Synthesis, crystal structures and isomerization kinetics of a series of [Co(dtc)2{P(OMe)3-nPhn}2]+ (n = 0-2) complexes (dtc- = N,N-dimethyldithiocarbamate): role of ?-donicity, ?-acidity, and cone angle of the P-ligands in the trans influence and trans effect

Article abstract of DOI:10.1039/b301712e

A series of geometrical isomers of the new cobalt(III) complexes containing P(OMe)3-nPhn (n = 0, 1 or 2), trans- and cis-[Co(dtc)2{P(OMe)3-nPhn}2]⁺ (dtc^- = N,N-dimethyldithiocarbamate), have been synthesized and the crystal structures and spectrometric properties were obtained. By comparison with reported structures of related [Co(dtc)2(P-ligand)2]⁺ complexes, it was shown that the Co-P bond lengths in [Co(dtc)2)P-ligand)2]⁺ were linearly dependent on the Tolman's cone angle, θ_T, of the P-ligands. On other hand, the Co-P bond lengths as well as the transition energies of the splitting component (¹A_1g -> a¹E_g:D_4h) of the first d-d band of trans-[Co(dtc)2(P-ligand)2]⁼ were not linearly dependent on the Tolman's electronic parameter (χ) of the P-ligands. In the trans-isomers, it was found that the mutual electronic trans influence was not significant in the [Co(dtc)2{P(OMe)3-nPhn}2]⁺ species while it was far more evident in the complexes with strong ?-donating phosphines such as PMe3 and PMe2Ph. A reversible cis-trans inter-conversion was observed for the [Co(dtc)2{P(OMe)3-nPhn}2]⁺ complexes: cis to trans isomerization was induced by irradiation of light while slow thermal trans to cis isomerization was observed as a dark reaction. Kinetic measurements of thermal trans to cis isomerization reactions revealed that the absorption change with time followed multi-exponential kinetics when no free P-ligand was added to the reaction mixture: by addition of excess amounts of free P-ligand, first-order kinetic traces were observed for all reactions. The apparent first-order rate constants were independent of the concentration of added free ligands for the isomerization reactions of [Co(dtc)2{P(OMe)2Ph}2]⁺ and [Co(dtc)2{P(OMe)Ph2}2]⁺, while saturation kinetics were observed for the reaction of [Co(dtc)2{P(OMe)3}2]⁺: all isomerization involved pre-dissociation of one of the coordinated P-ligands. It was indicated that a significant amount of square pyramidal [Co(dtc)2{P(OMe)3}]⁺ exists in equilibrium with [Co(dtc)2{P(OMe)3}2]⁺, while merely a trace amount of such a 5-coordinate species was involved in the reactions of [Co(dtc)2{P(OMe)2Ph}2]⁺ and [Co(dtc)2{P(OMe)Ph2}2]⁺. A contradictory phenomenon observed in these complexes, a small static trans influence with a significant kinetic trans effect, was explained by the combination of (1) a lack of strong ?-donicity of P(OMe)3-nPhn and (2) stabilization of the 5-coordinate square-pyramidal geometry through the electron sponge effect by the dtc^- ligand.

Full text of DOI:10.1039/b301712e

View Article Online / Journal Homepage / Table of Contents for this issue
Syntheses, crystal structures and isomerization kinetics of a series
of [Co(dtc)2{P(OMe)3 ؊ nPhn}2]؉ (n ؍ 0–2) complexes
(dtc؊ ؍ N,N-dimethyldithiocarbamate): role of ꢀ-donicity,
ꢁ-acidity, and cone angle of the P-ligands in the trans influence
and trans effect†
Satoshi Iwatsuki,a Shinsaku Kashiwamura,a Kazuo Kashiwabara,a Takayoshi Suzuki*b and
Hideo D. Takagi*a
a Department of Chemistry, Graduate School of Science and Research Center for Materials
Science, Nagoya University, Nagoya 464-8602, Japan
b Department of Chemistry, Graduate School of Science, Osaka University, Toyonaka 560-0043,
Japan
Received 12th February 2003, Accepted 10th April 2003
First published as an Advance Article on the web 29th April 2003
A series of geometrical isomers of the new cobalt() complexes containing P(OMe)3 Ϫ nPhn (n = 0, 1 or 2),
trans- and cis-[Co(dtc)2{P(OMe)3 Ϫ nPhn}2]ϩ (dtcϪ = N,N-dimethyldithiocarbamate), have been synthesized and the
crystal structures and spectrometric properties were obtained. By comparison with reported structures of related
[Co(dtc)2(P-ligand)2]ϩ complexes, it was shown that the Co–P bond lengths in [Co(dtc)2(P-ligand)2]ϩ were linearly
dependent on the Tolman’s cone angle, θT, of the P-ligands. On the other hand, the Co–P bond lengths as well as the
transition energies of the splitting component (1A1g
a1Eg: D4h) of the first d–d band of trans-[Co(dtc)2(P-ligand)2]ϩ
were not linearly dependent on the Tolman’s electronic parameter (χ) of the P-ligands. In the trans-isomers, it was
found that the mutual electronic trans influence was not significant in the [Co(dtc)2{P(OMe)3 Ϫ nPhn}2]ϩ species while
it was far more evident in the complexes with strong σ-donating phosphines such as PMe3 and PMe2Ph. A reversible
cis–trans inter-conversion was observed for the [Co(dtc)2{P(OMe)3 Ϫ nPhn}2]ϩ complexes: cis to trans isomerization
was induced by irradiation of light while slow thermal trans to cis isomerization was observed as a dark reaction.
Kinetic measurements of thermal trans to cis isomerization reactions revealed that the absorption change with
time followed multi-exponential kinetics when no free P-ligand was added to the reaction mixture: by addition of
excess amounts of free P-ligand, first-order kinetic traces were observed for all reactions. The apparent first-order
rate constants were independent of the concentration of added free ligands for the isomerization reactions of
[Co(dtc)2{P(OMe)2Ph}2]ϩ and [Co(dtc)2{P(OMe)Ph2}2]ϩ, while saturation kinetics were observed for the reaction
of [Co(dtc)2{P(OMe)3}2]ϩ: all isomerizations involved pre-dissociation of one of the coordinated P-ligands.
It was indicated that a significant amount of square pyramidal [Co(dtc)2{P(OMe)3}]ϩ exists in equilibrium with
[Co(dtc)2{P(OMe)3}2]ϩ, while merely a trace amount of such a 5-coordinate species was involved in the reactions of
[Co(dtc)2{P(OMe)2Ph}2]ϩ and [Co(dtc)2{P(OMe)Ph2}2]ϩ. A contradictory phenomenon observed in these complexes,
a small static trans influence with a significant kinetic trans effect, was explained by the combination of (1) a lack of
strong σ-donicity of P(OMe)3 Ϫ nPhn and (2) stabilization of the 5-coordinate square-pyramidal geometry through the
electron sponge effect by the dtcϪ ligand.
(PHPh2) complexes containing the dtcϪ ligand:9 it was indi-
cated that PHPh2 exhibits relatively strong π-acidity and that
Introduction
Preparations and characterization of mixed-ligand dithio-
carbamatocobalt() complexes with various P-ligands (ligands
the interplay of Tolman’s cone angle, σ-donicity and π-acidity,
together with the electron sponge effect of dtcϪ determines the
with phosphorus atoms at their coordination sites) have
stability of the complex and the bond length between cobalt()
and P, although Giering and co-workers had suggested that the
been reported to date.1–9 In particular, their cobalt()–
bis(phosphite) complexes (phosphite = P(OMe)3, P(OEt)3, and
P(OCH2)3CEt) are examples of rare air-stable compounds with
cobalt()–phosphite bonds.5 These phosphite ligands are
expected to exhibit weak σ-donicity, while π-acidity of these
phosphites are believed to be high.5 In the previously reported
π-acidity of PHPh2 is negligibly small.10 We also reported the
thermal trans to cis isomerization reaction of trans-[Co(dtc)2-
(PHPh2)2]ϩ in acetonitrile:9a although such thermal isomeriz-
ation reactions of octahedral complexes have been known
to take place either through the twist mechanism or through
dissociation of one of the coordinated ligands,11 it was clearly
shown that the activation enthalpies for the dissociative isomer-
ization process that involves stereochemical change from the
square pyramidal intermediate to the trigonal bipyramidal
transition state was much higher than that for the twist process
in the case of inert low-spin d6 metal complexes.12,13 In addition,
we also reported that the trans-[Co(dtc)2(PHPh2)2]ϩ complex
very slowly releases coordinated PHPh2.9a Such a tendency was
also explained by the relatively strong π-interaction between
Co() and P, as the σ-donicity of PHPh2, represented by the
pKa value of the conjugate acid as well as Tolman’s electronic
parameter, is very small.10 However, it should be noted that the
studies of various [Co(dtc)2(P-ligand)2]ϩ complexes (dtcϪ
=
N,N-dimethyldithiocarbamate),1–9 it was observed that the
bond lengths between cobalt() and P were not related to the
Tolman’s electronic parameter of each ligand.
We recently succeeded in the determination of crystallo-
graphic structures for cis- and trans-bis(diphenylphosphine)
† Electronic supplementary information (ESI) available: bond lengths
and angles for all of the trans- and cis-complexes are summarized in
Table S1–S6. The structures of the trans-isomers with different orien-
tations and conformations of the OMe and Ph substituents on the P
atoms are shown in Figs. S1–S3. See http://www.rsc.org/suppdata/dt/b3/
b301712e/
D a l t o n T r a n s . , 2 0 0 3 , 2 2 8 0 – 2 2 9 2
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
2280

View Article Online
Table 1 Results of the elemental analyses and the 1H NMR signals (δ) for the N–CH3 (dtcϪ) and O–CH3 (POMe) groups
Elemental analyses, found (calcd) (%)
1H NMR signals (δ)a
N–CH3 (dtcϪ)
O–CH3 (POMe)
Complex
C
H
N
cis-[Co(dtc)2{P(OMe)3}2]PF6
20.84 (20.81)
33.51 (33.68)
42.82 (43.84)
20.40 (20.81)
33.60 (33.68)
43.87 (43.84)
4.62 (4.37)
4.28 (4.37)
4.36 (4.37)
3.73 (4.37)
4.14 (4.37)
4.20 (4.37)
4.07 (4.05)
3.50 (3.57)
3.34 (3.20)
3.86 (4.06)
3.50 (3.57)
3.15 (3.20)
3.24, 3.28
2.96, 2.99
2.83, 2.88
3.31
2.99
2.72
3.86 (t)
cis-[Co(dtc)2{P(OMe)2Ph}2]PF6
cis-[Co(dtc)2{P(OMe)Ph2}2]PF6
trans-[Co(dtc)2{P(OMe)3}2]PF6
trans-[Co(dtc)2{P(OMe)2Ph}2]PF6
trans-[Co(dtc)2{P(OMe)Ph2}2]PF6
3.84 (t), 3.92 (t)
3.33 (filled-in d)
3.79 (t)
3.70 (t)
3.62 (t)
a In CDCl3 at 30 ЊC. The term “filled-in doublet” is used to describe a distorted triplet with a broad central peak.
dissociation tendency of PHPh2 in trans-[Co(dtc)2(PHPh2)2]ϩ,
although it is very slow (10Ϫ4–10Ϫ5 sϪ1 at 333 K), cannot be
expected from the relatively short bond length between
cobalt() and PHPh2,9a as it is known that trans-[Co(dtc)2-
(PMe3)2]ϩ in which the average Co()–P bond length (2.287 Å)
is longer than that in trans-[Co(dtc)2(PHPh2)2]ϩ (2.276 Å),
does not undergo either dissociation of coordinated PMe3 or
isomerization to the cis complex.8
Historically, the trans influence by which a metal to ligand
bond trans to the particular donor atom that is strongly bound
to the metal center suffers elongation has been suggested to be
related to the ground-state property inherent in such complexes,
while the trans effect by which lability of a ligand trans to
a particular donor atom increases has been related to both
ground state and transition-state characteristics.14 Therefore,
trans influence and trans effect are not necessarily related to
each other: to fully understand these two phenomena it
is necessary to examine the metal–ligand bond lengths and
lability of metal complexes with a series of ligands possessing
different donor/acceptor properties.
In this study, we successfully synthesized a series of [Co(dtc)2-
{P(OMe)3 Ϫ nPhn}2]ϩ complexes (n = 0–2), and the crystal struc-
tures of both cis- and trans-isomers for this series of complexes
were obtained. The spectrophotometric and NMR data were
also collected for the purpose of examining the static Co–P
interactions, while the trans to cis isomerization reactions of
these complexes were studied for the purpose of examining
the kinetic stability of the Co–P bonds. The pKa values of the
conjugate acids for P(OMe)3, P(OMe)2Ph, and P(OMe)Ph2 are
2.60, 2.64, and 2.69, while Tolman’s electronic parameters, χ,
for these ligands are 24.10, 19.45, and 16.30, respectively.10
Therefore, the σ-donicities of these three ligands are similar to
each other, while the χ parameters drastically decrease in the
order given. It is also known that the cone angle for P(OMe)Ph2
is the largest, 132Њ, and it decreases to 115Њ and 107Њ for
P(OMe)2Ph and P(OMe)3, respectively.15 Moreover, the χ
parameter for PHPh2 lies between those for P(OMe)2Ph and
P(OMe)Ph2, 17.35, while the pKa for PHPh2 is much smaller,
0.03, compared with those for these three phosphites.10
The structure and reactivity for trans- and cis-[Co(dtc)2-
{P(OMe)3 Ϫ nPhn}2]ϩ was compared with those for previously
reported trans- and cis-[Co(dtc)2(P-ligand)2]ϩ and trans- and
cis-[Co(acac)2(P-ligand)2]ϩ complexes to shed light on the exact
nature of the Co–P interactions and the role of the spectator
ligand.
Preparation of complexes
cis-[Co(dtc)2{P(OMe)3 ؊ nPhn}2](PF6 or BF4) (n ؍ 0, 1 and 2).
The literature method5 for the preparation of cis-[Co(dtc)2-
{P(OCH2)3CEt}2]BF4 was somewhat modified as follows. To an
ice-cold methanol solution (300 cm3) containing Co(BF4)2ؒ
6H2O (0.85 g, 2.5 mmol) and P(OMe)Ph2 (2.3 g, 10 mmol)
added dropwise was an ice-cold solution of tetramethylthiuram
disulfide (0.31 g, 1.3 mmol) in a mixture of dichloromethane
(10 cm3) and tetrahydrofuran (50 cm3) with stirring over 20 min.
The mixture was stirred in an ice bath for a further 2 h, and the
resulting red brown solution was evaporated to dryness under
reduced pressure. The residue was extracted with a small
amount of methanol, leaving a green precipitate of [Co(dtc)3]
behind. The filtered red brown extract was placed on a column
(7 × 35 cm) of Sephadex LH-20 resin. The adsorbed products
were eluted with methanol, affording a major red brown band
together with several minor bands. The reddish brown band
was collected and concentrated to 20 cm3 under reduced pres-
sure. Addition of excess NH4PF6 to the concentrated reddish
brown solution yielded red brown crystals, which were washed
with diethyl ether (20 cm3) and air-dried. Yield: 0.54 g (25%).
The corresponding P(OMe)2Ph and P(OMe)3 complexes were
also prepared by a similar method and the yield was 10–20%.
1
Elemental analyses are listed in Table 1, together with the H
NMR chemical shifts of the signals due to the CH3 group in
dtcϪ and P–OCH3.
trans-[Co(dtc)2{P(OMe)3 ؊ nPhn}2](PF6 or BF4) (n ؍ 0, 1 and
2). The trans-[Co(dtc)2{P(OMe)3 Ϫ nPhn}2]PF6 complexes were
deposited by irradiation of a methanol solution containing the
cis-isomer and excess NH4PF6 with a high-pressure mercury
lamp. The yield was 40–50%. The elemental analyses and the 1H
NMR data of the complexes are listed in Table 1. It was also
possible to prepare the trans-[Co(dtc)2{P(OMe)3 Ϫ nPhn}2]BF4
complexes by the reactions of P(OMe)3 Ϫ nPhn (n = 0, 1 or 2)
with trans-[Co(dtc)2(PPh3)2]BF4:8 the yield was ca. 50%. An
attempt to isolate trans-[Co(dtc)2{P(OMe)2Ph}2]BF4 and trans-
[Co(dtc)2{P(OMe)3}2]BF4 by an identical column separation
method to that employed for the preparation of cis-isomers was
also successful. The yield of trans-[Co(dtc)2{P(OMe)2Ph}2]BF4
was 78%, while the yield of trans-[Co(dtc)2{P(OMe)3}2]BF4 was
significantly reduced to 39% under the same conditions.
X-Ray crystallographic study
Single crystals of cis-[Co(dtc)2{P(OMe)3}2]BF4 were obtained
from methanol and those of cis-[Co(dtc)2{P(OMe)2Ph}2]PF6
and cis-[Co(dtc)2{P(OMe)Ph2}2]PF6 were deposited by slow
evaporation of a dichloromethane/methanol solution at room
temperature, while trans-[Co(dtc)2{P(OMe)3}2]BF4 and trans-
[Co(dtc)2{P(OMe)2Ph}2]BF4ؒ(CH3)2CO deposited as crystals
by diffusion of diethyl ether vapor into an acetone solution.
Thinplatecrystalsoftrans-[Co(dtc)2{P(OMe)Ph2}2]BF4ؒ0.5CH3-
OH were obtained by diffusion of diethyl ether vapor into a
methanol solution containing a few drops of acetone in a
refrigerator. Each crystal suitable for an X- ray diffraction study
was glued on top of a glass fiber with epoxy resin. Except for
Experimental
Syntheses
General procedures. The phosphorus ligands, P(OMe)3 Ϫ n
-
Phn, purchased from Aldrich Inc. or Wako Pure Chemicals Inc.
were used without further purification and handled under an
atmosphere of argon using standard Schlenk techniques until
they formed the air-stable cobalt() complexes. All of the sol-
vents used in the preparation were deaerated with argon
immediately before use.
D a l t o n T r a n s . , 2 0 0 3 , 2 2 8 0 – 2 2 9 2
2281

View Article Online
trans-[Co(dtc)2{P(OMe)Ph2}2]BF4ؒ0.5CH3OH, the X-ray dif-
fraction data were collected at 23(2) ЊC on a Rigaku AFC-5R
four circle diffractometer equipped with graphite-monochrom-
ated Mo-Kα radiation (λ = 0.71073 Å). Final lattice parameters
were determined by least-squares treatment using setting angles
of 25 reflections in the range of 27 < 2θ < 30Њ. The intensities
were corrected for Lorentz-polarization factors and for absorp-
tion effects either by the numerical integration method16 or
by the empirical method using three sets of Ψ-scan data.17
The diffraction data of trans-[Co(dtc)2{P(OMe)Ph2}2]BF4ؒ
0.5CH3OH were obtained at 23(2) ЊC on a Rigaku Mercury
CCD detector with graphite-monochromated Mo-Kα radiation
(λ = 0.71073 Å). A total of 1940 images with an oscillation
angle of ω = 0.2Њ were collected with 5 different goniometer
settings (8 < 2θ < 55Њ, exposure time = 15 s). Data were pro-
cessed by the CrystalClear program package,18 and the lattice
parameters were determined by least-squares treatment using
the setting angles of all observed reflections. Absorption correc-
tions were applied by the multi-scan method.19 The structures
were solved by direct methods using the SHELXS8620 or
SHELXS9721 programs, and refined on F 2 (with all independ-
ent reflections) using the SHELXL97 program21 with aniso-
tropic thermal parameters for all non-hydrogen atoms. H atoms
were introduced at the theoretically calculated positions and
treated with riding models. All calculations were carried out
using the TeXsan software package.22 Crystallographic data are
collected in Table 2, and selected bond lengths and angles
obtained for all complexes are summarized in Table 3.
did not produce the corresponding trans isomer, which is in
contrast to the case of analogous PMe2Ph and PMePh2
complexes.8
It was possible to efficiently prepare the trans-isomers by the
reactions of P(OMe)3 Ϫ nPhn with trans-[Co(dtc)2(PPh3)2]BF4
and by the photochemical conversions of the cis-isomers in the
presence of excess NH4PF6 (yield : 40–80%. The trans-isomers
thermally isomerized back to the original cis-isomers with a
very slow rate at room temperature. Such photochemical and
thermal inter-conversions between two isomers have also been
observed for other [Co(dtc)2(phosphite or PHPh2)2]ϩ com-
plexes,5,9 while no isomerization reaction was observed for
tertiary phosphine complexes such as [Co(dtc)2(PMe3 Ϫ nPhn)2]ϩ
(n = 0–3).8 As noted in the section on kinetic studies, the sub-
stituents on phosphorus play a critical role in the isomerization
process. The quantum yields, Φ, for the photochemical isomer-
ization of cis-[Co(dtc)2{P(OMe)Ph2 or P(OMe)2Ph}2]ϩ were
0.17–0.20 upon irradiation at 405 nm, which is comparable to
that for the corresponding P(OCH2)3CEt complex (Φ = 0.17).5
Crystal structures
In the crystals of trans-[Co(dtc)2{P(OMe)3}2]BF4 and trans-
[Co(dtc)2{P(OMe)2Ph}2]BF4ؒ(CH3)2CO, two Co atoms, Co(1)
and Co(2), were located at a crystallographic center of sym-
metry and the asymmetric unit contains two halves of the com-
plex cation and a whole BF4Ϫ anion (and an acetone molecule).
The molecular structures of one of the two independent com-
plex cations oftrans-[Co(dtc)2{P(OMe)3}2]ϩ andtrans-[Co(dtc)2-
{P(OMe)2Ph}2]ϩ are shown in Figs. 1a and 1b, respectively. The
other complex cations exhibited similar structures to these with
different orientations and conformations of the OMe and Ph
substituents on the P atoms (see Figs. S1 and S2†). For the trans-
[Co(dtc)2{P(OMe)Ph2}2]BF4ؒ0.5CH3OH complex, it was found
that there were three positions for the Co atoms: Co(1) at a
general position and Co(51) and Co(52) at a crystallographic
center of symmetry. Thus, the asymmetric unit consists of one
CCDC reference numbers 206913–206918.
See http://www.rsc.org/suppdata/dt/b3/b301712e/ for crystal-
lographic data in CIF or other electronic format.
Measurements
The NMR spectra were obtained in chloroform-d at 30 ЊC on a
JEOL Lambda 500 spectrometer. UV-Vis absorption spectra in
dichloromethane were measured with a Perkin- Elmer Lambda
19 spectrophotometer at room temperature. For the kinetic
measurements in acetonitrile-d3, a Bruker AMX400WB spec-
trometer was used at controlled temperatures. Kinetic meas-
urements of the isomerization reactions were also monitored
under an Ar atmosphere by JASCO V-560, V-570, Shimadzu
UV-1600, and Hitachi U-3400 UV-VIS-NIR spectrophoto-
meters. The temperature of all sample solutions was held con-
stant within 0.2 K. Acetonitrile used for the kinetic measure-
ments was obtained from Wako Pure Chemicals Inc., and puri-
fied by distillation from phosphorus pentoxide. The amount of
residual water in thus purified acetonitrile was examined by a
Mitsubishi Kasei CA01 Karl-Fisher apparatus, by which it was
determined to be ca. 1 mmol kgϪ1. Trace amounts of water
(ca. < 10 mmol kgϪ1) introduced into the sample solutions dur-
ing preparation are mentioned as “impure water” throughout
this article, otherwise the solvent is very pure. Tetra-n-butyl-
ammonium tetrafluoroborate (0.1 mol kgϪ1) was used to adjust
the ionic strength. All sample solutions were prepared under an
inert atmosphere of Ar.
Ϫ
whole and two half complex cations, two BF4 anions and a
methanol molecule. The molecular structures of the three crys-
tallographically independent complex cations were different in
the orientation of the phenyl rings of P(OMe)Ph2 (Fig. S3†).
In complex 1 with Co(1) at a general position, two crystallo-
graphically independent P(OMe)Ph2 ligands exhibited orien-
tation of their phenyl rings as found in either complex 51 or
complex 52 (Fig. 1c and Table 3).
The structures of the Co(dtc)2 moiety in trans-[Co(dtc)2-
{P(OMe)3 Ϫ nPhn}2]ϩ are similar to those found in the previously
reported trans-[Co(dtc)2(PMe3 Ϫ nPhn or PHPh2)2]ϩ com-
plexes:8,9 the Co–S bond lengths and the S–Co–S bite angles of
dtcϪ are in the range of 2.264–2.278 Å and 76.0–77.0Њ, respect-
ively (Table 4).
It is known that the Co–P bonds in trans-[CoIII(bidentate
monoanionic ligand)2(P-ligand) ]ϩ complexes are often elon-
2
gated by steric interactions (the mutual steric trans influence)
with the bidentate ligands in the equatorial plane: pentane-2,4-
dionate (acacϪ)23 and dimethylglyoximate (HdmgϪ)24 com-
plexes are included in this type of complex. However, such a
mutual steric trans influence does not seem to exist in the com-
plexes with dtcϪ as the bidentate monoanionic ligand. This is
probably caused by the small bulk of the dtcϪ ligand compared
with acacϪ and HdmgϪ,8 in addition to the electron sponge
action of dtcϪ as described in the latter section. The average
Co–P bond lengths in the trans-[Co(dtc)2(P-ligand)2]ϩ com-
plexes are summarized in Table 4, and the plots of Co–P length
against the Tolman’s cone angle15 of each P-ligand is shown in
Fig. 2, which clearly indicates that there is a linear relationship
between the cone angle and the Co–P bond length for the series
of trans-[Co(dtc)2(P-ligand)]ϩ complexes. As the σ-donicity for
P(OMe)3, P(OMe)2Ph, P(OMe)Ph2 and PPh3 is unequivocally
Results and discussion
Preparations
The cis-[Co(dtc)2{P(OMe)2Ph or P(OMe)Ph2}2]PF6 complexes
were prepared by a similar method to that previously reported
for cis-[Co(dtc)2 (phosphite)2]BF4 (phosphite
P(OEt)3 or P(OCH2)3CEt)5 and for cis-[Co(dtc)2(PHPh2)2]BF4.9
The reaction of a mixture of Co(BF4)2ؒ6H2O and P(OMe)3 Ϫ n
Phn (n = 0–2) in methanol with tetramethylthiuram disulfide
= P(OMe)3,
-
in a molar ratio of 2 : 8 : 1 afforded cis-[Co(dtc)2{P(OMe)3 Ϫ n
-
Ϫ
Phn}2]ϩ, which was isolated as the PF6 salt with yields of
10–25%. It was not possible to improve the yield because of the
preferential formation of [Co(dtc)3]. This preparative method
low and does not differ greatly from ligand to ligand (χd
=
D a l t o n T r a n s . , 2 0 0 3 , 2 2 8 0 – 2 2 9 2
2282

Table 2 Crystallographic data
cis-[Co(dtc)2{P(OMe)3}2]-
BF4
cis-[Co(dtc)2{P(OMe)2-
Ph}2]PF6
cis-[Co(dtc)2{P(OMe)-
Ph2}2]PF6
trans-[Co(dtc)2-
{P(OMe)3}2]BF4
trans-[Co(dtc)2{P(OMe)2Ph}2]-
BF4ؒ(CH3)2CO
2 × [trans-[Co(dtc)2{P(OMe)Ph2}2]-
BF4ؒ0.5CH3OHa
Complex
Chemical formula
Formula weight
C12H30BCoF4N2O6P2S4
634.30
C22H34CoF6N2O4P3S4
784.59
C32H38CoF6N2O2P3S4
876.72
C12H30BCoF4N2O6P2S4
634.30
C25H40BCoF4N2O5P2S4
784.51
C65H80B2Co2F8N4O5P4S8
834.59 × 2
Color and shape
Red, column
Red-orange, prism
Red, prism
Red, prism
Red, prism
Red, plate
of crystal
Crystal system
Space group
Monoclinic
P21/n
Monoclinic
P21/c
Orthorhombic
Pbca
Triclinic
P1
Monoclinic
P21/a
Triclinic
P1
¯
¯
a/Å
b/Å
c/Å
α/Њ
7.323(4)
13.349(3)
26.733(3)
90
95.56(2)
90
2601(2)
4
1.620
1304
1.163
0.041
7587
11.732(3)
19.827(5)
14.553(3)
90
94.91(2)
90
3373(1)
4
1.545
1608
0.963
0.107
7728
18.893(6)
25.288(3)
16.688(2)
90
90 109.07(1)
90
7973(3)
8
1.461
3600
0.820
0.027
11635
11.170(2)
12.212(2)
12.319(2)
109.67(2)
94.41(1)
107.36(2)
1328.7(4)
2
1.585
652
1.138
0.020
11.025(2)
12.594(2)
25.581(2)
90
84.228(7)
90
3541.5(8)
4
1.471
1624
0.868
0.018
10336
9.6508(5)
10.9697(8)
37.114(3)
88.261(8)
β/Њ
γ/Њ
85.066(8)
3893.8(5)
2
1.424
1724
0.790
0.039
12588
U/Å3
Z
Dcalc/Mg mϪ3
F(000)
µ(Mo-Kα)/mmϪ1
Rint
No. of independent
reflections
10336
.
R1 (F 2: Fo2 > 2σ(Fo ))
0.044
0.134
0.057
0.132
0.047
0.161
0.054
0.199
0.045
0.188
0.093
0.188
2
wR2 (F 2: all data)
a Two trans-[Co(dtc)2{P(OMe)Ph2}2]BF4in the unit cell.

D
Table 3 Selected bond lengths (Å) and angles (Њ) for the trans- and cis-[Co(dtc)2{P(OMe)3 Ϫ nPhn}2]ϩ complexes
trans-isomers
cis-isomers
trans-[Co(dtc)2{P(OMe)3}2]BF4
cis-[Co(dtc)2{P(OMe)3}2]BF4
Co(1)–P(1)
Co(1)–S(1)
Co(1)–S(2)
S(1)–Co(1)–S(2)
2.239(1)
2.273(1)
2.274(1)
76.42(4)
Co(2)–P(2)
Co(2)–S(3)
Co(2)–S(4)
S(3)–Co(2)–S(4)
2.243(1)
2.265(1)
2.280(1)
76.76(5)
Co–P(1)
Co–S(1)
Co–S(2)
P(1)–Co(1)–P(2)
S(1)–Co(1)–S(2)
2.200(1)
Co–P(2)
Co–S(3)
Co–S(4)
2.200(1)
2.266(2)
2.297(1)
2.275(2)
2.294(1)
97.16(4)
75.59(4)
S(3)–Co(1)–S(4)
75.96(4)
trans-[Co(dtc)2{P(OMe)2Ph}2]BF4ؒ(CH3)2CO
cis-[Co(dtc)2{P(OMe)2Ph}2]PF6
Co(1)–P(1)
Co(1)–S(1)
Co(1)–S(2)
2.263(2)
2.259(1)
2.268(1)
76.89(5)
28.5(1)
Co(2)–P(2)
Co(2)–S(3)
Co(2)–S(4)
2.269(1)
2.268(1)
2.265(1)
76.99(5)
26.8(2)
Co–P(1)
Co–S(1)
Co–S(2)
S(1)–Co(1)–S(2)
P(1)–Co(1)–P(2)
2.222(3)
2.260(2)
2.285(2)
76.53(8)
94.4(1)
Co–P(2)
Co–S(3)
Co–S(4)
S(3)–Co(1)–S(4)
2.214(3)
2.265(2)
2.293(3)
76.35(9)
S(1)–Co(1)–S(2)
S(3)–Co(2)–S(4)
a
b
ꢀ1
ꢀ2
trans-[Co(dtc)2{P(OMe)Ph2}2]BF4ؒ0.5CH3OH
cis-[Co(dtc)2{P(OMe)Ph2}2]PF6
Co(1)–P(1)
Co(1)–S(1)
Co(1)–S(2)
2.269(2)
2.279(2)
2.280(2)
173.71(7)
76.23(7)
99.96(7)
25.8(2)
Co(1)–P(2)
Co(1)–S(3)
Co(1)–S(4)
2.309(2)
2.281(2)
2.275(2)
Co–P(1)
Co–S(1)
Co–S(2)
2.244(1)
2.270(1)
2.288(1)
Co–P(2)
Co–S(3)
Co–S(4)
2.246(1)
2.271(1)
2.297(1)
P(1)–Co(1)–P(2)
S(1)–Co(1)–S(2)
S(1)–Co(1)–S(4)
S(3)–Co(1)–S(4)
S(2)–Co(1)–S(3)
76.48(7)
107.42(7)
26.2(3)
P(1)–Co(1)–P(2)
S(1)–Co(1)–S(2)
92.68(4)
76.13(4)
S(3)–Co(1)–S(4)
75.73(4)
c
c
ꢀ1 c
ꢀ3 c
ꢀ2
89.2(2)
ꢀ4
18.6(2)
Co(51)–P(51)
Co(51)–S(51)
Co(51)–S(52)
S(51)–Co(51)–S(52)
2.285(2)
2.285(2)
2.280(2)
76.08(6)
21.1(2)
Co(52)–P(52)
Co(52)–S(53)
Co(52)–S(54)
S(53)–Co(52)–S(54)
2.301(2)
2.270(2)
2.271(2)
76.90(6)
21.0(2)
d
e
ꢀ51
ꢀ53
d
e
ꢀ52
89.0(2)
ꢀ54
31.6(3)
a Dihedral angle between the CoS4(1) plane (defined by Co(1), S(1) and S(2)) and the phenyl ring consisted of C(11)–C(16). b Dihedral angle between the CoS4(2) plane (defined by Co(2), S(3) and S(4)) and the phenyl ring
consisted of C(31)–C(36). c Dihedral angle between the Co(1)S4 plane (defined by Co(1), S(1), S(2), S(3) and S(4)) and the four phenyl rings of complex 1: phenyl(1) C(11)–C(16); phenyl(2) C(17)–C(22); phenyl(3) C(31)–
C(36); phenyl(4) C(37)–C(42). d Dihedral angle between the Co(51)S4 plane (defined by Co(51), S(51) and S(52)) and the two crystallographically independent phenyl rings of complex 51: phenyl(51) C(61)–C(66);
phenyl(52) C(67)–C(72). e Dihedral angle between the Co(52)S4 plane (defined by Co(52), S(53) and S(54)) and the two crystallographically independent phenyl rings of complex 52: phenyl(53) C(81)–C(86); phenyl(54)
C(87)–C(92).

View Article Online
Table 4 Structural and electronic parameters for P-ligands and average bond lengths (Å) and angles (Њ) in trans- and cis-[Co(dtc)2-
(P-ligand)2]ϩcomplexes
trans-[Co(dtc)2(P-ligand)2]ϩ
cis-[Co(dtc)2(P-ligand)2]ϩ
χb
χd
Co–P
Co–S
S–Co–S
Co–P
Co–S(transS)
Co–S(transP)
P–Co–P
S–Co–S
a
b
P-ligand
θT
P(OMe)3
P(OMe)2Ph
P(OMe)Ph2
PPh3
PMePh2
PMe2Ph
PMe3
107
115
132
145
136
122
118
128
24.10
19.45
16.30
13.25
12.10
10.60
8.55
16.70
15.73
14.82
13.25
12.10
10.60
8.55
2.241
2.266
2.291
2.316
2.303
2.284
2.287
2.276
2.273
2.265
2.277
2.264
2.278
2.270
2.270
2.269
76.25
76.91
76.42
76.2
76.18
76.75
76.86
77.03
2.200
2.218
2.245
–
2.270
2.263
2.271
2.295
2.289
2.292
97.16
94.4
92.68
75.78
76.44
75.93
–
2.272
2.200
2.230
2.268
2.255
2.263
2.291
2.310
2.283
95.14
96.8
90.51
76.31
75.07
76.52
PHPh2
17.35
17.35
a Tolman’s cone angle from ref. 15. b From ref. 10.
16.70–13.25), the linear relation in Fig. 2 indicates that the
Co–P bond lengths depend primarily on the steric requirements
(Tolman’s cone angle, θT) of the P-ligands. The fact that a very
weak σ-donor, PHPh2, is also located on the same straight line
in Fig. 2 confirms the validity of the above assertion. On
the other hand, when the P-ligand is either PMe3 or PMe2Ph,
the Co–P bond length is longer than that predicted from the
straight line in Fig. 2. We presume that this deviation is the net
elongation caused by the electronic mutual trans influence.
Therefore, we conclude that the electronic trans influence of the
P-ligand, that seems to be parallel to the σ-donicity of the
P-ligand, is the secondary factor that determines the Co–P
bond lengths in the series of trans-[Co(dtc)2(P-ligand)2]ϩ com-
plexes: the primary factor to govern the bond lengths is the
cone angle of each ligand.
Fig. 2 Plots of the Co–P bond lengths (Å) against the Tolman’s cone
angle of the P-ligands in trans- and cis-[Co(dtc)2(P-ligand)2]ϩ (ᮀ: trans
complex, ᭹: cis complex). P-ligand = 1: P(OMe)3, 2: P(OMe)2Ph, 3:
P(OMe)Ph2, 4: PPh3, 5: PMePh2, 6: PMe2Ph, 7: PMe3, 8: PHPh2.
The structures of cis-[Co(dtc)2{P(OMe)3 Ϫ nPhn}2]ϩ (n = 0, 1
or 2) are shown in Fig. 3, and the structural parameters are
listed in Table 3. The Co–S bond lengths trans to the P-ligand
are longer by 0.02–0.03 Å than those for the mutually trans
Co–S (Table 4). Such a tendency was also common in the other
cis-[Co(dtc)2(P-ligand)2]ϩ complexes.5,8,9 The small elongation
of the Co–S bond may be attributed to the larger trans influence
of the P-ligand compared with that of dtcϪ. Moreover,
the elongations of the Co–S bonds trans to the P-ligands in
the bis(dtc) complexes are smaller than those observed for
the Co–O bonds at the trans position of the P-ligands in cis-
[Co(acac)2(P-ligand)2]ϩ complexes,25,26 indicating the larger
electron sponge effect of dtcϪ compared with acacϪ. The Co–P
bonds in cis-[Co(dtc)2{P(OMe)3 Ϫ nPhn}2]ϩ were shorter by
0.04–0.05 Å than those in the corresponding trans-isomers,
which is due to a combination of the reduced trans influence,
the reduced steric congestion around the Co–P bond, and the
electron-buffer effect by the dtcϪ ligand in the cis-complexes.
Fig. 1 Perspective views of one of the crystallographically inde-
pendent complex cations in (a) trans-[Co(dtc)2{P(OMe)3}2]BF4 (40%
probability level), (b) trans-[Co(dtc)2{P(OMe)2Ph}2]BF4ؒ(CH3)2CO
(50% probability level) and (c) trans-[Co(dtc)2{P(OMe)Ph2}2]BF4ؒ
0.5CH3OH (40% probability level). Hydrogen atoms are omitted for
clarity.
D a l t o n T r a n s . , 2 0 0 3 , 2 2 8 0 – 2 2 9 2
2285

View Article Online
Table 5 Results of the Gaussian curve fitting for the absorption band maxima, 10Ϫ3σ/cmϪ1 (ε/dm3 molϪ3 cmϪ1),a of [Co(dtc)2(P-ligand)2]ϩ complexes
Complex
d–d band
CT band
[Co(dtc)3]
15.50 (435), 20.62 (589)
19.41 (704), 23.73 (1552)
19.12 (714), 23.45 (1947)
18.33 (691), 23.20 (2382)
17.91 (743), 22.96 (1716)
18.33 (741), 23.31 (1340)
18.43 (654), 23.43 (2611)
20.23 (415)
18.20 (335)
17.66 (448)
15.74 (342)
16.58 (355), 19.97 (581)
17.17 (225)
25.87 (8437), 27.73 (2794), 31.21 (19490)
29.03 (9840), 34.08 (26980), 39.32 (19530)
28.64 (10430), 32.25 (10810), 35.88 (28540)
28.43 (10820), 31.33 (11000), 35.27 (27990)
28.89 (15200), 32.04 (21800)
cis-[Co(dtc)2{P(OMe)3}2]BF4
cis-[Co(dtc)2{P(OMe)2Ph}2]PF6
cis-[Co(dtc)2{P(OMe)Ph2}2]PF6
cis-[Co(dtc)2(PMe2Ph)2]PF6
cis-[Co(dtc)2(PMe3)2]BF4
cis-[Co(dtc)2(PHPh2)2]BF4
trans-[Co(dtc)2{P(OMe)3}2]BF4
trans-[Co(dtc)2{P(OMe)2Ph}2]BF4
trans-[Co(dtc)2{P(OMe)Ph2}2]BF4
trans-[Co(dtc)2(PPh3)2]BF4
trans-[Co(dtc)2(PMePh2)2]BF4
trans-[Co(dtc)2(PMe2Ph)2]BF4
trans-[Co(dtc)2(PMe3)2]BF4
trans-[Co(dtc)2(PHPh2)2]BF4
29.25 (10560), 32.62 (24990)
27.89 (5354), 31.17 (7791), 34.59 (26770)
27.76 (9477), 29.59 (13120), 33.15 (10800)
26.42 (9743), 29.15 (16100), 31.91 (12160)
25.04 (8615), 28.47 (10220), 30.94 (17240)
23.08 (9488), 30.18 (22290)
24.77 (10950), 29.63 (21750)
25.57 (12090), 29.70 (25120)
17.90 (258)
17.30 (369), 21.12 (617)
27.38 (13540), 30.11 (21970)
24.92 (10910), 29.63 (21130)
a In dichloromethane at room temperature.
1
As shown in Fig. 2, the relation between the Co–P bond
length and θT for the cis complexes is somewhat dispersed but
almost linear. However, there is no apparent relationship
between the P–Co–P bond angle and θT for the cis-complexes as
seen in Table 4: it seems that the electron sponge effect of dtcϪ
causes alteration of the P–Co–P angle as described in the latter
section, although it is also possible to consider that a suitable
orientation of the substituents on the P atom in the crystal
alters the configuration around the cobalt center.27
It was expected that the A1g
a1Eg transition energy, that
is parallel to the ligand-field strength, is linearly related to
Tolman’s χ parameter, as this parameter includes the effect of
π-acceptability. As seen from Fig. 5a, however, a V-shaped rel-
ation was found in such a plot for the series of trans-complexes.
A similar tendency was also found for the series of cis-com-
plexes. Such a V-shaped relation indicates that Tolman’s χ par-
ameter fails to predict the exact nature of the σ and π-back
bonding interactions between cobalt() and these P-ligands.
However, it is more rational to conclude that the steric factor
represented by the cone angle overrules the contribution of the
electronic factors represented by Tolman’s χ parameter (Fig.
5b). A similar conclusion was also derived for the trans influ-
ence (Fig. 2): the σ-donicity (= electronic factor) of the P-ligand
is the secondary factor that determines the Co–P bond lengths
in the trans-[Co(dtc)2(P-ligand)2]ϩ complexes, while the primary
factor to govern the bond lengths is the cone angle of each
P-ligand.
Absorption spectra
The UV-Vis absorption spectra of cis- and trans-[Co(dtc)2-
{P(OMe)3 Ϫ nPhn}2]ϩ complexes in dichloromethane are shown
in Fig. 4, which are similar to those for the analogous cis- and
trans-[Co(dtc)2(PMe3 Ϫ nPhn)2]ϩ complexes.8 For the cis-com-
plexes (Fig. 4a), there exists a weak (ε < 750 dm3 molϪ1 cmϪ1
)
shoulder around 19000 cmϪ1 and an absorption band (ε < 2500
dm3 molϪ1 cmϪ1) around 23500 cmϪ1, that were assigned to the
first (1A1g
1T1g: Oh) and the second (1A1g
1T2g) d–d transi-
Isomerization reactions
tion bands, respectively.8 The distinct splitting of the d–d band
due to the low molecular symmetry, C2 (or holohedrized D4h)
was not observed, as was the case for the other analogous
P-ligand complexes.5,8,9 The peak positions of these transition
bands estimated by the Gaussian curve fitting analyses are
listed in Table 5, from which the ligand-field strength, ∆, and
the Racah’s inter-electronic repulsion parameter, B, were esti-
mated by assuming C = 4B: ∆ = 20490 cmϪ1 and B = 270 cmϪ1
for cis-[Co(dtc)2{P(OMe)3}2]ϩ, ∆ = 20200 cmϪ1 and B = 271
cmϪ1 for cis-[Co(dtc)2{P(OMe)2Ph}2]ϩ, and ∆ = 19550 cmϪ1 and
B = 304 cmϪ1 for cis-[Co(dtc)2{P(OMe)Ph2}2]ϩ. It should be
noted that the ligand-field strength gets larger in the order
P(OMe)Ph2< P(OMe)2Ph < P(OMe)3 of these P-ligands, while
the σ-donicity of these P-ligands slightly decreases in this order,
indicative of the importance of π-back bonding between Co()
and P-ligands: the π-acidity of the P-ligand increases in this
order. The ligand-field strengths for the P(OMe)3 and
P(OMe)2Ph complexes are larger but the B parameters for these
complexes are smaller than those for the PMe3 (∆ = 19570 and
NMR measurements. The observed 1H NMR spectral
changes are shown in Fig. 6 for the trans to cis isomerization
reaction of [Co(dtc)2{P(OMe)Ph2}2]ϩ (0.01 mol kgϪ1
) in
acetonitrile at 303 K. It is shown that the proton signals corre-
sponding to the trans-species decreased with time while the
signals corresponding to the cis-species increased. Although the
low sensitivity of the NMR method did not allow us to observe
decomposition products during the experiment, it was shown
that a small amount of gradual decomposition did take place
causing a slight deviation in the spectral intensity with time
from the first-order kinetics, when no free P(OMe)Ph2 was
added to the reaction mixture. A slow decomposition of
the complex was also indicated for the isomerization reaction
of trans-[Co(dtc)2{P(OMe)2Ph}2]ϩ in the absence of free
P(OMe)2Ph.
1
On the other hand, the H NMR spectral changes for the
isomerization reaction of trans-[Co(dtc)2{P(OMe)3}2]ϩ were
somewhat different from those for the other two complexes. In
addition to the deviation from the first-order kinetics, the NMR
signals that indicate decomposition and/or intramolecular
isomerization of the free P(OMe)3 ligand were observed when
an excess amount of free P(OMe)3 was added to the solution:
the solution without free P(OMe)3 exhibited complicated
NMR signals with time, indicating gradual decomposition of
the compound in a complex manner. Fig. 7 shows the 1H
NMR spectrum of a sample solution containing [Co(dtc)2-
{P(OMe)3}2]ϩ (0.01 mol kgϪ1) with an excess (ca. 4 times more
than the complex) amount of free P(OMe)3. This spectrum was
observed after completion of the isomerization reaction at
303 K. The N–CH3 and P–OCH3 signals in the figure are those
B = 311 cmϪ1) and PMe2Ph (∆ = 19170 and B = 316 cmϪ1
)
complexes.8 The cis-complexes also exhibited two intense
absorption bands in the UV region, which were assigned to
ligand-to-metal charger transfer (LMCT) bands.
The absorption spectra of trans-[Co(dtc)2{P(OMe)3 Ϫ n
-
Phn}2]ϩ (Fig. 4b), exhibited only a weak (ε < 500 dm3 molϪ1
cmϪ1) shoulder in the 17000–20000 cmϪ1 region, that was
assigned to the splitting component (1A1g
a1Eg: D4h) of the
first d–d transition. The trans-isomers also exhibited some
intense bands at higher energy (> 23000–27000 cmϪ1), that were
assigned to the LMCT band in agreement with other related
trans-complexes.8,23
D a l t o n T r a n s . , 2 0 0 3 , 2 2 8 0 – 2 2 9 2
2286

View Article Online
Fig. 4 UV-Vis absorption spectra of (a) cis- and (b) trans-[Co(dtc)2-
{P(OMe)3 Ϫ nPhn}2]ϩ (n = 0 (ؒ ؒ ؒ); 1 (—); 2 (---) and 3 (- -)) in
ؒ
dichloromethane at room temperature.
Fig. 3 Perspective views (40% probability level for each crystal) of the
cationic part in (a) cis-[Co(dtc)2{P(OMe)3}2]BF4, (b) cis-[Co(dtc)2-
{P(OMe)2Ph}2]PF6 and (c) cis-[Co(dtc)2{P(OMe)Ph2}2]PF6. Hydrogen
atoms are omitted for clarity.
for the cis-isomers. Beside the doublet signals for free P(OMe)3,
small signals located at 3.71, 5.80, and 7.55 ppm and a series of
signals at 3.28 and around 3.68 ppm were observed. The latter
two signals, exhibiting a 12 : 9 ratio in terms of the proton
numbers, were assigned to the N–CH3 and P–OCH3 protons in
the 5-coordinate [Co(dtc)2{P(OMe)3}]ϩ species (the validity of
this assignment is shown in the latter section), while the other
signals correspond to the decomposition products of P(OMe)3
in the bulk: it is known that free P(OMe)3 tends to either iso-
Fig. 5 Plots of the d–d band positions (σ/cmϪ1) against (a) Tolman’s
electronic parameter (χ) and (b) Tolman’s cone angle (θT/Њ) for the P-
ligands in trans- and cis-[Co(dtc)2(P-ligand)2]ϩ. ᮀ : d–d (a1Eg) band of
the trans-isomers; ᭹ : d–d (1T1g: Oh) band of the cis-isomers. P-ligand =
1: P(OMe)3, 2: P(OMe)2Ph, 3: P(OMe)Ph2, 4: PPh3, 5: PMePh2, 6:
PMe2Ph, 7: PMe3, 8: PHPh2.
merize to P(᎐O)(Me)(OMe) or decompose to produce a series
᎐
2
of HPMex(O)(OMe)y 28 and the latter species exhibits 1H NMR
signals in this region.29 It seems that the coordinated P(OMe)3 is
D a l t o n T r a n s . , 2 0 0 3 , 2 2 8 0 – 2 2 9 2
2287

View Article Online
Fig. 6 1H NMR spectral change for trans to cis thermal isomerization reaction of [Co(dtc)2{P(OMe)Ph2}2]ϩ at 303 K in acetonitrile-d3. [trans-
Co(dtc)2{P(OMe)Ph2}2ϩ]0 = 0.01 mol kgϪ1. CF3SO3Na (0.1 mol kgϪ1) was used as the supporting electrolyte.
Fig. 7 1H NMR spectrum after completion of the thermal trans to cis
isomerization reaction of [Co(dtc)2{P(OMe)3}2]ϩ with excess free
P(OMe)3 at 303 K in acetonitrile-d3. [trans-Co(dtc)2{P(OMe)3}2ϩ
] =
0.01 mol kgϪ1. The concentration of free P(OMe)3 in the bulk is ca. 00.04
mol kgϪ1. CF3SO3Na (0.1 mol kgϪ1) was used as the supporting
electrolyte.
not subject to the intramolecular isomerization to P(᎐O)(Me)-
᎐
(OMe)2 and/or decomposition, while the sample solution con-
taining only P(OMe)3 exhibited similar NMR signals to those
in Fig. 7 after several hours: a similar difference in the reactivity
of the coordinated and free phosphine ligand was observed for
the isomerization reaction of the PHPh2 complex in a previous
study9a where impure water abstracts the P–H proton from free
PHPh2 but not from coordinated PHPh2. It seems, therefore,
that an acid–base interaction between free P(OMe)3 and impure
water (ca. 1 mmol kgϪ1) provides a variety of decomposition
processes for free P(OMe)3. Otherwise, the excess amount of
added free P(OMe)3 (> ca. 5 mmol) significantly suppresses
unfavorable side reactions initiated by the dissociation of
coordinated P(OMe)3.
Fig. 8 Absorption spectral change for the thermal trans to cis
isomerization reaction of [Co(dtc)2{P(OMe)Ph2}2]ϩ at 298
in
Spectrophotometric measurements
K
acetonitrile (a) without and (b) with excess free P(OMe)Ph2 in the bulk.
The inset in each figure shows the observed absorbance change with
time at 470 nm and the corresponding best-fit curve analyzed by the
single-exponential function.
As shown in Fig. 8a, although trans-[Co(dtc)2{P(OMe)Ph2}2]ϩ
and trans-[Co(dtc)2{P(OMe)2Ph}2]ϩ undergo thermal isomeriz-
ation to the corresponding cis-isomers, it was not possible to
describe the change in the absorption with time by simple first-
order kinetics. However, by addition of excess free ligand
(P(OMe)Ph2 or P(OMe)2Ph), the reaction traces obeyed
simple first-order kinetics (Fig. 8b) for up to three half-lives.
The observed first-order rate constants were independent of
the concentration of free ligand (Table 6). On the other hand,
the dependence of the conditional first-order rate constants on
D a l t o n T r a n s . , 2 0 0 3 , 2 2 8 0 – 2 2 9 2
2288

View Article Online
the concentration of excess free P(OMe)3 exhibited saturation
kinetics as shown in Fig. 9, while the absorption change with
time for each given [P(OMe)3]free was excellently first-order for
up to 3 half-lives. It seems obvious that the degree of dissoci-
ation of coordinated P(OMe)3 is large for trans-[Co(dtc)2-
{P(OMe)3}2]ϩ in acetonitrile while such a tendency is small for
the trans-P(OMe)2Ph and -P(OMe)Ph2 complexes.
decomposition of the 5-coordinate intermediate through
further dissociation of the P-ligands that certainly explain the
bi-exponential nature of the kinetic trace when no free ligand is
added to the solution.
(1)
(2)
(3)
By assuming
a steady-state for the trans-[Co(dtc)2-
(P-ligand)]ϩ (A) species, the following rate law is derived.
(4)
The kϪ1 process should be large when large amounts of free
P-ligand exist in the solution. Therefore, eqn. (4) reduces to
eqn. (5) under the experimental conditions.
Fig. 9 The dependence of the conditional rate constants for the
thermal trans to cis isomerization reaction of [Co(dtc)2{P(OMe)3}2]ϩ
on the concentration of excess free P(OMe)3 at 313 and 303 K. [trans-
(5)
Co(dtc)2{P(OMe)3}2ϩ
]
= 1.0 mmol kgϪ1. Tetra-n-butylammonium
0
tetrafluoroborate (0.1 mol kgϪ1) was used as the supporting electrolyte.
It is obvious, from Table 6, that the apparent rate constant
for the trans to cis isomerization reaction is independent of
the concentration of added free P-ligand, when [P(OMe)2Ph]free
or [P(OMe)Ph2]free > 10 mmol kgϪ1. Lack of the dependence
of kobs on [P(OMe)2Ph]free or [P(OMe)Ph2]free strongly indicates
that the dissociation mechanism expressed by reactions (1)
and (2) is not important when sufficient amounts of free
P(OMe)2Ph or P(OMe)Ph2 exist in the bulk. This result is also
consistent with the NMR observation: there is no appreciable
dissociation of coordinated P(OMe)2Ph or P(OMe)Ph2.
The situation is somewhat different in the case of the isomer-
ization reaction of trans-[Co(dtc)2{P(OMe)3}2]ϩ. The depend-
ence of the first-order rate constant on the concentration of free
P(OMe)3 exhibited saturation behavior (Fig. 9). Such a result is
explained by assuming a rapid equilibrium for eqn. (1).11
Possible mechanisms for the trans to cis isomerization
reactions are either (1) the dissociative mechanism through a
5-coordinate intermediate and (2) the intramolecular twist
mechanism through the intermediate with trigonal-prismatic
structure.11 It was shown in a previous study,9a that the dissoci-
ative isomerization that takes place with the structural change
from the square-pyramidal intermediate (A) (see Scheme 1) to
the cis-product via the trigonal-bipyramidal transition state (B)
cannot compete with the intramolecular twist mechanism
because the structural change from (A) to (B) is accompanied
by a very large activation enthalpy when the ligand-field is
strong.11–13 Therefore, although the overall reaction mechanism
including dissociation of a coordinated P-ligand is expressed
by the following equations, the k2 process may only lead to
Scheme 1 Proposed mechanism for the thermal trans to cis isomerization reactions of [Co(dtc)2{P(OMe)3 Ϫ nPhn}2]ϩ.
D a l t o n T r a n s . , 2 0 0 3 , 2 2 8 0 – 2 2 9 2
2289

View Article Online
(1Ј)
As the ligand-field of P(OMe)3 is larger than those of
P(OMe)2Ph or P(OMe)Ph2, the k2 process that leads to isomer-
ization is not important for this reaction either.12,13 Further-
more, the saturation behavior up to [P(OMe)3]free = 40 mmol
kgϪ1 indicates that the dissociation of coordinated P(OMe)3 is
significant: the NMR result shown in Fig. 7 also indicates the
significant degree of dissociation of coordinated P(OMe)3.
However, it should be noted that the reaction was first-order
even under the conditions where dissociation of P(OMe)3 is not
completely suppressed ([P(OMe)3]free < 40 mmol kgϪ1). Such an
observation leads to the conclusion that the decomposition of
5-coordinate intermediate (A) takes place through further dis-
sociation of coordinated P(OMe)3, which is completely sup-
pressed by the addition of excess free P(OMe)3. Therefore, we
suggest here that a dead-end mechanism involving (i) the form-
ation of the non-reactive 5-coordinate [Co(dtc)2{P(OMe)3}]ϩ
and (ii) the intramolecular twist process by reaction (3): the
observed pseudo-first-order rate constant is described by the
following equation.
(6)
The dissociation constant, K, was estimated as 2.5 × 10Ϫ3 mol
kgϪ1 at 313 K from the plots in Fig. 9. This value of K estimates
that 76% dissociation of coordinated P(OMe)3 is expected in a
1 mmol kgϪ1 solution of trans-[Co(dtc)2{P(OMe)3}2]ϩ, and that
ca. 6% dissociation takes place even under the condition of
[P(OMe)3]free = 40 mmol kgϪ1
.
The rate constants and activation parameters for the isomer-
ization reactions of trans-[Co(dtc)2{P(OMe)3 Ϫ nPhn}2]ϩ are
summarized in Table 6. The activation parameters of these
reactions are consistent with those reported for the ordinary
intramolecular process (Ray–Dutt type twist mechanism30,31),
including the previously reported intramolecular isomerization
reaction of trans-[Co(dtc)2(PHPh2)2]ϩ:9a a large activation
enthalpy with near-zero activation entropy.32,33 The AOM
calculation for the twist mechanism also indicates that the reac-
tion proceeds via a spin state change: the process through the
spin-quintet state of the trigonal prismatic transition state is
energetically more favorable compared with the process in
which the spin-singlet state is maintained during the activation
process.9a,11,13
The difference between eqns. (5) and (6) depends solely on
the following assumptions: when k1 Ӷ kϪ1[P(OMe)3]free, eqn. (5)
is valid as the steady state approximation is effective for the
5-coordinate species; while the rapid equilibrium in eqn. (1Ј)
ensures the relation, k3 Ӷ k1ϩ kϪ1[P(OMe)3]free, where K is
defined by k1/kϪ1.
11,34 As K is 2.5 × 10Ϫ3 kg molϪ1 at 313 K, the
inequality is expressed by k3 Ӷ kϪ1(2.5 × 10Ϫ3ϩ [P(OMe)3]free).
This inequality is fulfilled when k3 is small, which is the case
observed in this study: k3 = 4.22 × 10Ϫ4 sϪ1 at 313 K.35 Moreover,
we can safely infer that the dissociation rate constant, k1, for
P(OMe)3 is the largest of the three reactions examined in this
study (K is the largest for trans-[Co(dtc)2{P(OMe)3}2]ϩ while kϪ1
is assumed to be the same for all reactions): the kinetic trans
effect is in the order of P(OMe)3 ӷ P(OMe)Ph2, P(OMe)2Ph,
and PHPh2.
The dissociation rate constants for the P(OMe)Ph2 and
P(OMe)2Ph complexes, k1, were roughly estimated as ≈10Ϫ4 and
≈10Ϫ6 sϪ1 at 298 K, respectively, by the double-exponential
analyses of the kinetic trace when no free P-ligand was added to
each sample solution. The previously obtained rate constant for
the dissociation of coordinated PHPh2 from trans-[Co(dtc)2-
(PHPh2)2]ϩ, ≈10Ϫ5 sϪ1 at 298 K,9a is exactly in between those for
D a l t o n T r a n s . , 2 0 0 3 , 2 2 8 0 – 2 2 9 2
2290

View Article Online
the P(OMe)Ph2 and P(OMe)2Ph complexes. This order in the
dissociation rate constant is consistent with the expected order
order of ∆ for these complexes, the electron sponge effect of
the dtcϪ ligand seems to equally affect the energy levels of the
of π-acidity for these P-ligands: P(OMe)2Ph > PHPh2
>
ground state and the trigonal prismatic transition state.
P(OMe)Ph2, by considering the small pKa values of the conju-
gate acids and the decreasing value of the χ parameters for
these P-ligands. Such a result indicates that the π-acidity of the
P-ligand plays a significant role in the kinetic trans effect
observed for trans-[Co(dtc)2(P-ligand)2]ϩ in solution.
Conclusion
In this study, a series of trans- and cis-[Co(dtc)2{P(OMe)3 Ϫ n
-
Phn}2]ϩ complexes were synthesized and the structures of
related trans- and cis-[Co(dtc)2(P-ligand)2]ϩ complexes were
examined. It was found that the Co–P bond lengths as well as
the ligand-field parameters were primarily governed by the
steric bulk of the P-ligand, the cone angle. It was also shown
that the complexes of P-ligands with strong σ-donicity (PMe3)
and strong π-acidity (P(OMe)3) are equally stabilized. Such a
tendency was clearly seen in Figs. 5a and b with the inflection
point at the PPh3 ligand, for which the cone angle is the largest
of all P-ligands while both σ-donicity and π-acidity are not
eminent. The very labile nature of the PPh3 complex, which
was exploited for the efficient syntheses of the other trans-bis-
(P-ligand) complexes, may be explained by the relatively weak
Co–PPh3 bond caused by the very large cone angle of this
ligand. Otherwise, the almost identical stability of the Co–P
bonds for the series of trans- and cis-[Co(dtc)2(P-ligand)2]ϩ
complexes is explained by the electron sponge effect of the
spectator ligand, dtcϪ.
The static trans influence was examined on the basis of the
Co–P bond length and the ligand-field parameter, ∆, as men-
tioned in the previous section. It was shown in Figs. 2 and
5b that the trans influence in the series of trans-[Co(dtc)2-
(P-ligand)2]ϩ is essentially explained by Tolman’s cone angle:
P(OMe)Ph2 > P(OMe)2Ph > P(OMe)3. The kinetic results, how-
ever, indicate that the trans effect is the largest for P(OMe)3.
Therefore, it seems that the Co–P bond length is not directly
related to the kinetic trans effect: it seems that the kinetic trans
effect for this series of P-ligand complexes is governed by
the stability of the 5-coordinate species (A). On the basis of the
AOM calculations, it has already been reported that the energy
difference between the d6 cobalt() ions in the pseudo-
octahedral coordination geometry and in the pseudo-square
pyramidal coordination geometry is essentially described by the
π-interaction energy, ∆E ≈ Ϫ4eπ:36 the larger the π-acidity of
the P-ligand, the less stabilization of the 5-coordinate species is
indicated. Such a tendency is opposite to the order of the
observed kinetic trans effect. Therefore, a comprehensive
understanding of the kinetic trans effect observed for the series
trans-[Co(dtc)2(P-ligand)2]ϩ requires consideration of not only
the effect of the P-ligands but also the effect of the spectator
ligands.
The kinetic studies of the thermal trans to cis isomerization
reactions of these complexes revealed that the reactions proceed
through the intramolecular twist mechanism. A significant
degree of dissociation of coordinated P(OMe)3 was observed
for trans-[Co(dtc)2{P(OMe)3}2]ϩ. This strong (kinetic) trans
effect in trans-[Co(dtc)2{P(OMe)3}2]ϩ was attributed to the
stabilization of the 5-coordinate [Co(dtc)2{P(OMe)3}]ϩ caused
by the combination of strong π-acidity inherent in P(OMe)3
and the electron sponge effect of the spectator ligand, dtcϪ.
The chemical shift of the P–OCH3 proton (Table 1) does not
seem to directly reflect the electron density on the P-ligand
1
in the 6-coordinate species. On the other hand, the H NMR
signal of N–CH3 moiety as a function of the P-ligands were
observed at 2.39, 2.72, 2.99, and 3.31 ppm for PPh3,
P(OMe)Ph2, P(OMe)2Ph, and P(OMe)3 complexes, respectively
(Table 1), which may reflect the gradual decrease of electron
density on dtcϪ in this order. Therefore, it seems that the specta-
tor ligand, dtcϪ, supplies more electron density to the cobalt()
center as the π-acidity of the P-ligand increases. Such a capabil-
ity of the dtcϪ ligand to act as an electron sponge certainly
stabilizes cobalt() complexes with π-acids by altering the elec-
tron density on Co. This seems to be the main reason why the
Co–P bond lengths as well as the ligand-field strengths, ∆,
do not linearly depend on the electronic parameter, χ, of the
P-ligand. However, such an effect of dtcϪ as an electron sponge
may not be strong enough to compensate for the decrease in
the electron density on cobalt() in the trans-[Co(dtc)2-
{P(OMe)3}2]ϩ complex, since the π-acidity of P(OMe)3 is much
larger than those for P(OMe)2Ph and P(OMe)Ph2: χ = 24.10,
19.45, and 16.30 for P(OMe)3, P(OMe)2Ph, and P(OMe)Ph2,
respectively. As a result, the 5-coordinate species, [Co(dtc)2-
{P(OMe)3}]ϩ, is largely stabilized compared with the other
5-coordinate species in which the P-ligand is P(OMe)2Ph or
P(OMe)Ph2. In such a case, the P–OCH3 1H NMR signal of
the 5-coordinate species should be observed at a higher-field
than that for the 6-coodinate species since the loss of one
of the P(OMe)3 ligands stabilizes another Co–P bond through
References and notes
1 T. Ohishi, K. Kashiwabara and J. Fujita, Bull. Chem. Soc. Jpn.,
1987, 60, 583.
2 M. Kita, A. Okuyama, K. Kashiwabara and J. Fujita, Bull. Chem.
Soc. Jpn., 1990, 63, 1994.
3 M. Adachi, M. Kita, K. Kashiwabara, J. Fujita, N. Iitaka,
S. Kurachi and S. Ohba, Bull. Chem. Soc. Jpn., 1992, 65, 2037.
4 M. Kita, M. Okuno, K. Kashiwabara and J. Fujita, Bull. Chem. Soc.
Jpn., 1992, 65, 3042.
5 H. Matsui, M. Kita, K. Kashiwabara and J. Fujita, Bull. Chem. Soc.
Jpn., 1993, 66, 1140.
6 K. Kashiwabara, M. Watanabe, M. Kita and T. Suzuki, Bull. Chem.
Soc. Jpn., 1996, 69, 1947.
7 K. Kashiwabara, N. Taguchi, H. D. Takagi, K. Nakajima and
T. Suzuki, Polyhedron, 1998, 17, 1817.
8 T. Suzuki, S. Kashiwamura and K. Kashiwabara, Bull. Chem. Soc.
Jpn., 2001, 74, 2349.
9 (a) S. Iwatsuki, T. Suzuki, A. Hasegawa, S. Funahashi,
K. Kashiwabara and H. D. Takagi, J. Chem. Soc., Dalton Trans.,
2002, 3593; (b) T. Suzuki, S. Iwatsuki, H. D. Takagi and
K. Kashiwabara, Chem. Lett., 2001, 1068.
10 (a) H.-Y. Liu, K. Eriks, A. Prock and W. P. Giering, Organometallics,
1990, 9, 1758; (b) Md. M. Rahman, H.-Y. Liu, K. Eriks, A. Prock
and W. P. Giering, Organometallics, 1989, 8, 1.
11 R. B. Jordan, Reaction Mechanisms of Inorganic and Organometallic
Systems, Oxford University Press, Oxford, 2nd edn., 1998.
12 L. G. Vanquickenborne and K. Pierloot, Inorg. Chem., 1981, 20,
3673.
1
the enhancement of the π-interaction. In fact, the H NMR
signal corresponding to the coordinated P(OMe)3 appeared at
3.68 ppm in Fig. 7. The lower-field shift of the N–CH3 protons
on dtcϪ in Fig. 7 indicates stabilization of the cobalt() center
in the 5-coordinate [Co(dtc)2{P(OMe)3}]ϩ by the electron
sponge effect of this spectator ligand.
The AOM calculation also implies that the activation
enthalpy for the twist process increases with increasing ligand-
field parameter, ∆, when the electron pairing energy is the same
for all complexes.9a,37 As the activation enthalpies for the twist
processes of P(OMe)3 Ϫ nPhn complexes roughly followed the
13 L. G. Vanquickenborne and K. Pierloot, Inorg. Chem., 1984, 23,
1471.
14 (a) N. N. Greenwood and A. Earnshaw, Chemistry of the Elements,
Butterworth-Heinemann, Oxford, 2nd edn., 1997; (b) J. E. Huheey,
Inorganic Chemistry, Harper & Row, New York, 2nd edn., 1978.
15 C. A. Tolman, Chem. Rev., 1977, 77, 313.
16 P. Coppens, L. Leiserowitz and D. Rabinovich, Acta Crystallogr.,
Sect. A, 1965, 18, 1035.
17 A. C. T. North, D. C. Phillips and F. S. Mathews, Acta Crystallogr.,
Sect. A, 1968, 24, 351.
D a l t o n T r a n s . , 2 0 0 3 , 2 2 8 0 – 2 2 9 2
2291

View Article Online
18 Rigaku Co. Ltd., CrystalClear, Control program for Rigaku
Mercury CCD diffractometer, Akishima, Tokyo, Japan, 2001.
19 T. Higashi, abscor, empirical absorption correction based on Fourier
series approximation, Rigaku Corp., Tokyo, Japan, 1995.
20 G. M. Sheldrick, Acta Crystallogr., Sect. A, 1990, 46, 467.
21 G. M. Sheldrick, SHELXS97 and SHELXL97, University of
Göttingen, Germany, 1997.
22 Molecular Structure Corporation and Rigaku Co. Ltd., TeXsan,
Single crystal structure analysis software, ver. 1.9, The Woodlands,
TX, USA and Akishima, Tokyo, Japan, 1998.
28 O. Dahl and R. S. Edmundson, in The Chemistry of Organo-
phosphorus Compounds, ed. F. R. Hartley, John Wiley & Sons,
Chichester, 1996, vol. 4, ch. 1 and 2.
29 C. J. Pouchert and J. R. Campbell, The Aldrich Library of NMR
Spectra, Aldrich Chemical Co. Inc., Milwaukee, 1974, vol. 10.
30 P. C. Ray and N. K. Dutt, J. Indian Chem. Soc., 1941, 18,
289.
31 P. C. Ray and N. K. Dutt, J. Indian Chem. Soc., 1943, 20, 81.
32 F. F.-L. Ho and C. N. Reilley, Anal. Chem., 1969, 41, 1835.
33 R. G. Wilkins and M. J. G. Williams, J. Chem. Soc., 1957,
1763.
23 T. Suzuki, S. Kaizaki and K. Kashiwabara, Inorg. Chim. Acta, 2000,
298, 131.
34 G. I. Gellene, J. Chem. Educ., 1995, 72, 196.
24 (a) N. Bresciani-Pahor, M. Forcolin, L. G. Marzilli, L. Randaccio,
M. F. Summers and P. J. Toscano, Coord. Chem. Rev., 1985, 63,
1; (b) N. Bresciani-Pahor, L. Randaccio and P. J. Toscano,
J. Chem. Soc., Dalton Trans., 1982, 1559; (c) N. Bresciani-Pahor,
L. Randaccio, P. J. Toscano, A. C. Sandercock and L. G. Marzilli,
J. Chem. Soc., Dalton Trans., 1982, 129.
35 In this case, k1 is not necessarily very large. However, we observed
1H NMR signals corresponding to N–CH3 and P–OCH3 of the
5-coordinate species at ca. 3.28 and 3.68 ppm, 15 minutes after
sampling the trans-[Co(dtc)2{P(OMe)3}2]ϩ solution without free
P(OMe)3. Therefore, k1 is roughly estimated as 8 × 10Ϫ4 sϪ1
.
Accordingly, kϪ1 is estimated as ca. 3 kg molϪ1
inequality, k3 Ӷ k1 ϩ kϪ1[P(OMe)3]free as [P(OMe)3]free > 5 × 10Ϫ3 mol
kgϪ1
s
Ϫ1, that fulfills the
25 T. Suzuki, K. Kashiwabara, M. Kita, J. Fujita and S. Kaizaki,
Inorg. Chim. Acta, 1998, 281, 77.
.
26 T. Suzuki, T. Imamura, S. Kaizaki and K. Kashiwabara, Polyhedron,
2002, 21, 835.
27 (a) J. M. Smith, N. J. Coville, L. M. Cook and J. C. A. Boeyens,
Organometallics, 2000, 19, 5273; (b) T. E. Müller and D. M. P.
Mingos, Transition Met. Chem., 1995, 20, 533; (c) J. T. DeSanto,
J. A. Mosbo, B. N. Storhoff, P. L. Bock and R. E. Bloss,
Inorg. Chem., 1980, 19, 3086.
36 E. Larsen and G. N. La Mar, J. Chem. Educ., 1974, 51, 633.
37 The spin-pairing energy for a d6 configuration is somewhat reduced
to a value less than 20000 cmϪ1 depending on the decrease in the
inter-electronic repulsion. See: (a) F. A. Cotton, G. Wilkinson and
P. L. Gauss, Basic Inorganic Chemistry, Wiley, New York, 3rd edn.,
1995; (b) A. B. P. Lever, Inorganic Electronic Spectroscopy, Elsevier,
Amsterdam, 2nd edn., 1984.
D a l t o n T r a n s . , 2 0 0 3 , 2 2 8 0 – 2 2 9 2
2292