Structure and reactivity of the zero-valent ruthenium complex Ru(1,2-bis(diphenylphosphino)ethane)(CO)3 and the dicationic ruthenium dimer [Ru2(1,2-bis(diphenylphosphino)ethane)2(CO) 6]2+

Article abstract of DOI:10.1016/S0022-328X(97)00731-6

The structure of Ru(dppe)(CO)₃ was characterized by single crystal X-ray diffraction [monoclinic crystal system, space group P2₁/n, a=12.2353(2) A, b=16.0803(3) A, c=14.2451(3) A, β=111.109(1)°, V=2614.62(9) A³, Z=4] and found to be intermediate between trigonal bipyramidal and square pyramidal. One electron oxidation of Ru(dppe)(CO)₃ using [(η⁵-C₅H₅)₂Fe][PF₆] produced [Ru₂(dppe)₂(CO)₆][PF₆]₂. The dicationic complex was fully characterized by solution spectroscopic methods and by single crystal X-ray diffraction [trigonal crystal system, space group P3₁21, a=20.566(2) A, c=13.871(2) A, V=5080.7(8) A³, Z=3] and found to have a dimeric structure with two octahedral units sharing a common apex via a Ru(I)-Ru(I) single bond. One octahedral unit is rotated approximately 45° relative to the other, and the chelating phosphine ligand occupies sites cis and trans to the Ru(I)-Ru(I) bond. All of the carbonyl ligands were found to be terminal. The dimer was fluxional in solution and line-shape analysis of the ³¹P{¹H} and ¹³C{¹H} variable-temperature NMR spectra was used to investigate the exchange mechanism and evaluate the rate constants. The mechanism involved an intramolecular, two-site exchange involving pairwise bridging carbonyls and had activation parameters of 11.8±0.15 kcal mol^-1 and -6.7±0.6 eu for ΔH^? and ΔS^? respectively. The dimer exhibited similar thermal and photochemical reactivity patterns; disproportionation occurred in CH₃CN and halogen atom abstraction occurred in halogenated solvents.

Full text of DOI:10.1016/S0022-328X(97)00731-6

Journal of Organometallic Chemistry 557 (1998) 13–28
Structure and reactivity of the zero-valent ruthenium complex
Ru(1,2-bis(diphenylphosphino)ethane)(CO)3 and the dicationic
ruthenium dimer [Ru2(1,2-bis(diphenylphosphino)ethane)2(CO)6]2+
Steven J. Skoog, Amy L. Jorgenson, John P. Campbell, Michelle L. Douskey, Eric Munson,
Wayne L. Gladfelter *
Department of Chemistry, The Uni6ersity of Minnesota, 207 Pleasant St. SE, Minneapolis, MN 55455, USA
Received 25 May 1997
Abstract
The structure of Ru(dppe)(CO)3 was characterized by single crystal X-ray diffraction [monoclinic crystal system, space group
3
˚
˚
˚
˚
P21/n, a=12.2353(2) A, b=16.0803(3) A, c=14.2451(3) A, i=111.109(1)°, V=2614.62(9) A , Z=4] and found to be
intermediate between trigonal bipyramidal and square pyramidal. One electron oxidation of Ru(dppe)(CO)3 using [(p5-
C5H5)2Fe][PF6] produced [Ru2(dppe)2(CO)6][PF6]2. The dicationic complex was fully characterized by solution spectroscopic
˚
˚
methods and by single crystal X-ray diffraction [trigonal crystal system, space group P3121, a=20.566(2) A, c=13.871(2) A,
V=5080.7(8) A , Z=3] and found to have a dimeric structure with two octahedral units sharing a common apex via a
3
˚
Ru(I)–Ru(I) single bond. One octahedral unit is rotated approximately 45° relative to the other, and the chelating phosphine
ligand occupies sites cis and trans to the Ru(I)–Ru(I) bond. All of the carbonyl ligands were found to be terminal. The dimer was
fluxional in solution and line-shape analysis of the 31P{1H} and 13C{1H} variable-temperature NMR spectra was used to
investigate the exchange mechanism and evaluate the rate constants. The mechanism involved an intramolecular, two-site
exchange involving pairwise bridging carbonyls and had activation parameters of 11.890.15 kcal mol−1 and −6.790.6 eu for
DH‡ and DS‡ respectively. The dimer exhibited similar thermal and photochemical reactivity patterns; disproportionation
occurred in CH3CN and halogen atom abstraction occurred in halogenated solvents. © 1998 Elsevier Science S.A. All rights
reserved.
Keywords: Ruthenium; Carbonyl; Crystal structure; Dynamic NMR; Photochemistry; Metal-centered radicals
1. Introduction
During the course of our study of the activation of
nitroarenes by Ru(dppe)(CO)3 (1), where dppe is 1,2-
bis(diphenylphosphino)ethane, we became interested in
the properties of the radical cation formed by oxidation
of 1 due to its importance as a potential intermediate in
species and are important in both stoichiometeric and
catalytic reactions [5]. Iron-centered radicals formed
from bis(phosphine) carbonyl complexes generated elec-
trochemically [6] or chemically [7] were found to un-
dergo a variety of radical coupling [8], substitution [9],
and disproportionation reactions [10,11], and have been
Eq. (1) [1–4]. Radical processes play a large role in the
reaction mechanisms of organometallic
* Corresponding author. Tel.: +1 612 6244391; fax: +1 612
6267541; e-mail: gladfelt@chemsun.chem.umn.edu
0022-328X/98/$19.00 © 1998 Elsevier Science S.A. All rights reserved.
PII S0022-328X(97)00731-6

14
S.J. Skoog et al. / Journal of Organometallic Chemistry 557 (1998) 13–28
extensively studied [12,13]. Bis(phosphine) substituted
ruthenium(I) radical carbonyl complexes exhibit similar
reactivity to that of their iron analogs [2,14].
quartz NMR tubes were cooled during photolysis with
a water-jacketed quartz Schlenk tube in an ethanol
bath.
2.2. Synthesis of Ru(dppe)(CO)3 (1)
2. Experimental
A slightly modified version of the original procedure
was used [19]. Ruthenium carbonyl (1.99 g, 3.13 mmol)
and dppe (3.74 g, 9.40 mmol) were placed in a 250 ml
Parr autoclave, toluene (20 ml) was added, purged three
times with CO, charged to 68 atm, and finally heated at
125°C for 18 h. The resulting yellow solution was
filtered through a plug of glass wool and transferred to
a 100 ml flask capped with a septum, purged with N2
and layered with 30 ml of hexane. Large yellow crystals
were grown at room temperature and collected on a
glass frit. The crystals were washed with hexane and
dried yielding 4.86 g of Ru(dppe)(CO)3 (89% yield).
Samples were stored in a glove box. mp: 120° (dec). IR
2.1. General
Standard Schlenk techniques were used in handling
all organometallic compounds [15]. A nitrogen-filled
Vacuum Atmospheres dry box equipped with a Dri-
Train Model 40-1 inert gas purifier or a MBraun 150-M
dry box were used for manipulations requiring inert
atmosphere conditions. The compounds 1,2-bis(-
diphenylphosphino)ethane and Ru3(CO)12 were pur-
chased from Strem and the former was purified by
recrystallization from hexane/toluene. Ferrocenium
hexafluorophosphate was prepared according to litera-
ture procedures [16]. Solvents were distilled prior to use
under pre-purified nitrogen; toluene and hexane were
distilled from purple sodium benzophenone ketyl where
small amounts of triglyme were used to solubilize the
1
(cm−1, CH2Cl2): wCO 1910 (s), 1931 (s), 2001 (s). H-
NMR (ppm, CH2Cl2, RT) l 2.41 (d, JHP=20.4 Hz,
CH2); 7.42 (m, ArH); 7.63 (m, ArH). 31P{1H}-NMR
(ppm, CH2Cl2, RT) l 74.6 (s, dppe, T1=2.2390.03 s
at 23°C). 13C{1H}-NMR (ppm, CD2Cl2, RT) l 30.99 (t,
anion
in
hexane.
Methylene
chloride,
1,2-
dichloroethane and deuterated methylene chloride were
distilled from calcium hydride and stored under a nitro-
gen atmosphere over 4 A molecular sieves. Acetonitrile
JCP=24.7 Hz, CH2); 128.97 (t, JCP=4.86 Hz, m,
C6H5); 130.52 (s, p, C6H5); 132.62 (t, JCP=6.1 Hz, o,
C6H5); 137.52 (m, JCP=18.8 Hz, ipso, C6H5); 211.84 (t,
˚
and deuterated acetonitrile were either distilled from
CaH2 under a nitrogen atmosphere and either used
immediately or stored under nitrogen over 4 A molecu-
JCP=10.67 Hz, CO).
2.3. Synthesis of Ru(dppe)(CO)3−x(13CO)x
˚
lar sieves. Deuterated dimethyl ether was prepared us-
ing literature procedures [17]. Carbon monoxide (CP
grade) and 13CO (Isotech 99.2 atom % 13C) were used
without further purification. All infrared spectral mea-
surements were carried out on a Mattson Polaris FTIR
spectrometer. The 1H-, 13C{1H}-, and 31P{1H}-NMR
spectral studies were performed using a Varian VXR-
300 spectrometer operating at 300, 75.4 and 121.4 MHz
respectively. All 31P{1H}-NMR spectra were externally
referenced to 85% D3PO4 (0.0 ppm) or (wherever possi-
ble) internally referenced to [PF6]− (−141 ppm). Pro-
ton and 13C{1H}-NMR spectra were referenced to
resonances corresponding to residual protons or carbon
atoms of the solvent. Electron paramagnetic resonance
(EPR) spectral measurements were made on a ESP 300
spectrometer using a quartz sample tube. Ultraviolet
spectra were obtained on a HP 8452A diode array
spectrometer at various concentrations and the molar
absorptivity was obtained from a Beer’s law plot. Pho-
tolysis experiments were carried out using a 175 W
medium pressure Hg vapor lamp. High pressure NMR
experiments were carried out in a sapphire NMR tube
[18]. Thermal and photochemical reactivity experiments
were carried out in glass or quartz NMR tubes, respec-
tively, capped with a septum under a nitrogen atmo-
sphere, and sealed with Teflon tape and Parafilm. The
2.3.1. Method A
A 1 ml volumetric flask was loaded with 18.9 mg
(0.0324 mmol) of Ru(dppe)(CO)3 and filled to 1 ml with
C6D6 in the glove box. One-half of this solution was
transferred via syringe into a high-pressure sapphire
NMR tube. The NMR tube was attached to a stainless
steel manifold and evacuated. The head space of the
NMR tube was pressurized to 155 psig with 13CO. The
tube was heated at 100°C for 2 h. Analysis of the
31P{1H}-NMR spectrum showed that the ratio of non-
labeled to mono-labeled to di-labeled 1 was 3:5:2.
2.3.2. Method B
A thick-walled 15 ml flask with an attached ground
glass joint and Kontes valve was loaded with 1.01 g
(1.73 mmol) of Ru(dppe)(CO)3 in a glove box. The
flask was removed from the box, 4 ml of toluene was
added, three freeze-pump-thaw cycles were completed
and after cooling to −196°C, one equivalent of 13CO
was condensed into the flask using a calibrated gas
ballast at a known pressure. After warming to room
temperature, the flask was heated to 80°C for 22 h.
Analysis by 31P{1H}-NMR indicated ca. 6:4:1 ratio of
non-labeled to mono-labeled to di-labeled 1. The sam-
ple was then transferred to a 50 ml Erlenmeyer flask

S.J. Skoog et al. / Journal of Organometallic Chemistry 557 (1998) 13–28
15
and layered with 20 ml of hexanes to give 900 mg of
crystalline product (90% yield) that was collected on a
glass frit and dried under vacuum.
(m, –CH2, 2H); 3.22 (m, –CH2, 2H); 7.59–7.65 (ArH,
20H). 31P{1H}-NMR (ppm, CD2Cl2, RT) l 64.75 (s,
dppe); −141.0 (septet, JPF=710.6 Hz, PF6). 13C{1H}-
NMR (ppm, CD2Cl2, RT) l 4.16 (s, NCCH3); 26.59
(m, CH2), 130.28 (t, JCP=4.9 Hz, aryl); 130.41 (t,
2.4. Synthesis of [Ru2(dppe)2(CO)6][PF6]2 (2)
JCP=5.4 Hz, aryl); 131.72 (t, JCP=4.6 Hz, aryl);
A 50 ml round-bottom Schlenk flask was loaded with
314 mg (0.538 mmol) of 1, and a solid addition flask
was loaded with 158 mg (0.477 mmol) of [Cp2Fe][PF6].
The flask was removed from the glove box and 9 ml of
CH2Cl2 was added via syringe and cooled to −78°C.
After the solution was cool, the [Cp2Fe][PF6] was
slowly added over a 45 min period and allowed to stir
at −78°C for 2 h after complete addition of the
oxidant. The reaction solution was allowed to warm
slowly to 0°C, and the CH2Cl2 was reduced to ca. 3 ml.
The concentrated solution was then cooled again to
−78°C and 30 ml of hexanes was added via syringe to
the rapidly stirring solution, to precipitate 2 as a yellow
solid. The solid material was filtered at −78°C on a
glass frit, washed with toluene to remove excess 1 and
finally dried under vacuum for several hours to give 330
mg of 2 (84% yield). This same procedure was used
with 13CO-labeled 1 as starting material to afford 13CO-
labeled 2. The reaction was also successfully scaled up
to 1.0 g of 1 with appropriate increases of oxidant and
carried out in 20 ml of CH2Cl2. IR (cm−1, CH2Cl2) wCO
2102 (w); 2062 (m); 2039 (vs). UV (CH2Cl2); u=386
131.99 (t, JCP=4.9 Hz, aryl); 132.87 (s, aryl); 133.19 (s,
aryl); 195 (t, JCP=15.5 Hz, CO). Anal. Calcd for
3 · CH3CN: C, 42.16; H, 3.43; N, 4.33. Found: C,
42.95; H, 3.92; N, 4.03.
2.6. Synthesis of [Ru(dppe)(CO)3Cl][PF6] (5)
A water-jacketed quartz Schlenk flask was charged
with 200 mg (0.137 mmol) of 2 in a glove box followed
by addition of 20 ml of 1,2-C2H2Cl2 and photolyzed for
16 h at 25°C. The resulting solution was transferred via
filter cannula to a round-bottom flask, and the solvent
was completely removed under vacuum. 3 ml of CH2Cl2
were added to dissolve the residue, followed by 30 ml of
hexanes, and a light brown oil formed on the bottom of
the flask. The flask was cooled to −78°C, and the walls
were scraped to loosen the solidified oil. The precipitate
was filtered at −78°C, yielding 146 mg of crude micro-
crystalline tan product (70% yield). Subsequent re-
peated recrystallization from CH2Cl2/pentane (1:10)
provided the analytically pure material. IR (cm−1
,
1
CH2Cl2) wCO 2135 (vs), 2097.5 (s), 2053 (s). H-NMR
(ppm, CD2Cl2, RT) l 3.13 (m, –CH2, 2H); 3.45 (m,
–CH2, 2H); 7.52–7.59 (m, ArH, 20H). 31P{1H}-NMR
(ppm, CD2Cl2, RT) l 54.51 (s, dppe); −141.0 (septet,
1
nm (‡=17500 M−1 cm−1). H-NMR (ppm, CH2Cl2,
RT) l 2.88 (br, CH2, 2H); 2.95 (br, CH2, 2H); 7.39–
7.61 (ArH, 20H). 31P{1H}-NMR (ppm, CH2Cl2, RT) l
56.76 (br, dppe, T1=1.7490.14 s at 23°C); 46.47 (br,
dppe, T1=1.6290.07 s at 23°C); −141.0 (septet,
J
PF=711 Hz, PF6). 13C{1H}-NMR (ppm, CD2Cl2, RT)
l 25.55 (m, CH2); 129.72 (t, JCP=5.5 Hz, aryl); 130.79
(t, JCP=5.1 Hz, aryl); 132.34 (t, JCP=4.9 Hz, aryl);
132.81 (t, JCP=4.5 Hz, aryl); 132.98 (s, aryl); 133.64 (s,
aryl); 184.64 (dd, JCP1=95.16 Hz; JCP2=13.42 CO);
187.02 (t, JCP=11.76 Hz, CO). Anal. Calcd for 4: C,
45.59; H, 3.15; Cl, 4.64. Found: C, 45.73; H, 3.25; Cl,
4.46.
J
PF=710.6 Hz, PF6−, T1=5.4890.07 s at 23°C).
13C{1H}NMR (see Results). Anal. Calcd. for 2: C,
47.81; H, 3.32. Found: C, 47.72; H, 3.15.
2.5. Synthesis of [Ru(dppe)(CO)2(CH3CN)2][PF6]2 (3)
A Schlenk tube was charged with 300 mg (0.206
mmol) of 2 followed by addition of 10 ml of CH3CN
and refluxed for 45 min. Analysis by IR spectroscopy
indicated a 1:1 mixture of 1 and 3. The solvent
(CH3CN) was completely removed by vacuum and 3 ml
of CH2Cl2 was added to the red oil. 30 ml of toluene
was added yielding a red oily precipitate and a yellow
supernatant. The mixture was cooled to −78°C and a
red/brown precipitate formed. The solid was filtered
and washed with toluene until the washings were color-
less. The material was recrystallized from CH2Cl2/ hex-
ane (1:10) to give 106 mg (56% yield) of 3. In order to
obtain analytically pure material the compound was
dissolved in CH3CN, evaporated to dryness, and finally
2.7. Variable-temperature NMR spectroscopy
For variable-temperature NMR experiments the
thermocouple was calibrated against a neat ethylene
glycol standard at elevated temperature, neat methanol
near room temperature [20]. The low-temperature (B
−20°C) calibration utilized a mixture of CH3OH with
6% HCl (by volume) to −40, and CH3OH/HCl/
CH2Cl2 in a 10:1:20 ratio down to −110. Both of these
solutions were calibrated in a higher temperature range.
Measurements of T1 were made using the inversion
recovery method (180–~–90), and the data processed
using Varian Software.
recrystallized from CH2Cl2/pentane (1:10). IR (cm−1
,
Unless otherwise noted samples for the dynamic
NMR experiments were loaded in a glove box into
NMR tubes with ground glass joints and attached to a
CH2Cl2) wCO 2032 (vs); wCN 2299 (w), 2323 (w). 1H-
NMR (ppm, CD2Cl2, RT) l 2.58 (s, NCCH3, 6H); 2.92

16
S.J. Skoog et al. / Journal of Organometallic Chemistry 557 (1998) 13–28
Schlenk line. Solvent was vacuum distilled into the tube,
which was cooled to −78°C. The mixture was cooled to
−198°C and sealed under vacuum. To prepare the NMR
sample of Ru(dppe)(CO)3−x(13CO)x in (CD3)2O, 0.1 ml
of the solution of labeled 1, prepared as described in
Method A was dried in vacuo for 30 min, and 0.2 ml of
(CD3)2O was vacuum transferred to give a 16 mM
solution.
rotation frame. A search was performed by taking a
series of frames in three orthogonally related regions
of reciprocal space. Each region investigated was com-
prised of 20 ten second frames, 0.3° apart in omega.
Table 1
Summary of crystallographic data for 1 and 2
Crystal Parameters
Variable-temperature NMR experiments with 2 uti-
lized 37 mM solutions. Different solvents were used in
different temperature regimes: CD2Cl2 was used as the
solvent in a sealable NMR tube below 30°C; 1,2-
C2H4Cl2/CD2Cl2 (3:1 by volume) was used as the solvent
in a sealable NMR tube up to 60°C; and 1,2-C2H4Cl2/
CD2Cl2 (3:1) was used as the solvent in a high-pressure
sapphire NMR tube up to 90°C. Experiments run at
elevated temperatures(\40°C) were recorded as rapidly
as possible to minimize decomposition of 2.
Spin simulations were carried out using the DSYMPC
program [21], available from Du¨sseldorf University,
spectral deconvolution and adjusted weighting of broad-
ened peaks were accomplished using the Varian software.
Complete line-shape analysis for the variable-tempera-
ture 31P{1H}- and 13C{1H}-NMR spectra of 2 was
carried out using DNMR5 (QCMP 059) available from
the QCPE at Indiana University.
Compound
Crystal system
Space group
Formula
1
2
monoclinic
P21/n
C29H24O3P2Ru
trigonal
P3121
C58H48F12O6P6Ru2
Formula weight (g
583.49
1456.92
mol−1
)
Unit cell dimensions
˚
a (A)
12.2353(2)
16.0803(3)
14.2451(3)
90
111.109(1)
90
2614.62(9)
4
1.482
20.566(2)
20.566(2)
13.871(2)
90
90
120
5080.7(8)
3
1.761
174(2)
0.664
0.468×0.26
˚
b (A)
˚
c (A)
h (°)
i (°)
k (°)
3
˚
V (A )
Z
z(calcd.) (g cm−3
Temperature (K)
)
173(2)
0.751
0.5×0.25×0.12
Abs coeff (mm−1
)
Cryst dimens (mm)
×0.156
Trans. factors, max
to min %
0.623–0.554
0.752–0.642
2.8. Solid-state NMR spectroscopy
Measurement of in-
tensity data
Diffractometer
All solid-state NMR experiments were performed on
a Chemagnetics CMX Infinity 300 MHz spectrometer
operating at 75.4 MHz for 13C and 121.4 MHz for 31P
using cross-polarization (CP) [22], magic-angle spinning
(MAS), and high-power 1H decoupling. A standard
Chemagnetics pencil probe equipped with a variable-
temperature accessory was used to spin a zirconia rotor
at 4–5 kHz. Single-contact cross polarization with high-
power proton decoupling (90° pulse width=4 ms, pulse
delay=180.0 s, 128 transients) was used to acquire the
13C spectrum used in this study. The methyl resonance
of hexamethylbenzene was used as an external reference
and set to 17.35 ppm. A single-contact cross polarization
sequence was also used to acquire the 31P spectrum (90°
pulse width=4 ms, pulse delay=10.0 s, 48 transients.)
A fresh sample of triphenylphosphine was used as a
reference and set at −11.9 ppm.
Siemens SMART Plat-
form CCD
Mo-Kh (0.71073 A)
graphite
Semiempirical
11008
˚
Radiation
Monochromator
Abs corr applied
No. of reflns measd
25234
5414
5412
No. of unique reflns 4240
No. of reflns used
Solution and refine-
ment
Programs used
Method of structure Direct methods
solution
4240
SHELXTL-V5.0
Adjusted weighting
scheme
w=[|2(F2o)+(AP)2
+(BP)]−1
,
where P=(F2o+2Fc2))/
3
Constants for
weighting
A=0.0397, B=2.9317 A=0.0248, B=
4.19
Final Ra indices [I\ R1=0.0327, wR2=
R1=0.0281,
wR2=0.0611
R1=0.0313,
wR2=0.0655
1.053
2.9. X-ray structural analysis of 1 and 2
2|(I)]
0.0808
Ra indices (all data) R1=0.0362, wR2=
Single crystals were grown from hexane/toluene for 1
and hexane/CH2Cl2 for 2. Table 1 includes a summary
of the details of the crystallographic data. A suitable
crystal was selected and mounted on the tip of a glass
fiber with STP oil. It was transferred to the goniometer
of a Siemens SMART CCD System and cooled to −100°
with a Siemens LT-2 low-temperature device. Crystal
quality and centering was confirmed by taking a 60s
0.0836
1.067
Goodness-of-fit on
F2
Æ
Ç1/2
Ã
Ã
Ã
ꢀ ꢁFo ꢂ − ꢂ Fcꢁ
ꢀ[w(F2o−F2c)2]
a R1=
ꢀꢂFoꢂ
wR2=Ã
ꢀ[w(F2o)2]
Ã
Ã
È
É

S.J. Skoog et al. / Journal of Organometallic Chemistry 557 (1998) 13–28
17
Table 2
These were harvested to yield a total of about 100
˚
Selected bond distances (A) and angles (°) for Ru(P(C6H11)3)2(CO)3
reflections with intensities greater than 10 s. An initial
unit cell was obtained, and checked for centering and
higher symmetry. None was found. The initial cell
constants were refined, and an initial orientation matrix
was found. Data were collected by examining a ran-
domly oriented region of reciprocal space in three
segments; the frames collected in a given segment were
0.3° apart in omega. The highest resolution data col-
Metal-ligand distances
Ru(1)–C(1)
Ru(1)–C(3)
Ru(1)–P(2)
1.904 (3) Ru(1)–C(2)
1.935 (3) Ru(1)–P(1)
2.3641 (8)
1.901 (4)
2.3612 (7)
Intraligand distances
C(1)–O(1)
C(3)–O(3)
P(1)–C(21)
P(2)–C(12)
1.149 (4) C(2)–O(2)
1.147 (4) P(1)–C(11)
1.845 (3) P(1)–C(31)
1.840 (3) P(2)–C(41)
1.827 (3)
1.162 (4)
1.846 (3)
1.828 (3)
1.833 (3)
˚
lected was 0.87 A. The total shutter open time for each
double-correlated frame was 60 s; two 30 s frames were
collected, and the data summed, doubling the dynamic
range of the detector. The default gain on the detector
signal was 4×, which was automatically dropped to
1× when the detector range was exceeded. Final cell
constants were determined during integration of the
data using 8192 intense, well-centered reflections.
P(2)–C(51)
Ligand-metal-ligand
C(2)–Ru(1)–C(1) 89.34 (13) C(2)–Ru(1)–C(3) 115.23 (14)
C(1)–Ru(1)–C(3) 100.00 (12) C(2)–Ru(1)–P(1) 91.35 (10)
C(1)–Ru(1)–P(1) 163.55 (9) C(3)–Ru(1)–P(1) 94.57 (9)
C(2)–Ru(1)–P(2) 143.49 (11) C(1)–Ru(1)–P(2) 87.71 (9)
C(3)–Ru(1)–P(2) 101.10 (9) P(1)–Ru(1)–P(2)
81.89(3)
Metal carbonyls
The structures were solved and refined using the
SHELXTL-plus 5.0 series of programs [23,24]. The space
group for 1 was determined to be P21/n (c14) and for
2 to be P3121 (c152) based on systematic extinctions
and intensity statistics. The space group of 2 is chiral;
its enantiomer is P3221. The low absolute structure
factor (Flack parameter) of 0.01(3) confirms the correct
choice of enantiomers. The structures were solved using
the direct methods program XS. Hydrogen atoms were
placed in calculated positions and refined with a riding
model with B values 20% larger than those on the
attached carbon atoms. Anisotropic refinement of 1
was then performed and converged to a final R1=
0.0358 (4|, based on F2). For 2, the hexafluorophos-
phate ion is rotationally disordered over two positions
in a 70/30% ratio. Each fluorine was refined anisotropi-
cally with the appropriate occupancy. The ion as a
whole was refined with bond length and rigid body
constraints. The unit cell contains large open channels
running parallel to the c-axis that were filled with
completely disordered solvent. The contribution of the
solvent to the scattering power was determined using
the program PLATON; it was found to be equal to 404
electrons. The total volume occupied by the disordered
Ru(1)–C(1)–O(1) 177.3 (3)
Ru(1)–C(3)–O(3) 175.3 (3)
Ru(1)–C(2)–O(2) 178.2 (3)
factors and anomalous scattering terms were taken the
usual sources [26], and the effects of anomalous disper-
sion were included for the non hydrogen atoms. For 1
selected bond angles and bond distances are listed in
Table 2, final atomic coordinates and equivalent
isotropic thermal parameters, along with their esti-
mated standard deviations are listed in Table 3. For 2
selected bond angles and bond distances are listed in
Table 4, final atomic coordinates and equivalent
isotropic thermal parameters, along with their esti-
mated standard deviations are listed in Table 5. The
molecular structure of 1 is provided in Fig. 1 and the
molecular structure of 2 in Fig. 2.
3. Results
3.1. Structure of Ru(dppe)(CO)3 (1)
3
˚
solvent was 959 A . The structure factors were cor-
rected for this solvent using the SQUEEZE option of
PLATON [25], reducing R from 0.035 to 0.028. The
density calculated from the SHELEX software package
used a entire volume of the unit cell including the
disordered solvent. The reported density was corrected
for the extra space occupied by the disordered solvent
(see Table 1).
Yellow crystals of 1 were grown from a toluene
solution by layering with hexane at room temperature
and were characterized by single crystal X-ray diffrac-
tion. The structure was intermediate between trigonal
bipyramidal with apical phosphorus and CO ligands
and square pyramidal with an apical carbonyl ligand.
The P(1)–Ru–P(2) bond angle of 81.98(3)° was com-
parable to related ruthenium complexes [3,27], and the
remaining bond angles were similar to those determined
for the iron analog, Fe(dppe)(CO)3 [28]. The Ru–C and
Ru–P bond distances in 1 were ca. 6–7% longer than
those observed in Fe(dppe)(CO)3. The five-membered
ring of the chelating phosphine was puckered with a
dihedral P(1)–C(11)–C(12)–P(2) angle of 50.3°, which
falls well within the mean range for torsion angles of
Full anisotropic refinement of all non hydrogen
atoms was performed using full-matrix least squares on
F2 using the program XL. A semiempirical psi-scan
absorption correction was applied to the data prior to
solution and refinement. The final difference map was
essentially flat, with maximum and minimum peaks
corresponding to 0.71 and −0.33 e A−3. Scattering
˚

18
S.J. Skoog et al. / Journal of Organometallic Chemistry 557 (1998) 13–28
Table 4
45–55° for various transition metal dppe compounds
[28].
˚
Bond distances (A) and angles (°) in [Ru2(dppe)2(CO)6](PF6)2 (2)
In solution the 31P{1H}-NMR spectrum contained a
sharp singlet for the two phosphorus nuclei owing to
the fluxional behavior of the complex at room tempera-
Metal-ligand distances
Ru(1)–C(11)
Ru(1)–C(12)
Ru(1)–P(11)
1.907 (4)
1.961 (4)
2.4457 (10) Ru(1)–Ru(1)%
Ru(1)–C(13)
Ru(1)–P(12)
1.948 (4)
2.3742 (9)
3.0413 (6)
1
ture. The H-NMR spectrum exhibited a doublet for
Intraligand distances
C(11)–O(11)
C(13)–O(13)
P(11)–C(61)
P(12)–C(31)
P(12)–C(71)
the equivalent methylene units of the dppe bridge due
to splitting by phosphorus. This resonance was shifted
slightly downfield by 0.3 ppm from the free dppe
ligand. The 13C{1H}-NMR spectrum exhibited triplets
for the methylene and phenyl carbon atoms. These
triplets were thought to be due to virtual coupling to
both phosphorus atoms and were observed in other
compounds containing the dppe ligand [29]. During the
13CO enrichment of 1 via Method A the observed rate
for the isotopic substitution of the carbonyl ligand was
1.2×10−2 s−1 measured from the intensity of the
Ru–CO resonance during the first half-life of the reac-
1.144 (4)
1.131 (4)
1.824 (4)
1.818 (4)
1.832 (4)
C(12)–O(12)
P(11)–C(51)
P(11)–C(72)
P(12)–C(41)
1.128 (4)
1.817 (4)
1.846 (3)
1.829 (4)
Ligand-metal-ligand
C(11)–Ru(1)–C(13) 91.9 (2)
C(13)–Ru(1)–C(12) 171.90 (14)
C(13)–Ru(1)–P(12)
C(11)–Ru(1)–P(11) 172.52 (10)
C(12)–Ru(1)–P(11) 93.38 (10)
C(11)–Ru(1)–Ru(1)% 86.28 (10)
C(12)–Ru(1)–Ru(1)% 87.23 (10)
P(11)–Ru(1)–Ru(1)% 101.15 (2)
C(11)–Ru(1)–C(12) 87.8 (2)
C(11)–Ru(1)–P(12) 89.57 (10)
C(12)–Ru(1)–P(12) 99.84 (10)
C(13)–Ru(1)–P(11) 87.94 (11)
P(12)–Ru(1)–P(11) 82.95 (3)
C(13)–Ru(1)–Ru(1)% 84.68 (9)
P(12)–Ru(1)–Ru(1)% 171.67 (3)
88.25 (10)
Table 3
Metal carbonyls
Ru(1)–C(12)–O(12) 175.9 (3)
Ru(1)–C(13)–O(13) 175.5 (3)
Atomic coordinates (×104) and equivalent isotropic displacement
Ru(1)–C(11)–O(11) 174.7 (3)
parameters (A ×103) for 1
2
˚
Atom
x
y
z
Ueq
SOF
tion and followed an apparent zero-order dependence
on the concentration of 1.
Ru(1)
C(1)
O(1)
C(2)
O(2)
C(3)
O(3)
P(1)
P(2)
12181 (1)
11542 (3)
11196 (2)
12559 (3)
12759 (3)
10780 (3)
9906 (2)
−1227 (1)
−135 (2)
534 (1)
10216 (1)
10091 (2)
10046 (2)
9052 (3)
8325 (2)
9995 (2)
27 (1)
34 (1)
47 (1)
42 (1)
61 (1)
33 (1)
46 (1)
27 (1)
28 (1)
32 (1)
34 (1)
31 (1)
35 (1)
40 (1)
46 (1)
49 (1)
41 (1)
30 (1)
35 (1)
41 (1)
43 (1)
47 (1)
41 (1)
32 (1)
37 (1)
45 (1)
48 (1)
49 (1)
42 (1)
33 (1)
42 (1)
52 (1)
53 (1)
52 (1)
44 (1)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
−1033 (2)
−918 (2)
−1878 (2)
−2220 (1)
−2381 (1)
−1044 (1)
−2664 (2)
−1863 (2)
−2227 (2)
−1473 (2)
−1344 (2)
−1963 (2)
−2712 (2)
−2847 (2)
−3340 (2)
−3292 (2)
−4007 (2)
−4779 (2)
−4837 (2)
−4125 (2)
−91 (2)
3.2. Fluxional beha6ior of 1
The line shape of the carbonyl resonance was studied
as a function of temperature by 13C{1H}-NMR spec-
troscopy in order to estimate the energy barrier of the
fluxional process exhibited by 1. At room temperature
in solution, the carbonyl environments were averaged
to yield a triplet at 211.84 ppm (JCP=10.67 Hz). As the
temperature was lowered, the peaks broadened (Fig. 3).
The slow-exchange limit was not observed, however,
owing to the freezing point of the dimethyl ether sol-
vent (−141.5°C) and instrumental limitations. At tem-
peratures below −133°C further broadening was
observed and was attributed, at least in part, to in-
creased viscosity of the solvent close to its freezing
point.
Solid-state NMR spectroscopy was used to observe
the slow-exchange spectrum of 1. A sample of 13CO-en-
riched 1 was prepared as described in Method B, and
the CPMAS 13C{1H}-NMR spectrum at room tempera-
ture showed three carbonyl environments with peaks at
202, 215, and 221 ppm in a 1:1:1 ratio. The average of
these resonances (212.3) was close to the resonance
observed in solution. No dynamic behavior was ob-
served in the 13C{1H}-NMR spectrum up to 100°C.
Spectral acquisition at higher temperatures was not
attempted owing to decomposition of the complex at
120°C. The CPMAS 31P{1H}-NMR spectrum gave two
resonances at 77.1 and 59.5 ppm and no broadening
was observed at elevated temperature.
9822 (2)
13458 (1)
12861 (1)
13652 (3)
13931 (3)
14989 (3)
15344 (3)
16503 (3)
17329 (3)
16995 (3)
15834 (3)
13015 (2)
12313 (3)
11982 (3)
12344 (3)
13027 (3)
13362 (3)
13638 (2)
14171 (3)
14794 (3)
14875 (3)
14353 (3)
13747 (3)
11731 (3)
10853 (3)
9945 (3)
10672 (1)
11980 (1)
11978 (2)
12594 (2)
10768 (2)
10492 (2)
10588 (3)
10973 (3)
11266 (3)
11154 (3)
9945 (2)
C(11)
C(12)
C(21)
C(22)
C(23)
C(24)
C(25)
C(26)
C(31)
C(32)
C(33)
C(34)
C(35)
C(36)
C(41)
C(42)
C(43)
C(44)
C(45)
C(46)
C(51)
C(52)
C(53)
C(54)
C(55)
C(56)
8931 (2)
8349 (2)
8776 (3)
9785 (3)
10368 (2)
12556 (2)
12038 (2)
12460 (3)
13412 (3)
13944 (3)
13526 (2)
12534 (2)
12285 (2)
12653 (3)
13287 (3)
13540 (3)
13170 (3)
391 (2)
1102 (2)
1343 (2)
871 (2)
148 (2)
−1126 (2)
−521 (2)
−571 (3)
−1214 (2)
−1814 (2)
−1771 (2)
9916 (3)
10768 (3)
11683 (3)
Ueq is defined as one third of the trace of the orthogonalized Uij
tensor.

S.J. Skoog et al. / Journal of Organometallic Chemistry 557 (1998) 13–28
19
CH Cl
2
3.3. Synthesis of [Ru2(dppe)2(CO)6][PF6]2 (2)
1+[Cp2Fe][PF6] “ 21/2[Ru2(dppe)2(CO)6][PF6]2
−78°C
+Cp2Fe
(2)
Complex 2 was prepared from 1 by oxidation using
one equivalent of [Cp2Fe][PF6] in CH2Cl2 at low tem-
perature (Eq. (2)). In the solid state 2 was apparently
stable in air for several days, but it decomposed in
solution when
exposed to air and at temperatures above 80°C in
CH2Cl2, and 1,2-C2H4Cl2. Complex 2 was insoluble in
hydrocarbon solvents, CHCl3, and CCl4. The IR spec-
trum of the carbonyl region of 2 in CH2Cl2 was shifted
to higher energy relative to 1, which was consistent with
decreased y-back-bonding of the terminal carbonyls to
a Ru(I) metal center. The absence of IR stretches in the
1750–1850 cm−1 region suggested that all the car-
bonyls were terminal in solution. The electronic absorp-
tion spectrum exhibited an absorption maximum at 386
nm (‡=17500 M−1 cm−1), which was assigned to the
M–M |“|* transition (Fig. 4) [30] and compares
favorably with the previously isolated Ru(I)–Ru(I)
dimer [Ru2(v-bbmb)2(CO)6][PF6]2 where umax=388 nm
(‡=13900 M−1 cm−1) and bbmb is 2,2%-bis[(1,1%-
biphenyl-2,2%-diyl)phosphite]-3,3%-di-t-butyl-5,5%-dimeth-
oxy-1,1%-biphenyl} [14]. No EPR signal was detected for
a 40 mM sample of 2 in CH2Cl2 at room temperature.
Table 5
Atomic coordinates (×104) and equivalent isotropic displacement
2
parameters (A ×103) for 2
˚
Atom
x
y
z
Ueq
23 (1)
SOF
Ru(1) 6000 (1)
C(11) 5280 (2)
O(11) 4859 (2)
C(12) 5172 (2)
O(12) 4668 (2)
C(13) 6789 (2)
O(13) 7261 (2)
5625 (1)
5864 (2)
5991 (2)
4598 (2)
4020 (2)
6675 (2)
7277 (2)
5305 (1)
5516 (1)
4872 (2)
4286 (2)
3786 (2)
3876 (2)
4456 (2)
4949 (2)
6433 (2)
6711 (2)
7428 (3)
7871 (3)
7606 (3)
6880 (2)
5776 (2)
6526 (2)
6885 (3)
6501 (3)
5775 (3)
5410 (2)
4316 (2)
4027 (2)
3287 (2)
2811 (2)
3084 (2)
3832 (2)
5249 (2)
5589 (2)
4023 (1)
3342 (4)
985 (1)
1486 (2)
1855 (2)
784 (3)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
30 (1)
41 (1)
30 (1)
38 (1)
31 (1)
43 (1)
27 (1)
28 (1)
30 (1)
37 (1)
42 (1)
42 (1)
41 (1)
37 (1)
33 (1)
43 (1)
52 (1)
61 (1)
58 (1)
43 (1)
31(1)
45 (1)
55 (1)
53 (1)
45 (1)
38 (1)
29 (1)
34 (1)
38 (1)
39 (1)
41 (1)
33 (1)
32 (1)
32 (1)
37 (1)
691 (2)
1004 (2)
1072 (2)
562 (1)
P(11)
P(12)
C(31)
C(32)
C(33)
C(34)
C(35)
C(36)
C(41)
C(42)
C(43)
C(44)
C(45)
C(46)
C(51)
C(52)
C(53)
C(54)
C(55)
C(56)
C(61)
C(62)
C(63)
C(64)
C(65)
C(66)
C(71)
C(72)
P(13)
F(1)
6960 (1)
6347 (1)
5685 (2)
4985 (2)
4515 (3)
4753 (3)
5450 (3)
5919 (2)
6777 (2)
6307 (3)
6602 (3)
7367 (3)
7836 (3)
7540 (2)
7636 (2)
8190 (2)
8733 (2)
8708 (3)
8162 (2)
7638 (2)
6672 (2)
7102 (2)
6897 (2)
6263 (2)
5833 (2)
6036 (2)
7073 (2)
7562 (2)
2864 (1)
2328 (4)
2584 (1)
3467 (2)
3225 (3)
3921 (3)
4866 (3)
5113 (3)
4428 (3)
3176 (2)
3450 (3)
3815 (3)
3904 (3)
3646 (3)
3290 (3)
−400 (3)
−292 (3)
3.4. Structure of [Ru2(dppe)2(CO)6][PF6]2 (2)
Single crystals of 2 were grown from a saturated
CH2Cl2 solution by careful layering with hexanes at room
temperature. Single crystal X-ray diffraction demon-
strated that in the solid state complex 2 exists as a
Ru(I)–Ru(I) dimer held together with a long metal–
˚
metal bond (3.0413(6) A). The structure is similar to that
of [Ru2(PPh3)2(CO)4(CH3CN)4]2+, which has a Ru(I)–
˚
Ru(I) bond length of 2.8731(8) A and a similar staggered
ligand arrangement [31]. The extreme length of the M–M
bond in 2 can be attributed mostly to steric repulsion
involving the equatorial phosphines, a smaller contribu-
tion may result from Coulombic repulsion between the
cationic metal centers. The structure of 2 exhibited a
slightly distorted octahedral arrangement of the symme-
try-equivalent ruthenium atoms. The bite angle
(82.95(3)°) of the dppe ligand was similar to that of 1
(vide supra), and the Ru–CO bond of the carbonyl trans
to the equatorial phosphine (1.907(4) A) is shorter than
either of the two M–CO bonds cis to the phosphines
(1.961(4) and 1.948(4) A). The dihedral angle between
P(11)–Ru(1)–Ru(1)%–P(11)% was 148.14(5)°. This large
angle reduced the unfavorable steric interactions between
the bulky equatorial phosphines. The inward leaning of
the trans carbonyls toward M–M bond (C(13)–Ru(1)–
(C(12)=171.90(14)°) was observed in similar M–M
bonded dimers (viz. (Me3P)(CO)4OsW(CO)5 [32],
Mn2(CO)10 and Re2(CO)10) [33].
−1001 (3)
−1833 (3)
−1959 (3)
−1241 (3)
410 (2)
801 (3)
682 (3)
162 (3)
˚
−237 (3)
−115 (3)
2531 (3)
1646 (2)
2112 (1)
1504 (5)
992 (8)
2789 (4)
3191 (9)
1228 (4)
2289 (12)
2232 (2)
3112 (6)
2017 (15)
2003 (2)
˚
100 (2) 0.70
95 (4) 0.30
83 (2) 0.70
69 (3) 0.30
88 (2) 0.70
100 (4) 0.30
79 (1)
85 (2) 0.70
105 (4) 0.30
F(1B) 2971 (10) 4114 (11)
F(2) 3403 (4) 4722 (4)
F(2B) 2702 (10) 3934 (10)
F(3)
3341 (4)
4483 (3)
4801(8)
3663(2)
3631(4)
3214 (9)
4386(2)
F(3B) 3509 (8)
F(4)
F(5)
F(5B) 2130 (9)
F(6) 2337 (2)
3379 (2)
2453 (4)
1
3.5. NMR spectroscopy of [Ru2(dppe)2(CO)6][PF6]2 (2)
69 (1)
1
At ambient temperature 2 exhibited two broad
31P{1H} resonances (peak width W=100 Hz) between
Ueq is defined as one third of the trace of the orthogonalized Uij
tensor.

20
S.J. Skoog et al. / Journal of Organometallic Chemistry 557 (1998) 13–28
40 and 50 ppm and a septet at −141 ppm. The ratio
of the integration of each of the 31P{1H} peaks, corre-
sponding to two different dppe environments and
the PF6 anions, was 1:1:1 supporting the stoichiome-
try of a dicationic Ru(I)–Ru(I) dimer of the formula-
tion given in Eq. (2). In the low temperature
(−40°C) 31P{1H}-NMR spectrum the line width
narrowed and a pair of triplets were observed (JPC
=
5.4 Hz) (Fig. 5a).
At room temperature two sets of carbonyl reso-
nances were observed in the 13C{1H}-NMR spectrum;
a triplet at 200.2 ppm (J=10.4 Hz) and three broad
features between 197.9 and 199.5 ppm. At low-tem-
perature (−40°C) the natural-abundance 13C{1H}-
NMR spectrum of 2 in CD2Cl2 yielded a doublet of
triplets and a doublet of quartets in an integral ratio
of 2:1 (Fig. 5b). At lower temperatures these reso-
nances broadened. The NMR spectrum at (−40°C)
was consistent with the structure shown in Scheme 1.
The low-temperature 31P{1H}-NMR spectrum was
simulated (Fig. 5a) using lPa=43 ppm (JPaPb=7 Hz;
JPaP%=4.8 Hz) and lPb=54 ppm (JPbP=7 Hz; and
Fig. 1. Structure and labeling scheme for Ru(dppe)(CO)3 1. Atoms
are depicted as 50% probability ellipsoids, and hydrogen atoms are
omitted.
JPbP=4.8 Hz) and a 4 Hz natural line width. The
low-temperature natural-abundance 13C{1H}-NMR
spectrum was also simulated (Fig. 5b) using the
favored from other mechanistic criterion (vide infra)
and the rates assuming this model were used. The
following rates (s−1) were obtained for the tempera-
tures (°C) indicated in parenthesis: 146 (10); 300 (20);
610 (30); 1050 (40); 1900 (50); 3500 (60); 6650 (70);
11700 (80); 20000 (90). The Eyring plot shown in Fig.
7, yielded activation parameters DH‡=11.890.15 kcal
mol−1 and DS‡= −6.790.6 eu. The width of the
resonances in the spectra used for the line-shape analy-
sis were \30 Hz; sufficiently broad to provide well-
defined rate constants over the temperature range used
for the analysis [34].
parameters lCa=199.33 ppm (JCaP=13.4 Hz; JCaP
=
5.5 Hz; JCaPb%=3.3 Hz); lCb=197.95 ppm (JCbPb=73
Hz; JCbPb=5.5 Hz; JCbPb%=2.5 Hz) with 5.3 Hz natu-
ral line width. The low-temperature (−60°C)
31P{1H}-NMR spectrum of 13CO-labeled 2 showed the
resonance at 43 ppm partially split by 70 Hz into
broad peaks by a trans carbonyl and was consistent
with the low temperature 13C{1H}-NMR spectra.
3.6. Dynamic NMR spectroscopy of 2
The variable-temperature 13C{1H}-NMR spectra are
shown in Fig. 8. The doublet of triplets corresponding
to Ca and Ca% (−40°C) coalesced near −2°C and
formed a triplet at 27°C. The doublet of quartets
Fig. 6 displays representative experimental and calcu-
lated 31P{1H}-NMR spectra between 10 and 90°C. A
simultaneous two-site exchange mechanism (Eq. (3)) was
used to model the exchange process. The analysis used
four different nuclei with two different configurations;
(3)
the remaining parameters were obtained from the simu-
lations of the low-temperature 31P{1H}-NMR spectra
(Fig. 5a). A covariant set of rate constants were ob-
tained from the line-shape analysis of the same spectra
assuming a four-site exchange mechanism (i.e. non-cor-
related phosphine exchange) and were related to the
first set by a factor of two. The two-site mechanism is
(−40°C) corresponding to Cb and Cb% coalesced at
27°C into a broad feature centered at 198.5 ppm.
Further heating to 57°C transformed them into a
triplet. A noteworthy feature of these spectra was that
the two unique carbonyl environments do not exchange
with each other at the fast-exchange limit.

S.J. Skoog et al. / Journal of Organometallic Chemistry 557 (1998) 13–28
21
Fig. 2. Structure and labeling scheme for the cation [Ru2(dppe)2(CO)6]2+ present in crystalline [Ru2(dppe)2(CO)6][PF6]2. Atoms are depicted as
50% probability ellipsoids and hydrogen atoms are omitted for clarity. The unlabeled atoms are related to the labeled ones by a two-fold rotation
axis.
Line-shape analysis of the Cb and Cb% portion of the
13C{1H}-NMR spectra was performed on the spectra
obtained at −2, 27, and 47°C (Fig. 8). The mechanism
used to simulate the exchange process is given in Eq (4).
This mechanism was modeled by using three different
3.8. Disproportionation of [Ru2(dppe)2(CO)6][PF6]2 (2)
in acetonitrile
When 2 was dissolved in acetonitrile and heated to
(4)
nuclei (Pa, Pb, and Cb) in two different configurations
where Pa exchanged with Pb along with parameters based
on the simulation of the low temperature spectrum. The
following rates (s−1) were obtained for the temperatures
(°C) indicated in parenthesis: 50 (−2), 550 (27), 1800
(47) and are plotted along side the rates obtained from
the 31P{1H}-NMR line-shape analysis (Fig. 7).
reflux for 30 min a quantitative disproportionation
produced 1 and a new compound, [Ru(dppe)(CO)2
(CH3CN)2][PF6]2 (3) in a 1:1 ratio as determined by
31P{1H}-NMR spectroscopy in CD3CN. The singlets at
74.3 ppm and 64.6 ppm were assigned to 1 and 3
respectively. The reaction was monitored by IR spec-
troscopy in CH3CN and yielded four strong stretches;
three at 2000, 1926, and 1903 cm−1 were assigned to 1
and the fourth (at 2030 cm−1) was assigned to 3. The
same reaction was carried out with a 13CO-enriched
sample of 2 in CD3CN and two triplets were observed
in the carbonyl region of the 13C{1H}-NMR spectrum;
one centered at 212.8 ppm (JCP=10.3 Hz) assigned to
1, and the other centered at 195.6 ppm (JCP=15.9 Hz).
The triplets were observed in a 3:2 ratio and the only
product consistent with the NMR data was trans-
[Ru(dppe)(CO)2(CH3CN)2][PF6]2 (3) (Eq. (5)).
3.7. Attempted labeling of 2
A sample of 2 in CH2Cl2 was placed under 80 psig of
13CO in a high-pressure sapphire NMR tube. No car-
bonyl exchange was observed at room temperature
during a 48 h period by 13C{1H}-NMR spectroscopy and
the resonance at 185 ppm corresponding to free 13CO in
solution remained narrow (W=4.5 Hz) during the
course of the experiment. The 31P{1H}-NMR resonances
were not effected by the presence of 13CO.

22
S.J. Skoog et al. / Journal of Organometallic Chemistry 557 (1998) 13–28
(5)
Complex 3 was isolated preparatively as an off-white
solid. A methyl resonance was observed at 2.6 ppm in
1
the H-NMR spectrum integrating to six protons rela-
tive to the protons of the dppe backbone. The 13C{1H}-
NMR spectrum in CD2Cl2 obtained from the same
isolated sample yielded a triplet at 195 ppm (JCP=15.5
Hz) and was assigned to the CO resonance of 3.
exhibiting characteristic cis coupling to the phosphorus
atoms. These observations were consistent with the
structure drawn in Eq. (5).
The same disproportionation reaction was observed
to occur photochemically when 2 was placed in a
quartz NMR tube and photolyzed for ca. 4 h at 25°C.
Three resonances were observed by 31P{1H}-NMR
spectroscopy at 74.3, 64.6, and 72.6 ppm, correspond-
ing to 1, 3 and a new complex (4) in a 2:1:1 ratio
respectively. No reaction (B5%) occurred when an
acetonitrile solution of 2 was allowed to sit for a 15 h
period in a glass NMR tube exposed to laboratory
light.
A photolysis experiment was carried out using a
13CO-enriched sample of 2, and the same three reso-
nances appeared in the 31P{1H}-NMR spectrum after
ca. 6 h. The phosphorus resonances at 74.3 and 64.8
ppm had characteristic splitting due to isotopic enrich-
ment of 13CO, while the third resonance was a singlet.
Fig. 3. Variable-temperature (75.4 MHz) 13C{1H} NMR spectra in
the carbonyl region of a 16 mM solution of a 13CO-labeled sample of
Ru(dppe)(CO)3 (1) in (CD3)2O at the indicated temperatures.
Fig. 4. Electronic absorption spectrum of a 4.53×10−5 M solution
of [Ru2(dppe)2(CO)6][PF6]2 (2) in CH2Cl2 at ambient temperature.

S.J. Skoog et al. / Journal of Organometallic Chemistry 557 (1998) 13–28
23
(6)
3.9. Halogen atom abstraction reactions of
[Ru2(dppe)2(CO)6][PF6]2 (2)
When 2 was dissolved in 1,2-C2H4Cl2 and heated to
ca. 100°C for 30 min, a singlet in the 31P{1H}-NMR
spectrum grew in at 54.6 ppm. When the same reaction
was carried out using 13CO-enriched 2 the phosphorus
resonance observed at 54.6 ppm was partially split by
13CO into a doublet of doublets (JPCtrans=95.2 Hz,
JPCcis=15.29 Hz). Three different sets of resonances
were observed in the carbonyl region by 13C{1H}-NMR
spectroscopy: a triplet at 186.04 ppm (JCP=11.5 Hz), a
doublet of doublets at 184.64 ppm (JCPtrans=94.6 Hz,
JCPcis=14.1 Hz), and characteristic exchanged-broad-
ened resonances at 198–200 ppm corresponding to 2.
The first two sets of resonances were observed in a 1:2
ratio. From the these NMR experiments, this new
species was assigned as fac-[Ru(dppe)(CO)3Cl][PF6] (5)
(Eq. (7)).
This structural assignment
(7)
was also consistent with the IR spectrum (wCO; 2135 (s),
2097 (s), and 2053 (s) cm−1). From a sample isolated
via photolysis (vide infra) four sets of triplets were
observed in the aryl region of the 13C{1H} spectrum
and was consistent with the NMR spectral patterns
Fig. 5. Actual and simulated natural abundance NMR spectra of a 37
mM solution of [Ru2(dppe)2(CO)6][PF6]2 (2) at −40°C in CD2Cl2.
The experimental spectra are located above the simulated spectra. (a)
(121 MHz) 31P{1H} spectrum, (b) (75.4 MHz) natural-abundance
13C{1H} spectrum.
observed
in
the
isostructural
species
fac-
Re(dppe)(CO)3(OMe) [29].
Complex 5 was unstable at elevated temperatures.
When heated to ca. 100°C for 1 h, two new unknown
species grew in at the expense of 5 by as monitored by
31P{1H}-NMR spectroscopy. These species exhibited a
singlet at 69.8 ppm, and a pair of doublets at 68.9 ppm
and 68.3 ppm (JPC=14.6 Hz) in a 3:7 ratio. The same
experiment was carried out using 13CO-enriched 2 and
three sets of resonances were observed in the carbonyl
region of the 13C{1H} spectrum: a multiplet at 199.8
ppm, a triplet at 196.0 (JCP=16.1 Hz), and a singlet at
184.5 ppm (assigned as free CO). No further attempts
were made to identify these two compounds.
A similar halogen atom abstraction reaction was
observed between 2 and 1,2-C2H4Cl2 carried out at
25°C under photochemical conditions. Complex 5
would slowly grow in upon photolysis of 2 in a quartz
NMR tube. After 4 h of photolysis 2 was completely
After additional photolysis the resonances at 72.6 ppm
became larger at the expense of the resonance at 64.6
ppm. Gas evolution was observed in the quartz NMR
tube, and the 13C{1H}-NMR spectrum yielded a triplet
at 212.8 ppm (JCP=10.4 Hz) assigned to 1, a triplet at
195.6 ppm (JCP=15.5 Hz) assigned to 3, and a singlet
located at 185.5 assigned to free gaseous CO.
Complex 3, which was isolated on a preparative
scale, was placed in a quartz NMR tube and pho-
tolyzed for 4 h. The reaction was monitored periodi-
cally by 31P{1H}-NMR spectroscopy, and the initial
resonance corresponding to 3 at 64.8 ppm was found to
be replaced by a resonance at 72.6 ppm. Since no 13CO
resonance was detected in 4 by NMR spectroscopy, and
because free CO was observed, we assign the 31P{1H}
singlet at 72.6 ppm to [Ru(dppe)(CH3CN)4]2+ (Eq. (6)).

24
S.J. Skoog et al. / Journal of Organometallic Chemistry 557 (1998) 13–28
Scheme 1.
converted to 5 in addition to a small amount of an-
other unidentified species as evidenced by the appear-
ance of singlets at 54.58 ppm and 57.94 ppm (7:3
ratio) in the 31P{1H}-NMR spectrum.
would be required to form the octahedral product.
This conformational change was related to the flux-
ional process which interconverts the two inequivalent
phosphorus atoms of the dppe ligand. Based on previ-
ous studies of the fluxional behavior of five coordi-
nate complexes (viz. Fe(CO)5, Ru(CO)5 and
Fe(dppe)(CO)3) [36] we felt that the barrier might be
measurable by a combination of low temperature and
solid state NMR techniques.
In contrast to 3, complex 5 was photochemically
stable, but thermally unstable. A sample of 5 gener-
ated under thermal conditions was transferred to a
quartz NMR tube and photolyzed for 2 h. There was
no change in the 31P{1H} spectrum during the course
of the experiment. This same sample was then heated
to 100°C for 45 min, and 5 was converted to the two
unidentified species described above. These findings
provided a method to isolate 5 at room temperature
on a preparative scale via photolysis.
The combination of the data collected from the so-
lution and solid state NMR studies allowed an esti-
mation of the energy barrier of the fluxional process
for 1. The solid state NMR studies provided the
chemical shift difference between the axial and equa-
torial carbonyls (lw ca. 1200 Hz) in complex 1 at the
slow-exchange limit. From the solution state 13C{1H}-
NMR studies the coalescence temperature was esti-
mated to be −135910°C. Using an expression
relating the rate constant to the peak separation of
the two sites at the coalescence temperature (k=
ylw2−1/2), we estimated the rate constant to be 2660
s−1 [17]. From this value we estimate DG‡138 for the
4. Discussion
4.1. Fluxional beha6ior of Ru(dppe)(CO)3
From previous work we discovered that a rate de-
termining electron-transfer step occurred in the reac-
tion between 1 and nitroarenes (Eq. (1)). A large
intrinsic barrier was measured (DG‡(0)=14 kcal
mol−1 at 25°C) for the electron-transfer reaction [35].
Part of this reorganization energy was thought to in-
volve a conformational change of 1 from trigonal
bipyramidal (tbp) to square pyramidal (spy), which
exchange process to be 6 kcal mol−1
.
4.2. Fluxional beha6ior of [Ru2(dppe)2(CO)6][PF6]2
Several different mechanisms for the fluxional be-
havior of 2 can be postulated: phosphine or carbonyl

S.J. Skoog et al. / Journal of Organometallic Chemistry 557 (1998) 13–28
25
dissociation, trigonal twist, Ru(I)–Ru(I) bond homol-
ysis, and exchange through bridging carbonyls. At the
fast-exchange limit the 13C{1H}-NMR spectrum shows
that Ca and Cb do not exchange with one another.
Each of the resonances retain coupling to the 31P
nuclei under conditions of both slow and fast ex-
change (Fig. 8) indicating that the exchange process is
intramolecular [37]. The quantitative agreement be-
tween the line-shape analysis of the 31P{1H}- and
13C{1H}- dynamic NMR (DNMR) spectra implies
that the rate of exchange of the phosphorus atoms
was the same as that of the carbon atoms. The obser-
vation that no exchange was observed between 2 and
gaseous 13CO rules out a mechanism involving ligand
dissociation. A trigonal-twist mechanism, often in-
Fig. 7. Eyring plot of the rates of exchange using the model described
in the text at different temperatures: (“) rates obtained from 31P{1H}
NMR line-shape analysis, (ꢀ) rates obtained from 13C{1H} NMR
line-shape analysis.
voked for carbonyl exchange in M(CO)3 units of clus-
ters [38], can be eliminated because it would require
exchange between Ca and Cb. Exchange involving
Ru(I)–Ru(I) bond homolysis would involve genera-
tion of a five-coordinate metal-centered radical inter-
mediate that could undergo rapid exchange followed
by radical recombination. This mechanism can also be
ruled out because it would involve scrambling of Ca
and Cb.
The bridging carbonyl mechanism is depicted in
Scheme 2 [39]. It involves migration of the carbonyls
trans to the phosphines to form a bridged intermedi-
ate, which affords the exchanged product upon open-
ing of the bridges. This mechanism does not involve
exchange between the different carbonyl environments
and was consistent with the agreement between the
rates obtained from the line-shape analysis of the
13C{1H} and 31P(1H} DNMR spectra (Fig. 7). Al-
though other stereochemistries of the bridged interme-
diate could be drawn these exchange pathways lead to
exchange between Ca and Cb.
Exchange of terminal carbonyls through bridged in-
termediates is well-known in examples including
Os3(CO)12−xLx clusters (where L can be various com-
binations of CO or PR3) [40–42] as well as in
(Me3P)(CO)4OsW(CO)5 [32], and MnRe(CO)10 [43].
Alex and Pomeroy have measured the activation en-
ergy for this process for Os3(CO)11[P(OMe)3] where
DG‡293=14 kcal mol−1 and compares favorably to the
activation energy measured for the exchange process
of 2 where DG‡293=13.8 kcal mol−1. Schmidt et al.
measured the activation parameters for the exchange
of 13CO between the two metal centers of
Re(13CO)5Mn(CO)5 and found DH‡=12.791.4 kcal
mol−1 and DS‡= −4196 eu [43]. This value for the
Fig. 6. Representative spectra of the experimental and simulated (121
MHz) 31P{1H} NMR spectra of
a
37 mM solution of
[Ru2(dppe)2(CO)6][PF6]2 (2) at the indicated temperatures. The exper-
imental spectra are in the left column, and the simulated spectra are
in the right column with corresponding rate constants used for the
simulation. At elevated temperatures 2 would decompose in CD2Cl2
and 1,2-C2H2Cl2 to form fac-[Ru(dppe)(CO)3Cl][PF6] (see results),
which exhibits a resonance at 54.6 ppm and can be seen in the various
spectra and was omitted in the spectrum at 50°C for clarity.

26
S.J. Skoog et al. / Journal of Organometallic Chemistry 557 (1998) 13–28
enthalpy of reaction compared favorably to DH‡ for
the exchange process for 2 (11.8 kcal mol−1) but the
DS‡ for 2 (−6.7 eu) was much larger than that for
Re(13CO)5Mn(CO)5.
4.3. Thermal and photochemical disproportionation of
[Ru2(dppe)2(CO)6][PF6]2
Although other mechanistic possibilities exist, based
on ample precedent [10,11,44–47] the disproportiona-
tion of 2 in CH3CN is thought to involve initial
metal–metal bond rupture to form a 17e− species
(Eq. (8)), followed by substitution of a carbonyl (Eq.
(9)). A second CH3CN addition forms a 19e− species
(Eq. (10)) prompting an electron transfer to another
radical cation (Eq. (11)) to yield products 1 and 3.
hw or D
’+
[Ru2(dppe)2(CO)6]2+ “ 2[Ru(dppe)(CO)3]
(8)
[Ru(dppe)(CO)3]’+ +MeCN
“[Ru(dppe)(CO)2MeCN]’+ +CO
(9)
[Ru(dppe)(CO)2MeCN]’+MeCN
’+
“[Ru(dppe)(CO)2(MeCN)2]
(10)
’+
[Ru(dppe)(CO)2(MeCN)2]’+ +[Ru(dppe)(CO)3]
“[Ru(dppe)(CO)2(MeCN)2]2+ +Ru(dppe)(CO)3
(11)
The bond dissociation energy for dimers containing
M–M bonds was reported to be 28 kcal mol−1 for
Mn2(CO)8(PPh3)2 [48], 25 kcal mol−1 for Fe(p5-
C5H5)(CO)2]2, and 22 kcal mol−1 for [Mo(p5-
C5H5)(CO)3]2 [49]. Substitution of a phosphine for a
CO ligand weakens M–M bonds for both steric and
electronic reasons [50]. For example Mn2(CO)10 fails
to undergo thermal M–M bond homolysis [51,52],
but Mn2(CO)8(PPh3)2 [48] and Mn2(CO)6(depe)2
(where depe is 1,2 bis(diethylphosphino)ethane) [53]
undergo spontaneous M–M bond homolysis upon
heating. Similar findings were observed by Brown and
coworkers by synthesizing persistent metal centered
radicals with bulky phosphine ligands Mn(CO)3(P(i-
Bu)3)2 [54]. Within a group the strength of M–M
bonds increases upon descending the triad [55]. The
effects of substitution and metal size are compensating
in the case of 2. In addition to the above mentioned
effects, the dicationic nature of 2 may further weaken
the Ru(I)–Ru(I) bond due to electrostatic effects.
If we assume that the thermal reactions exhibited
by 2 occur by rate determining bond cleavage, we
would be able to provide a rough measure of the
energetics of M–M bond rupture. This assumption is
consistent with the observed correlation between
products of the thermal and photochemical reactions
[56]. A first-order rate of 1.8×10−5 s−1 was mea-
Fig. 8. Variable-temperature (75.4 MHz) 13C{1H} NMR spectra in
the carbonyl region of
a
37 mM solution 13CO-labeled
[Ru2(dppe)2(CO)6][PF6]2 (2) at the indicated temperatures. The simu-
lated spectra are shown above the experimental spectra with corre-
sponding rate constants.

S.J. Skoog et al. / Journal of Organometallic Chemistry 557 (1998) 13–28
27
Scheme 2.
sured (at 50°C by IR spectroscopy of a 3.4 mM solu-
tion of 2) for the disproportionation reaction corre-
sponding to DG‡323=26 kcal mol−1. This represents a
reasonable value for the M–M bond homolysis pro-
cess [50].
DG‡343 ca. 26 kcal mol−1. This activation barrier was
similar to that of the disproportionation reaction and
consistent with the proposed M–M bond homolysis
mechanism.
4.4. Halogen atom abstraction
Acknowledgements
Irradiation of M2(CO)8L2 (where L is a monoden-
tate or bidentate phosphine and M=Mn, Re) causes
M–M bond homolysis and in the presence of alkyl
halides halogen atom abstraction occurs [57]. The
predominant isomer formed from halogen atom ab-
straction reactions in the Mn systems containing
bidentate phosphines is fac-Mn(CO)3L2X [53].
We thank Ken Caulton for helpful conversations.
The research was supported by a grant from the Na-
tional Science Foundation (CHE-9223433)
References
Of the two possible isomers for the halogen atom
abstraction of 2 only the fac-[Ru(dppe)(CO)3Cl]+ iso-
mer was formed. One possible reaction sequence in-
voking intermediates A and B to rationalize the
observed products is depicted in Scheme 3. The equi-
librium between intermediates A and B would be nec-
essary to form the observed products. Similar
equilibria have been proposed to account for the ob-
served preference for phosphine ligands to occupy the
basal position of sqpy Mn complexes generated upon
photolysis of the corresponding dimers [54,58].
[1] S.J. Skoog, J.P. Campbell, W.L. Gladfelter, Organometallics 13
(1994) 4137.
[2] S.J. Sherlock, D.C. Boyd, B. Moasser, W.L. Gladfelter, Inorg.
Chem. 30 (1991) 3626.
[3] J.D. Gargulak, A.J. Berry, M.D. Noirot, W.L. Gladfelter, J.
Am. Chem. Soc. 114 (1992) 8933.
[4] A.J. Kunin, M.D. Noirot, W.L. Gladfelter, J. Am. Chem. Soc.
111 (1989) 2739.
[5] J.K. Kochi, Organometallic Mechanism and Catalysis, Aca-
demic Press, New York, 1978.
[6] S.W. Blanch, A.M. Bond, R. Colton, Inorg. Chem. 20 (1981)
755.
[7] P.K. Baker, N.G. Connelly, B.M.R. Jones, J.P. Maher, K.R.
Somers, J. Chem. Soc. Dalton Trans. (1980) 579.
[8] P.K. Baker, K. Broadley, N.G. Connelly, J. Chem. Soc. Dal-
ton Trans. (1982) 471.
[9] W.C. Trogler, in: W.C. Trogler (Ed.), Organometallic Radical
Processes, Elsevier, Amsterdam 1990, pp. 306–337.
[10] M.J. Therien, C.-L. Ni, F.C. Anson, J.G. Osteryoung, W.C.
Trogler, J. Am. Chem. Soc. 108 (1986) 4037.
[11] L. Song, W.C. Trogler, J. Am. Chem. Soc. 114 (1992) 3355.
[12] N.G. Connelly, W.E. Geiger, Adv. Organomet. Chem. 23
(1984) 1.
The rate of the halogen atom abstraction reaction
at 70°C was estimated (from the half-life assuming a
first-order dependence on 2 by 31P{1H}-NMR spec-
troscopy) to be ca. 4×10−4 s−1 corresponding to
[13] N.G. Connelly, Chem. Soc. Rev. 18 (1989) 153.
[14] B. Moasser, C. Gross, W.L. Gladfelter, J. Organomet. Chem.
471 (1994) 201.
[15] B.J. Burger, J.E. Bercaw, in: A.L. Wayda, M.Y. Darensbourg
(Eds.), Experimental Organometallic Chemistry, vol. 357,
American Chemical Society, Washington D.C. 1987, pp. 79–
98.
[16] D.N. Hendrickson, Y.S. Sohn, H.B. Gray, Inorg. Chem. 10
(1971) 1559.
[17] J. Sandstrsm, Dynamic NMR Spectroscopy, Academic Press,
London, 1982.
[18] D.C. Roe, J. Mag. Res. 63 (1985) 388.
[19] R.A. Sanchez-Delgado, J.S. Bradley, G. Wilkinson, J. Chem.
Soc. Dalton Trans. (1976) 399.
Scheme 3.

28
S.J. Skoog et al. / Journal of Organometallic Chemistry 557 (1998) 13–28
[38] R.F. Alex, R.K. Pomeroy, Organometallics 6 (1987) 2437.
[39] E. Band, E.L. Muetterties, Chem. Rev. 78 (1978) 639.
[40] A. Forster, B.F.G. Johnson, J. Lewis, T.W. Matheson, B.H.
Robinson, W.G. Jackson, J. Chem. Soc. Chem. Com. (1974)
1042.
[41] A.J. Demming, S. Donovan-Mtunzi, S.E. Kabir, J. Organomet.
Chem. 281 (1985) C43.
[42] R.F. Alex, R.K. Pomroy, Organometallics 6 (1987) 2437.
[43] S.P. Schmidt, F. Basolo, C.M. Jensen, W.C. Trogler, J. Am.
Chem. Soc. 108 (1986) 1894.
[20] A.J. Gordon, R.A. Ford, The Chemist’s Companion, John Wiley
and Sons, New York, 1972.
[21] G. HSegele, S. Goudestsidis, H.W. Hsffken, T. Lenzen, R.
Spiske, U. Weber, Phosphorus Sulphur Silicon 77 (1993) 262.
[22] A. Pines, M.G. Gibby, J.S. Waugh, J. Chem. Phys. 59 (1973)
569.
[23] SHELXTL-plus V5.0, Siemens Industrial Automation, Inc.,
Madison, WI, 1995.
[24] G.M. Sheldrick, Acta Crystallogr. A46 (1990) 467.
[25] A.L. Spek, Acta. Cryst. A46 (1990) C34.
[26] A.J.C. Wilson, International Tables for Crystallography, Vol. C,
Kluwer Academic Publishers, Dordrecht, 1995.
[27] J.D. Gargulak, W.L. Gladfelter, Inorg. Chem. 33 (1994) 253.
[28] L.P. Battaglia, D. Delledonne, M. Nardelli, C. Pelizzi, G.
Predieri, G.P. Chiusoli, J. Organomet. Chem. 330 (1987) 101.
[29] S.K. Mandal, D.M. Ho, M. Orchin, Inorg. Chem. 30 (1991)
2244.
[30] R.A. Levenson, H.B. Gray, J. Am. Chem. Soc. 97 (1975) 6042.
[31] W.G. Klemperer, B. Zhong, Inorg. Chem. 32 (1993) 5821.
[32] F.W.B. Einstein, T. Jones, R.K. Pomeroy, P. Rushman, J. Am.
Chem. Soc. 106 (1984) 2707.
[33] M.R. Churchill, K.N. Amoh, H.J. Wasserman, Inorg. Chem. 20
(1981) 1609.
[34] G. Binsch, H. Kessler, Angew. Chem. Int. Ed. Engl. 19 (1980)
411.
[35] S.J. Skoog, W.L. Gladfelter, J. Am. Chem. Soc. 119 (1997)
11049.
[36] M. Akhtar, P.D. Ellis, A.G. MacDiarmid, J.D. Odom, Inorg.
Chem. 11 (1972) 2917.
[44] M. Absi-Halabi, J.D. Atwood, N.P. Forbus, T.L. Brown, J. Am.
Chem. Soc. 102 (1980) 6248.
[45] A. Sisak, L. Mark-, J. Organomet. Chem. 330 (1987) 201.
[46] A.E. Stiegman, D.R. Tyler, Acc. Chem. Res. 17 (1984) 61.
[47] S.B. McCullen, T.L. Brown, Inorg. Chem. 20 (1981) 3528.
[48] A. Poe¨, C.V. Sekhar, J. Am. Chem. Soc. 107 (1985) 4874.
[49] J.R. Pugh, T.J. Meyer, J. Am. Chem. Soc. 114 (1992) 3784.
[50] R.A. Jackson, A. Poe¨, Inorg. Chem. 17 (1978) 997.
[51] D. Sonnenberger, J.D. Atwood, J. Am. Chem. Soc. 102 (1980)
3484.
[52] S.P. Schmidt, W.C. Trogler, F. Basolo, Inorg. Chem. 21 (1981)
1698.
[53] D.R. Tyler, A.S. Goldman, J. Organomet. Chem. 311 (1986)
349.
[54] S.B. McCullen, T.L. Brown, J. Am. Chem. Soc. 104 (1982) 7496.
[55] J.A. Connor, H.A. Skinner, in: M.H. Chisholm (Ed.), Reactivity
of Metal–Metal Bonds, Vol. ACS Symp. Ser. No 155, American
Chemical Society, Washington D.C., 1981, pp. 197–205.
[56] D.R. Tyler, Prog. Inorg. Chem. 36 (1988) 125.
[57] M.C. Baird, Chem. Rev. 88 (1988) 1217.
[37] M.A. Gallop, B.F.G. Johnson, J. Lewis, J. Chem. Soc. Chem.
Commun. (1987) 1831.
[58] H.W. Walker, G.B. Rattinger, R.L. Belford, T.L. Brown,
Organometallics 2 (1983) 775.
.