru(1)–P(1) 83.24(10)
C(1)–Ru(1)–P(2) 166.4(5)
Cl(1)–Ru(1)–P(2) 94.78(11)
P(1)–Ru(1)–P(2) 92.39(10)
C(1)–Ru(1)–Cl(1)
C(1)–Ru(1)–Cl(2)
Cl(1)–Ru(1)–Cl(2) 171.29(10)
P(3)–Ru(1)–P(1)
Cl(2)–Ru(1)–P(1)
P(3)–Ru(1)–P(2)
Cl(2)–Ru(1)–P(2)
98.3(5)
84.3(4)
2.3. X-ray crystallography for [P2-(P3 fc)2Ru(pyr)2Cl2]
(5b)
162.16(11)
105.30(10)
99.72(11)
83.35(11)
Data were measured on a Nonius KappaCCD area
detector equipped with a rotating anode generator
(Mo–Ka radiation) and controlled by COLLECT [8]
software.
Table 2
2.3.1. Crystal data
Formula C60H51Cl2Fe2N2P3Ru, 3C3H6O, FW 1350.84,
triclinic, space group P1, a=11.927(2), b=15.891(3),
,
Selected bond lengths (A) and bond angles (°) for [P2-
(P3fc)2Ru(pyr)2Cl2] (5a)
(
Bond lengths
N1–Ru1
P1–Ru1
,
c=17.912(4) A, h=111.29(3), i=92.55(3), k=
3
2.170(2)
2.3076(9)
2.4203(11)
2.020(3)
2.045(3)
2.032(3)
2.039(3)
2.038(3)
2.062(3)
2.044(3)
2.048(3)
2.066(3)
2.029(3)
N2–Ru1
P2–Ru1
2.175(2)
2.3312(11)
2.4224(10)
2.027(3)
2.044(3)
2.023(3)
2.037(3)
2.053(3)
2.048(3)
2.050(3)
2.051(3)
2.052(3)
2.038(3)
,
101.54(3)°, U=3073.7(11) A , Z=2, Dcalc=1.460 Mg
m
−3, v(Mo–Ka)=0.923 mm−1, F(000)=1392, crystal
Ru1–Cl2
C23–Fe1
C25–Fe1
C27–Fe1
C29–Fe1
C31–Fe1
C39–Fe2
C41–Fe2
C43–Fe2
C45–Fe2
C47–Fe2
Ru1–Cl1
C24–Fe1
C26–Fe1
C28–Fe1
C30–Fe1
C32–Fe1
C40–Fe2
C42–Fe2
C44–Fe2
C46–Fe2
C48–Fe2
size=0.16×0.14×0.06 mm3, T=150(2) K. Intensities
of 45139 reflections [2.925q527.5°; −155h515,
−205k520, −235l523] were recorded and pro-
cessed to obtain 13818 [Rint=0.0967] unique data with
I\2|(I). The data were corrected for absorption ef-
fects (SORTAV) [9] (min. and max. transmission factors
0.8664, 0.9467). The structure was solved by direct
methods and refined by full-matrix least-squares on F2
using all unique data using the SHELX-97 suite of
programs [10]. The non-hydrogen atoms were all an-
isotropic. There were three molecules of C3H6O solvate
per complex molecule. H atoms were included in calcu-
lated positions (riding model) with Uiso=1.2×Ueq of
the parent carbon. Final R {on 10096 data with I\
2|(I)}=0.0428 and wR2 (on F2, all 13818 data)=
0.1076, w=1/[|2(Fo2)]. Selected bond lengths and angles
are given in Table 2.
Bond angles
N1–Ru1–N2
N2–Ru1–P1
N2–Ru1–P2
N1–Ru1–Cl2
P1–Ru1–Cl2
N1–Ru1–Cl1
P1–Ru1–Cl1
Cl2–Ru1–Cl1
83.25(9)
173.30(6)
90.53(7)
86.20(7)
87.31(4)
85.20(7)
96.05(4)
170.86(3)
N1–Ru1–P1
N1–Ru1–P2
P1–Ru1–P2
N2–Ru1–Cl2
P2–Ru1–Cl2
N2–Ru1–Cl1
P2–Ru1–Cl1
91.87(7)
169.23(6)
94.97(4)
87.78(7)
102.40(4)
88.15(7)
85.82(4)
2.4. NMR reactions: general procedure
[1.915q525.07°; −235h523, −95k514, −255l
525] were recorded and processed to obtain 7395
[Rint=0.1307] unique data with I\2|(I). The
data were corrected for absorption effects (DIFABS) [5]
(min. and max. absorption correction factors 0.836,
1.018).
A solution of complex 2b (500 mg, 0.57 mmol) in
d-chloroform (15 cm3) was used as a standard. The
solution was subdivided into 1.5 cm3 fractions and to
each a tenfold excess of the appropriate ligand was
1
added. The H- and 31P-NMR spectra were recorded.

I.R. Butler et al. / Journal of Organometallic Chemistry 637–639 (2001) 538–548
541
Table 3
+44.61 (d, J=35 Hz). Anal. Calc. for C48H59Cl2-
Fe2N2P3Ru: C, 55.36; H, 5.71; N, 2.69. Found: C,
55.50; H, 5.53; N, 2.53% (as dried, non-crystalline
powder). Mass spectrum: no parent ion (1038–44 amu)
was observed, EI, CI or FAB. Crystals suitable for
X-ray diffraction were grown in acetone.
Co-ordination 31P-NMR chemical shifts on addition of ligands to
complex 2b to form new complexes of the general type 3b
Ligand
PiPr2(cis)
PPh(trans)
DPiPr2
DPPh
CO"
CH3CN
PhCH2CN
N(CH2CH2OH)3
21.13
21.96
22.77
46.38
−20.59
27.32
28.17
+37.7
+15.6
+14.8
−8.84
+102.8
+54.9
+54.1
+26.9
1
Complex 5b+Pd(COD)Cl2: H-NMR(CDCl3): 0.41
(dd, J=7 Hz, 12 Hz, 3H), 0.57 (dd, J=7 Hz, 12 Hz,
3H), 1.03–1.38 (overlapping dd’s, 18H), 2.35 (m, 2H),
2.82 (m, 2H), aliphatic protons, 3.20 (m, H), 3.64 (m,
1H), 3.94 (m, 1H), 4.02 (m, 1H), 4.17 (m, 1H), 4.26 (m,
1H), 4.33 (m, 1H), 4.50 (m, 1H), 5.03 (m, 1H), 5.23 (m,
1H), other multiplets are overlapped; ferrocenyl pro-
tons; 7.27–7.51 (br m, 5H) (aromatic protons), 7.13 (m,
2H), 7.52 (m, 2H), 8.50 (m, 2H), 8.72 (m, 2H). 31P-
NMR: +23.90 (s), +30.79, (br d), +44.69 (d).
55.33
2.5. Isolation of pyridine complexes 5a, 5b (general
procedure)
A solution of the complex (50 mg, 0.049 mmol), 2a
or (50 mg, 0.057mmol) 2b in dichloromethane (3cm3)
was treated with the appropriate pyridine base (10–30-
fold molar excess). The solution turned yellow or or-
ange from the initial violet colour. n-Hexane was added
dropwise until the solution became turbid. The mixture
was filtered and the solution was cooled to −20 °C for
10 h. The product, in each case, crystallises from solu-
tion as yellow or orange/yellow nodules. N.B.: resolu-
tion in CH2Cl2 results in partial reconversion to 2a or
2b, respectively. Spectroscopic data are presented in
Tables 3 and 4.
1
5a: H-NMR (CDCl3): 2.29 (m, 1H), 2.92 (m, 1H),
3.19 (m, 1H), 3.39 (m, 1H), 3.45 (m, 1H), 3.46 (m, 1H),
3.61 (m, 1H), 3.70 (m, 1H), 3.79 (m, 1H), 3.99 (m, 1H),
4.29 (m, 1H), 4.37 (m, 1H), 4.47 (m, 1H), 5.00 (m, 1H),
5.80 (m, 1H), ferrocenyl protons; 6.70–7.00 (br m, 4H),
7.06–7.50 (m’s, 21H), (phenyl protons) 8.47 (br s), 8.80
(br s), pyridinyl protons; 31P-NMR (CDCl3): −18.71
(s), +32.41 (d), +50. 67 (d) JP–P=36.6 Hz (2a,
+90.84 (t), +38.30 (d), JP–P=30.5 Hz). Anal. Calc.
for C60H51Cl2Fe2N2P3Ru: C, 61.20; H, 4.37; N, 2.38.
Found: C, 60.78; H, 4.32; N, 2.25 (as dry, non solvated,
powder). Crystals may be grown in acetone.
5b: 1H-NMR (CDCl3): 0.36 (dd, 3H, JP–H=7, 12
Hz), 0.51(dd, JP–H=7, 12 Hz, 3H), 0.58–0.79 (series of
overlapping dd’s, 12H), 1.18–1.33 (2 overlapping dd’s,
JS=7 Hz, 13 Hz, 6H), 2.07 (br m, 2H), 2.66 (br m,
2H), aliphatic protons; 2.12 (m, 1H), 3.12 (m, 1H), 3.17
(m, 1H), 3.22 (m, 1H), 3.47 (m, 1H), 3.57 (m, 1H), 3.66
(m, 1H), 3.92 (m, 1H), 4.07 (m, 1H), 4.16 (m, 1H), 4.22
(m, 1H), 4.25 (m, 1H), 4.40 (m, 1H), 4.90 (m, 1H), 5.13
(m, 1H), 5.17 (m, 1H), ferrocenyl protons; 7.02–7.20
(ms, 5H). 31P-NMR: −1.13(s), +30.85 (d, J=35 Hz),
2.5.1. Reaction with [Pd(COD)Cl2]
1
[5a PdCl2]: H-NMR (CDCl3, d5-pyr) 2.66 (m, 1H),
3.00 (m, 1H), 3.45 (m, 1H), 3.92 (m, 1H), 3.95 (m, 1H),
4.04 (m, 1), 4.12 (m, 1H), 4.37 (m, 1H), 4.43 (2m, 2H),
4.50 (m, 1H), 4.60 (m, 1H), 4.68 (m, H), 5.30 (m, 1H),
5.35 (m, 1H), 5.90 (m, 1H), 5.53 (traces of COD).
6.50–7.75 (series of overlapping multiplets, 25H). 31P-
Table 4
Reaction of Complex 2b with pyridines and related ligands: 31P-NMR Chemical Shift Values
Ligand (pKa)
PPh
PiPrA
44.61
PiPrB
DPPh
DPiPrc
−6.9
Pyridine (5.25)
30.85
−1.13
+51.4
2-Methylpyridine (5.97) (a-picoline)
No reaction at ambient temperature
3-Methylpyridine (5.68) (b-picoline)
30.45
30.44
30.16
29.70
31.66
28.60
33.08
30.18
44.10
44.54
45.27
44.35
48.68
44.10
49.11
44.53
−0.85
−0.84
−1.21
−1.10
−1.85
−1.29
−0.99
−1.08
+51.8
+51.8
−6.6
−7.0
−7.7
−6.8
−9.1
−6.5
−10.2
−7.0
4-Methylpyridine (6.02) (g-picoline)
3-Acetylpyridine
4,4%-Dipyridine
Pyridazine (2.24)
Pyrazine (0.65)
Pyrazole
Isoquinoline (5.42)
+52.10
+52.53
+50.60
+53.85
+49.20
+52.10
Quinoline (4.90), diphenylacetylene, quinoxaline (0.56), pyrrole, 3-methylisoquinoline
No reaction"
Partial delegation
Acetonitrile
2,2%-dipyridine
34.90
28.50
28.09
23.50
51.78
−1.16
−1.16
−1.39
−1.39
+47.3
+53.7
+54.2
+58.7
−14.5
−1.6
−14.4
+5.4
35.60
51.66
31.88
D
+a-N,N-dimethylphenylethylamine
Benzylnitrile

542
I.R. Butler et al. / Journal of Organometallic Chemistry 637–639 (2001) 538–548
Fig. 2. The single crystal structure of compound 3c [18].
NMR: +23.24 (s), +32.41 (d), +50.73 (d), JP–P
36.6 Hz.
=
were obtained however using a revised isolation proce-
dure we have now been able to isolate crystalline sam-
ples. The single crystal structure of one of these
products, 3c, has been determined. The reaction
chemistry in this case is analogous to that of
2.6. Reaction of compound 2a with elemental sulfur
[Ru(triphos)Cl2], triphos=[Ph2(CH2)2]2PPh [3].
A
A mixture of compound 5a was prepared as de-
scribed above on a 1 g scale in dichloromethane to
which elemental sulfur 250 mg was added. The solution
was filtered and layered with hexane. The golden yel-
low/tan crystals of the product, which were obtained
after 24 h slow diffusion, were isolated and washed with
petrol. 7a: 1H-NMR (CDCl3, d5-pyr): 2.31 (bs, 1H),
3.06 (bs, 1H), 3.59 (bs, 1H), 3.76 (bs, 1H), 3.90 (bs,
1H), 3.93 (bs, 1H), 3.96 (bm’s 2H), 4.17 (bs, 1H), 4.24
(bs, 1H), 4.46 (bs, 1H), 4.51 (bs, 1H), 4.62 (bs, 1H),
5.08 (bs, 1H), 5.30 (bs, 1H), 5.89 (bs, 1H), (ferrocenyls).
6.40–7.70 (series of overlapping multiplets, 36H), 8.20,
(bm, 1H), 8.31 (bm, 1H), 8.82, (bm, 2H) 9.18 (bm,
1H).). 31P-NMR: +32.71 (br s), +40.91(s), +50.67
(br s). Anal. Calc. for C60H51Cl2Fe2N2P3RuS.CH2Cl2:
C, 56.64; H, 4.13; N, 2.17. Found: C, 56.78; H, 4.19; N,
2.25%.
correlation may be drawn between the 31P chemical
shifts of the ligand phosphorus resonances with the
carbonyl stretching frequencies for these complexes.
Thus, as expected, the more basic phosphine ligands
have the effect of increasing the strength of the metal–
carbonyl bond. For example the carbonyl stretching
frequencies of these complexes 3a–3d are observed at
1988, 1962, 1974 and 1954 cm−1, respectively, indicat-
ing that the stronger metal carbonyl bonding occurs
when more electron density is provided by the ferro-
cenylphosphine ligands. The 31P-NMR chemical shifts
for the phosphine trans to the carbonyl group in com-
plexes 3a–3d are respectively are at −8.7, −20.6,
+18.05 and +2.77 ppm [2].
3.1. X-ray structure of [(P3 fc)2Ru(CO)Cl2] (3c)
The molecular structure of (3c) is shown in Fig. 2,
which also indicates the crystallographic atom number-
ing scheme used for the non-carbon atoms. The disor-
dered solvate species (CHCl3) in the lattice are omitted
for clarity. Selected molecular geometry parameters are
presented in Table 1.
There are numerous reports on the structures of
metal complexes containing PPP donor ligands [11–17],
but complexes of tridentate PPP ligand with two ferro-
3. Results and discussion
In a recent paper the synthesis of a series of Ru(II)
complexes of the diferrocenyltriphosphine ligands 1 was
described. These complexes react with carbon monox-
ide rapidly at 25 °C to form the pseudo octahedral
product complexes 3, (Scheme 1). In the initial investi-
gation only powdered microcrystalline samples of 3

I.R. Butler et al. / Journal of Organometallic Chemistry 637–639 (2001) 538–548
543
cenyl units are relatively rare, the only such example,
[RuCl2{Ph2P(C5H4)2Fe}2P(Ph)]·2CHCl3, (2a), being re-
ported very recently [2]. This compound was shown to
have a five co-ordinate square-pyramidal geometry with
the bridging phosphorus at the apical position, and a
phenyl hydrogen blocking the vacant ‘sixth’ co-ordina-
tion site on the Ru atom. The present structure is very
similar, with a carbonyl group occupying the ‘sixth’
co-ordination site and thus completing the typical six
co-ordinate octahedral geometry around the metal
atom. The main deviations from ideal geometry are the
angles (a) P(1)–Ru(1)–Cl(2) which, instead of being ca.
90°, has a value of 105.30(10)°, and (b) P(1)–Ru(1)–
P(3) and P(2)–Ru(1)–Cl(1) which, instead of 180°, are
162.16(11) and 166.4(5)°, respectively. The Cl(1)–
Ru(1)–Cl(2) angle is close to linear [171.29(10)°], and
shows less distortion compared with 2a [153.19(4)°].
The much wider P(1)–Ru(1)–Cl(2) angle may be ex-
singlet. The position and multiplicity of the latter reso-
nances is indicative of deligation of a phosphine group.
An identical observation was made in the reaction of
complex 2b with phenylacetonitrile, under identical ex-
perimental conditions. The data are summarised in
Table 3 and in part in Table 4. Interestingly in the case
of triethanolamine the terminal phosphorus resonances
are shifted upfield, i.e. stronger bonding or a marked
change in geometry is indicated, possibly due to an
interaction with the hydroxy group(s).
The next stage of the investigation progressed to
examine the co-ordination behaviour of pyridine. When
the 31P-NMR spectrum was examined it was clear that
in this case one of the terminal isopropyl-substituted
phosphines had become detached from the metal centre
i.e. deligated, however it was not clear if the product
complex was a mono or bis-pyridinyl complex. It could
be envisaged that the pyridine entered the ‘vacant’
co-ordination site, which subsequently triggered the
deligation of the geometrically strained phosphine. The
free site could then be filled by a second pyridine,
(Scheme 1b).
,
plained by the short Cl(2)···H(26) 2.521 A steric
interaction.
3.2. Co-ordination studies
To extend the scope of the coordination study and to
continue with the original objective it was decided to
examine the co-ordination behaviour of a number of
substituted pyridines and related ligands. Of these lig-
ands pyrazine, pyradazine, 3-acetylpyridine and iso-
quinoline reacted in an analogous fashion to pyridine
while quinoline, quinoxaline, 3-methyl-iso-quinoline,
pyrrole and 1,2-diphenylacetylene did not react at am-
bient temperature (although several of these ligands did
show evidence of reaction after standing for 20 h). It is
clear that the steric constraints inhibited the co-ordina-
tion of 2-methylpyridine, quinoline, 2,2%-dipyridine
and quinoxaline. The relevant data are summarised in
Table 4.
It was of interest to examine the possibility of using
the pseudo five co-ordinate complexes 2 as a means of
obtaining a scale of ligand basicity, i.e. to react the
complexes with a range of related ligands to use the
31P-NMR shifts as a basis for a correlation. The colour
change (violet to yellow), on going from five to six
co-ordination, provides a rapid visual measure. The
first complex chosen to be examined was complex 2b
because of its high solubility in a range of solvents. The
reactions were investigated using 31P-NMR using the
ligand chemical shifts as a means of correlation. The
reaction of complex 2b in acetonitrile was first
considered.
There was a considerable upfield shift in the ligand
resonances observed on addition of acetonitrile to a
CDCl3 solution of complex 2b. The solution changed
colour from violet to yellow as expected. The 31P ligand
resonances are observed at +82.23 and +37.54 ppm,
respectively, in compound 2b, the latter being due to
the two equivalent isopropyl substituted phosphorus
nuclei while in the product complex the new resonances
were observed at +27.32 and +21.96 ppm, respec-
tively. The larger upfield shift (+54.9 ppm) is observed
for the phosphorus trans to the incoming acetonitrile
ligand. This chemical shift is considerably smaller than
that previously observed in the case of carbon monox-
ide co-ordination (+102.8) an indication of its weaker
ligand field strength (i.e. lack of back donation). More
interestingly, however, was the observation that the
reaction is not clean in that three other weaker reso-
nances were also observed at chemical shift values of
+51.78, +34.90 and −1.57 ppm, respectively, the
former two resonances being doublets and the latter a
At this point it was clearly evident from the spectro-
scopic data that (i.e. the similarity of the phosphorus
chemical shift data) it would be impossible to use the
coordination reaction as a means of obtaining a ligand
basicity scale because of the electronic buffering effect
of the metal–ligand interaction. However more inter-
esting, as a potential pH trigger, was the pyridine-as-
sisted metal–phosphine cleavage reaction, which may
be able to be useful in a switchable catalysis regime.
Thus this reaction was explored in more detail. To this
point only 31P data have been analysed therefore it
1
became evident that H-NMR data were also required.
A d-chloroform solution of complex 2a or 2b was
treated with a drop of pyridine to obtain the yellow
solution. The solvent was then removed under vacuum
to leave a yellow oil, which was hexane washed. The
product was redissolved in d-chloroform whereupon
the violet colour was partially re-established demon-
1
strating the reversibility of the reaction. The H-NMR

544
I.R. Butler et al. / Journal of Organometallic Chemistry 637–639 (2001) 538–548
spectrum obtained was that of a mixture of the starting
compound and a pyridine-containing complex pre-
sumed to be the product. No evidence for the forma-
tion of any intermediate complex was observed, i.e. no
evidence for the 5-coordinate complex. A pure sample
of the yellow product complex 3a was however ob-
tained on cooling the original hexane washings to yield
yellow nodules. The yellow product was redissolved in
was not possible to obtain a sample for crystallographic
analysis. The H-NMR spectrum of such samples, indi-
cate that the reaction is not clean, hence the difficulty
on obtaining a crystalline sample. This is despite appar-
ently clean 31P-NMR spectra, e.g. Fig. 3C.
Finally another method available for trapping the
deligated complexes was to add sulfur to the solution
obtained on pyridine addition. Compound 2a in a
range of solvents may be thus trapped to provide
crystalline samples of the complex 7a. The pendant
phosphine sulfide shifts to lower field than than that of
the metal-bound phosphine. This can be seen in Fig.
3D.
1
1
d-chloroform and the H-NMR was re-recorded. Again
the starting complex was regenerated and the pyridine
was liberated to solution. Integration of the pyridine
resonances indicates the present of two pyridine
molecules per complex.
A solution of 2b in CDCl3 was treated with one drop
of 4-methylpyridine to effect delegation then a CDCl3
solution containing [Pd(1,5-COD)Cl2] was added drop-
wise. It was evident from the results obtained that the
pendant phosphine was co-ordinated to the palladium
centre; the free phosphine resonance shifted downfield
to +42.37 ppm from −0.84 ppm while the remaining
resonances shifted only slightly from +30.44 to
+30.62 ppm and from +44.54 to +44.48 ppm,
respectively.
The product complex 6b is red in contrast to the
yellow 5b. A similar result was observed for the pyri-
dine complex 5b. Although it was possible to isolate
samples of 6b directly by precipitation on attempted
recrystallisation partial decomposition occurred and it
The phosphine, which is unbound is thus oxidised
and is not able to reco-ordinate on removal of pyridine.
This complex should be able to react with a range of
bidentate ligand to displace the pyridine groups and
thus extend the synthetic possibilities. This, in part, will
be the subject of the next paper in this series.
Returning to the quest for absolute structural infor-
mation regarding compounds 5a and 5b, it was subse-
quently found it was possible to crystallise compound
5a, which is slightly less soluble in organic solvents.
Crystals could be obtained from dichloromethane–hex-
ane solutions or from acetone. A crystal of the product
5a, which was obtained from acetone, was subjected to
single crystal X-ray diffraction and the results are dis-
cussed below.
Fig. 3. The 31P-NMR spectra of (A) complex 2a, (B) complex after addition of pyridine-formation of 5a, (C) the complex after addition of
Pd(COD)Cl2 and (D) after the addition of excess elemental sulfur (formation of 7a).

I.R. Butler et al. / Journal of Organometallic Chemistry 637–639 (2001) 538–548
545
be closer, and hence bind more strongly, to the central
metal acceptor. In addition the absence of other com-
peting p-bonding ligands, such as the carbonyl group in
3c, will cause a shortening of the phosphine–metal
bond. The remaining bond lengths about the Ru centre
are in accordance with those previously observed [11–
17].
The geometric parameters observed for the coordi-
nated part of the {Ph2P(C5H4)2Fe}2PPh ligand are sim-
ilar to those previously observed in 3c. The orientation
of the ferrocenyl moiety involved in the chelate (Fe(1))
is very similar to that observed in 3c, with Ru(1)–P(1)–
C(23)–Fe(1)=40.4(6)°
and
Ru(1)–P(2)–C(32)–
Fe(1)=19.6(5)° [41.3(9) and 19.6°, respectively, in 3c].
This indicates a preferred orientation of the ferrocene
group within this ligand when it bonds in a chelating
mode. In contrast however, the pendant ferrocenyl
donor arm (Fe(2)) has a torsion angle, Ru(1)–P(2)–
C(39)–Fe(2)=162.2(6)°, indicating complete freedom
to adopt the most energetically favourable orientation.
The angles about the uncoordinated phosphorus (P(3))
are all approximately 100°, which is typical of the
pyramidal geometry adopted by free phosphines with
relatively bulky functionality.
3.4. Electrochemistry
Fig. 4. Crystal structure of compound 5a.
Figs. 5 and 6 summarize the changes in cyclic and
differential pulse voltammetric responses recorded upon
adding pyridine to dichloromethane solutions of com-
plexes 2a and 2b, respectively.
3.3. X-ray structure of [P2-(P3 fc)2Ru(pyr)2Cl2] (5a)
The molecular structure and atomic labelling scheme
of non-carbon atoms of 5a are depicted in Fig. 4. To
further increase the clarity of the figure the three ace-
tone solvent molecules are omitted. Selected molecular
geometry parameters are given in Table 2.
Considering Fig. 5 initially, responses a, b, which
show the electrochemical behaviour of 2a before addi-
tion of pyridine, show two first reversible oxidations
already discussed in a previous paper [2] which are
assigned to the sequential one-electron oxidation of the
two ferrocene subunits (E%= +0.58 V and +0.85 V,
respectively). We now confidently assign the third most
anodic step (E°%= +1.23 V) to the Ru(II)/Ru(III) re-
dox change, which is followed by chemical complica-
tions (ipc/ipa=0.6 at 0.1 V s−1). As deducible from
responses c and d, a dramatic change takes place upon
addition of pyridine. The Ru(II)/Ru(III) oxidation
shifts by about 0.2 V towards less positive potential
values (E°%= +1.02 V). Concomitantly, the two fer-
rocene-centred oxidations tend to become irreversible.
Taking into account that the diphenylphosphine-ligand,
in the absence of the RuCl2 coordination exhibits an
irreversible two-electron oxidation, whereas in the pres-
ence of the ferrocenophane-type binding by RuCl2 two
reversible one-electron oxidations occur, we assign the
behaviour of 2a in the presence of pyridine to the
release by the Ru(II) centre of one of the phosphino-cy-
clopentadienyl ligands, in agreement with the equi-
librium shown in Scheme 2, which in the present case
must clearly lie to the right.
The structure of 5a is comprised of a similar PPP
ligand, with two ferrocenyl units, coordinated to an
octahedral ruthenium centre as that of 3c (vide supra).
However in this instance the third phosphorus coordi-
nation site about the ruthenium has been displaced and
is occupied by a pyridyl ligand, thus making the PPP-
2fc ligand bidentate. The coordination environment of
the Ru centre is very similar to that of 3c, with the two
chloride ligands trans to each other in axial positions
and two phosphorus donors adjacent to each other in
the equatorial plane but with the remaining two equa-
torial positions occupied by two pyridyl ligands as
opposed to the third phosphorus and a carbonyl moi-
ety. The coordination geometry is slightly distorted
octahedral. This is much closer to idealised geometry
than the complex 3c (see angles about Ru(1) in Table
1), which is due to the greater strain on the coordina-
tion sphere imposed by binding the third chelate arm.
The Ru–P distances are significantly shorter than those
in 3c, which is due to the reduced steric influence of a
bidentate versus a tridentate ligand allowing donors to

546
I.R. Butler et al. / Journal of Organometallic Chemistry 637–639 (2001) 538–548
In the case of complex 2b, responses a, b parallel
those discussed above with two first ferrocene-centred
oxidations (E°%= +0.29 and +0.85 V, respectively)
followed by the Ru(II)/Ru(III) oxidation (E°%= +1.12
V), complicated by following reactions (ipc/ipa=0.5 at
0.1 V s−1). As a consequence of the electron-donating
i
ability of the Pr phosphino-substituents, all the oxida-
tion processes are easier than in complex 2a. As re-
sponses c and d show, addition of pyridine slightly
perturbs the voltammetric patterns. Essentially, the
Ru(II)/Ru(III) oxidation splits into two parts, one of
which is shifted towards more positive potential values
by about 0.1 V (E%= +1.03 V). The two ferrocene
oxidations maintain their features of chemical re-
versibility, with a very slight shift (because of the more
emphasised roundness in DPV peaks), if any, towards
more positive potential values. If an allowance is made
Fig. 6. Cyclic and differential pulse voltammograms recorded at
platinum electrode in CH2Cl2 solution containing 2b (4.5×10−4 mol
dm−3) and NBu4PF6 (0.2 mol dm−3). (a, b) Before addition of
pyridine; (c, d) after addition of pyridine in bout 1:2 (2b:pyridine)
ratio. Scan rates: (a, c) 0.004 V s−1; (b, d) 0.1 V s−1
.
to assign the partial shift of the Ru-centred oxidation to
the pyridine coordination, the results show the equi-
librium shown in Scheme 3 lies more to the left in this
case.
In conclusion, electrochemical measurements indicate
that the nature of the electron-donating groups in the
phosphine ligands influences pyridine attachment to the
ruthenium(II) centre, which is favoured by electron-
withdrawing groups.
3.5. Molecular modelling
Molecular dynamics techniques have been used to
study the influence of the addition of a second pyridine
ligand on the conformation of the metal complexes 3a
and 3b. Fig. 7 shows the conformation of the com-
pound a as well as the conformation after addition of
CO or pyridine. Geometry optimization was performed.
A simple molecular modelling investigation was carried
out on complexes 2a and 3b. The carbonyl complex 3c,
reported previously [2], was modelled and it was found
that the geometry observed on the crystallographic
study concurred with the minimum energy conforma-
tion predicted by the computer model (Fig. 7).
Fig. 5. Cyclic and differential pulse voltammograms recorded at a
platinum electrode in CH2Cl2 solution containing compound 2a
−4
(3.5×10
mol dm −3) and NBu4PF6 (0.2 mol dm−3). (a, b)
Before addition of pyridine; (c, d) after addition of pyridine in about
1:2 (17:pyridine) ration. Scan rates: (a, c) 0.004 V s−1 (b, d) 0.1 V
s−1
.

I.R. Butler et al. / Journal of Organometallic Chemistry 637–639 (2001) 538–548
547
Scheme 2. Trapping the deligated phosphine with Pd(II).
Scheme 3. Trapping the deligated phosphine with elemental sulfur.
Fig. 7. The steric congestion is apparent on attempted pyridine coordination complex 2 and the deligation of a phosphine in complex 2a showing
the minimised conformation of complex 5a.
Interestingly, when it was attempted simulate the
binding of the pyridine to the five-coordinate complex
it was clearly obvious that, due to the steric crowding,
the six-coordinate complex was unstable. If a phos-
phine was deligated the pyridine was able to bind
readily and indeed it was possible to coordinate a
second pyridine. It became clear that before deligation
takes place, a minimum concentration of pyridine in
solution is required, i.e. equilibrium according to
Scheme 1 exists between the pyridine and the deligated
complex. However no intermediate was detected using
NMR, although both products 5 and the starting com-
plex 2 may be observed in solution simultaneously.
In conclusion the reaction chemistry of these ruthe-
nium complexes is non-trivial: the reaction chemistry is
dependent on the nature of the incoming ligand. With
non-sterically demanding p-acceptor ligands the vacant
sixth coordination site is filled, while with weaker field
ligands partial deligation of one of the phosphines is
also observed. Finally non-sterically demanding p-
donor ligands lead to only an equilibrium where phos-
phine deligation is observed. Clearly the coordination
chemistry should be further studied with other sub-
strates and its catalytic reactions should investigated be
a matter of priority. The electrochemical results also
indicate the importance of the nature of the phosphine
ligands in solution equilibria. These studies will be the
subject of the next paper in the series.
4. Supplementary material
Crystallographic data for the structural analysis have
been deposited with the Cambridge Crystallographic

548
I.R. Butler et al. / Journal of Organometallic Chemistry 637–639 (2001) 538–548
(b) C. Bianchini, P.J. Perez, M. Peruzzini, F. Zanobini Avacca,
Inorg. Chem. 30 (1991) 279.
[4] J.A. Darr, S.R. Drake, M.B. Hursthouse, K.M.A. Malik, Inorg.
Chem. 32 (1993) 5704.
[5] N.P.C. Walker, D. Stuart, DIFABS Program for Absorption
Correction, Acta Crystallogr. Sect. A 39 (1983) 158; adapted for
FAST geometry by A.I. Karaulov, University of Wales Cardiff,
1991.
Data Centre, CCDC Nos. 154080 for 3c and 154081 for
5b. Copies of this information may be obtained free of
charge from The Director, CCDC, 12 Union Road,
Cambridge CB2 1EZ, UK (fax: +44-1233-336033; e-
mail: deposit@ccdc.cam.ac.uk or www: http://www.
ccdc.cam.ac.uk).
[6] G.M. Sheldrick, SHELXS-86, Program for Crystal Structure Solu-
tion, Acta Crystallogr. Sect. A 46 (1990) 467.
Acknowledgements
[7] G.M. Sheldrick, SHELXL-93, Program for Crystal Structure
Refinement, University of Go¨ttingen, Germany, 1993.
[8] R. Hooft, B.V. Nonius, COLLECT: data collection software, 1998.
[9] R.H. Blessing, J. Appl. Crystallogr. 30 (1997) 421.
[10] G.M. Sheldrick, SHELXL-97, Program for Crystal Structure Solu-
tion and Refinement, University of Gottingen, Germany, 1997.
[11] A. Albinati, Q. Zhong Jiang, H. Ruegger, L.M. Venanzi, Inorg.
Chem. 32 (1993) 4940.
[12] R.R. Guimerans, E.C. Hernandez, M.M. Olmstead, A.L. Balch,
Inorg. Chim. Acta 165 (1989) 45.
[13] G. Delgado, A.V. Rivera, T. Suarez, B. Fontal, Inorg. Chim.
Acta 233 (1995) 145.
IRB gratefully acknowledges the support of the Uni-
versity of Wales for sponsoring this research. Mass
spectroscopic data were obtained at the University of
Wales, Swansea, central EPSRC service whose aid we
gratefully acknowledge. PZ gratefully acknowledges the
financial support of the University of Siena (PAR
1999). A range of supporting data and figures may be
obtained on request (authors@ferrocene.com).
[14] L. Solujic, E.B. Milosavijevic, J.H. Nelson, N.W. Alcock, J.
Fischer, Inorg. Chem. 28 (1989) 3453.
[15] W.S. Sheldrick, K. Brandt, Inorg. Chim. Acta 217 (1994) 51.
[16] G. Jia, I. Lee, D.W. Meek, J.C. Gallucci, Inorg. Chim. Acta 177
(1990) 81.
[17] I. Dahlenburg, S. Kerstan, D. Werner, J. Organomet. Chem. 411
(1991) 457.
[18] K. Davies, SNOOPI, Program for Crystal Structure Drawing,
University of Oxford, UK, 1983.
References
[1] I.R. Butler, R.L. Davies, Synthesis (1996) 1350.
[2] I.R. Butler, U. Griesbach, P. Zanello, D.E. Hibbs, M.B. Hurst-
house, K.M.A. Malik, J. Organomet. Chem. 565 (1998) 243.
[3] (a) M.M. Taquikhan, R. Mohiuddina, J. Coord. Chem. 6 (1977)
171;