Synthesis and reactivity studies of the diphosphine ligand 2-(ferrocenylidene)-4,5-bis(diphenylphosphino)-4-cyclopenten-1,3-dione (fbpcd): Bridge-to-chelate ligand isomerization kinetics in Os3(CO)10(fbpcd) and X-ray diffraction structure of the ortho-metalated cluster HOs3(CO)9[μ-PhP(C6H4)C{do uble bond, long}C...

Article abstract of DOI:10.1016/j.poly.2007.03.057

Full title: Synthesis and reactivity studies of the diphosphine ligand 2-(ferrocenylidene)-4,5-bis(diphenylphosphino)-4-cyclopenten-1,3-dione (fbpcd): Bridge-to-chelate ligand isomerization kinetics in Os₃(CO)₁₀(fbpcd) and X-ray diffraction structure of the ortho-metalated cluster HOs₃(CO)₉[μ-PhP(C₆H₄)C{do uble bond, long}C(PPh₂)C(O)C{double bond, long}CH(C₅H₄FeCp)C(O)] Knoevenagel condensation of ferrocenecarboxaldehyde with 4,5-bis(diphenylphosphino)-4-cyclopenten-1,3-dione (bpcd) yields the new diphosphine ligand 2-(ferrocenylidene)-4,5-bis(diphenylphosphino)-4-cyclopenten-1,3-dione (fbpcd) in near quantitative yield. The reaction of fbpcd with the activated cluster 1,2-Os₃(CO)₁₀(MeCN)₂ has been examined, with the diphosphine-bridged cluster 1,2-Os₃(CO)₁₀(fbpcd) (2b) observed as the major product. Cluster 2b is unstable and transforms into the corresponding chelating isomer 1,1-Os₃(CO)₁₀(fbpcd) (2c) on heating. The kinetics associated with the conversion of 2b → 2c have been studied over the temperature range of 313-343 K by ¹H NMR and UV-vis spectroscopies. On the basis of the observed activation parameters, a nondissociative isomerization process that involves a transient μ₂-bridged phosphine moiety is presented. Near-UV irradiation of cluster 2c leads to CO loss and ortho-metalation of the one of the ancillary phenyl groups to give the hydride cluster HOs₃(CO)₉[μ-PhP(C₆H₄)C{do uble bond, long}C(PPh₂)C(O)C{double bond, long}CH(C₅H₄FeCp)C(O)] (3) as a 45:55 mixture of diastereomers. The fbpcd ligand and clusters 2b, 2c, and 3 have been isolated and characterized in solution by IR and NMR (¹H and ³¹P) spectroscopies, and by X-ray diffraction analysis in the case of the hydride-bridged cluster 3.

Full text of DOI:10.1016/j.poly.2007.03.057

Polyhedron 26 (2007) 3585–3594
www.elsevier.com/locate/poly
Synthesis and reactivity studies of the diphosphine ligand
2-(ferrocenylidene)-4,5-bis(diphenylphosphino)-4-cyclopenten-1,3-dione
(fbpcd): Bridge-to-chelate ligand isomerization kinetics
in Os3(CO)10(fbpcd) and X-ray diffraction structure of the
ortho-metalated cluster HOs3(CO)9[l-PhP(C6H4)-
C@C(PPh2)C(O)C@CH(C5H4FeCp)C(O)]
a,
b
b,
*
William H. Watson *, Bhaskar Poola , Michael G. Richmond
a
Department of Chemistry, Texas Christian University, Fort Worth, TX 76129, USA
b
Department of Chemistry, University of North Texas, Denton, TX 76203, USA
Received 10 January 2007; accepted 25 March 2007
Available online 13 April 2007
Abstract
Knoevenagel condensation of ferrocenecarboxaldehyde with 4,5-bis(diphenylphosphino)-4-cyclopenten-1,3-dione (bpcd) yields the
new diphosphine ligand 2-(ferrocenylidene)-4,5-bis(diphenylphosphino)-4-cyclopenten-1,3-dione (fbpcd) in near quantitative yield.
The reaction of fbpcd with the activated cluster 1,2-Os3(CO)10(MeCN)2 has been examined, with the diphosphine-bridged cluster 1,2-
Os3(CO)10(fbpcd) (2b) observed as the major product. Cluster 2b is unstable and transforms into the corresponding chelating isomer
1,1-Os3(CO)10(fbpcd) (2c) on heating. The kinetics associated with the conversion of 2b!2c have been studied over the temperature
range of 313-343 K by 1H NMR and UV–vis spectroscopies. On the basis of the observed activation parameters, a nondissociative isom-
erization process that involves a transient l2-bridged phosphine moiety is presented. Near-UV irradiation of cluster 2c leads to CO loss
and ortho-metalation of the one of the ancillary phenyl groups to give the hydride cluster HOs3(CO)9[l-PhP(C6H4)C@C(PPh2)-
C(O)C@CH(C5H4FeCp)C(O)] (3) as a 45:55 mixture of diastereomers. The fbpcd ligand and clusters 2b, 2c, and 3 have been isolated
and characterized in solution by IR and NMR (1H and 31P) spectroscopies, and by X-ray diffraction analysis in the case of the
hydride-bridged cluster 3.
Ó 2007 Elsevier Ltd. All rights reserved.
Keywords: Diphosphine ligand; Isomerization kinetics; Osmium clusters; C–H Bond activation; ortho-Metalation; Photochemistry
1. Introduction
sphino)-N-p-tolylmaleimide (bmi), and 4,5-bis(diphenyl-
phosphino)-4-cyclopenten-1,3-dione (bpcd) at polynuclear
metal clusters have remained under active investigation by
our groups [1–3]. These particular diphosphine ligands
have provided crucial insight into the thermally and photo-
chemically promoted C–H and P–C bond activation
manifolds exhibited by such diphosphines at polynuclear
metal clusters. Moreover, the same ligands have also been
shown to undergo nondissociative isomerization about
various cluster polyhedra via l2-phosphine moieties in a
The coordination chemistry and redox reactivity of
the rigid, unsaturated diphosphine ligands 2,3-bis(diphenyl-
phosphino)maleic anhydride (bma), 2,3-bis(diphenylpho-
*
Corresponding authors. Tel.: +1 817 257 7195 (W.H. Watson); tel.: +1
940 565 3548 (M.G. Richmond).
E-mail addresses: w.watson@tcu.edu (W.H. Watson), cobalt@unt.edu
(M.G. Richmond).
0277-5387/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.poly.2007.03.057

3586
W.H. Watson et al. / Polyhedron 26 (2007) 3585–3594
manner analogous to that displayed by CO ligands [4]. The
structures of these three diphosphine ligands are shown
below.
Herein we present our results on the condensation of
bpcd with ferrocenecarboxaldehyde to give the new diphos-
phine ligand 2-(ferrocenylidene)-4,5-bis(diphenylphosph-
ino)-4-cyclopenten-1,3-dione (fbpcd). The reactivity of the
fbpcd ligand with the triosmium cluster 1,2-Os3(CO)10-
(MeCN)2 has been examined, and the kinetics for the iso-
merization of the fbpcd ligand in 1,2-Os3(CO)10(fbpcd)
(2b) to 1,1-Os3(CO)10(fbpcd) (2c) are discussed. Near-UV
irradiation of cluster 2c leads to CO loss and production
of the hydride-bridged cluster HOs3(CO)9[l-PhP(C6H4)-
C@C(PPh2)C(O)C@CH(C5H4FeCp)C(O)] (3). The ortho-
metalation of one of the aryl groups of the fbpcd ligand
in 3 has been structurally confirmed.
O
N
O
O
O
O
O
O
Ph2P
PPh2
Ph2P
PPh2
Ph2P
PPh2
bma
bmi
bpcd
2. Experimental
Wishing to explore further the chemistry of organometal-
lic compounds with structurally more diverse diphosphine
ligands that possess highly conjugated p arrays and well-
defined redox and photochemically sensitive moieties, we
have instituted studies on the modification of the bpcd
ligand. The successful synthesis of such second-generation
diphosphine ligands would allow us the opportunity to study
heretofore novel cluster-ligand systems as redox relays,
energy-storage devices, and photochemical and/or redox
sensing components [5,6]. Early results using bpcd as a plat-
form for the construction of new diphosphine ligands have
been encouraging, and significant structural modifications
of the parent diphosphine with the aldehydes 9-anthracene-
carboxaldehyde and thiophene-2-carboxaldehyde have
produced the new ligands 2-(anthracenyl-9-ylidene)-
4,5-bis(diphenylphosphino)-4-cyclopenten-1,3-dione and
2-(2-thienylidene)-4,5-bis(diphenylphosphino)-4-cyclopen-
ten-1,3-dione, respectively, through a Knoevenagel conden-
sation route [7,8]. Scheme 1 depicts the synthetic protocol
and structures of these new diphosphine ligands.
2.1. General
The starting cluster 1,2-Os3(CO)10(MeCN)2 was pre-
pared from Os3(CO)12 [9], which was in turn was obtained
from the carbonylation of OsO4 using a 450 mL Parr Ser-
ies 4560 bench-top mini reactor [10]. The bpcd ligand was
synthesized from hexachlorocyclopentadiene [11]. The
chemicals hexachlorocyclopentadiene, ferrocenecarboxal-
dehyde, and Me3NO Æ 2H2O were purchased from Aldrich
Chemical Co. The latter chemical was dried by azeotropic
distillation from benzene and stored under argon when
not in use. All reaction and NMR solvents were distilled
from an appropriate drying agent using Schlenk tech-
niques and stored under argon in storage vessels equipped
with high-vacuum Teflon stopcocks [12]. The combustion
analysis was performed by Atlantic Microlab, Norcross,
GA.
All IR spectra were recorded on a Nicolet 20 SXB FT-
IR spectrometer in sealed 0.1 mm NaCl cells, while the
Ph2P
O
PPh2
bpcd
O
H
O
H
molecular sieves
CH2Cl2/-H2O
O
S
O
O
Ph2P
Ph2P
Ph2P
Ph2P
S
O
O
Scheme 1.

W.H. Watson et al. / Polyhedron 26 (2007) 3585–3594
3587
1H spectra were recorded at 200 MHz on a Varian Gemini-
2.4. Thermal isomerization of 2b to 2c
200 spectrometer. The 31P NMR spectra were collected in
the proton-decoupled mode on a Varian 300-VXR spec-
trometer at 121 MHz, with the reported chemical shifts
referenced to external H3PO4 (85%), taken to have d = 0.
The MALDI-TOF mass spectra for the fbpcd ligand and
cluster 3 were recorded at the University of the Pacific in
the positive ion mode.
To a Schlenk vessel under argon was charged 0.10 g of
the previous cluster mixture, followed by 25 mL of toluene
via cannula. The solution was next saturated with CO,
after which the vessel was sealed and removed from the
vacuum line and placed in a thermostated bath at 315 K.
The solution was heated for several days and the progress
of the reaction was periodically checked by NMR spectros-
copy, which revealed the complete consumption of 2b and
a concomitant increase in the amount of cluster 2c and only
a trace amount of cluster 3. Heating the same sample for an
additional 45 days showed only the presence of cluster 2c.
The solvent was removed and cluster 2c was run across a
short column of silica gel and then recrystallized from
benzene/CH2Cl2. Yield: 95 mg (95%). IR (CH2Cl2):
m(CO) 2095 (s), 2046 (s), 2010 (vs), 1975 (s), 1961 (m),
1926 (m), 1719 (w, symm dione), 1675 (m, antisymm dione)
2.2. Synthesis of the diphosphine ligand fbpcd
To a Schlenk tube containing 0.20 g (0.43 mmol) of
bpcd dissolved in 30 mL of CH2Cl2 at room temperature
was added 92 mg of ferrocenecarboxaldehyde (0.43 mmol)
˚
and ca. 5 g of 4 A molecular sieves. The introduction of the
molecular sieves was accompanied by an immediate color
change from the orange–yellow color of the two starting
reagents to that of purple, signifying the presence of the
desired product. The vessel was sealed and stirring was
continued overnight. TLC examination of the reaction
solution the following day revealed the complete consump-
tion of the starting materials and the presence of a single
purple spot at Rf = 0.50 using CH2Cl2/hexane (1:1) as a
mobile phase. The crude reaction was filtered to remove
the molecular sieves and the solvent concentrated down
to ca. 0.5 mL, after which the product was run across a
short column of silica gel using a 40:60 mixture of
CH2Cl2/hexane as the eluent. Yield of purple fbpcd:
0.28 g (97%). IR (CH2Cl2): m(CO) 1717 (w, symm dione),
1
cmꢀ1. H NMR (toluene-d8): d 3.71 (s. 5H, Cp), 4.26 (br,
2H, Cp), 4.95 (br, 2H, Cp), 7.05–7.29 (m, 12H, aryl),
7.63 (s, 1H, vinyl), 7.83–8.00 (m, 8H, aryl). 31P NMR (tol-
uene-d8): d 19.8 (AB quartet, JP–P = 21 Hz). Anal. Calc. for
C50H30FeO12Os3P2 Æ CH2Cl2: C, 38.38; H, 2.02. Found: C,
37.99; H, 1.95%.
2.5. Photochemical preparation of 3 from 2c
To 50 mg (0.033 mmol) of 2c in a Schlenk tube was
added 20 mL of toluene, after which the vessel was
freeze–pump–thaw degassed three times and sealed. The
solution was irradiated at room temperature using two
GE Blacklights having a maximum output of 366 20 nm
for an extended period of time, with periodic removal of
the liberated CO through additional freeze–pump–degas
cycles. Depending on the thickness of the Schlenk tube, near
quantitative conversion to the desired hydride-bridged clus-
ter 3 was typically achieved after two weeks of continuous
irradiation. The sample of 3 was passed across a short col-
umn of silica gel using CH2Cl2 and then recrystallized from
benzene/CH2Cl2. Yield: 40 mg (82%). IR (CH2Cl2): m(CO)
2085 (vs), 2041 (vs), 2015 (vs), 1986 (vs), 1967 (sh), 1720
1671 (vs, antisymm dione) cmꢀ1 1H NMR (CDCl3): d
.
4.13 (s, 5H, Cp), 4.67 (br, 2H, Cp), 5.16 (br, 2H, Cp),
7.27–7.44 (m, 21H, aryls and alkenyl). 31P NMR (CDCl3):
3
d ꢀ23.49 (AB quartet, JP–P = 60 Hz). MALDI-TOF MS
(m/z): 660.7 [fbpcd]+.
2.3. Preparation of 1,2-Os3(CO)10(fbpcd) (2b) from
1,2-Os3(CO)10(MeCN)2 and fbpcd
To 0.25 g (0.27 mmol) of 1,2-Os3(CO)10(MeCN)2 in
30 mL of CH2Cl2 was added 0.18 g (0.27 mmol) of fbpcd
under argon. The solution was stirred at room tempera-
ture, and the progress of the reaction was monitored by
TLC analysis, which confirmed the consumption of both
starting materials after 3 h and the presence of a new dark
blue spot corresponding to the clusters 2b, 2c, and 3. The
solvent was evaporated under vacuum and the residue
passed over a short column of silica gel using a 70:30 mix-
ture of hexane/CH2Cl2 to furnish 0.30 g of a dark blue
solid. The composition of the isolated cluster mixture was
(w, symm dione), 1676 (m, antisymm dione) cmꢀ1 1H
.
NMR (toluene-d8): d ꢀ15.87 (t, hydride, JP–H = 14 Hz),
3.61 and 3.65 (s, 5H each, Cp), 4.19 and 4.32 (br, 2H each,
Cp), 4.56 and 4.85 (br, 1H each, Cp), 4.93 and 4.91 (br, 1H
each, Cp), 6.90–8.30 (overlapping m, aryls and vinyl). 31P
NMR (toluene-d8): d 21.40 (d, JP–P = 13 Hz), 22.11 (d,
JP–P = 13 Hz), 30.64 (t, JP–P = 13 Hz). MALDI-TOF MS
(m/z): 1457.7 [MꢀCO]+.
1
analyzed by using H and 31P NMR spectroscopies, which
revealed the product distribution: 2b (80%), 2c (19%), and 3
2.6. X-ray diffraction data for cluster 3
(1%).
1
NMR spectroscopic data for 2b: H NMR (toluene-d8):
Tables 1 and 2 contain the X-ray data and processing
parameters and selected bond distances and angles, respec-
tively, for cluster 3. Single crystals of 3 suitable for X-ray
diffraction analysis were grown at room temperature from
a CH2Cl2 solution containing 3 that had been layered with
d 3.54 (s, 5H, Cp), 4.20 (br, 2H, Cp), 4.64 (br, 2H, Cp),
6.95–7.25 (m, 12H, aryl), 7.55 (s, 1H, vinyl), 7.70–7.95
(m, 8H, aryl). 31P NMR (toluene-d8): d ꢀ18.0 (AB quartet,
JP–P = 17 Hz).

3588
W.H. Watson et al. / Polyhedron 26 (2007) 3585–3594
Table 1
using PLATON [15] and solved by direct methods with all
nonhydrogen atoms refined anisotropically. All carbon-
bound hydrogen atoms were assigned calculated positions
and allowed to ride on the attached heavy atom, unless
otherwise noted. The hydride ligand that bridges the
Os(1)–Os(2) vector could not be located and the presence
of a disordered pentane solvent molecule that resisted
refinement was also noted. Convergence for 3 was achieved
at R = 0.0455 and Rw = 0.0992 for 11383 independent
reflections with I > 2r(I).
X-ray crystallographic data and processing parameters for cluster 3
CCDC Entry no.
Crystal system
Space group
628577
triclinic
ꢀ
P1
˚
a (A)
11.275(3)
14.681(4)
16.557(4)
100.085(4)
91.507(4)
101.870(4)
2635(1)
˚
b (A)
˚
c (A)
a (°)
b (°)
c (°)
3
˚
V (A)
Molecular formula
Fw
C49H30FeO11Os3P2 + disordered C5H12
1555.26
2
1.959
0.71073
7.598
2.7. Kinetic studies
Formula units per cell (Z)
Dcalc (Mg/m3)
The UV–vis studies were carried out using a mixture of
clusters 2b and 2c at a concentration of ca. 10ꢀ4 M using
1.0 cm quartz UV–vis cells that were equipped with a
high-vacuum Teflon stopcock to facilitate handling on
the vacuum line. Solutions containing the isomeric mix-
ture of clusters 2b and 2c were prepared under 1 atm
CO and used immediately before each kinetic measure-
ment. The role of CO is to suppress the formation of
the hydride-bridged cluster 3 from cluster 2c at elevated
temperature. The Hewlett–Packard 8452A diode array
spectrometer employed in our studies was configured with
a variable-temperature cell holder and was connected to a
VWR constant temperature circulator, allowing for the
quoted temperatures to be maintained within 0.5 K.
The UV–vis kinetics for the transformation of 2b!2c
were monitored by following the increase of the 360 nm
absorbance band as a function of time typically for 4–6
half-lifes. The rate constants quoted in Table 3 were deter-
mined by non-linear regression analysis using the single
exponential function1:
˚
k (Mo Ka) ( A)
l (mmꢀ1
Rmerge
)
0.0603
Absorption correction
Absorption correction factor
Total reflections
Independent reflections
Data/restraints/parameters
R
empirical
0.9223/0.2490
21587
11383
11383/0/617
0.0455
0.0992
0.959
2.503, ꢀ1.662
Rw
Goodness-of-fit on F2
Dq(max), Dq(min) (e/A )
3
˚
Table 2
˚
Selected bond distances (A) and angles (°) in cluster 3
Bond distances
Os(1)–Os(3)
Os(2)–Os(3)
Os(1)–P(2)
Os(2)–C(24)
C(1)–C(5)
2.9132(7)
2.8666(8)
2.305(2)
2.165(9)
1.54(1)
Os(1)–Os(2)
Os(1)–P(1)
P(1)ꢂ ꢂ ꢂP(2)
C(1)–C(2)
2.9441(8)
2.365(2)
3.184(3)
1.32(1)
C(2)–C(3)
1.50(1)
C(3)–C(4)
C(4)–C(5)
Cp(centroid)–Fea
1.47(1)
1.45(1)
1.646(4)
C(4)–C(6)
C(6)–C(7)
Cp0(centroid)–Feb
1.37(1)
1.44(1)
1.655(6)
AðtÞ ¼ A1 þ DA ꢁ eðꢀktÞ
1
The H NMR kinetic experiment was conducted under
Bond angles
C(41)–Os(1)–C(42)
C(42)–Os(1)–P(2)
C(42)–Os(1)–P(1)
C(42)–Os(1)–Os(3)
P(1)–Os(1)–Os(3)
P(1)–Os(1)–Os(2)
C(43)–Os(2)–Os(1)
C(1)–P(1)–Os(1)
C(6)–C(4)–C(5)
C(5)–C(4)–C(3)
C(24)–C(23)–P(1)
a
92.4(4)
94.7(3)
96.7(3)
C(41)–Os(1)–P(2)
C(41)–Os(1)–P(1)
P(2)–Os(1)–P(1)
P(2)–Os(1)–Os(3)
P(2)–Os(1)–Os(2)
C(44)–Os(2)–Os(3)
C(23)–P(1)–Os(1)
C(2)–P(2)–Os(1)
C(6)–C(4)–C(3)
C(4)–C(6)–C(7)
87.0(3)
168.9(3)
85.95(7)
172.47(6)
120.63(5)
87.1(3)
115.0(3)
107.3(3)
131.2(7)
133.0(8)
CO (1 atm) in a 5 mm NMR tube possessing a J. Young
valve and was monitored for slightly over 2 half-lifes.
The concentration of the initial ca. 70/30 mixture of 2b/
2c was ca. 10ꢀ2 M and the solvent used was toluene-d8.
The extent of the reaction was determined by following
the decrease in the intensity of the Cp singlet of 2b at d
3.54, as quantified against the methoxy resonance of the
internal standard p-dimethoxybenzene.
The activation parameters for the isomerization of 2b to
2c were calculated from a plot of ln(k/T) versus Tꢀ1 [16],
with the error limits representing deviation of the data
points about the least-squares line of the Eyring plot.
86.5(3)
86.52(5)
86.66(5)
117.2(3)
105.1(3)
121.0(7)
107.8(6)
119.7(6)
Cp centroid defined by the atoms C(7)–C(11).
Cp0 centroid defined by the atoms C(12)–C(16).
b
pentane. The reported X-ray data were collected on a Bru-
ker SMARTä 1000 CCD-based diffractometer at 211 K. The
frames were integrated with the available SAINT software
package using a narrow-frame algorithm, [13] and the
structure was solved and refined using the SHELXTL program
package. [14] The molecular structure was checked by
1
The rate calculations were performed by using the Agilent ChemSta-
tion software package configured to the spectrometer or the commercially
available program ORIGIN6.0. Here the initial (A0) and final (A
absorbances and the rate constant (k) were floated to give the quoted
least-squares value for first-order rate constant k.
)
1

W.H. Watson et al. / Polyhedron 26 (2007) 3585–3594
3589
ing to the dione moiety at 1717 and 1671 cmꢀ1 [17]. These
carbonyl stretching bands are 25 and 34 cmꢀ1 lower in
energy, respectively, relative to the m(CO) bands of the
starting bpcd ligand [11,18], supporting the fact that the
fbpcd ligand is more electron rich than the parent diphos-
phine ligand. The 1H NMR spectrum confirmed the loss of
the methylene hydrogens of the bpcd ligand and the formyl
hydrogen of the ferrocenecarboxaldehyde substrate. A sin-
glet at d 4.13 (5H) and two broadened resonances at d 4.67
and 5.16, each of which integrates for 2H, are readily
assigned to the intact ferrocene Cp ring and the functional-
ized Cp ring that is attached to the dione platform. The
vinyl hydrogen associated with the ferrocenylidene moiety
and the aromatic hydrogens from the four phenyl groups
appear as a multiplet from d 7.27–7.44. The two inequiva-
lent 31P groups in the fbpcd ligand appear as a classic AB
quartet centered at d ꢀ23.49 with a J/Dm value of 0.30 [19].
The MALDI-TOF mass spectrum of the diphosphine
ligand revealed a strong m/z peak at 660.7 for the singly
charged parent [fbpcd]+ ion consistent with the proposed
formulation of the fbpcd ligand.
Table 3
Experimental rate constants for the isomerization of 2b to 2ca
Entry number
Temperature (K)
Solvent
105k (sꢀ1
)
1
2
3
4
5
6
7
8
9
a
313.0
318.0
323.0
323.0
323.0
328.0
333.0
338.0
343.0
toluene
toluene
toluene
toluene
toluene-d8
toluene
toluene
toluene
toluene
2.07 0.04
3.21 0.08
4.90 0.09
5.9 0.1b
7.0 0.5c
9.8 0.1
18.5 0.4
29
1
54.1 0.2
The UV–vis kinetic data were collected in ca. 10ꢀ4 M solution con-
taining a ca. 70/30 mixture of clusters 2b and 2c under 1 atm CO by
following the increase in the absorbance of the 360 nm band.
b
Reaction performed in the presence of PPh3 (25 equivalents).
Reaction analyzed by 1H NMR spectroscopy using ca. 10ꢀ2 M solu-
c
tion containing a ca. 70/30 mixture of 2b/2c.
3. Results and discussion
3.1. Synthesis and spectroscopic characterization of the
diphosphine ligand fbpcd
3.2. Reaction of 1,2-Os3(CO)10(MeCN)2 with the ligand
fbpcd
Knoevenagel condensation of ferrocenecarboxaldehdye
with the diphosphine ligand bpcd proceeds rapidly in
CH2Cl2 at room temperature in the presence of added
Treatment of the activated cluster 1,2-Os3(CO)10-
(MeCN)2 (1) with an equimolar amount of the fbpcd ligand
at 293 K in CH2Cl2 leads to the rapid consumption of 1
and the fbpcd ligand (<1 h) and formation of a new dark
blue spot (Rf = 0.60 using a 1:1 mixture of CH2Cl2/hexane
as the eluent), as determined by TLC analysis. NMR
analysis (1H and 31P) of the reaction solution before a for-
mal work-up revealed the presence of three species assign-
able to 1,2-Os3(CO)10(fbcpd) (2b), 1,1-Os3(CO)10(fbpcd)
(2c), and HOs3(CO)9[l-PhP(C6H4)C@C(PPh2)C(O)C@CH-
(C5H4FeCp)C(O)] (3). These three products showed no dis-
cernable separation on silica gel and neutral alumina when
analyzed with sundry different mobile phases. The diphos-
phine-bridged cluster 2b is formed as the major product in
ca. 80% yield, with cluster 2c present in ca. 18–19% and the
hydride-bridged cluster 3 accounting for the reminder of
the ternary mixture. The observed product distribution is
reminiscent of that found by us in the reaction between
cluster 1 and the related diphosphine ligand bmi [2a]. These
data are in accord with cluster 1 serving as an independent
platform for the formation of clusters 2b and 2c and not a
process involving the isomerization of 2b to 2c since our
extrapolated rate data for such an isomerization at 293 K
would occur with a half-life on the order of 42 h (vide
infra). The bimodal substitution behavior exhibited in the
reaction of cluster 1 and fbpcd is not without precedent,
as we have found that the reaction of cluster 1 with the
diphosphine ligand 1,2-bis(dimethylphosphino)ethane
(dmpe) affords the isomeric clusters 1,2-Os3(CO)10(dmpe)
and 1,1-Os3(CO)10(dmpe) as a noninterconverting mixture
[20]. Scheme 2 illustrates the course of this substitution
reaction using the fbpcd ligand.
˚
molecular sieves (4 A) to afford the new diphosphine ligand
2-(ferrocenylidene)-4,5-bis(diphenylphosphino)-4-cyclo-
penten-1,3-dione (fbpcd) in 97% yield. The role of the
molecular sieves is to drive the condensation reaction to
completion through the removal of the accompanying
H2O. This latter aspect was verified by a control experi-
ment where the ferrocenecarboxaldehdye and bpcd
reagents were stirred for a period of several days in the
absence of molecular sieves. No fbpcd ligand was observed,
as assessed by TLC and IR analyses. However, the addition
of the molecular sieves to this solution led to the immediate
formation of the desired ligand. The fbpcd ligand was sub-
sequently isolated as an air-sensitive purple solid after
chromatographic separation over silica using CH2Cl2/
petroleum ether (3:2) as the eluent. The following equation
illustrates the reaction leading to the fbpcd ligand.
O
H
Fe
O
Ph2P
O
PPh2
O
Ph2P
Ph2P
molecular sieves
CH2Cl2/-H2O
O
Fe
bpcd
fbpcd
ð1Þ
The fbpcd ligand was fully characterized in solution by
IR and NMR spectroscopies, and by MALDI-TOF mass
spectrometry. The IR spectrum showed the vibrationally
coupled symmetric and antisymmetric m(CO) bands belong-

3590
W.H. Watson et al. / Polyhedron 26 (2007) 3585–3594
Os
MeCN
cluster 1
Os
Os
NCMe
fbpcd
CH2Cl2
H
O
Fc
Ph2
P
Os
H
Os
Os
O
Ph2
P
O
Os
Os
PPh2
Ph2P
H
Os
Os
Os
Os
P
P
O
Ph
Ph2
Fc
O
H
O
cluster 2b
cluster 3
cluster 2c
Fc
Scheme 2.
Given the nature of the product mixture and the fact
that both 2b and 2c possess overlapping carbonyl bands,
the IR spectral data for pure 2b cannot be reported with
certainty.2 Fortunately, the NMR data for 2b could be
extracted from the crude reaction mixture due to sufficient
as part of our interest in the coordinative flexibility exhib-
ited by rigid unsaturated diphosphine ligands [2a,3c,22].
Heating the initially formed cluster mixture under 1 atm
1
CO produced clear spectroscopic changes in the H and
31P NMR spectra in concert with the conversion of 2b
and 3 to the chelating cluster 2c. Provided these reactions
were conducted in the dark or under red-light conditions,
the formation of 2c from 2b was found to be quantitative.
Pure 2c was subsequently isolated via recrystallization and
characterized in solution by traditional spectroscopic meth-
ods and combustion analysis.
1
resolution of the two major components. The H NMR
spectrum of 2b recorded in toluene-d8 exhibits three distinct
sets of cyclopentadienyl resonances at d 3.54 (5H), 4.20
(2H), and 4.64 (2H) that are similar to the pattern dis-
played by the free fbpcd ligand. The aromatic hydrogens
appear as two sets of multiplets from d 6.95–7.25 (m,
12H, aryl) and 7.70–7.95 (m, 8H, aryl), with a vinyl singlet
The IR spectrum of 2c recorded in CH2Cl2 revealed ter-
minal Os–CO carbonyl stretching bands at 2095 (s), 2046
(s), 2010 (vs), 1975 (s), 1961 (m), and 1926 (m) cmꢀ1, with
the two dione carbonyl bands appearing at 1719 (w) and
1675 (m) cmꢀ1. The symmetry and frequencies of the termi-
nal carbonyl stretching bands closely match those IR data
reported by us for the diphosphine-chelated clusters 1,1-
Os3(CO)10(bpcd) and 1,1-Os3(CO)10(bmi) [2a,3c]. The
chelation of the fbpcd ligand in 2c is accompanied by a
for the ferrocenylidene moiety appearing at d 7.55. The 31
P
NMR spectrum of 2b reveals the presence of an AB quartet
centered at d ꢀ18.0 with JP–P = 17 Hz. The high-field loca-
tion of the bridging fbpcd ligand in 2b is in keeping with
the general trend found in other diphosphine-bridged clus-
´
ters vis-a-vis the corresponding chelating isomer (vide
infra) [21].
1
3.3. Synthesis of 2c and isomerization kinetics
down-field shift of the signature cyclopentadienyl H reso-
nances relative to 2b and appear at d 3.71 (5H), 4.26 (2H),
and 4.95 (2H). The resonances belonging to the aromatic
hydrogens [d 7.05–7.29 (12H) and 7.83–8.00 (8H)] and
the lone ferrocenylidene hydrogen (d 7.63) are relatively
unchanged from that of cluster 2b. Finally, the 31P NMR
spectrum displays an AB quartet at d 19.8 for the diphos-
phine ligand. The large nuclear deshielding exhibited by
The thermally promoted isomerization of the fbpcd
ligand about the triosmium frame in 2b was next explored
2
The spectral contributions from the hydride cluster 3 (<1%) did not
offer significant interference in comparison to the carbonyl frequencies
from cluster 2c.

W.H. Watson et al. / Polyhedron 26 (2007) 3585–3594
3591
Fig. 1. 1H NMR kinetic data for the isomerization of 2b!2c in toluene-d8 at 323 K in the cyclopentadienyl region at (A) 0 h, (B) 3 h, and (C) 20 h.
the fbpcd ligand in 2c relative to the bridging analogue is
consistent with the adoption of a chelating coordination
mode [21].
The kinetics for the isomerization of 2b!2c were also
investigated by UV–vis spectroscopy in toluene solution
over the temperature range 313–343 K. The first-order rate
constants for these reactions are reported in Table 3, while
Fig. 2a shows the spectral changes that accompany the
isomerization at 323 K. The rate of the reaction was easily
assessed by following the increase in the absorbance of the
360 nm belonging to 2c. The isomerization was judged to
be free from kinetic complications and negligible material
loss based on the isosbestic point at 390 nm that is main-
tained throughout the course of the reaction. Fig. 2b shows
the fit between the absorbance data and the least-squares
regression curve that affords the quoted rate constant in
Table 3. The effect of added PPh3 on the reaction at
323 K was also examined (Table 3, entry 4), yielding a rate
constant of 5.9 0.1 sꢀ1 that is essentially indistinguish-
able with the rate constant measured under CO (entry 3).
These data argue against a rate-limiting dissociative pro-
cess that yields a long-lived unsaturated intermediate.
The kinetic data for the conversion of 2b!2c strongly sup-
port the involvement of a unimolecular process whereby
one of the phosphine moieties and two CO groups undergo
an equatorial permutation about the cluster polyhedron via
a l-phosphine/l-CO intermediate. The Eyring plot of ln(k/
T) versus Tꢀ1 shown in Fig. 3 afforded activation parame-
ters (DHꢀ = 23.1 0.3 kcal/mol and DSꢀ = ꢀ7 1 eu)
consistent with the structural rearrangement exemplified
by the ‘‘merry-go-process’’ that is depicted in the following
equation [2a,3c,22]:
The isomerization kinetics for 2b!2c were initially
examined by 1H NMR spectroscopy in toluene-d8 at
323 K. Here a cluster mixture that contained ca. a 70/30
composition of 2b to 2c was employed.3 Representative
spectral changes in the cyclopentadienyl region from d
3.00–5.20 are illustrated in Fig. 1. Here the growth of the
three cyclopentadienyl resonances belonging to the chelat-
ing isomer was found to be proportional to the rate of
consumption of the Cp resonances belonging to 2b. A tra-
ditional plot of ln [2b] versus time furnished the first-order
rate constant quoted in Table 3 (entry 5), which agrees
nicely with the two other rate constants measured by
UV–vis spectroscopy at the same temperature (entries 3
and 4). The clean conversion to 2c found by NMR spec-
troscopy underscores the fact that the isomerization is
not accompanied by the formation of extraneous mate-
rial(s) and is unidirectional unlike the isomerization reac-
tions in the diphosphine-substituted clusters Os3(CO)10-
[(Z)PPh2CH@CHPPh2] and H4Ru4(CO)10(dppe) that give
rise to an equilibrium mixture of bridging and chelating
isomers [22,23].
3
The trace amount of cluster 3, which is initially present and ultimately
converted to 2c under CO, can be ignored as it does not adversely affect
the data analysis.

3592
W.H. Watson et al. / Polyhedron 26 (2007) 3585–3594
1.46
1.5
1.0
0.5
0.0
1.44
1.42
1.40
1.38
400
600
0
25000
50000
75000
Wavelength (nm)
Time (s)
Fig. 2. UV-vis spectral changes for the isomerization of 2b!2c in toluene at 323 K (left) and the absorbance changes at 360 nm versus time curve for the
experimental data (n) and the least-squares fit of k ( , right).
H
3.4. Synthesis of the hydride-bridged cluster 3 from 2c
O
Fc
Ph2
P
O
Ph2
P
The photochemical reactivity of 2c was examined given
our interest in diphosphine ligand activation at metal
clusters through ortho-metalation [2a,3c,22]. We have previ-
ously demonstrated the photoefficient CO loss and
C–H bond activation of an ancillary aryl ligand in
the diphosphine-chelated clusters Os3(CO)10(bpcd) and
Os3(CO)10(bmi) to give the hydride-bridged clusters HOs3-
(CO)9[l-Ph2PC@C(PhP{C6H4})C(O)CH2C(O)] and HOs3-
(CO)9[l-Ph2PC@C(PhP{C6H4})C(O)N(p-tolyl)C(O)],
respectively. Irradiation of cluster 2c using 366 nm light
leads to the slow loss of CO and ortho-metalation of one of
the phenyl ligands to give the cluster HOs3(CO)9-
[l-PhP(C6H4)C@C(PPh2)C(O)C@CH(C5H4FeCp)C(O)] (3).
In comparison to our earlier studies on Os3(CO)10(bpcd)
and Os3(CO)10(bmi), the photochemically promoted ortho-
metalation of 2c is exceeding slow, as assessed by NMR anal-
ysis. Depending upon the amount of 2c employed, good
yields of 3 could be achieved by prolonged photolysis for sev-
eral weeks with frequent removal of the liberated CO. Opti-
cal excitation of Os3(CO)10(bpcd) has been shown to furnish
the unsaturated cluster Os3(CO)9(bpcd), followed by an
intermediate p-aryl complex that precedes the formal
ortho-metalation step. The p-aryl complex is sensitive to
extraneous CO, which upon capture regenerates the starting
cluster Os3(CO)10(bpcd) [3c]. The same scenario is assumed
to be operative in the case of cluster 2c, which accounts of
the observed sensitivity to CO. Photolysis of a CO-saturated
solution of 2c shows no evidence for the formation of the
desired hydride-bridged cluster 3. Controlled thermolysis
of 3 with CO (1 atm) regenerates 2c in quantitative yield,
as determined by NMR spectroscopy. Cluster 3 was isolated
by column chromatography, followed by recrystallization,
as an air-sensitive, blue solid. Eq. (3) illustrates the reaction
of 2c to 3:
H
Os
Os
O
PPh2
Fc
Os
Os
Os
Os
O
P
Ph2
cluster 2b
Os
Ph2
P
O
O
H
Os
Os
P
Ph2
Fc
cluster 2c
ð2Þ
-13
-14
-15
-16
-17
0.0029
0.0030
0.0031
1/Temp(K-1
0.0032
)
Fig. 3. Eyring plot for the conversion of 2b!2c in toluene over the
temperature range of 313–343 K.

W.H. Watson et al. / Polyhedron 26 (2007) 3585–3594
3593
Fig. 4. Thermal ellipsoid plot of cluster 3 showing the thermal ellipsoids at the 50% probability level. The pentane solvent molecule has been omitted for
clarity.
MALDI-TOF mass spectrum showed a weak peak centered
Os
H
Os
on m/e 1457.7 for the loss of one CO group from of 3, as
well as ions for the loss of seven additional CO groups.
The molecular composition of one of the diastereomers
of 3 was established by X-ray diffraction analysis. Fig. 4
shows the thermal ellipsoid plot 3 and ascertains the activa-
tion of one of the aryl rings via ortho-metalation. The
Ph2
P
hν/-CO
O
Os
H
Os
Os
Os
Ph2P
heat + CO
P
Ph2
Fc
O
P
O
Ph
cluster 2c
H
O
cluster 3
˚
Os–Os bond distances range from 2.866(8) A [Os(2)–
Fc
˚
Os(3)] to 2.9441(8) A [Os(1)–Os(2)], with an mean distance
ð3Þ
˚
of 2.908 A that parallels those Os–Os single-bond distances
Cluster 3 was characterized in solution by IR and NMR
reported for other Os3 clusters. Despite the fact that the
bridging hydride was not located during the crystallo-
graphic refinement, its location may be confidently
assigned to the Os(1)–Os(2) vector [24]. Here the formation
of an edge-bridging hydride leads to an elongation of the
Os(1)–Os(2) bond length relative to the other two Os–Os
bonds in 3. Both phosphorus atoms of the fbpcd ligand
coordinate to the Os(1) center in a chelating fashion and
exhibit distinct axial [P(1)] and equatorial [P(2)] disposi-
tions. This represents a change in the phosphine disposition
from 2c where both phosphine moieties were initially situ-
ated at equatorial sites. Presumably, the stereochemical
change facilitates the ensuing ortho-metalation. The P(1)–
Os(1)–P(2) angle of 85.95(7)° and the two Os–P distances
spectroscopies. The ortho-metalation attendant on photol-
1
ysis was confirmed by H NMR spectroscopy. The obser-
vation of a high-field triplet at d ꢀ15.87 provides
unequivocal evidence for the presence of an edge-bridged
cluster hydride. Particularly informative were the NMR
data that revealed the existence of a 45:55 mixture diaste-
reomers for 3, whose origin derives from the proximal
and distal orientations of the ferrocene moiety relative to
the ortho-metalated aryl group. The nonfunctionalized
Cp ring in the 1H NMR spectrum of 3 appears as two sing-
lets at d 3.61 and 3.65, while the other Cp ring exhibits
three sets of broadened multiplets at d 4.19 and 4.32 (br,
2H each), 4.56 and 4.85 (br, 1H each), and 4.93 and 4.91
(br, 1H each). The essentially equimolar composition of
diastereomers was corroborated by 31P NMR spectros-
copy. The presence of two distinct high-field doublets
recorded at d 21.40 and 22.11 and one triplet at d 30.64,
which results from overlapping doublets, is consistent with
this formulation. No attempts have been made to assign
these resonances to specific phosphine moieties. The
˚
˚
of 2.305(2) A [Os(1)–P(2)] and 2.365(2) A [Os(1)–P(1)]
exhibit values unremarkable to those angles and distance
found by us in related compounds [1–3]. The Os(2)–C(24)
˚
bond distance of 2.165(9) A is in excellent agreement
with those metal-carbon sigma bond distances found in
di- and trinuclear osmium compounds containing an
ortho-metalated aryl ligand [25]. The nine CO groups are

3594
W.H. Watson et al. / Polyhedron 26 (2007) 3585–3594
(b) S.G. Bott, H. Shen, R.A. Senter, M.G. Richmond, Organomet-
allics 22 (2003) 1953;
(c) W.H. Watson, G. Wu, M.G. Richmond, Organometallics 25
(2006) 930;
linear and exhibit distances and angles consistent with such
linkages. The ferrocene and dione ring moieties are unex-
ceptional and require no comment.
(d) W.H. Watson, S.G. Bodige, K. Ejsmont, J. Liu, M.G. Richmond,
J. Organomet. Chem. 691 (2006) 3609;
4. Conclusions
(e) W.H. Watson, B. Poola, M.G. Richmond, J. Organomet. Chem.
691 (2006) 5579.
[4] (a) F.A. Cotton, B.E. Hanson, in: P. de Mayo (Ed.), Rearrangements
in Ground and Excited States, Academic Press, New York, 1980
[Chapter 12];
(b) S. Aime, L. Milone, Prog. NMR Spectrosc. 11 (1977) 183.
[5] (a) M. Hissler, J.E. McGarrah, W.B. Connick, D.K. Geiger, S.D.
Cummings, R. Eisenberg, Coord. Chem. Rev. 208 (2003) 115;
(b) C.J. McAdam, J.L. Morgan, B.H. Robinson, J. Simpson, P.H.
Rieger, A.L. Rieger, Organometallics 22 (2003) 5126.
[6] For a report on Knoevenagel condensation reactions using 4,5-
dichloro-4-cyclopenten-1,3-dione, the direct precursor in the synthesis
of the diphosphine ligand bpcd, see: R. Alfred, Z. Hellmut, Chem.
Ber. 94 (1961) 1800.
The synthesis and properties of the new diphosphine
ligand fbpcd are described. The kinetics for the bridge-
to-chelate isomerization of the fbpcd ligand in Os3(CO)10-
(fbpcd) have been measured, with a sequence involving a
nondissociative ‘‘merry-go-round’’ migration of the ancil-
lary fbpcd ligand and two CO groups supported by a mod-
erately negative entropy of activation. The chelating isomer
of Os3(CO)10(fbpcd) losses CO upon 366 nm irradiation,
leading to the ortho-metalation of the one of the aryl
groups and formation of the hydride-bridged cluster
HOs3(CO)9[l-PhP(C6H4)C@C(PPh2)C(O)C@CH(C5H4Fe-
Cp)C(O)], whose molecular structure has been crystallo-
graphically determined.
[7] W.H. Watson, B. Poola, M.G. Richmond, J. Chem. Crystallogr. 36
(2006) 715.
[8] W.H. Watson, D. Wiedenfeld, A. Pingali, B. Poola, M.G. Richmond,
Polyhedron (2007) in press.
[9] J.N. Nicholls, M.D. Vargas, Inorg. Synth. 26 (1989) 289.
[10] S.R. Drake, P.A. Loveday, Inorg. Synth. 28 (1990) 230.
[11] (a) D. Fenske, H. Becher, Chem. Ber. 107 (1974) 117;
(b) D. Fenske, Chem. Ber. 112 (1979) 363.
[12] D.F. Shriver, The Manipulation of Air-Sensitive Compounds,
McGraw-Hill, New York, 1969.
[13] Saint Version 6.02, Bruker Analytical X-ray Systems, Inc. Copyright
1997–1999.
[14] SHELXTL Version 5.1, Bruker Analytical X-ray Systems, Inc.
Copyright 1998.
[15] PLATON - A Multipurpose Crystallographic Tool, Utrecht Univer-
sity, Utrecht, The Netherlands, A.L. Spek (2001).
[16] B.K. Carpenter, Determination of Organic Reaction Mechanisms,
Wiley-Interscience, New York, 1984.
[17] N.B. Coltrup, L.H. Daly, S.E. Wiberley, Introduction to Infrared and
Raman Spectroscopy, Academic Press, New York, 1990.
[18] R. Meyer, D.M. Schut, K.J. Keana, D.R. Tyler, Inorg. Chim. Acta
240 (1995) 405.
[19] F.A. Bovey, Nuclear Magnetic Resonance Spectroscopy, Academic
Press, New York, 1969.
[20] W.H. Watson, B. Poola, M.G. Richmond, J. Chem. Crystallogr. 36
(2006) 123.
Acknowledgements
Financial support from the Robert A. Welch Founda-
tion (Grants P-0074-WHW and B-1093-MGR) is much
appreciated, and Professor. Andreas H. Franz (University
of the Pacific) is thanked for recording the MALDI-TOF
mass spectra.
Appendix A. Supplementary material
CCDC 628577 contains the supplementary crystallo-
graphic data for 3. These data can be obtained free of charge
via http://www.ccdc.cam.ac.uk/conts/retrieving.html, or
from the Cambridge Crystallographic Data Centre, 12
Union Road, Cambridge CB2 1EZ, UK; fax: +(44) 1223-
336-033; or e-mail: deposit@ccdc.cam.ac.uk. Supplemen-
tary data associated with this article can be found, in the
online version, at doi:10.1016/j.poly.2007.03.057.
[21] (a) P.E. Garrou, Chem. Rev. 81 (1981) 229;
(b) M.G. Richmond, J.K. Kochi, Organometallics 6 (1987) 254.
[22] W.H. Watson, G. Wu, M.G. Richmond, Organometallics 24 (2005)
5431.
References
[23] S. Kandala, M.G. Richmond, Inorg. Chem. 45 (2006) 5976.
[24] D.M.P. Mingos, D.J. Wales, Introduction to Cluster Chemistry,
Prentice-Hall, Englewood Cliffs, NJ, 1990.
[25] (a) A.J. Deeming, S.E. Kabir, N.I. Powell, P.A. Bates, M.B.
Hursthouse, J. Chem. Soc., Dalton Trans. (1987) 1529;
(b) A.J. Deeming, K.I. Hardcastle, S.E. Kabir, J. Chem. Soc., Dalton
Trans. (1988) 827;
[1] (a) For bma ligand reactivity studies, see: K. Yang, J.M. Smith, S.G.
Bott, M.G. Richmond, Organometallics 12 (1993) 4779;
(b) S.G. Bott, K. Yang, K.A. Talafuse, M.G. Richmond, Organo-
metallics 22 (2003) 1383;
(c) S.G. Bott, K. Yang, M.G. Richmond, J. Organomet. Chem. 690
(2005) 3067; 691 (2006) 3771.
[2] (a) For bmi and related ligand reactivity studies, see: W.H. Watson,
B. Poola, M.G. Richmond, J. Organomet. Chem. 691 (2006) 4676;
(b) K. Ejsmont, W.H. Watson, J. Liu, M.G. Richmond, J. Chem.
Crystallogr. 33 (2003) 541.
(c) S.E. Kabir, K.M.A. Malik, E. Molla, M.A. Mottalib, J.
Organomet. Chem. 616 (2000) 157;
(d) R.D. Adams, J.E. Cortopassi, J. Cluster Sci. 6 (1995) 437;
(e) S.B. Colbran, P.T. Irele, B.F.G. Johnson, F.J. Lahoz, J. Lewis,
P.R. Raithby, J. Chem. Soc., Dalton Trans. (1989) 2023.
[3] (a) For bpcd ligand reactivity studies, see: H. Shen, S.G. Bott, M.G.
Richmond, Organometallics 14 (1995) 4625;