Chloride Transfer across the Liquid-Liquid Interface Facilitated by a Mo

Article abstract of DOI:10.1246/bcsj.74.2349

A series of new cobalt(III) phosphine complexes of trans-[Co(dtc)2(PMe3-nPhn)2]BF4 (dtc = N,N-dimethyldithiocarbamate; n = 0, 1, 2 or 3) and cis-[Co(dtc)2(PMe3 or PMe2Ph)2](BF4 or PF6) have been prepared and their crystal structures and spectroscopic properties have been investigated. It is found that the Co-S bond lengths vary with steric and electronic factors of the P-ligands; i.e. (1) the intramolecular ?-? stacking interaction between the dtc plane and the phenyl ring of the P-ligand, and (2) the electronic trans influence of the trans-positioned P-ligand. The strength of electronic trans influence decreases as PMe3 > PMe2Ph > P(OCH2)3CEt in accordance with the order of ?-donicity strengths. In the series of trans-isomers, the electronic trans influence is competitive with the steric requirement of the phosphine to elongate the mutually trans Co-P bonds. The steric trans influence via the equatorial dtc ligands for such an elongation of the Co-P bonds seems to be negligible, which is in sharp contrast to the situation for via pentane-2,4-dionate (acac) ligands in the analogous complexes, trans-[Co(acac)2(PMe3-nPhn)2]PF6. This is probably due to the compactness of dtc and the resulting open space at the Co atom. The fact that the Co-P bond lengths in the dtc complexes are shorter than those in the acac complexes is reflected in the larger stability toward hydrolysis of the dtc complexes. In the UV-vis absorption spectra, the degenerate splitting component (a¹E_g) of the first d-d transition band of trans-[Co(dtc)2(PMe3-n-Phn)2]⁺ is observed at almost the same position (within 300 cm^-1) as that of the corresponding acac complexes, while the transition energies of the P-to-Co LMCT of these two series of complexes are rather different (at least 2700 cm^-1) from each other. Furthermore, the first and the second d-d transition bands of cis-[Co(dtc)2(PMe3 or PMe2Ph)2]⁺ are observed at lower energy than those of cis-[Co(dtc)2{P(OMe)3}2]⁺ in spite of a weaker ?-donor of phosphite. The separation of the first and the second d-d transition bands of the P(OMe)3 complex is remarkably smaller than the separation of the bands of the PMe3 one, being indicative of a further reduction of the interelectronic repulsion in the P(OMe)3 complex.

Full text of DOI:10.1246/bcsj.74.2349

Bull. Chem.
Soc. Jpn.
74
12
© 2001 The Chemical Society of Japan
Bull. Chem. Soc. Jpn., 74, 2349–2359 (2001) 2349
The Chemical
Society of Ja-
pan
The Chemical
Society of Ja-
pan
AI
12
15
Preparation, Crystal Structures and Spectroscopic Properties of a Series
of Cobalt(Ⅲ) Phosphine Complexes: trans- and cis-[Co(dtc)2(PMe3−nPhn)2]+
(dtc = N,N-Dimethyldithiocarbamate; n = 0, 1, 2 or 3)
Cobalt(Ⅲ) Complexes with dtc and PMe3−nPhn
T. Suzuki et al.
2001
2349
2359
BCSJA8
0009-2673
01210
6
Takayoshi Suzuki,* Shinsaku Kashiwamura,† and Kazuo Kashiwabara*,†
Department of Chemistry, Graduate School of Science, Osaka University, Toyonaka 560-0043.
†Department of Chemistry, Graduate School of Science, Nagoya University, Chikusa-ku, Nagoya 464-8602.
(Received June 22, 2001)
22
2001
8
27
A series of new cobalt(Ⅲ) phosphine complexes of trans-[Co(dtc)2(PMe3−nPhn)2]BF4 (dtc = N,N-dimethyldithio-
carbamate; n = 0, 1, 2 or 3) and cis-[Co(dtc)2(PMe3 or PMe2Ph)2](BF4 or PF6) have been prepared and their crystal struc-
tures and spectroscopic properties have been investigated. It is found that the Co–S bond lengths vary with steric and
electronic factors of the P-ligands; i.e. (1) the intramolecular π–π stacking interaction between the dtc plane and the phen-
yl ring of the P-ligand, and (2) the electronic trans influence of the trans-positioned P-ligand. The strength of electronic
trans influence decreases as PMe3 > PMe2Ph > P(OCH2)3CEt in accordance with the order of σ-donicity strengths. In
the series of trans-isomers, the electronic trans influence is competitive with the steric requirement of the phosphine to
elongate the mutually trans Co–P bonds. The steric trans influence via the equatorial dtc ligands for such an elongation
of the Co–P bonds seems to be negligible, which is in sharp contrast to the situation for via pentane-2,4-dionate (acac)
ligands in the analogous complexes, trans-[Co(acac)2(PMe3−nPhn)2]PF6. This is probably due to the compactness of dtc
and the resulting open space at the Co atom. The fact that the Co–P bond lengths in the dtc complexes are shorter than
those in the acac complexes is reflected in the larger stability toward hydrolysis of the dtc complexes. In the UV-vis ab-
2001
2.4.2
2.5
sorption spectra, the degenerate splitting component (a1Eg) of the first d–d transition band of trans-[Co(dtc)2(PMe3−n
-
Phn)2]+ is observed at almost the same position (within 300 cm−1) as that of the corresponding acac complexes, while the
transition energies of the P-to-Co LMCT of these two series of complexes are rather different (at least 2700 cm−1) from
each other. Furthermore, the first and the second d–d transition bands of cis-[Co(dtc)2(PMe3 or PMe2Ph)2]+ are observed
at lower energy than those of cis-[Co(dtc)2{P(OMe)3}2]+ in spite of a weaker σ-donor of phosphite. The separation of
the first and the second d–d transition bands of the P(OMe)3 complex is remarkably smaller than the separation of the
bands of the PMe3 one, being indicative of a further reduction of the interelectronic repulsion in the P(OMe)3 complex.
In previous studies,1–8 we have prepared several series of
mixed-ligand phosphine complexes of cobalt(Ⅲ) with pentane-
2,4-dionate (acac) and investigated their crystal and molecular
structures, spectroscopic and electrochemical properties, and
reactivities toward isomerization or hydrolysis. While oxida-
tion of an ethanolic mixture of [Co(acac)2(H2O)2] and PPh3
with PbO2/AcOH yielded trans-[Co(acac)2(PPh3)2]+ exclu-
sively, a similar reaction using PMe3 afforded only cis-isomer.4
Such a difference in the preference of a particular geometrical
isomer would result from competition between electronic and
steric requirements of phosphines; strong Lewis bases such as
PMe3 tend to form a mutually cis configuration, which is unfa-
vorable for bulky phosphines like PPh3. However, trans-
[Co(acac)2(PMe3)2]+ was synthesized by a reaction of PMe3
with trans-[Co(acac)2(PPh3)2]+ in methanol.2 Once isolated,
the trans-isomer is stable toward isomerization in dry organic
solvents, but readily hydrolyzed in wet organic solvents to
form trans-[Co(acac)2(PMe3)(H2O)]+, while the cis-[Co-
(acac)2(PMe3)2]+ is stable toward isomerization and hydrolysis
in the same condition.2,4 In order to interpret these properties
of the mixed-ligand acac–phosphine complexes from structural
points of view, we performed the X-ray crystal structure analy-
ses for several series of complexes,1–3,6,8 and found that the dif-
ferences in Co–P bond lengths are closely correlated to these
properties. For example, the strong electronic trans influence
of PMe3, which was indicated by comparison of the Co–O
bond lengths between cis- and trans-[Co(acac)2(PMe3)2]+,
gave mutual large elongation of Co–P bonds in the trans-iso-
mer.2 In addition, the steric trans influence of PPh3 was real-
ized from the structural analysis of trans-[Co(acac)2(PPh3)2]+.1
These electronic and steric trans influences of phosphines
would lead to high reactivities of the trans-isomers for substi-
tution and/or hydrolysis.
We have also prepared a number of mixed-ligand N,N-di-
methyldithiocarbamato (dtc) complexes of cobalt(Ⅲ) incorpo-
rating phosphines, phosphinites, phosphonites, or phosphites
(P-ligand).8–14 In the case of monodentate phosphite (P(OMe)3,
P(OEt)3, P(OCH2)3CEt), the cis-isomers of [Co(dtc)2(phos-
phite)2]+ were readily converted photochemically to the trans-
isomers, which isomerized thermally to the original cis-iso-
mers.10,13 From these previous findings, we concluded that it
would be interesting to prepare a series of mixed-ligand dtc–
monodentate phosphine complexes: [Co(dtc)2(PMe3−nPhn)2]+
(n = 0, 1, 2 or 3), and to compare their molecular structures

2350 Bull. Chem. Soc. Jpn., 74, No. 12 (2001)
Cobalt(Ⅲ) Complexes with dtc and PMe3-nPhn
methanol (1:1, 40 cm3) with stirring. The mixture was stirred for
2 h at room temperature, and the resulting brown solution was fil-
tered to remove a green precipitate of [Co(dtc)3]. The filtrate was
evaporated to dryness under reduced pressure. The residue was
washed with diethyl ether (50 cm3), and then dissolved in a mix-
ture of methanol and dichloromethane (10:1, 55 cm3). The fil-
tered solution was divided into three portions, and each portion
was placed on a column of Sephadex LH-20 (φ 4.5 × 40 cm). The
adsorbed products were eluted with methanol, separating into four
colored bands: the first, a minor brown band; the second, a major
red purple band; the third, a minor yellow brown band; the fourth,
a minor green band. The major red purple band was collected, and
evaporated to dryness under reduced pressure. This residue was
and chemical and spectroscopic properties to those of the relat-
ed complexes above. However, there were only two complexes
so far for the monodentate phosphines: trans-[Co(dtc)2(PPh3 or
PEt3)2]+.9 Here, we will describe preparation, X-ray structural
analyses and spectroscopic characterization of trans-
[Co(dtc)2(PMe3)2]BF4 (t-0), trans-[Co(dtc)2(PMe2Ph)2]BF4 (t-
1), trans-[Co(dtc)2(PMePh2)2]BF4 (t-2), trans-[Co(dtc)2-
(PPh3)2]BF4 (t-3), cis-[Co(dtc)2(PMe3)2]BF4 (c-0), and cis-
[Co(dtc)2(PMe2Ph)2]PF6 (c-1ꢀ).
Experimental
Phosphines were handled under an atmosphere of argon until
such time as they formed air-stable cobalt(Ⅲ) complexes. All of
the solvents used in the preparation of complexes were deaerated
with argon for 20 min immediately before use. [Co(dtc)3] was
prepared by the literature method15 and recrystallized from dichlo-
romethane and methanol.
found to be
a mixture of t-1 (42%) and cis-[Co(dtc)2-
1
(PMe2Ph)2]BF4 (c-1: 58%) by H NMR. In order to separate the
isomers, the residue was dissolved in a mixture of methanol and
dichloromethane (4:1, 50 cm3), the filtered solution was evaporat-
ed (to 30 cm3) in the open air, and the concentrate was cooled in a
refrigerator overnight. Red brown needle crystals of pure t-1 were
deposited, which were collected by filtration and dried in air. The
product was recrystallized from CH2Cl2/CH3OH to give dark red
columnar crystals. Yield: 512 mg (17.7%). Anal. Found: C,
39.75; H, 5.17; N, 4.23%. Calcd for C22H34BCoF4N2P2S4: C,
39.89; H, 5.17; N, 4.23%.
trans-[Co(dtc)2(PPh3)2]BF4 (t-3): In order to achieve a high-
er yield and a better purity of this complex, the previous method9
was modified as follows. To a pink-colored ethanolic solution
(200 cm3) containing Co(BF4)2•6H2O (2.01 g, 5.90 mmol) and
PPh3 (3.11 g, 11.8 mmol) was added a pale yellow solution of tet-
ramethylthiuram disulfide (1.41 g, 5.86 mmol) in a mixture of
dichloromethane and ethanol (1:1, 60 cm3) with stirring. The
mixture was stirred for 1 h at room temperature, and a brown pre-
cipitate which formed was collected by filtration. To remove a by-
product of [Co(dtc)3], the product was dissolved in dichlo-
romethane (30 cm3) and reprecipitated by addition of diethyl ether
(70 cm3). The precipitate was filtered off, and dissolved in a mix-
ture of dichloromethane and ethanol (1:1, 140 cm3). Slow evapo-
ration of the solution to ca. 60 cm3 in the open air afforded brown
needle crystals of t-3. Yield: 2.42 g (45.0%). Purple hexagonal
plate crystals of t-3•H2O suitable for X-ray analysis were obtained
by repeated recrystallization from CH2Cl2/CH3OH in the open air.
From the filtrate of red brown needle crystals of t-1, crystals of
the pure cis-isomer was obtained as the PF6− salt (c-1ꢀ) as follows.
Solid NH4PF6 (4.5 g) was added to the filtrate with stirring and the
mixture was stirred for a while, to result in a red brown precipi-
tate. The precipitate was collected by filtration, dried in vacuo,
and dissolved in dichloromethane (10 cm3). To the filtered solu-
tion was diffused diethyl ether vapor in a closed vessel; the red
plate crystals which deposited were collected by filtration. The
product was recrystallized from CH2Cl2/CH3OH. Yield: 1.17 g
(37.1%). Anal. Found: C, 36.65; H, 4.70; N, 3.92%. Calcd for
C22H34CoF6N2P3S4: C, 36.67; H, 4.76; N, 3.89%.
Anal. Found: C, 54.22; H, 4.66; N, 3.08%.
C42H44BCoF4N2OP2S4: C, 54.32; H, 4.78; N, 3.02%.
Calcd for
trans-[Co(dtc)2(PMe3)2]BF4 (t-0): To a suspension of com-
plex t-3 (1.37 g, 1.50 mmol) in methanol (100 cm3) was added a
toluene solution of PMe3 (1.0 mol dm−3, 3.0 cm3) with stirring.
The mixture was stirred for 1 h at room temperature, giving a dark
red solution. The solution was evaporated to dryness under re-
duced pressure, and the residue was thoroughly washed with di-
ethyl ether (200 cm3). The product was found to be pure t-0 by 1H
NMR. Yield: 723 mg (93.7%). Further purification was done by
recrystallization from CH3OH to give red columnar crystals.
trans-[Co(dtc)2(PMePh2)2]BF4 (t-2):
To a pink-colored
methanol solution (80 cm3) of Co(BF4)2•6H2O (977 mg, 2.87
mmol) was added PMePh2 (1.15 g, 5.74 mmol) with stirring; the
color of the mixture turned to orange immediately. A solution of
tetramethylthiuram disulfide (689 mg, 2.87 mmol) in a mixture of
dichloromethane and methanol (1:2, 60 cm3) was added with stir-
ring to the orange solution, and the mixture was stirred for 2 h at
room temperature. The reaction mixture was then filtered to re-
move a green precipitate of [Co(dtc)3]. The filtrate was concen-
trated nearly to dryness under reduced pressure, and the residue
was washed with diethyl ether (50 cm3). This crude product was
Anal. Found: C, 26.70; H, 5.66; N, 5.18%.
C12H30BCoF4N2P2S4: C, 26.77; H, 5.62; N, 5.20%.
Calcd for
cis-[Co(dtc)2(PMe3)2]BF4 (c-0), trans-[Co(dtc)2(PMe3)2]PF6
(t-0ꢀ) and cis-[Co(dtc)2(PMe3)2]PF6 (c-0ꢀ): To a methanol solu-
tion (70 cm3) of Co(BF4)2•6H2O (1.02 g, 3.00 mmol) was added a
toluene solution of PMe3 (1.0 mol dm−3, 6.0 cm3) with stirring;
the color of the solution turned to green immediately. A solution
of tetramethylthiuram disulfide (726 mg, 3.02 mmol) in a mixture
of dichloromethane and methanol (1:1, 40 cm3) was added with
stirring to the green solution. After stirring for 2 h at room tem-
perature, the mixture was evaporated to dryness under reduced
pressure. The residue was washed with diethyl ether (100 cm3),
and dissolved in methanol (40 cm3). Undissolved green precipi-
tate was filtered off, and the filtrate was again evaporated to dry-
ness under reduced pressure. The residue was dissolved in a mix-
ture of methanol and dichloromethane (4:1, 20 cm3), and the fil-
tered solution was placed on a column of Sephadex LH-20 (φ 4.5
× 40 cm). The adsorbed products were eluted with methanol,
found to be
a mixture of t-2 (ca. 92%), cis-[Co(dtc)2-
(PMePh2)2]BF4 (c-2: ca. 5%) and [Co(dtc)3] (ca. 3%) by 1H NMR.
The crude product was dissolved in hot methanol (60 °C, 70 cm3),
and the filtered solution was cooled in a refrigerator overnight to
give dark red needle crystals, which were collected by filtration
and dried in air. The product was recrystallized from CH2Cl2/
CH3OH, affording dark red columnar crystals of pure t-2. Yield:
1.35 g (59.9%). Anal. Found: C, 48.85; H, 4.81; N, 3.56%. Calcd
for C32H38BCoF4N2P2S4: C, 48.86; H, 4.87; N, 3.56%.
trans-[Co(dtc)2(PMe2Ph)2]BF4 (t-1) and cis-[Co(dtc)2(PMe2-
Ph)2]PF6 (c-1ꢀ): To a greenish brown methanol solution (50
cm3) containing Co(BF4)2•6H2O (1.48 g, 4.36 mmol) and PMe2Ph
(1.20 g, 8.69 mmol) was added a solution of tetramethylthiuram
disulfide (1.05 g, 4.36 mmol) in a mixture of dichloromethane and

T. Suzuki et al.
Bull. Chem. Soc. Jpn., 74, No. 12 (2001) 2351
separating into four colored bands: the first, a major red purple
band; the second, a minor brown band; the third, a minor green
band; the fourth, a minor pink band. The major red purple band
was collected, and evaporated to dryness under reduced pressure.
This residue was found to be a mixture of t-0 (35%) and c-0 (65%)
by 1H NMR. In order to separate the isomers, the residue was dis-
solved in hot methanol (50 °C, 15 cm3), and the filtered solution
was slowly cooled to room temperature to form red prismatic
crystals of c-0. Further purification was done by recrystallization
from CH2Cl2/CH3OH. Yield: 315 mg (19.5%). Anal. Found: C,
26.70; H, 5.65; N, 5.10%. Calcd for C12H30BCoF4N2P2S4: C,
26.77; H, 5.62; N, 5.20%.
group was either I42d or I41md. The structure was solved reason-
ably on the assumption of I42d, but not on that of I41md. Both Co
and B atoms were located on a crystallographic two-fold axis.
Two of the F atoms of BF4− anion showed positional disorder that
was related to the crystallographic two-fold symmetry.
Compound t-3•H2O was found to crystallize in a monoclinic
space group P21 with Z = 4, indicating that two crystallographi-
cally independent complex cations, two BF4− anions and two sol-
vated water molecules exist in an asymmetric unit. Assumption of
the corresponding centrosymmetric space group P21/m did not
give any resonable structure solution.
Since the crystal of c-0 decomposed gradually during the data
collection, a linear decay correction was applied, although no sol-
vent molecules of crystallization were found in the structure anal-
ysis. The space group was assumed to be a centrosymmetric C2/c,
which gave a resonable structure solution with Z = 4. The Co
atom located on a crystallographic C2 axis. The counter anion was
located close to a crystallographic center of symmetry, so that the
B and two F atoms were treated as positionally disordered ones.
The structure of compound c-1ꢀ could be solved without any
difficulty on the assumption of centrosymmetric space group P1
with Z = 2.
Tables of crystallographic data (excluding structure factors),
atomic coordinates, thermal parameters, full lists of bond lengths
and angles, and some additional figures showing the molecular
structures of the complexes are deposited as Document No. 74063
at the Office of the Editor of Bull. Chem. Soc. Jpn. Crystallo-
graphic data have been deposited at the CCDC, 12 Union Road,
Cambridge CB2 1EZ, UK and copies can be obtained on request,
free of charge, by quoting the publication citation and deposition
numbers CCDC 169101–169106.
From the filtrate of red crystals of c-0, the PF6− salts of the iso-
mers, cis-[Co(dtc)2(PMe3)2]PF6 (c-0ꢀ: ca. 40%) and trans-
[Co(dtc)2(PMe3)2]PF6 (t-0ꢀ: ca. 60%), were precipitated by addi-
tion of a methanol solution (5 cm3) of NH4PF6 (5 g). The precipi-
tates were dissolved in a mixture of dichloromethane and metha-
nol (1:2, 20 cm3) at 30 °C, and the filtered solution was cooled in
a refrigerator overnight to give red prismatic crystals of pure t-0ꢀ.
Yield: 272 mg (15.2%). Anal. Found: C, 24.16; H, 5.02; N,
4.75%. Calcd for C12H30CoF6N2P3S4: C, 24.16; H, 5.07; N,
4.70%. The filtrate was evaporated to 10 cm3 in the open air to af-
ford red prismatic crystals of c-0ꢀ, which were recrystallized from
CH2Cl2/CH3OH to give pure crystals of c-0ꢀ. Yield: 181 mg
(10.1%). Anal. Found: C, 24.21; H, 4.89; N, 4.77%. Calcd for
C12H30CoF6N2P3S4: C, 24.16; H, 5.07; N, 4.70%.
Measurements. The 1H and 13C NMR spectra were obtained
at 30 °C on a JEOL GSX-400 spectrometer using tetramethylsil-
ane as an internal reference. The 31P and 59Co NMR spectra were
recorded at 30 °C on a JEOL Lambda 500 spectrometer using
85% H3PO4 and [Co(dtc)3] (in CDCl3) as an external reference for
31P and 59Co NMR, respectively. The chemical shift of [Co(dtc)3]
was set to be δ 6704.16 UV-vis absorption spectra in CH2Cl2 were
measured on a Perkin-Elmer Lambda 19 spectrophotometer at
room temperature.
X-ray Crystallographic Study. The X-ray diffraction data
were measured at 23 °C on an automated Rigaku AFC-5R four-
circle diffractometer equipped with a graphite-monochromated
Mo Kα radiation (λ = 0.71073 Å). For t-3•H2O, the intensities
collected up to 2θ = 55° by the ω scan mode were corrected for
Lorentz-polarization factors and for absorption effects by the nu-
merical integration method.17 For the others, the 2θ–ω scan meth-
od was applied to obtain the data up to 2θ = 60°, and the absorp-
tion corrections were made by an empirical method using three
sets of ψ-scan data.18 The structures were solved by the direct
method using SHELXS86 program,19 and refined on F2 (with all
independent reflections) by means of SHELXL97 program.20 All
calculations were carried out using a TeXsan software package.21
Crystal data are collected in Table 1; the details for each analysis
are given below.
Results and Discussion
Preparation and Structural Characterization of Com-
plexes. In previous papers, we have described the synthesis
of mixed-ligand cobalt(Ⅲ) complexes containing dtc and
phosphines9 or phosphites10 by oxidation of an ethanolic mix-
ture of Co(BF4)2•6H2O and either phosphine or phosphite with
tetramethylthiuram disulfide. While such a reaction with phos-
phite (P(OMe)3, P(OEt)3, or P(OCH2)3CEt) gave cis-[Co-
(dtc)2(phosphite)2]+ and [Co(dtc)(phosphite)4]2+, a similar re-
action with PPh3 afforded only trans-[Co(dtc)2(PPh3)2]BF4 (t-
3); neither the cis-isomer nor the tetrakis(PPh3) complexes
were obtained from the reaction mixture.9 When the reaction
was performed in a molar ratio of Co2+:PPh3:disulfide =
1:2:1, the isolated yield of complex t-3 was improved to 45%.
A similar reaction of Co(BF4)2•6H2O, PMePh2 and the di-
sulfide in a mixture of methanol and dichloromethane yielded
trans-[Co(dtc)2(PMePh2)2]BF4 (t-2) as a main product (60%
isolated yield). The geometrical structure of the product was
Compounds t-0 and t-1 were found to crystallize in a triclinic
space group P1 with Z = 1 and in a monoclinic space group P21/n
with Z = 2, respectively. In each crystal the Co atom was located
on a crystallographic center of symmetry. The B atom of BF4− an-
ion was also located on a center of symmetry, and the F atoms
showed severe positional disorder. The apparent structure of the
counter anion was an octahedral B(2/3F)6− with the F atoms having
1
determined by the H and 13C NMR spectra, both of which
showed a singlet resonance for N–CH3 (Table 2), and this
structure was confirmed by X-ray analysis (vide infra). The
corresponding cis-isomer, cis-[Co(dtc)2(PMePh2)2]BF4, could
not be isolated, but the existence (ca. 5%) in the crude reaction
1
−
product was detected by H NMR spectroscopy: a filled-in
a large thermal elipsoid, although the existence of BF4 in the
doublet resonance at δ 1.850 for P–CH3 and two singlet ones at
δ 2.789 and 2.985 for N–CH3.
In the case of PMe2Ph, both trans- and cis-isomers could be
isolated from the reaction mixture. After the chromatographic
crystals was confirmed by their infrared spectra and elemental
analyses.
For compound t-2, the Laue group (4/mmm) and the systematic
absences (h + k + l = odd, 2h + l = 4n) indicated that the space

2352 Bull. Chem. Soc. Jpn., 74, No. 12 (2001)
Cobalt(Ⅲ) Complexes with dtc and PMe3-nPhn
Table 1. Crystallographic Data
Compounds
t-0
t-1
t-2
t-3•H2O
c-0
c-1ꢀ
C12H30BCoF4N2P2S4 C22H34BCoF4N2P2S4 C32H38BCoF4N2P2S4 C42H44BCoF4N2OP2S4 C12H30BCoF4N2P2S4 C22H34CoF6N2P3S4
Chemical
formula
Formula weight 538.30
662.43
786.56
928.71
538.30
720.59
Color and shape red brown,
red, prism
dark red, prism
purple, plate
red, prism
red, plate
of crystal
Size of specimen
(mm)
column
0.50×0.30×0.20 0.60×0.40×0.30 0.20×0.20×0.20 0.35×0.28×0.26
0.20×0.20×0.14 0.40×0.35×0.20
Crystal system
Space group
a / Å
b / Å
c / Å
triclinic
P1
monoclinic
P21/n
6.630(3)
15.689(3)
14.447(3)
90
92.23(3)
90
1501.6(7)
2
tetragonal
I42d
21.622(3)
21.622
15.620(5)
90
90
90
7303(3)
8
1.431
3248
monoclinic
P21
9.970(5)
32.176(5)
14.230(3)
90
91.18(2)
90
4564(2)
4
monoclinic
C2/c
20.883(5)
11.681(6)
12.351(10)
90
123.01(3)
90
2526(2)
4
triclinic
P1
8.667(4)
11.170(3)
6.425(1)
90.86(2)
100.66(3)
78.43(3)
598.7(3)
1
14.481(2)
14.715(2)
7.433(1)
94.78(1)
94.26(1)
92.40(1)
1572.3(3)
2
α / °
β / °
γ / °
U / Å3
Z
Dcalc / Mg m−3
F(000)
1.493
278
1.465
684
1.352
1920
1.415
1112
1.522
740
µ(Mo Kα)
1.230
0.997
0.833
0.679
1.166
1.016
/ mm−1
Transmission
factors
Rint
0.732–1.000
0.013
0.893–1.000
0.018
0.946–1.000
0.053
0.818–0.853
0.035
0.766–1.000
0.166
0.857–1.000
0.014
No. of
independent
reflections
No. of
3480
4371
2888
10696
3738
9189
135
179
222
1034
137
351
parameters
R1 (F2:
0.045
0.037
0.044
0.074
0.062
0.036
2
2
Fo >2σ(Fo ))
wR2
0.136
1.027
0.119
1.010
0.107
1.008
0.218
1.019
0.186
1.046
0.109
1.031
(F2: all data)
GOF
Table 2. Selected NMR Dataa) in δ from TMS for 1H and 13C; 85% H3PO4 for 31P; K3[Co(CN)6] for 59Co
Complex
1H
13C
31P
59Co
P–CH3
N–CH3
3.296 (t)
2.874 (t)
2.659
2.387
3.243
3.304
3.256
3.297
3.253
P–CH3
N–CH3
39.07
38.56
38.29
37.70
38.06
38.37
38.10
38.54
37.67
S2CN
t-0
t-1
t-2
t-3
c-0
1.477 (vt)
1.823 (vt)
2.104 (vt)
—
12.93 (vt)
12.68 (vt)
11.50 (vt)
—
202.91
201.85
199.32
198.00
201.49
−2.77
4.79
14.21
25.17
7.27
3560
3575
3713
—
1.524 (fd)
15.16 (vt)
3144
c-1ꢀ
1.482 (fd)
1.574 (fd)
—
9.69 (vt)
16.21 (vt)
—
201.24
7.45
3368
[Co(dtc)3]
204.90
—
6704
a) at 30°C in CDCl3, except for 59Co NMR of t-0, t-1, and t-2 which are in CD3CN. The symbols in
parenthesis, fd, vt and t denote filled-in doublet, virtual triplet and triplet, respectively, and the other
signals are singlet.
separation from by-products, the obtained product was a mix-
ture of trans-[Co(dtc)2(PMe2Ph)2]BF4 (t-1: 42%) and cis-
[Co(dtc)2(PMe2Ph)2]BF4 (c-1: 58%), which was determined by
the 1H NMR spectrum. The isomers were separated from each
similar to those of t-2, except for the observation of a weak
coupling between N–CH3 and P nuclei to give a triplet reso-
1
nance for N–CH3. The H and 13C NMR spectra of c-1ꢀ give
two filled-in doublet and two virtual triplet signals, respective-
ly, for P–CH3 groups, which is consistent to the C2 symmetry
of the cationic complex.
−
−
other by fractional recrystallization of the BF4 or PF6 salt
1
(see Experimental). The H and 13C NMR spectra of t-1 are

T. Suzuki et al.
Bull. Chem. Soc. Jpn., 74, No. 12 (2001) 2353
Scheme 1. A plausible equilibrium in an acetonitrile solution
of complex t-3.
For PMe3 complexes, the product obtained by a similar
method to the above PMe2Ph complexes was found to be a
mixture of trans-[Co(dtc)2(PMe3)2]BF4 (t-0: 35%) and cis-
[Co(dtc)2(PMe3)2]BF4 (c-0: 65%). The formation of the trans-
isomer is in contrast to the preparation of [Co(acac)2(PMe3)2]+
by oxidation of an ethanolic mixture of [Co(acac)2(H2O)2] and
PMe3 by PbO2/AcOH, which gave only the cis-isomer.4 The
isomers were separated by fractional recrystallization of the
−
−
BF4 or PF6 salt (see Experimental), and the isolated yields
of the complexes are: c-0, 20%; cis-[Co(dtc)2(PMe3)2]PF6 (c-
0ꢀ), 10%; and trans-[Co(dtc)2(PMe3)2]PF6 (t-0ꢀ), 15%. We
have also attempted to prepare t-0 by a reaction of t-3 with
PMe3, analogously to the preparative method of trans-
[Co(acac)2(PMe3)2]PF6 from trans-[Co(acac)2(PPh3)2]PF6.2
This method was successful, and complex t-0 could be ob-
tained in 94% isolated yield without any formation of c-0 and
the other impurities. This indicates that complex t-3 may be a
good starting material for the complexes having a trans-
[Co(dtc)2(ligand)2]+ moiety.
1H, 13C, and 31P NMR spectroscopy suggest that, although
the PPh3 complex of t-3 is stable in dichloromethane and chlo-
roform, in acetonitrile, it dissociates one of the PPh3 ligands
and exists as an equilibrium mixture together with the dissoci-
ated product, which can probably be assigned as trans(P,
µ(S))-[{Co(dtc-κ2S)(PPh3)}2(µ(S)-dtc-κ2S,Sꢀ)2]2+ (Scheme 1)
by analogy of the 1H NMR spectrum to that of the correspond-
ing P(OMe)2Ph complex.13 All of the other isolated bis(dtc)
complexes are fairly stable in dichloromethane, chloroform,
acetonitrile, and methanol, even when the solvents are wet.
The stability of trans-[Co(dtc)2(PMe3, PMe2Ph or PMe-
Ph2)2]BF4 toward hydrolysis is in contrast to the analogous
bis(acac) complexes, trans-[Co(acac)2(PMe3−nPhn)2]PF6,
which are easily hydrolyzed in wet organic solvents.2,4
Fig. 1. Perspective drawings of the cationic complexes in (a)
cis-[Co(dtc)2(PMe3)2]BF4 (c-0: 40% probability level) and
(b) cis-[Co(dtc)2(PMe2Ph)2]PF6 (c-1ꢀ: 50% probability lev-
el). For both figures hydrogen atoms are omitted for clari-
ty.
trans influences of the P-ligands, unless there exists a stacking
interaction between the dtc plane and the phenyl ring of P-
ligand (vide infra). In the PMe3 complex (c-0), the Co–S(2)
bond is longer by 0.055 Å than the Co–S(1) bond (see Table
3), owing to the electronic trans influence of PMe3. For the
PMe2Ph complex (c-1ꢀ), there is an intramolecular π–π stack-
ing interaction between one of the dtc ligands and the phenyl
ring of C(31)–C(36), but the other dtc has no such interaction,
as seen from Fig. 1b. Therefore, we will compare only the
Co–S(1) and Co–S(2) bond lengths for evaluating the electron-
ic trans influence of PMe2Ph. The Co–S(2) bond trans to
P(PMe2Ph) is longer by 0.032 Å than the Co–S(1) bond trans
to S(dtc). In the P(OCH2)3CEt complex, cis-[Co(dtc)2{P-
(OCH2)3CEt}2]BF4, the difference between two kinds of Co–S
bond lengths is only 0.016 Å.10 Thus, the order of the
strengths of electronic trans influence in cis-[Co(dtc)2(P-
ligand)2]+ is suggested to be PMe3 > PMe2Ph > P(OCH2)3-
CEt. This order coincides with the strength of the σ-donicity
of the P-ligands: PMe3 (χd = 8.55) > PMe2Ph (10.60) >
P(OCH2)3CEt (18.39).22
Crystal Structures. The cis-Isomers: Electronic trans
Influences of PMe3 and PMe2Ph. The X-ray crystallo-
graphic analyses for c-0 and c-1ꢀ confirmed that the complexes
have cis configuration for the phosphine ligands. The molecu-
lar structures of the cationic complexes in c-0 and c-1ꢀ are
shown in Fig. 1, and selected bond lengths and angles are list-
ed in Table 3. In the PMe3 complex the Co atom is sited on a
crystallographic C2 axis, which bisects the P(1)–Co–P(1 ) and
S(2)–Co–S(2 ) angles. Both complexes have a similar struc-
ture with respect to the Co(dtc)2 moiety to those in cis-
[Co(dtc)2{P(OCH2)3CEt}2]BF4,10 [Co(dtc)2(dmpf)]BPh4 (dmpf
= 1,1 -bis(dimethylphosphino)ferrocene)8 and [Co(dtc)3],15
except for a slight deviation of the Co–S bond lengths de-
scribed below. The bite angles of dtc are 75.07(7)° and aver-
age 76.15(2)° for c-0 and c-1ꢀ, respectively, which are a little
smaller than the angles in [Co(dtc)3] (average 76.4°).
The Co–S bond lengths in cis-[Co(dtc)2(P-ligand)2]+ pro-
vide a good measure to elucidate the strength of electronic
Two Co–P bond lengths in the PMe2Ph complex (c-1ꢀ) are
2.2795(6) and 2.2637(7) Å, showing a relatively large discrep-

2354 Bull. Chem. Soc. Jpn., 74, No. 12 (2001)
Cobalt(Ⅲ) Complexes with dtc and PMe3-nPhn
Table 3. Selected Bond Lengths (Å) and Angles (°) of cis-[Co(dtc)2(PMe3)2]BF4
and cis-[Co(dtc)2(PMe2Ph)2]PF6
cis-[Co(dtc)2(PMe3)2]BF4 (c-0)
Co–P(1)
S(1)–Co–S(2)
P(1)–Co–S(2 ) 164.60(7)
2.200(2)
75.07(7)
Co–S(1) 2.255(2)
Co–S(2)
2.310(2)
P(1)–Co–P(1 ) 96.8(1)
S(1)–Co–S(1 ) 167.4(1)
cis-[Co(dtc)2(PMe2Ph)2]PF6 (c-1ꢀ)
Co–P(1)
Co–P(2)
S(1)–Co–S(2)
P(1)–Co–P(2)
P(1)–Co–S(4)
2.2795(6) Co–S(1) 2.2578(7)
2.2637(7) Co–S(3) 2.2791(7)
Co–S(2)
Co–S(4)
75.99(2)
164.80(2)
173.46(3)
2.2902(7)
2.2921(7)
76.31(2)
95.14(2)
173.03(2)
S(3)–Co–S(4)
S(1)–Co–S(3)
P(2)–Co–S(2)
Fig. 2. Perspective drawing (50% probability level) of the
cationic complex in trans-[Co(dtc)2(PMe3)2]BF4 (t-0).
ancy (0.016 Å) between them. Such a difference in the Co–
P(PMe2Ph) bond lengths may be caused by the above-men-
tioned stacking interaction between dtc and PMe2Ph, which
makes the interacting Co–S(dtc) bond shorter, as clarified in
the next section. In fact, the Co–P(2) bond of the interacting
PMe2Ph is shorter than the Co–P(1) bond of the non-interact-
ing one. However, we can not claim any generality of such a
relationship between the Co–P bond lengths and the stacking
interaction at present, because there are no other examples rel-
evant to the relationship.
The Co–P bond length in the PMe3 complex (c-0) is
2.220(2) Å, which is shorter than those in the PMe2Ph complex
(c-1ꢀ). The corresponding Co–P bond lengths in cis-
[Co(dtc)2{P(OCH2)3CEt}2]BF4, 2.169(1) and 2.172(1) Å,10 are
even shorter, in accordance with the cone angles of P-ligands:
PMe3, 118°; PMe2Ph, 122°; P(OCH2)3CEt, 101°.23 However,
for PHPh2 having a larger cone angle (126°), the correspond-
ing Co–P bond lengths in cis-[Co(dtc)2(PHPh2)2]BF4•CH3CN•
0.5Et2O are unexpectedly short, 2.2340(6) and 2.2258(7) Å.14
Moreover, the P–Co–P bond angles in the PMe3, PMe2Ph,
P(OCH2)3CEt and PHPh2 complexes are 96.8(2), 95.14(2),
92.55(5) and 90.51(2)°, respectively. These observations sug-
gest that, although the cone angles of P-ligands would be im-
portant as well, the other steric effects such as mutual orienta-
tion of their substituent groups (including the above-men-
tioned intramolecular stacking interaction with a dtc moiety)
and the electronic (σ-donor/π-acceptor) character of P-ligands
must be taken into consideration to specify the Co–P bond
Fig. 3. (a) Perspective and (b) top views (50% probability
level; hydrogen atoms are omitted) of the cationic complex
in trans-[Co(dtc)2(PMe2Ph)2]BF4 (t-1). For the figure of
(b) the PMe2Ph ligand below the equatorial CoS4 plane is
omitted for clarity.
lengths and P–Co–P angles in cis-[Co(dtc)2(P-lagand)2]+-type
complexes.
The trans-Isomers: Steric trans Influence via the Equato-
rial dtc Ligands for the Elongation of Co–P Bond.
Per-
spective drawings of the complex cations in trans-
[Co(dtc)2(PMe3)2]BF4 (t-0) and trans-[Co(dtc)2(PMe2Ph)2]BF4
(t-1) are shown in Figs. 2 and 3, respectively, and selected
bond lengths and angles are listed in Table 4. In both complex-
es the Co atom is located at a crystallographic inversion center,
and the complex cation has Ci molecular symmetry. The Co–S
bond lengths in t-0 (average 2.270 Å) are comparable to those
in [Co(dtc)3] (average 2.264 Å).15 Although the average Co–S

T. Suzuki et al.
Bull. Chem. Soc. Jpn., 74, No. 12 (2001) 2355
Table 4. Selected Bond Lengths (Å) and Angles (°) of trans-[Co(dtc)2(PMe3−nPhn)2]BF4(•H2O)
trans-[Co(dtc)2(PMe3)2]BF4 (t-0)
Co–P(1)
S(1)–Co–S(2)
2.287(1)
76.86(3)
Co–S(1)
2.2681(9)
Co–S(2)
2.2715(8)
2.2782(9)
2.293(1)
trans-[Co(dtc)2(PMe2Ph)2]BF4 (t-1)
Co–P(1)
S(1)–Co–S(2)
2.2843(8) Co–S(1)
76.75(3)
2.2626(7)
Co–S(2)
trans-[Co(dtc)2(PMePh2)2]BF4 (t-2)
Co–P(1)
S(1)–Co–S(2)
S(1)–Co–S(1 )
2.303(1)
76.18(4)
94.38(6)
Co–S(1)
2.264(1)
P(1)–Co–P(1 )
S(2)–Co–S(2 )
Co–S(2)
169.66(7)
113.29(6)
trans-[Co(dtc)2(PPh3)2]BF4•H2O (t-3•H2O)
Co(1)–P(1)
Co(1)–P(2)
Co(51)–P(51)
Co(51)–P(52)
S(1)–Co(1)–S(2)
S(1)–Co(1)–S(3)
2.319(3)
2.303(3)
2.331(4)
2.311(3)
76.2(1)
Co(1)–S(1)
Co(1)–S(3)
Co(51)–S(51) 2.254(4)
Co(51)–S(53) 2.245(3)
2.242(3)
2.251(3)
Co(1)–S(2)
Co(1)–S(4)
Co(51)–S(52) 2.271(3)
Co(51)–S(54) 2.279(3)
76.6(1)
2.285(3)
2.283(3)
S(3)–Co(1)–S(4)
S(2)–Co(1)–S(4)
94.0(1)
113.4(1)
S(51)–Co(51)–S(52) 76.0(1)
S(51)–Co(51)–S(53) 94.3(1)
S(53)–Co(51)–S(54) 76.1(1)
S(52)–Co(51)–S(54) 113.6(1)
P(51)–Co(51)–P(52) 167.1(1)
P(1)–Co(1)–P(2)
168.3(1)
bond length in t-1 (2.270 Å) is just the same as that in t-0, there
is a rather large discrepancy between two Co–S bond lengths
in t-1 (Co–S(1) = 2.2626(7) and Co–S(2) = 2.2782(9) Å).
This discrepancy would be related to the intramolecular π–π
stacking interaction between the dtc moiety and the phenyl
substituent of PMe2Ph. As seen from Fig. 3b, the phenyl ring
is located just above the S(1) atom, and the dihedral angle be-
tween the dtc mean plane and the phenyl ring is 18.0(1)°. Such
a stacking interaction between dtc plane and phenyl ring, ac-
companied with a shortening of the interacting Co–S bond
lengths, is also observed in the analogous PMePh2 (t-2) and
PPh3 (t-3) complexes (vide infra and Table 4). Thus, it should
be noted that the intramolecular π–π stacking makes the inter-
acting Co–S bond shorter.
symmetry; the Co atom is sited on a crystallographic C2 axis,
which bisects the S(1)–Co–S(1 ) and P(1)–Co–P(1 ) angles.
One of the phenyl rings, C(11)–C(16), of PMePh2 is located
above the S(1) atom, similar to the PMe2Ph complex t-1, with
the dihedral angle of 5.24(6)° between the dtc mean plane and
the phenyl ring. The Co–S(1) bond is shorter by 0.029 Å than
the Co–S(2) bond. The other phenyl ring, C(17)–C(22), is lo-
cated in the cleft between S(2) and S(2 ) atoms, and oriented
perpendicularly to the equatorial CoS4 plane, the angle be-
tween the phenyl ring and the CoS4 mean plane being 87.1(1)°.
This is also the case in another side of the equatorial CoS4
plane, as suggested from molecular C2 symmetry (Fig. 4b).
Such a conformation of the phenyl rings is remarkably differ-
ent from that in trans-[Co(acac)2(PMePh2)2]PF6.3 Since dtc is
a sterically more compact didentate ligand than acac, as indi-
cated by the smaller bite angle, the cleft in the equatorial
Co(dtc)2 coordination plane is much larger than that of the
Co(acac)2 plane, as illustrated in Fig. 5. Moreover, since the
coordination of dtc is rather flexible as far as the deviation
The Co–P bond length in the PMe3 complex (t-0: 2.287(1)
Å) is comparable to (or slightly longer than) that in the
PMe2Ph one (t-1: 2.2843(8) Å), in contrast to those of the cis-
complexes. The differences in Co–P bond lengths between the
trans- and cis-isomers are 0.087 Å for the PMe3 complexes
and 0.021 Å24 for the PMe2Ph ones. This is also due to the fact
that the electronic trans influence of PMe3 is stronger than that
of PMe2Ph, which elongates the mutually trans Co–P bonds.
Since there would be no difference in the electronic cis influ-
ence,25 a smaller steric requirement of PMe3 would make the
Co–P(PMe3) bonds shorter than the Co–P(PMe2Ph) bonds,
while a stronger σ-donicity of PMe3 induces the stronger elec-
tronic trans influence, leading to mutual elongation of the Co–
P(PMe3) bonds. Therefore, the comparable Co–P bond lengths
observed in t-0 and t-1 result from the competition of these
opposite effects. We have previously found a similar result
for the Co–P bonds in the analogous acac complexes: trans-
[Co(acac)2(PMe3 or PMe2Ph)2]+.1
Ⅲ
from the regular octahedron of the Co coordination sphere,
the cleft of the S(2)–Co–S(2 ) side (113.29(6)°) is larger than
the other S(1)–Co–S(1 ) side (94.38(6)°). As indicated in Fig.
4b by the space filling model, the o-hydrogen atom of each
phenyl ring oriented perpendicularly is just fitted into the larg-
er cleft between S(2) and S(2 ) atoms.
Similar structural characteristics to the above PMePh2
complex are also found in the PPh3 complex, trans-
[Co(dtc)2(PPh3)2]BF4 (t-3). There are two crystallographically
independent complex cations: cation A with Co(1) atom and
cation B with Co(51). A perspective view of the cation A is
depicted in Fig. 6, and the molecular structure of cation B is
very similar to that of A. Both complex cations have, not a
crystallographically imposed, but a pseudo molecular C2 sym-
metry. One of three phenyl rings of PPh3 is perpendicularly
The molecular structure of trans-[Co(dtc)2(PMePh2)2]+ in t-
2 is shown in Fig. 4a. The complex cation has molecular C2

2356 Bull. Chem. Soc. Jpn., 74, No. 12 (2001)
Cobalt(Ⅲ) Complexes with dtc and PMe3-nPhn
Fig. 5. Space filling models of (a) the Co(dtc)2 moiety in
trans-[Co(dtc)2(PMePh2)2]BF4 (t-2) and (b) the Co(acac)2
moiety in trans-[Co(acac)2(PMePh2)2]PF6.
Fig. 4. (a) Perspective drawing (50% probability level; hy-
drogen atoms are omitted) of the cationic complex in
trans-[Co(dtc)2(PMePh2)2]BF4 (t-2). (b) Space filling
model of the cationic complex viewed from the bisector of
the S(2)–Co–S(2 ) angle.
oriented to the larger cleft in the equatorial Co(dtc)2 plane; the
dihedral angles between the phenyl ring and the CoS4 mean
plane are 83.8(6)–89.6(3)°; the angles of S(1)–Co(1)–S(3) and
S(51)–Co(51)–S(53) average 94.2°, while those of S(2)–
Co(1)–S(4) and S(52)–Co(51)–S(54) average 113.5°.
Fig. 6. Perspective drawing (40% probability level; hydro-
gen atoms are omitted) of one of the complex cations in
trans-[Co(dtc)2(PPh3)2]BF4•H2O (t-3•H2O).
The compactness and flexibility of dtc coordination result in
reduction of the steric interaction from the equatorial ligands
length in t-3•H2O is 2.316 Å, which is extremely shorter (by
0.073 Å) than that in trans-[Co(acac)2(PPh3)2]PF6.1 As seen
from Fig. 7, the Co–P bond lengths in t-1, t-2 and t-3 are lin-
early correlated to the crystallographic cone angle of PMe2Ph,
PMePh2 and PPh3, although the plot for t-0 deviates from the
linearity due to the strong electronic trans influence of PMe3
(vide intra). Therefore, what the plots in Fig. 7 indicate is an
exceptionally long Co–P bond in trans-[Co(acac)2(PPh3)2]PF6.
In the previous paper,1 we have concluded that such long Co–P
bonds in trans-[Co(acac)2(PPh3)2]PF6 result from the severe
steric interaction between the equatorial acac and the axial
to elongate the axial Co–P bonds in trans-[Co(dtc)2(PMe3−n
-
Phn)2]BF4, when compared to those in trans-[Co-
(acac)2(PMe3−nPhn)2]PF6. In Fig. 7, the Co–P bond lengths in
the above two series of complexes, together with those in the
related complexes, are plotted against the crystallographic
cone angle26 of PMe3−nPhn. The Co–P bond lengths in t-0, t-1
and t-2 are shorter by 0.020, 0.017 and 0.026 Å, respectively,
than those of the corresponding acac complexes; this fact is in-
dicative of reduction of the steric interaction from the equatori-
al ligands. The shortening of the Co–P bond lengths is quite
remarkable for the PPh3 complexes. The average Co–P bond

T. Suzuki et al.
Bull. Chem. Soc. Jpn., 74, No. 12 (2001) 2357
Fig. 7. Comparison of the Co–P bond lengths against the
crystallographic cone angles26 of PMe3−nPhn: trans-
[Co(dtc)2(PMe3−nPhn)2]+ (ꢀ); trans-[Co(acac)2(PMe3−n
-
Phn)2]+ (ꢁ); trans-[Co(acac)2(PMe3−nPhn)(H2O)]+ (ꢂ);
trans(P,P),cis(C,C)-[Co(acac)(CN)2(PMePh2)2] (ꢃ).
PPh3 ligands in both sides; that is, the steric trans influence via
equatorial acac ligands, where the steric interaction is mainly
associated with the orientation and conformation of three phe-
nyl rings of PPh3. In the present dtc complex, such a steric in-
teraction between the equatorial dtc and the axial PPh3 ligands
would become less effective because of the compactness and
flexibility of dtc coordination, as mentioned above. In fact,
one of the phenyl rings of PPh3 (and PMePh2) can orient per-
pendicularly to fit the cleft of the equatorial CoS4 plane, and
the other phenyl rings can orient rather freely to minimize the
steric interaction. Hence, it is stated that the steric trans influ-
ence via equatorial dtc ligands is negligible in trans-
[Co(dtc)2(PMe3−nPhn)2]BF4.
A shorter Co–P bond in the related trans-bis(PMePh2) type
complex has been found in trans(P,P),cis(C,C)[Co(acac)-
(CN)2(PMePh2)2] (2.2698(4) Å; see Fig. 7),3 which is as stable
as the present dtc complexes for hydrolysis. Although the
electronic cis influence25 should be taken into consideration for
the comparison of their reactivities, it is difficult to evaluate the
influence experimentally. However, the Co–P bond lengths are
affected by not only the steric but also the electronic influenc-
es, so that the length would be a good experimental measure to
consider the difference in reactivity, when complexes having
the same phosphine are compared. In conclusion, the short
Co–P bonds in the dtc complexes are the structural counterpart
for the stability toward hydrolysis.
Spectroscopic Properties. The absorption spectra of c-0
and c-1ꢀ, together with those of cis-[Co(dtc)2{P(OMe)3}2]-
BF410 and [Co(dtc)3], are shown in Fig. 8a. The spectra of c-0
and c-1ꢀ are very similar to each other in the region up to
34000 cm−1; there are two bands with medium intensity
around 18000 and 23000 cm−1, a shoulder feature around
29000 cm−1, and a very intense band around 32500 cm−1. The
two lowest energy bands are assignable as the first and the sec-
ond d–d transition ones, respectively, similar to those of the
analogous cis-[Co(dtc)2(P-ligand)2]+ and [Co(dtc)2(P–P)]+ (P–
P = Me2P(CH2)nPMe2 (n = 1: dmpm, 2: dmpe, or 3: dmpp) or
(MeO)2P(CH2)2P(OMe)2) complexes.9–11 Due to the lowering
Fig. 8.
UV-vis absorption spectra of (a) cis-
[Co(dtc)2(PMe3)2]+ ( ——— ), cis-[Co(dtc)2(PMe2Ph)2]+
( - - - - ), cis-[Co(dtc)2{P(OMe)3}2]+ ( – • • – ), and
[Co(dtc)3] ( – – – ); and (b) trans-[Co(dtc)2(PMe3−nPhn)2]+
(n = 0 ( ——— ); 1 ( – • – ); 2 ( – – – ); and 3 ( - - - - )) in
dichloromethane at room temperature.
molecular symmetry of the complexes from holohedrized Oh
symmetry, the splitting of the d–d transition bands is expected,
but the first d–d bands of the complexes (c-0 and c-1ꢀ) are
found to be as symmetrical as the first d–d band of [Co(dtc)3].
The positions of the first and the second d–d bands of the com-
plexes are estimated by the Gaussian curve fitting analysis, and
are listed in Table 5. It is found that the first and the second d–
d transition energies of the PMe3 complex are slightly higher
than those of the PMe2Ph complex, but are lower than those of
the didentate diphosphine (dmpm, dmpe, and dmpp) complex-
es.9 It is also interesting to compare these energies to those of
the corresponding phosphite (P(OMe)3, P(OEt)3, and
P(OCH2)3CEt) complexes.10 The first d–d band of the PMe3
complex is observed to be lower in energy by 1100 cm−1 than
that of the P(OMe)3 complex, but the difference in the second
d–d transition energy between these complexes is only 400
cm−1 (Table 5). This fact indicates that the interelectronic re-
pulsion parameter, B, of the P(OMe)3 complex is reduced re-
markably from that of the PMe3 complex. The reduction of B
parameter in the phosphite complex would result from π back-
bonding, which is negligible for Co -phosphine bonds.7,22
Ⅲ
The absorption spectra of the trans series of complexes are
shown in Fig. 8b, and the positions of the absorption bands
evaluated by the Gaussian curve fitting analysis are listed in
Table 5. In the region of 12000–20000 cm−1, an absorption
band or shoulder is observed with the intensity of ε ~ 300 dm3

2358 Bull. Chem. Soc. Jpn., 74, No. 12 (2001)
Cobalt(Ⅲ) Complexes with dtc and PMe3-nPhn
Table 5. Gaussian Curve Fitting Results for Absorption Spectra of the Complexes, 10−3 σ/cm−1
(ε / dm3 mol−3 cm−1
)
Complex
d–d Band
CT-band
c-0
c-1ꢀ
t-0
t-1
t-2
t-3
18.33(740.8)
17.91(742.6)
17.90(258.1)
17.17(225.4)
16.58(354.5)
15.74(342.3)
15.50(435.2)
23.31(1340)
22.96(1716)
29.25(10560) 32.62(24990)
28.89(15200) 32.04(21800)
22.81(587.0)a) 27.38(13540) 30.11(21970)
22.12(665.8)a) 25.57(12090) 29.70(25120)
19.97(580.5)
24.77(10950) 29.63(21750)
20.20(868.7)a) 23.08(9488)
30.18(22290)
27.73(2794)
[Co(dtc)3]
cis-[Co(dtc)2-
{P(OMe)3}2]BF4
20.62(588.9)
23.73(1552)
25.87(8437)
29.03(9840)
31.21(19490)
19.41(704.2)
34.08(26980)
a) Very broad shoulder in the observed spectra.
Ⅲ
mol−1 cm−1. This band (or shoulder) is regularly blue-shifted
in the order of PPh3 < PMePh2 < PMe2Ph < PMe3. Previous-
ly, for the PPh3 complex,9 we have assigned this band as the
degenerate splitting component (a1Eg) of the first d–d transi-
tion band under the holohedrized D2h symmetry, and the same
assignment resulted from the analogous acac complexes,
trans-[Co(acac)2(PMe3−nPhn)2]+.1 Interestingly, the transition
energies of the dtc and acac complexes are approximately the
same; the difference being less than 300 cm−1. This fact sug-
gests that the ligand-field strengths of these two series of com-
plexes, trans-[Co(dtc)2(PMe3−nPhn)2]+ and trans-[Co(acac)2-
(PMe3−nPhn)2]+, are nearly the same, when the same phos-
phine is concerned. There is a very broad shoulder feature ob-
served in the spectra of dtc complexes (Table 5), but we could
not assign this transition to either the non-degenerate splitting
component (1A2g) or the second d–d transition ones (1B2g and
b1Eg).
gies is smaller than that of χ(Co (acac)2) parallels the electro-
chemical (reduction potential) data of the related dtc and acac
complexes.7,8,12 The details in electrochemistry of the present
complexes will be reported elsewhere, togther with those of
the related P(OMe)3−nPhn complexes.13
The data of 31P and 59Co NMR chemical shifts of the com-
plexes are also given in Table 2. Except for t-3, all of the com-
plexes gave a broad 59Co resonance in the region of δ 3000–
4000, but these values are very different from those of the cor-
responding acac complexes (δ 12200–13000).1 Because at
least the a1Eg transition energies of the dtc and acac complexes
are nearly equal, the difference in the 59Co NMR chemical
shifts may indicate a large difference in reduction parameter,
α,1,29 between these two series of complexes, but the details are
still unkown at present.
Conclusion
Each complex of trans-[Co(dtc)2(PMe3−nPhn)2]+ shows an
intense CT band around 30000 cm−1. In the lower energy re-
gion, there is another CT band, which is regularly blue-shifted
as the number of methyl groups of phosphine increased: from
23080 cm−1 for t-3 to 27380 cm−1 for t-0. Thus, it seems like-
ly to assign these CT bands to S-to-Co and P-to-Co LMCT
transitions, respectively. A P-to-Co LMCT band is also ob-
served for the series of trans-[Co(acac)2(PMe3−nPhn)2]+ com-
plexes (18970, 21420, 22830, and 24640 cm−1 for the PPh3,
PMePh2, PMe2Ph, and PMe3 complexes, respectively),1 but the
CT transition energy is lower (by more than 2700 cm−1) than
the present dtc complex. The P-to-Co LMCT transition energy
is estimated by
The compactness and flexibility of dtc coordination have
given rise to remarkable differences in the structures of the
mixed-ligand dtc–phosphine cobalt(Ⅲ) complexes from those
of the corresponding acac–phosphine complexes. The most in-
triguing differences have been found in the structures of trans-
[Co(dtc)2(PMePh2 or PPh3)2]BF4 (t-2 or t-3) and trans-
[Co(acac)2(PMePh2 or PPh3)2]PF6,1,3 which contain sterically
rather bulky phosphines. While all complexes of trans-
[Co(acac)2(PMe3−nPhn)2]+ and trans-[Co(dtc)2(PMe3 or
PMe2Ph)2]+ have crystallographically imposed Ci molecular
symmetry with the inversion center at the Co atom, the com-
plexes of trans-[Co(dtc)2(PMePh2 or PPh3)2]+ have (crystallo-
graphically imposed or pseudo) C2 molecular symmetry, where
the size of two clefts in the equatorial Co(dtc)2 plane is re-
markably different (Fig. 5a), as exemplified by two angles of
S(1)–Co–S(1 ) and S(2)–Co–S(2 ) for t-2. Toward the larger
cleft, one of the phenyl rings of PMePh2 or PPh3 is oriented
perpendicularly to minimize the steric interaction between the
equatorial dtc and the axial phosphine ligands (Fig. 4b). Such
a perpendicular orientation of phenyl ring would be prohibited
for the bis(acac) complexes because of the larger steric re-
quirement of acac, so that a strong steric trans influence via the
equatorial acac ligands for the elongation of mutually trans
Co–P bonds would be expected, especially for the PPh3 com-
plex. In contrast, the steric trans influence via the equatorial
dtc ligands would, therefore, be negligibly weak. This is re-
Ⅲ
σ
LMCT = 30000{χ(P-ligand) − χ(Co )} + ∆ − 7.6B,
Ⅲ
where χ(P-ligand) and χ(Co ) are the optical electronegativi-
Ⅲ 1,27
ties of P-ligand and Co .
The ∆ values of the dtc and acac complexes are nearly equal
to each other, as suggested in the above. Although it is known
Ⅲ
that many Co –dtc complexes show rather large reduction of
the interelectronic repulsion,28 the reduction (expected to be
less than 100 cm−1) of B value alone can not explain the ob-
served large difference in the LMCT transition energy. There-
fore, it is indicative of a different optical electronegativity be-
Ⅲ
Ⅲ
tween Co (dtc)2 and Co (acac)2 moieties. The fact that the
Ⅲ
value of χ(Co (dtc)2) obtained by the LMCT transition ener-
flected in the Co–P bond lengths in trans-[Co(dtc)2(PMe3−n-

T. Suzuki et al.
Bull. Chem. Soc. Jpn., 74, No. 12 (2001) 2359
Phn)2]+ being shorter than those in the corresponding trans-
[Co(acac)2(PMe3−nPhn)2]+; also the shorter Co–P bonds ob-
served can interpret the stability of the dtc complexes toward
hydrolysis.
Kinoshita, K. Kashiwabara, and J. Fujita, Bull. Chem. Soc. Jpn.,
55, 3179 (1982).
7
8
M. Kita and K. Kashiwabara, Polyhedron, 16, 2771 (1997).
M. Adachi, M. Kita, K. Kashiwabara, J. Fujita, N. Iitaka, S.
Kurachi, and S. Ohba, Bull. Chem. Soc. Jpn., 65, 2037 (1992).
M. Kita, A. Okuyama, K. Kashiwabara, and J. Fujita, Bull.
It has also been found that the Co–S bond lengths in the dtc–
phosphine complexes vary with two effects: (1) the intramo-
lecular π–π stacking interaction between the dtc plane and the
phenyl ring of phosphine, and (2) the electronic trans influence
of the phosphine. The π–π stacking interaction makes the in-
teracting Co–S bond length shorter, as illustrated most clearly
in the structure of trans-[Co(dtc)2(PMe2Ph)2]+. Comparison
of the Co–S bond lengths which do not have the stacking inter-
action in cis-[Co(dtc)2(PMe3, PMe2Ph, or P(OCH2)3CEt)2]+
gives the order of the strength of the electronic trans influence
as PMe3 > PMe2Ph > P(OCH2)3CEt, which is in accordance
with the order of their σ-donicity. In contrast, the first d–d
transition energies of cis-[Co(dtc)2(PMe3, PMe2Ph, or
P(OMe)3)2]+ do not follow the order of σ-donicity of the P-
ligands. The first d–d transition band of the PMe3 complex is
observed at an energy lower by 1100 cm−1 than that of the
P(OMe)3 complex, but the difference in energy of the second
d–d bond between these complexes is only 400 cm−1. This in-
dicates that the interelectronic repulsion parameter, B, in the
P(OMe)3 complex (B = 270 cm−1) is much reduced from that
in the PMe3 complex (311 cm−1).
9
Chem. Soc. Jpn., 63, 1994 (1990).
10 H. Matsui, M. Kita, K. Kashiwabara, and J. Fujita, Bull.
Chem. Soc. Jpn., 66, 1140 (1993).
11 M. Kita, M. Okuno, K. Kashiwabara, and J. Fujita, Bull.
Chem. Soc. Jpn., 65, 3042 (1992).
12 M. Okuno, M. Kita, K. Kashiwabara, and J. Fujita, Chem.
Lett., 1989, 1643.
13 K. Kashiwabara, T. Suzuki et al., to be published; S.
Kashiwamura, K. Kashiwabara, M. Kita, and J. Fujita, The 43rd
Symposium on Coordination Chemistry of Japan, Sendai (1993),
Abstract 3A07.
14 T. Suzuki, S. Iwatsuki, H. D. Takagi, and K. Kashiwabara,
Chem. Lett., 2001, 1068.
15 H. Iwasaki and K. Kobayashi, Acta Crystallogr., Sect. B,
36, 1657 (1980).
16 A. Yamasaki, J. Coord. Chem., 24, 211 (1991).
17 P. Coppens, L. Leiserowitz, and D. Rabinovich, Acta Crys-
tallogr., 18, 1035 (1965).
18 A. C. T. North, D. C. Phillips, and F. S. Mathews, Acta
Crystallogr., Sect. A, 24, 351 (1968).
19 G. M. Sheldrich, Acta Crystallogr., Sect. A, 46, 467 (1990).
20 G. M. Sheldrich, SHELXL97, University of Göttingen,
Germany (1997).
The authors wish to thank Prof. Shigeru Ohba (Keio Univer-
sity) and Mr. Seiji Adachi (Osaka University) for the structural
analysis of compound t-2 and measurements of 31P and 59Co
NMR spectra, respectively.
21 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).
22 H.-Y. Liu, K. Eriks, A. Prock, and W. P. Giering, Organo-
metallics, 9, 1758 (1990); Md. M. Rahman, H.-Y. Liu, K. Eriks,
A. Prock, and W. P. Giering, Organometallics, 8, 1 (1989).
23 C. A. Tolman, Chem. Rev., 77, 313 (1977).
24 For this comparison, only Co–P(2) = 2.2637(7) Å of c-1ꢀ
is used as the relevant Co–P bond length in the cis-isomer, be-
cause the PMe2Ph ligands in t-1 have a stacking interaction with
dtc as the P(2)Me2Ph in c-1ꢀ.
25 N. Bresciani-Pahor, M. Forcolin, L. G. Marzilli, L.
Randaccio, M. F. Summers, and P. J. Toscano, Coord. Chem. Rev.,
63, 1 (1985); N. Bresciani-Pahor, M. Calligaris, G. Nordin, and L.
Randaccio, J. Chem. Soc., Dalton Trans., 1982, 2549.
26 T. E. Müller and D. M. P. Mingos, Transition Met. Chem.,
20, 533 (1995).
References
1
T. Suzuki, S. Kaizaki, and K. Kashiwabara, Inorg. Chim.
Acta, 298, 131 (2000).
2
T. Suzuki, K. Kashiwabara, M. Kita, J. Fujita, and S.
Kaizaki, Inorg. Chim. Acta, 281, 77 (1998).
3
K. Kashiwabara, M. Kita, H. Masuda, S. Kurachi, and S.
Ohba, Bull. Chem. Soc. Jpn., 68, 883 (1995).
K. Katoh, H. Sugiura, K. Kashiwabara, and J. Fujita, Bull.
4
Chem. Soc. Jpn., 57, 3580 (1984); K. Kashiwabara, K. Katoh, T.
Ohishi, J. Fujita, and M. Shibata, Bull. Chem. Soc. Jpn., 55, 149
(1982).
5
S. Hayakawa, H. Nishikawa, K. Kashiwabara, and M.
Shibata, Bull. Chem. Soc. Jpn., 54, 3593 (1981).
M. Kita, K. Kashiwabara, and J. Fujita, Bull. Chem. Soc.
6
27 A. B. P. Lever, “Inorganic Electronic Spectroscopy,” 2nd
Ed., Elsevier, Amsterdam (1984), Chap. 5.
28 T. Ohishi, K. Kashiwabara, and J. Fujita, Bull. Chem. Soc.
Jpn., 60, 583 (1987).
Jpn., 61, 3187 (1988); M. Kita, K. Nemoto, K. Kashiwabara, J.
Fujita, M. Goto, T. Miyamoto, and S. Ohba, Bull. Chem. Soc. Jpn.,
66, 1687 (1993); M. Takata, K. Kashiwabara, H. Ito, T. Ito, and J.
Fujita, Bull. Chem. Soc. Jpn., 58, 2247 (1985); M. Atoh, I.
29 Y. Shimura, Bull. Chem. Soc. Jpn., 61, 693 (1988).