Isocyanide derivatives of manganese cobalt nonacarbonyl: Synthesis and characterization

Article abstract of DOI:10.1016/S0020-1693(99)00081-X

The synthesis and spectral characterization of isocyanide derivatives of MnCo(CO)₉ of the type MnCo(CO)_9-n(CNR)_n (n=1,2; R=(p-MeOC₆H₄)NC, (2,6-xylyl)NC, (t-butyl)NC, (n-butyl)NC, (2,6-diisopropylphenyl)NC) are reported. The X-ray structure of MnCo(CO)₈((2,6-xylyl)NC) (1) and MnCo(CO)₇((2,6-xylyl)NC)₂ (2) are reported herein as well. The structures of 1 and 2 contain a single metal-metal bond and no bridging ligands. The length of the Mn-Co bond is 2.870(1) and 2.9035(7) A in 1 and 2, respectively. The IR spectra of Mn(CO)₅Br, Mn₂(CO)₁₀, MnCo(CO)₉ and MnCo(CO)_9-n(CNR)_n (n=1,2; R=CH₃) were calculated using density functional theory. The calculated IR spectra were employed to assign the structure of the MnCo(CO)_9-n(CNR)_n (n=1,2) compounds.

Full text of DOI:10.1016/S0020-1693(99)00081-X

Inorganica Chimica Acta 288 (1999) 159–173
Isocyanide derivatives of manganese cobalt nonacarbonyl:
synthesis and characterization
Karsten Beck 1, John J. Alexander *, Jeanette A. Krause Bauer, Jeffrey L. Nauss 2
Department of Chemistry, Uni6ersity of Cincinnati, Cincinnati, OH 45221, USA
Received 12 June 1998; accepted 1 February 1999
Abstract
The synthesis and spectral characterization of isocyanide derivatives of MnCo(CO)9 of the type MnCo(CO)9−n(CNR)n (n=1,2;
R=(p-MeOC6H4)NC, (2,6-xylyl)NC, (t-butyl)NC, (n-butyl)NC, (2,6-diisopropylphenyl)NC) are reported. The X-ray structure of
MnCo(CO)8((2,6-xylyl)NC) (1) and MnCo(CO)7((2,6-xylyl)NC)2 (2) are reported herein as well. The structures of 1 and 2 contain
˚
a single metal–metal bond and no bridging ligands. The length of the Mn–Co bond is 2.870(1) and 2.9035(7) A in 1 and 2,
respectively. The IR spectra of Mn(CO)5Br, Mn2(CO)10, MnCo(CO)9 and MnCo(CO)9−n(CNR)n (n=1,2; R=CH3) were
calculated using density functional theory. The calculated IR spectra were employed to assign the structure of the MnCo(CO)9−n
-
(CNR)n (n=1,2) compounds. © 1999 Elsevier Science S.A. All rights reserved.
Keywords: Crystal structures; Manganese complexes; Cobalt complexes; Carbonyl complexes; Isocyanide complexes
1. Introduction
gated. In the case of M2(CO)10 (M=Mn, Re) com-
plexes of the type M2(CO)10−n(CNR)n (n=1–4)
(R=alkyl, aryl) were obtained. In no case could more
than four carbonyls be substituted by isocyanides. By
contrast, reaction of Co2(CO)8 with CNR gave the
disproportionation product [Co(CNR)5]+[Co(CO)4]−
at room temperature and the completely substituted
product Co2(CNR)8 (R=2,6-xylyl) at 80°C.
The structures of the Mn and Re compounds show a
single metal–metal bond, no bridging ligands, and stag-
gered arrangement of the eight equatorial ligands [4,5].
The isocyanide ligands are always equatorial and, if on
the same metal, cis to each other. The structures of
Co2(CNR)8 (R=2,6-xylyl, t-butyl) resemble that of
Co2(CO)8, with six terminal isocyanides and two bridg-
ing isocyanides [6,7].
Since the chemistry of Co2(CO)8 towards CNR is so
different compared to that of M2(CO)10 (M=Mn, Re),
it seemed interesting to investigate the reactions of
CNR with MnCo(CO)9 [8]. Especially, we were inter-
ested in whether substitution occurs preferentially on
the manganese site or on the cobalt site. Another point
of interest was whether CNR-substituted Mn–Co car-
bonyl complexes would contain any bridging ligands,
similar to the structure of Co2(CNR)8, or only terminal
Transition metal carbonyl complexes are of interest
as homogeneous catalysts [1]. Multimetallic complexes
should theoretically have advantages over mononuclear
complexes. For example, they have the ability to form
multicenter metal-to-ligand bonds to a substrate, thus
assisting in the activation of that species toward further
reactions. Multimetallic complexes also offer the possi-
bility for the selective and sequential activation of two
(or more) different substrate species [2].
Isocyanides, CNR, are isolobal with CO and often
are able to displace CO in transition metal complexes.
Isocyanides contain an organic group R, which can be
altered to fine tune the metal complex electronically and
sterically. Isocyanides are considered to be stronger
s-donors, and weaker p-acceptors than CO [3].
The reactions of Mn2(CO)10 [4], Re2(CO)10 [5] and
Co2(CO)8 [6] with CNR have previously been investi-
* Corresponding author. Tel.: +1-513-556 9200; fax: +1-513-556
9239.
1 Present address. Department of Chemistry, University of Michi-
gan, Ann Arbor, MI 48103-1055.
2 Present address. Molecular Simulations, Inc., 9685 Scranton
Road, San Diego, CA 92121-3752.
0020-1693/99/$ - see front matter © 1999 Elsevier Science S.A. All rights reserved.
PII: S0020-1693(99)00081-X

160
K. Beck et al. / Inorganica Chimica Acta 288 (1999) 159–173
anhydrous grade (Fisher) and used as received. All
other solvents were reagent grade and used as received.
Mn2(CO)10 and Co2(CO)8 were purchased from Strem.
Me3NO was purchased from Eastman, benzene-d6 from
Acros, acetone-d6 from Cambridge Isotope Laborato-
ries and CDCl3 from Aldrich. (2,6-xylyl)NC was pur-
chased from Fluka. (t-butyl)NC [12], (p-MeOC6H4)NC
[12], (2,6-diisopropylphenyl)NC [12], (n-butyl)NC [13],
Fig. 1. Structure of MnCo(CO)9.
ligands similar to M2(CO)10 and M2(CO)10−n(CNR)n
(n=1–4) (M=Mn, Re).
NaCo(CO)4
[14],
NaCo[(2,6-xylyl)NC]4
[15],
Mn(CO)5Br [16] and Mn(CO)5−n(CNR)nBr (n=1–4)
[17] were synthesized by published methods. Mn-
Co(CO)9 was prepared according to a literature method
[18], except that diethyl ether was used instead of THF
as the solvent. Silica gel 60, 230–400 mesh, from EM
science and Aluminum Oxide 90, active basic grade 1,
(alumina) were used for column chromatography.
Infrared spectra were recorded on a Perkin–Elmer
1600 Series FTIR spectrophotometer. IR cells were 1.0
mm solution cells with NaCl windows. For spectral
intensities the abbreviations vw=very weak, s=
MnCo(CO)9 is believed to have Cs symmetry (Fig. 1),
although no X-ray data have been reported for Mn-
Co(CO)9 or any unbridged derivative. The proposed
structure is based on the compound’s IR spectrum, and
the 55Mn and 59Co NMR. The local site symmetry is
C36 at the cobalt site, and C46 at the manganese site [9].
In contrast to Co2(CO)8 and M2(CO)10 (M=Mn,
Re), the metal–metal bond in MnCo(CO)9 is polarized,
with a partial negative charge on cobalt [9]. This in-
creases the electron density around Co in MnCo(CO)9
compared to Co2(CO)8. Consistent with this interpreta-
tion is the fact that the carbonyl stretching force con-
stants on the Co(CO)4 moiety in MnCo(CO)9 are
calculated to be lower than the values found in
Co2(CO)8 (non-bridged isomer), HCo(CO)4, and
CH3Co(CO)4, indicating increased back donation
[10,11].
This publication reports the synthesis and characteri-
zation of isocyanide-substituted derivatives of Mn-
Co(CO)9. The calculated IR spectra of MnCo(CO)9 and
of its CNR substituted derivatives using density func-
tional theory (DFT) methods are presented herein as
well. Furthermore, the crystal structures of Mn-
Co(CO)8((2,6-xylyl)NC) (1) and MnCo(CO)7((2,6-xy-
lyl)NC)2 (2) are reported.
1
strong, vs=very strong are used. H NMR data were
recorded using a Bruker AM-250 spectrometer. The
following abbreviations are used: m=multiplet, s=
singlet.
The complexes MnCo(CO)9−n(CNR)n (n=1, 2) are
temperature sensitive. They slowly decompose at
room temperature and must be stored in the freezer at
−25°C. Therefore satisfactory elemental analysis could
not be obtained for all complexes. Still, since the IR
1
and H NMR data of all MnCo(CO)8(CNR) complexes
1
are consistent with 1, and the IR and H NMR data of
all MnCo(CO)7(CNR)2 complexes are consistent with 2,
we feel confident that the synthesized complexes were
correctly identified as MnCo(CO)9−n(CNR)n (n=1, 2).
Elemental analyses were performed by Quantitative
Analysis, Whitehouse, NJ, and the results are presented
in Table 1.
2. Experimental
2.1. General experimental details, materials, and
2.2. Synthesis of MnCo(CO)9−n(CNR)n (n=1, 2)
analysis of the products
2.2.1. Procedure A (R=(p-MeOC6H4)NC,
All reactions were done under an argon atmosphere.
Tetrahydrofuran (THF), toluene and hexane were dis-
tilled from potassium benzophenone ketyl. Benzene and
CH2Cl2 were distilled from P2O5. Diethyl ether was
(2,6-xylyl)NC, (t-butyl)NC, (n-butyl)NC)
MnCo(CO)9 (0.15 mmol, 54.9 mg) was dissolved in
benzene (10 ml) in the presence of PdO (5% mol). The
appropriate number of equivalents of CNR, dissolved
Table 1
Elemental analysis
Calc. (found) C
Calc. (found) H
Calc. (found) N
MnCo(CO)7((2,6-xylyl)NC)2 (2)
52.47 (51.57)
40.80 (40.10)
48.02 (46.99)
68.77 (67.78)
3.15 (3.12)
1.50 (1.62)
3.26 (3.44)
5.37 (5.00)
4.89 (4.57)
2.97 (2.89)
2.67 (2.74)
8.30 (8.01)
MnCo(CO)8(p-MeOC6H4)NC (9)
MnCo(CO)8((2,6-diisopropylphenyl)NC) (8)
[Mn((2,6-xylyl)NC)6]+[Co(CO)4]− (5)

K. Beck et al. / Inorganica Chimica Acta 288 (1999) 159–173
161
in 3 ml of benzene, was then added dropwise. The
reaction mixture was stirred at room temperature and
monitored by IR. The reaction was terminated when
all of the MnCo(CO)9 was consumed. The solvent
was removed in vacuo. Column chromatography (alu-
mina, eluent pentane) gave the desired products. The
yields of the crude products were ꢀ25%. Multiple
recrystallizations from pentane at −70°C yielded
highly pure MnCo(CO)9−n(CNR)n as orange solids,
or dark red oils (n=1; R=(t-butyl)NC (3), (n-
butyl)NC) (4)).
described by DNP basis sets, comparable in quality
to Gaussian 6-31G** basis sets [19b]. The calculations
were carried out at different levels of theory: the local
density approximation by Vosko–Wilk–Nusair (VWN)
[20], the gradient-corrected exchange functional by
Becke (B) [21], gradient-corrected correlation func-
tionals of Perdew and Wang (PW) [22] and Lee et al.
(LYP) [23]. The COSMO model by Klamt et al. was
used to account for solvation effects [24]. The follow-
ing combinations of functionals were used: VWN,
BPW, BYLP, BVWN, VWN-COSMO (CCl4, cyclo-
hexane, diethyl ether, CHCl3) for Mn(CO)5Br; VWN
and VWN-COSMO (CCl4) for Mn2(CO)10, and
VWN-COSMO (CCl4) for all other compounds. Har-
monic vibrational frequencies are computed by diago-
nalizing the mass-weighted second-derivative matrix F.
2.2.2. Procedure B (R=(t-butyl)NC, (2,6-xylyl)NC,
(p-MeOC6H4)NC, (2,6-diisopropylphenyl)NC)
To a suspension of Mn(CO)5−n(CNR)nBr (1.00
mmol) in diethyl ether, freshly prepared NaCo(CO)4
(ꢀ1.5 mmol) in diethyl ether was added via cannula.
The reaction mixture was stirred overnight at room
temperature. The solvent was removed in vacuo and
the residue was chromatographed on alumina. Elution
with pentane gave the products in ꢀ30% yield. Mul-
tiple recrystallizations from pentane at −70°C
yielded highly pure MnCo(CO)9−n(CNR)n as orange
solids, or dark red oils (3 and 4).
All intensities are reported in units of km mol−1
.
2.6. X-ray diffraction study
For X-ray examination and data collection, a suit-
able crystal was mounted in a glass capillary. The
approximate dimensions were 0.21×0.20×0.18 mm
(1) and 0.45×0.20×0.25 mm (2).
Crystallographic data for 1 and 2 are presented in
Table 2.
2.3. Synthesis of MnCo(CO)9−n((2,6-xylyl)NC)n
(n=1,2)
Intensity data were collected at room temperature
on a Siemens SMART [25]3 CCD diffractometer
(platform goniostat with  fixed at 54.70°, sealed-tube
generator, graphite-monochromated Mo Ka radiation,
The reaction was run according to procedure B.
Instead of alumina, silica gel was used for column
chromatography. Elution with hexane/THF (5%) gave
two fractions in order of elution. 55.0 mg of 1 (6%)
and 43.0 mg of 2 (4%) were isolated as orange solids.
˚
u=0.71073 A). The detector was set at a calibrated
distance of 4.959 cm from the crystal. A series of 20-s
data frames measured at 0.3° increments of ꢀ were
collected with three different 2q and ƒ values to cal-
culate a preliminary unit cell.
2.4. Synthesis of [Mn((2,6-xylyl)NC)6]+[Co(CO)4]− (5)
For data collection 2082 data frames were mea-
sured at 0.3° intervals of ꢀ for a 20-s duration each.
In order to correct for high-energy backgrounds in
the images, data frames were collected as the sum of
two 10-s exposures and non-correlating events were
eliminated. The data frames were collected in four
distinct shells (2q= −30.0°, ꢀ= −30.0°, ƒ=0.0,
90.0, 180.0, 270.0°, x=54.70°), which combined mea-
sured nearly one sphere (completeness: 93.56% (1);
91.4% (2)) of intensity data (reflections: 10 523 (1);
13 696 (2)) with a maximum q value of 28.29°. The
initial 100 frames of the first data shell were recol-
lected at the end of the data collection to correct for
crystal decay (negligible amount in the case of both
samples).
To a solution of MnCo(CO)9 (50.0 mg; 0.137 mmol)
in benzene (10.0 ml), in the presence of PdO (5% mol),
11 equivalents of (2,6-xylyl)NC were added. The
reaction mixture was stirred at reflux temperature for
two hours. The white precipitate was filtered off and
washed with pentane. Recrystallization from CH2Cl2/
pentane yielded 81 mg (59%) of 5 as a white solid.
IR (CH2Cl2): w(CN) 2080 cm−1 (vs); w(CO) 1889
cm−1 (s). [17]: [Mn(2,6-xylyl-NC)][PF6]: IR (CH2Cl2):
1
2075 cm−1 (vs); 1996 cm−1 (vw) H NMR (CDCl3):
7.1 (m) C6H3 (3H), 2.44(s) Me (6H).
2.5. Computational details
Calculations were carried out by employing the
DMol program through the INSIGHT II 95 (Version
4.0.0) graphical interface, both by Molecular Simula-
tions [19a]. Geometries were optimized before vibra-
tional frequencies were calculated. Atomic orbitals were
3 The SMART and SAINT programs were used for data collection
and processing, respectively.

162
K. Beck et al. / Inorganica Chimica Acta 288 (1999) 159–173
form kFc[1+0.0001Fc u3/sin(2q)]−1/4 where k=
0.28435, =0.0021(11) was applied. Also for 1, hydro-
gen atom positions were located directly from the
difference map. One methyl group exhibits disordered
hydrogen positions and occupancies were fixed at 0.5. In
the case of 2, hydrogen atom positions were calculated
based on a geometric criterion for the electron density
map. All methyl hydrogen groups have disordered hy-
drogen positions and occupancies were fixed at 0.5.
All hydrogen atoms were allowed to ride on their
respective atoms. Hydrogen atom isotropic temperature
factors were refined as U(C)a=U(H) where a=1.5 for
methyl hydrogens and a=1.2 for the remaining hydro-
gens. The refinement converged with crystallographic
agreement factors of R1=5.63%, wR2=9.98% and S=
1.112 for 2373 reflections for 1, and R1=5.23%, wR2=
10.71% and S=1.065 for 4013 reflections for 2, with
I]2|(I) and 325 variable parameters. A final difference
Fourier map showed maximum residual electron density
2
The data frames were processed using the program
SAINT [25]. The data were corrected for decay, Lorentz
and polarization effects. A semi-empirical absorption
correction was applied using SADABS [26] (correction:
min. transmission 0.7357, max. transmission 0.9489 (1);
min. transmission 0.5718, max. transmission 0.9280
(2)).
The structures were solved and tables and figures
generated by a combination of direct methods using
SHELXTL v5.03 [27] and the difference Fourier technique
and refined by full-matrix least-squares on F2 for the
unique reflections (4661 (1); 6360 (2)) diffracting to 0.75
˚
A resolution. Non-hydrogen atoms were refined with
anisotropic displacement parameters. The phenyl moi-
eties exhibit large thermal motion associated with a
number of carbon atoms; however, possible dis-
order was not addressed. Weights were assigned as
2
2
w
−1=|2(Fo )+(aP)2+bP where P=0.33333Fo
+
0.66667Fc 2 and a=0.0450, b=0.0000 (1) or a=0.0428,
b=0.6186 (2). For 1 an extinction correction of the
−3
−3
˚
˚
of 0.304 e A
(1) and 0.522 e A
(2).
Table 2
Crystallographic data and structure refinement for 1 and 2
Empirical formula
Formula weight
Temperature (K)
C17H9CoMnNO8 (1)
469.12
297(2)
C25H18CoMnN2O7 (2)
572.28
273(2)
˚
Wavelength (A)
0.71073
0.71073
Crystal system
Space group
triclinic
triclinic
(
(
P1
P1
Unit cell dimensions
˚
a (A)
9.786(1)
8.618(1)
h (°)
b (A)
72.11(1)
10.669(1)
85.18(1)
10.743(1)
64.95(1)
965.6(2)
90.55(1)
9.374(1)
˚
i (°)
100.51(1)
17.425(1)
106.56(1)
1323.8(2)
2
1.436
1.148
580
˚
c (A)
k (°)
Volume
3
˚
Z (A )
2
Density (calc.) (Mg m−3
)
1.613
1.556
486
Absorption coefficient (mm−1
F(000)
)
Crystal size (mm)
q Range for data collection (°)
Limiting indices
0.21×0.20×0.18
2.21–28.29
−135h512
−145k514
−145l514
10523
0.45×0.25×0.20
2.50–28.29
−115h511
−125k512
−235l523
13696
Reflections collected
Independent reflections
Refinement method
Data/restraints/parameters
Goodness-of-fit on F2
Final R indices [I\2|(I)]
R1
4661 (Rint=0.0546)
6360 (Rint=0.0310)
full-matrix least-squares on F2
full-matrix least-squares on F2
4661/0/254
1.006
6360/0/325
1.065
0.0563
0.0998
0.0523
0.1071
wR2
R indices (all data)
R1
0.1382
0.0960
wR2
0.1269
0.304, −0.320
0.0021(11)
0.1280
0.522, −0.284
−3
˚
Largest difference peak, hole (e A
Extinction coefficient
)

K. Beck et al. / Inorganica Chimica Acta 288 (1999) 159–173
163
spectrum of the reaction mixture. Attempts to purify
the product were unsuccessful. A H NMR spectrum
3. Results and discussion
1
could not be obtained, due to paramagnetic impurities
which led to line broadening. In an attempt to achieve
CO substitution on the Mn site of [Mn(CO)3((n-
butyl)NC)3]+[Co(CO)4]− by [Co(CO)4]−, the reaction
mixture was irradiated with UV light for 2.5 h. Only
decomposition of the carbonyl complex could be de-
tected by IR.
All carbonyls in Co2(CO)8 can be substituted by
CNR [6]. In an attempt to achieve complete substitu-
tion in MnCo(CO)9, 11 equivalents of (t-butyl)NC or
(2,6-xylyl)NC were allowed to react with MnCo(CO)9
in benzene/PdO under reflux conditions. (2,6-xylyl)NC
yielded 5 as a white powder in high yield, whereas
(t-butyl)NC gave only unidentified decomposition
products.
3.1. Synthesis of MnCo(CO)9−n(CNR)n (n=1,2)
3.1.1. The compounds MnCo(CO)8(CNR) and
MnCo(CO)7(CNR)2
(R=p-MeOC6H4; 2,6-xylyl; t-butyl; n-butyl) were
obtained from the reaction of MnCo(CO)9 and CNR in
benzene (Eq. (1)), in the presence of PdO. PdO is
known to catalyze CO substitutions by CNR [4]. The
complexes are orange solids, except for MnCo-
(CO)8(CNR) (R=t-butyl (3), n-butyl (4)), which melt
at ꢀ15−20°C.
MnCo(CO)9+nCNR“(MnCo(CO)9−n(CNR)n+nCO
(n=1,2)
(1)
Other solvents did not give satisfactory results. In
CH2Cl2, toluene and acetonitrile metal–metal bond
cleavage occurred, indicated by the presence of a
strong, broad band at ꢀ1900 cm−1 in the IR spectrum
due to Co(CO)4−. Apparently these solvents promote
the nucleophilic substitution of Co(CO)4− by CNR. In
hexane, the reaction yielded an unidentified mixture of
compounds. In diethyl ether, using (p-MeOC6H4)NC,
To determine the substitution site of the CNR in the
complexes MnCo(CO)9−n(CNR)n, the following reac-
tion (Eq. (2)) was performed in diethyl ether:
Mn(CO)5−n(CNR)nBr+NaCo(CO)4
“(MnCo(CO)9−n(CNR)n+NaBr
(n=1–3) (R=t-butyl; 2,6-xylyl; p-MeOC6H4;
2,6-diisopropylphenyl)
(2)
only
a
very small amount of MnCo(CO)7((p-
MeOC6H4)NC)2 (6) could be isolated. It is somewhat
surprising that, out of the solvents tested, only benzene
gives a clean reaction; however, Coville et al. made
similar observations when they allowed Mn2(CO)10 to
react with CNR in the presence of PdO [4a].
For n=1,2 the IR spectra of MnCo(CO)9−n(CNR)n
were the same, independent of the synthetic pathway (A
or B). This shows that the substitution of CO by CNR
takes place at the Mn center. It is interesting to note
that substitution of CO in MnCo(CO)9 by PPh3 takes
place at the Co center. Furthermore, the following
reaction (Eq. (3)) did not yield the (PPh3)-substituted
MnCo carbonyl complex [18].
Me3NO is also known to promote the substitution of
CO by CNR in metal carbonyl complexes [28].
Nonetheless, all attempts to synthesize MnCo(CO)9−n
-
(CNR)n employing Me3NO failed. For R=(2,6-xy-
lyl)NC only 5 could be isolated, even after heating up
the reaction mixture to 50°C for 1 hour. In the case of
R=t-butyl an unidentified mixture of compounds, con-
taining Co(CO)4− (broad band at ꢀ1900 cm−1), could
be detected by IR. Attempts to separate this mixture
were unsuccessful. Apparently, terminal COs in Mn-
Co(CO)9 are not sufficiently electrophilic to be attacked
by Me3NO.
Attempts to synthesize complexes of the type Mn-
Co(CO)9−n(CNR)n with n\2 were equally unsuccess-
ful. The reaction of 3 or 4 equivalents of (2,6-xylyl)NC
with MnCo(CO)9 gave only the disubstituted product 2.
However, when 3 equivalents (n-butyl)NC were used,
cleavage of the metal–metal bond occurred even when
the reaction was run in benzene in the presence of PdO.
We suggest that [Mn(CO)3((n-butyl)NC)3]+[Co(CO)4]−
is the main product of the reaction. This conclusion is
based on the fact that three bands in the w(CN) (2222,
2188 and 2165 cm−1), three bands in the w(CO) region
(2033, 1973 and 1966 cm−1), and a strong, broad band
at 1890 cm−1, due to Co(CO)4−, are visible in the IR
Mn(CO)4(PPh3)Br+NaCo(CO)4
X
“(MnCo(CO)8(PPh3)+NaBr
(3)
Still, van Dijik et al. reported the successful synthesis
of (CO)4CoMn(CO)3(2,2%-bipyridine) via the reaction of
Mn(CO)3(2,2%-bipyridine)Br with NaCo(CO)4 [29].
In the reaction mixture of Eq. (2) only traces of
Mn2(CO)10, and no sign of the formation of Co2(CO)8
could be detected. This indicates that the reaction pro-
ceeds via nucleophilic substitution of Br− by Co(CO)4−
and not through radical intermediates.
When one equivalent of Mn(CO)2(CNR)3Br (R=(t-
butyl), (2,6-xylyl)) was allowed to react with 1.5 equiva-
lents of NaCo(CO)4, only MnCo(CO)7(CNR)2 could be
isolated. Coville et al. obtained similar results for
the reaction of M2(CO)10 (M=Mn,Re) with CNR
[4,5].
Reaction 2 gives better yields in general than Reac-
tion 1. Reaction 2 was run in THF and in diethyl ether.
The yields in diethyl ether were better than in THF,

164
K. Beck et al. / Inorganica Chimica Acta 288 (1999) 159–173
and the reaction time was much shorter. The precipita-
tion of NaBr, formed during the reaction, is pre-
sumably the main driving force. NaBr is seemingly less
soluble in diethyl ether than in THF, since THF is a
more polar solvent than diethyl ether. Therefore NaBr
presumably precipitates more rapidly and more com-
pletely in diethyl ether than in THF, speeding up the
reaction and increasing the yield.
The reaction of NaCo(CO)4 and Mn(CO)4((2,6-xy-
lyl)NC)Br not only yielded 1; surprisingly, disubstituted
2 was also isolated. The latter complex must have been
formed during workup, because its CO stretching bands
did not appear in the IR spectrum of the reaction
mixture. After testing different materials for column
chromatography, silica gel was selected for the workup
of this reaction. Apparently reaction occurred on the
column, since it got very hot. The color of the band
eluted with pentane changed from dark red to brown–
red upon elution. Very low yields of the mono- and
disubstituted complexes were isolated.
In an attempt to synthesize Mn–Co carbonyl com-
plexes CNR substituted at the Co site [Co((2,6-xylyl)-
NC)4]− was prepared from Co2((2,6-xylyl)NC)8
using the methods of Cooper and coworkers [15].
[Co((2,6-xylyl)NC)4]− was then allowed to react with
Mn(CO)5Br; the expected product was MnCo(CO)5-
((2,6-xylyl)NC)4. The IR spectrum of the reaction mix-
ture exhibited 13 mostly medium to weak bands in the
carbonyl region, indicating a mixture of compounds.
Attempts to separate the mixture proved to be unsuc-
cessful, and no product could be isolated.
tions agree well with the experiment. For bimetallic
complexes the number of IR bands predicted by group
theory and the number of bands observed experimen-
tally are often different. This is either due to the fact
that the bimetallic complexes deviate from the ideal
symmetry assumed in the predictions or due to overlap
of bands.
Density functional theory (DFT) has been used ex-
tensively to calculate vibrational frequencies of mono-
nuclear transition metal carbonyl complexes, because it
is computationally less expensive than ab initio meth-
ods [31]. For example, Fan and Ziegler calculated the
vibrational spectra of Ni(CO)4 and Cr(CO)6 [31a]. The
CO stretching frequencies were calculated by local den-
sity approximation (LDA) method and nonlocal (NL)
methods, including correction terms based on electron
density gradients. All calculated CO stretching frequen-
cies were lower than the observed frequencies by 12–53
cm−1. The CO stretching frequencies for Fe(CO)3(h4-
butadiene) and Fe(CO)3(h4-norbornadiene) calculated
by Thiel and Bu¨hl employing NL methods deviated
between 12 and 4 cm−1 from the experimental values
[31b]. Some of the calculated values were lower and
some were higher than the experimental values, with no
general pattern of deviation. Berce´s calculated the vi-
brational spectrum of W(CO)6 by LDA and NL meth-
ods [31c]. The calculated CO stretching frequencies
deviated between 2 and 83 cm−1 from the experimental
values and were, except for one value, too low. Jonas
and Thiel reported the calculated vibrational spectra
for several transition-metal hexacarbonyls, employing
NL methods [31d]. For Mn(CO)6+ and Re(CO)6+ the
calculated CO stretching frequencies were too low by
between 2 and 24 cm−1, compared to the experimental
values. In all cases, the numbers of bands were well
reproduced. For the calculations of Ni(CO)4, Cr(CO)6,
Mn(CO)6+ and Re(CO)6+ the relative IR intensities
tally with the experimental values [31a,d]. The relative
IR intensities for W(CO)6, Fe(CO)3(h4-butadiene), and
Fe(CO)3(h4-norbornadiene) were not reported [31b,c].
As a check on the structures of the parent and
substituted MnCo dimers, we used DMol software
module to calculate their IR spectra and compare with
experimental values. A convenient feature of the DMol
program is the graphical interface with Insight II,
whereby the calculated normal modes of vibration can
be visualized as arrow-style pictures. Mn(CO)5Br was
utilized to calibrate the program parameters, since this
molecule has been extensively studied by IR and since
the calculation of its IR spectrum was computationally
less expensive than for any binuclear complex. The
results are given in Table 4.
1
3.2. H NMR spectra
1
The H NMR spectra of the MnCo(CO)9−n(CNR)n
complexes agree with the proposed structures. The
chemical shifts of the coordinated isocyanide ligands
are listed in Table 3 and are only slightly different from
1
those of the uncooordinated ligands. However, the H
NMR signals of CH3–CH2–CH2–CH2–NC in com-
plex 4 are significantly shifted upfield (l= ꢀ1.7 ppm).
The 1H NMR spectra of 3 and MnCo(CO)7((t-
butyl)NC)2 (7) exhibit a singlet for the methyl protons,
unlike the free isocyanide, which exhibits a 1:1:1 triplet
due to nitrogen(14N)-hydrogen coupling (l=1.40; J=
2.1 Hz; acetone-d6) [30]. The fact that only one signal
for all methyl protons of (2,6-xylyl)NC and (t-butyl)NC
is seen in the spectrum, indicates that all protons are
equivalent; hence the isocyanide ligands rotate freely
around the R3C–NCMn bond.
3.3. IR spectra and computational studies
The frequencies calculated by gradient-corrected
Group theory is commonly employed to predict the
number of IR-active bands of molecules. For mononu-
clear transition metal carbonyl complexes the predic-
DFT methods were significantly lower (ꢀ75 cm−1
)
than the experimental values. However, the relative
intensities were reasonable. The calculations involving

K. Beck et al. / Inorganica Chimica Acta 288 (1999) 159–173
165
Table 3
1H NMR data
Compound
1H NMRa
Solvent
CDCl3
MnCo(CO)8(2,6-(xylyl)NC) (1)
7.12–7.22 (m) (C6H3)
2.45 (s) (Me)
MnCo(CO)8(t-(butyl)NC) (3)
MnCo(CO)8(n-(butyl)NC) (4)
1.53 (s) (Me)
CDCl3
−0.71 (m) (CH2)
−0.19 (m) (CH2)
−0.07 (m) (CH2)
3.64 (m) (Me)
acetone-d6
MnCo(CO)8(2,6-(diisopropyl)phenylNC) (8)
6.74–6.92 (m) (C6H3)
3.23 (septet) (CH)
1.09 (s) (Me)
benzene-d6
1.07 (s) (Me)
MnCo(CO)8(p-MeOC6H4NC) (9)
MnCo(CO)8(p-MeC6H4NC) (11)
6.64 (ABq, J=12 Hz) (C6H4)
6.22 (ABq, J=12 Hz) (C6H4)
3.00 (s) (Me)
benzene-d6
benzene-d6
6.63 (m) (C6H4)
6.48 (m) (C6H4)
1.78 (s) (Me)
MnCo(CO)7(2,6-(xylyl)NC)2 (2)
MnCo(CO)7(p-MeOC6H4NC)2 (6)
7.10–7.22 (m) (C6H3)
2.47 (s) (Me)
CDCl3
CDCl3
7.36 (ABq, J=12 Hz) (C6H4)
6.90(ABq, J=12 Hz) (C6H4)
3.84 (s) (Me)
MnCo(CO)8(t-(butyl)NC)2 (7)
MnCo(CO)7(n-(butyl)NC)2 (12)
1.53 (s) (Me)
CDCl3
0.98 (m) (CH2)
1.56 (m) (CH2)
1.80 (m) (CH2)
3.90 (m) (Me)
acetone-d6
MnCo(CO)7(2,6-(diisopropyl)phenylNC)2 (13)
[Mn(2,6-(xylyl)NC)6]+[Co(CO)4]− (5)
6.82–6.89 (m) (C6H3)
3.48 (septet) (CH)
1.15 (s) (Me)
benzene-d6
CD3CN
1.14 (s) (Me)
7.16 (s) (C6H3)
2.43 (s) (Me)
a Shifts reported in ppm downfield relative to TMS internal standard (CDCl3, CD3CN), or relative to the solvent peak of residual H
(benzene-d6, acetone-d6). Abbreviations: m, multiplet; s, singlet, q, quartet.
spectra are comparable to those in the other studies
reported in the literature.
VWN and COSMO gave the most satisfactory results.
However, the calculated frequencies for the A1 bands
were higher than observed while the E band was lower,
both in CCl4 and in CHCl3 solvents. To check the
parameters for a binuclear complex, the IR spectrum of
Mn2(CO)10 was calculated using VWN and COSMO.
The data are presented in Table 5, and the frequencies
show good agreement with the experimental values,
although they are all slightly too high (ꢀ30 cm−1).
The experimental CO stretching frequencies for
Mn2(CO)10 are on an average of 11 cm−1 higher in the
gas phase than in CCl4 solution. The DMol calculations
show a similar trend, as the CO stretching frequencies
in the VWN calculations are on an average of 9 cm−1
higher than in the VWN-COSMO (CCl4) calculations.
The discrepancies between calculated and observed
The IR spectrum of MnCo(CO)9 exhibits six bands in
the w(CO) region instead of the nine bands predicted by
group theory for Cs symmetry. The symmetry assign-
ments of the individual w(CO) for the experimental
values in Table 6 are taken from Ref. [18]. Employing
a so called ‘free rotational model’ [33] allows the bands
at 2025.8 and 1981.0 cm−1 to be labeled E modes. The
basic principle of this model is the accidental degener-
acy of all equatorial w(CO) force constants on each
separate metal, the vanishing of certain off-diagonal
elements of the F matrix and the fortuitous equality of
others. This approach reduces the number of predicted
w(CO) to seven (4A+B+2E), of which six (4A+2E)
are IR active.

166
K. Beck et al. / Inorganica Chimica Acta 288 (1999) 159–173
We wanted to test whether the experimental results
The IR spectrum was calculated for the Cs symmetry
agreed better with calculated results for the ‘correct’
predicted by Moosberry and Sheline [9] (Fig. 1), and
structure (Cs) than for the ‘wrong’ structure (C46) and
for a hypothetical C46 symmetry (Fig. 2).
Table 4
IR data for Mn(CO)5Bra,b
c
Band symmetry in C46
Experimental (CCl4/CHCl3)d
Calc. VWN
Calc. BPW
Calc. BYLP
A1
E
A1
2135/2138 (w)
2052/2054 (vs)
2001/2007 (m)
2151 (78)
2063 (1233)
2041 (680)
2095 (104)
2001 (1210)
1999 (647)
2067 (82)
1986 (1213)
1963 (659)
Calc. BVWN
Calc. VWN;COSMO
(CCl4)
(Cyclohexane)
(Diethyl ether)
(CHCl3)
mR=2.228
mR=2.015
mR=4.335
mR=4.806
2058 (84)
1978 (1213)
1953 (647)
2150 (142)
2029 (2696)
2007 (1496)
2152 (91)
2053 (1659)
2027 (931)
2152 (111)
2043 (2066)
2017 (1167)
2152 (115)
2041 (2116)
2019 (1191)
a Wavenumbers in cm−1
.
b Relative intensities in parentheses; w=weak, m=medium, vs=very strong. calc. intensities in km mol−1
.
c IR active CO stretching modes.
d This work.
Table 5
IR data for Mn2(CO)10
a,b
c
Band symmetry in D4d
Experimental (gas phase)d
Experimental (CCl4)e
Calc. VWN
Calc. VWN, COSMO (CCl4)
B2
E1
B2
2053.3
2026.1
1994.3
2044 (s)
2013 (vs)
1983 (s)
2078 (1262)
2041 (2092)
2026 (713)
2076 (1358)
2031 (2945)
2012 (1335)
a Wavenumbers in cm−1
.
b Relative intensities in parentheses; s=strong, vs=very strong. calc. intensities in km mol−1
.
c IR active CO stretching modes.
d Ref. [32].
e This work.
Table 6
IR data for MnCo(CO)9
a,b
Experimental (hexane)c/(CCl4)d
Calc. VWN, Cs-symmetry
Calc. VWN, C46-symmetry
COSMO (CCl4)
COSMO (CCl4)
A
A
B
E
E
A
A
E
E
2116.5 (0.004)/2117 (w)
2056.2 (0.347)/2056 (vs)
2045e
2142 (42)
2085 (1522)
2060 (34)
2042 (2371)
2040 (2415)
2033 (192)
2021 (878)
2011 (87)
A1
A1
B1
E
2148 (87)
2068 (1401)
2065 (10)
w1
w3
2025.8 (1.0)/2024 (vs)
2052 (2135)
w5
w5
w2
w4
w6
w6
2004.0 (0.004)/2004 (sh,w)
1996.0 (0.154)/1995 (m)
1981.0 (0.111)/1977 (m)
B1
A1
E
2019 (1)
2000 (648)
1995 (841)
2005 (196)
a Wavenumbers in cm−1
.
b Relative intensities in parentheses, sh=shoulder, m=medium, s=strong, vs=very strong. calc. intensities in km/mol.
c Ref. [18].
d This work.
e Ref. [10], calculated.

K. Beck et al. / Inorganica Chimica Acta 288 (1999) 159–173
167
Fig. 2. Structure of MnCo(CO)9 with hypothetical symmetry C46
.
Fig. 3. Possible structures for (CNR)(CO)4MnCo(CO)4, (CO ligands
omitted for clarity).
thus to test the ability of the calculations as a predictor
of structure. The results are given in Table 6.
puted from a complete set of force constants a B mode
localized on the Mn moiety to be at 2045 cm−1 [10].
For Cs, the band calculated at 2060 cm−1 is a B mode
localized on the Mn moiety, whereas for C46 symmetry
the B mode localized on the Mn moiety was calculated
to be at 2019 cm−1. The band calculated at 2065 cm−1
for C46 is a B mode localized on the Co moiety.
Keeping in mind that all calculated w(CO) values are
too high, the results for the Mn-localized B mode in Cs
symmetry tally better with Sbrignadello’s interpretation
than the results obtained for C46 symmetry.
Based on the assignments of the CO stretching fre-
quencies, it is reasonable to assume that for the calcu-
lated spectrum in Cs symmetry, the bands at 2042 and
2040 cm−1, as well as the bands at 2011 and 2005
cm−1, could overlap in the experimental spectrum. The
band calculated at 2060 (Cs) has an intensity of 34 and
may therefore be invisible in the experimental spectrum.
Hence, only six bands might be expected to be visible in
the w(CO) region.
All calculated wavenumbers are too high (ꢀ20–30
cm−1). The band patterns for both calculated spectra
were very similar. Bor et al. based their assignment of
the w(CO) of MnCo(CO)9 on a comparison of the IR
spectrum of MnCo(CO)9 to the IR spectra of
Mn2(CO)10 (D4d) and HgCo2(CO)8 (D3d) [18]. Assign-
ments of w(CO) for the spectra calculated in this work
were made based on the graphical display for each
normal mode of vibration. The assignments of the
w(CO) for both calculated spectra were then compared
to Bor’s assignments.
The assignments for the Cs-structure calculations
tally with Bor’s assignments, while the assignments for
C46-structure calculations differ from Bor’s assign-
ments. Frequency w2 for C46 is a B mode localized on
the Mn moiety, and not an A mode, corresponding to
the symmetric stretch of the three equatorial COs on
Co, as in Bor’s assignment. Also, Sbrignadello com-
For C46, the intensities of the bands at 2065 and 2019
cm−1 are 10 and 1, respectively, making them likely to
be invisible in the IR spectrum. Consequently only five
bands might be expected to be visible in the IR spec-
trum. Hence, the DMol calculations are consistent with
the Cs symmetry for MnCo(CO)9 suggested by Moos-
berry and Sheline [9].
The calculations suggest that accidental band overlap
is the reason only six bands are observed in the IR
spectrum of MnCo(CO)9. Therefore the ‘free rotational
model’ of Bor et al. is not the only possible explanation
of the compound’s IR spectrum.
The IR spectra of MnCo(CO)9−n(CNCH3)n (n=0−
2) were calculated using VWN as functionals, and
employing the COSMO (CCl4) model. The IR spectra
of the complexes MnCo(CO)8(CNR) (R=2,6-xylyl (1),
t-butyl (3), n-butyl (4), 2,6-diisopropylphenyl (8), p-
MeOC6H4 (9), p-ClC6H4 (10), p-MeC6H4 (11)) are
listed in Table 7.
All complexes show one band (w1) due to coordinated
CNR and 6 (3 and 4) or 7 (1, 8, 9, 10 and 11) bands
due to terminally coordinated CO. The reason why all
alkyl isocyanide containing complexes show one w(CO)
band fewer in the IR spectrum than all aryl isocyanide
containing complexes is probably overlap of w4 and w5.
The spectra show no evidence of any bridging ligands.
Group theory predicts 8 IR active w(CO), regardless of
Table 7
IR data for MnCo(CO)8(CNR)a,b,c
R=2,6-xylyl
(1)
R=t-butyl
(3)
R=n-butyl
(4)
R=2,6-
diisopropyl
(8)
2153 (m)
2080 (m)
2042 (s)
2015 (vs)
2010 (vs)
1974 (s)
2175 (m)
2084 (m)
2040 (s)
2189 (m)
2086 (m)
2041 (s)
2154 (m)
2080 (m)
2042 (s)
2016 (vs)
2009 (vs)
1974 (s)
w1
w2
w3
w4
w5
w6
w7
w8
2009 (vs)
1970 (s)
2010 (vs)
1972 (s)
1960 (m)
1935 (vw)
1956 (sh,m)
1930 (vw)
1952 (m)
1942 (vw)
1953 (m)
1935 (vw)
R=p-MeOC6H4 (9) R=p-ClC6H4 (10)
R=p-MeC6H4 (11)
2159 (m)
2082 (m)
2042 (s)
2015 (vs)
2012 (vs)
1974 (s)
2153 (m)
2080 (m)
2043 (s)
2013 (vs)
2000 (vs)
1976 (s)
2158 (m)
2082 (m)
2042 (s)
2011 (vs)
1996 (vs)
1974 (s)
w1
w2
w3
w4
w5
w6
w7
w8
1958 (m)
1937 (vw)
1962 (m)
1936 (vw)
1961 (m)
1935 (vw)
a Wavenumbers in cm−1
.
b Relative intensities in parentheses; vw=very weak, m=medium,
s=strong, vs=very strong.
c In pentane.

168
K. Beck et al. / Inorganica Chimica Acta 288 (1999) 159–173
In order to compare the calculated spectra with the
experimental spectra, the average deviations of all com-
puted frequencies from the experimental spectrum of 9
were calculated. w(CN) was not included in the calcula-
tion of the average deviation because its energy greatly
depends on the nature of R. Since CH3 was chosen as
R for the calculations, w(CN) is expected to be calcu-
lated at higher wavenumbers, than for any other R
group [4].
Fig. 4. Possible structures for (CO)5MnCo(CNR)(CO)3, (CO ligands
omitted for clarity).
The average deviations, given in Table 8, show that
structures A and B have similar deviations from the
experimental values, whereas structures C and D show
similar but significantly higher deviations. Therefore,
the calculations indicate that structures C and D are
unlikely structures for MnCo(CO)8(CNR).
whether the CNR ligand is in equatorial (A, Fig. 3) or
axial position (B, Fig. 3) on Mn.
Group theory also predicts eight bands due to w(CO)
in the IR spectrum for (CO)5MnCo(CO)3(CNR), CNR
in equatorial (C, Fig. 4) or axial position (D, Fig. 4).
However, the band pattern should be different for
each structure, and ideally only one should resemble the
pattern of the experimental spectrum. Therefore we
calculated the IR spectra for structures A, B, C and D
to determine whether calculations lead to a unique
assignment of structure. IR spectra for A, B, C and D
were calculated with R=CH3 to minimize computa-
tional effort. The calculated spectra were compared
with the measured spectrum of 9. An aryl isocyanide
complex was chosen as comparison because of the good
band resolution. The absolute values of the calculated
w(CO), given in Table 8, are all too high (ꢀ30 cm−1).
For structure A, the band calculated at 2016 cm−1
(w6), which corresponds to an A1g mode in point group
D3d, is probably not resolved in the experimental spec-
trum. Its intensity is small compared to the intensity of
the 2024 cm−1 band, so that it may be hidden in the
tail of the 2024 cm−1 band. Hence seven bands are
expected to be observed in the w(CO) region. For
structure B w6 was calculated at 2041 cm−1. Again,
keeping in mind that all calculated wavenumbers are
too high by 20–30 cm−1, one would expect to see this
band at ꢀ2010–2020 cm−1 in the experimental spec-
Table 8
Comparison of experimental and calculated IR data for MnCo(CO)8(CNR)a,b
R=p-MeOC6H4 (9) experimental
R=CH3, calc. VWN COSMO
R=CH3, calc. VWN COSMO
(pentane)
(CCl4) equatorial, A
(CCl4) axial, B
2159 (m)
2082 (m)
2042 (s)
2015 (vs)
2012 (vs)
2258 (529)
2110 (332)
2068 (1650)
2036 (1179)
2024 (2524)
2016 (92)
2000 (1246)
1993 (267)
1990 (221)
2246 (1161)
2108 (382)
2065 (840)
2019 (2728)
2017 (2311)
2041 (78)
2012 (553)
1988 (23)
1984 (69)
w1
w2
w3
w4
w5
w6
w7
w8
w9
1974 (s)
1958 (m)
1937 (vw)
Average deviation
23
22
R=CH3, calc. VWN COSMO
R=CH3, calc. VWN COSMO
(CCl4) equatorial, C
(CCl4) axial, D
2222 (649)
2119 (598)
2061 (1666)
2048 (521)
2029 (1149)
2019 (1796)
2012 (620)
2010 (974)
1989 (181)
2264 (955)
2122 (520)
2032 (850)
2027 (2246)
2025 (2225)
2045 (43)
2011 (689)
1989 (214)
1982 (358)
w1
w2
w3
w4
w5
w6
w7
w8
w9
Average deviation
a Wavenumbers in cm−1
31
36
.
b Relative intensities in parentheses; vw=very weak, m=medium, s=strong, vs=very strong. Calculated intensities in km mol−1
.

K. Beck et al. / Inorganica Chimica Acta 288 (1999) 159–173
169
Table 9
a,b,c
IR data for MnCo(CO)7(CNR)2
R=2,6-xylyl (2)
R=p-MeOC6H4 (6)
R=t-butyl (7)
R=n-butyl (12)
R=2,6-diisopropyl (13)
2152 (w)
2119 (m)
2106 (m)
2050 (s)
2027 (m)
2018 (vs)
1994 (vs)
1955 (s)
2159 (w)
2129 (m)
2115 (m)
2050 (s)
2028 (m)
2018 (vs)
1994 (vs)
1954 (s)
2172 (w)
2148 (m)
2133 (m)
2050 (s)
2023 (m)
2015 (vs)
1988 (vs)
2185 (w)
2162 (m)
2148 (m)
2053 (s)
2025 (m)
2017 (vs)
1989 (vs)
1957 (s)
1949 (s)
2153 (w)
2119 (m)
2108 (m)
2049 (s)
2028 (m)
2018 (vs)
1993 (vs)
1954 (s)
w1
w2
w3
w4
w5
w6
w7
w8
w9
1947 (s)
a In pentane.
b Wavenumbers in cm−1
.
c Relative intensities in parentheses; vw=very weak, w=weak, m=medium, s=strong, vs=very strong.
trum. This band would very likely be obscured in the
experimental spectrum by w4 and w5. Additionally the
band calculated at 1988 cm−1 (w8) has a very low
intensity, making it likely to be invisible in the IR
spectrum. Therefore, only 6 bands would be visible in
the w(CO) region.
that all of the samples of MnCo(CO)7(CNR)2 are con-
taminated with MnCo(CO)8(CNR). The synthesis of 1
has shown that the Mn–CNR bond is rather labile; still
it seems unlikely that an isomer with the CNR group in
axial position exists for MnCo(CO)7(CNR)2, but not
for MnCo(CO)8(CNR). Alternatively, one of the high-
energy bands might be a combination band.
The complexes show six (12) or five bands (2, 6, 7,
and 13) due to terminal CO stretching in the IR spec-
trum, instead of the seven bands group theory predicts.
The IR spectrum of MnCo(CO)7(CNCH3)2 was calcu-
lated for two possible structural isomers, eq,eq-cis-
Overall, the calculated spectrum for structure A tal-
lies best with the experimental spectrum of 9. The
DMol calculations lead to a unique assignment of the
structure for 9; this agrees with the structure of 1
determined by X-ray diffraction. For Mn-
Co(CO)8(CNR), w(CN) is shifted to higher wavenum-
bers compared to Mn2(CO)9(CNR) (R=t-butyl,
2,6-xylyl). The more positive the oxidation state of a
metal, the higher w(CN) is [34]. Since the Co(CO)4
moiety withdraws electron density from the Mn moiety,
w(CN) is expected to be at higher wavenumbers. Also,
the highest totally symmetric w(CO) (w2), symmetric on
Mn and Co, is shifted to lower wavenumbers, com-
pared to MnCo(CO)9. This is due to the fact that CNR
is a better overall electron-donor than CO. More elec-
tron density around the Mn center strengthens the
Mn–C bond; hence the CO bond is weakened. As a
result w(CO) is lowered.
(CNCH3)2(CO)3MnCo(CO)4
(E,
Fig.
5)
and
ax,eq-(CNCH3)2(CO)3MnCo(CO)4 (F, Fig. 5).
The results of the calculations are given in Table 10.
The results were compared to the IR spectrum of 12,
because the IR spectrum of this complex showed the
best resolution of w(CO). In all other IR spectra of
MnCo(CO)7(CNR)2 complexes w8 and w9 (Table 9)
probably overlap. For reasons mentioned above, w(CN)
were not included in the calculation of the average
deviation of the calculated wavenumbers from the ex-
perimentally obtained wavenumbers. All calculated
band positions are too high, except for w6. The average
deviation for structures E and F is 22 and 17 cm−1
,
The IR spectra of MnCo(CO)7(CNR)2 (R=2,6-xylyl
(2), p-MeOC6H4 (6), t-butyl (7), n-butyl (12), 2,6-diiso-
propylphenyl (13)) are listed in Table 9.
respectively. So, no choice between the structures is
possible on this basis. Both calculated spectra exhibit
an additional band (w10), not seen in the experimental
spectrum. w10 has a weak intensity for E (145), and a
very strong intensity for F (1368). For structure E, w10
might be obscured in the experimental spectra, since it
Surprisingly, the spectra exhibit two medium and one
weak absorptions due to the two terminal CNR groups
(\2100 cm−1). A possible explanation is that in solu-
tion different isomers exist. The two CNR ligands in
the second isomer could be in equatorial-trans position
to each other or in axial-equatorial position. The first
structure would make two trans CO ligands compete
for electrons, which is unlikely. The band position of
the highest w(CN) in the IR spectra of Mn-
Co(CO)7(CNR)2 is very close to w(CN) of the corre-
sponding MnCo(CO)8(CNR); however repeated
recystallization of MnCo(CO)7(CNR)2 did not change
the relative band intensities. Therefore it seems unlikely
Fig. 5. Possible structures for (CNR)2(CO)3MnCo(CO)4, (CO ligands
omitted for clarity).

170
K. Beck et al. / Inorganica Chimica Acta 288 (1999) 159–173
Table 10
Table 12
Comparison of experimental and calculated IR data for Mn-
Co(CO)7(CNR)2
Bond angles (°) for 1a
a,b
C(5)–Mn(1)–C(6)
C(6)–Mn(1)–C(8)
C(6)–Mn(1)–C(7)
C(5)–Mn(1)–C(9)
C(8)–Mn(1)–C(9)
91.6(2)
91.6(2)
92.0(2)
88.8(2)
86.7(2)
C(5)–Mn(1)–C(8)
C(5)–Mn(1)–C(7)
C(8)–Mn(1)–C(7)
C(6)–Mn(1)–C(9)
C(7)–Mn(1)–C(9)
92.4(2)
93.6(2)
172.9(2)
178.2(2)
89.8(2)
82.17(13)
82.56(13)
95.1(2)
96.9(2)
97.4(2)
82.7(2)
81.7(2)
176.1(4)
174.1(4)
177.9(4)
177.0(4)
176.4(4)
Experimental
R=n-butyl (12) calc. VWN,
R=CH3,
R=CH3,
calc. VWN,
COSMO (CCl4); E COSMO (CCl4); F
w1
w2
w3
w4
w5
w10
w6
w7
w8
w9
2185 (w)
2162 (m)
2148 (m)
2053 (s)
2025 (m)
2251 (511)
2222 (674)
2260 (1047)
2221 (930)
C(5)–Mn(1)–Co(1) 172.52(13) C(6)–Mn(1)–Co(1)
C(8)–Mn(1)–Co(1)
C(9)–Mn(1)–Co(1)
C(3)–Co(1)–C(2)
C(3)–Co(1)–C(4)
C(2)–Co(1)–C(4)
91.89(14) C(7)–Mn(1)–Co(1)
97.60(12) C(3)–Co(1)–C(1)
125.2(3)
115.4(2)
115.6(2)
2084 (683)
2051 (1676)
2023 (145)
2009 (1216)
1997 (492)
1992 (325)
1975 (390)
2082 (1081)
2044 (899)
2018 (1368)
2000 (319)
1995 (1760)
1975 (1157)
1962 (239)
C(1)–Co(1)–C(2)
C(1)–Co(1)–C(4)
C(3)–Co(1)–Mn(1)
C(2)–Co(1)–Mn(1)
2017 (vs)
1989 (vs)
1957 (s)
1949 (s)
C(1)–Co(1)–Mn(1) 176.0(2)
C(4)–Co(1)–Mn(1)
O(1)–C(1)–Co(1)
O(3)–C(3)–Co(1)
O(5)–C(5)–Mn(1)
O(7)–C(7)–Mn(1)
O(9)–C(9)–Mn(1)
C(11)–C(10)–N(1)
C(10)–C(11)–C(12) 116.7(4)
C(12)–C(11)–C(17) 121.6(4)
C(12)–C(13)–C(14) 120.8(4)
C(10)–C(15)–C(14) 116.7(4)
C(14)–C(15)–C(16) 121.9(4)
86.59(13) C(9)–N(1)–C(10)
178.9(5)
174.0(5)
178.9(5)
178.7(4)
172.5(4)
118.2(4)
O(2)–C(2)–Co(1)
O(4)–C(4)–Co(1)
O(6)–C(6)–Mn(1)
O(8)–C(8)–Mn(1)
C(11)–C(10)–C(15) 123.7(4)
C(15)–C(10)–N(1) 118.0(4)
C(10)–C(11)–C(17) 121.7(4)
C(13)–C(12)–C(11) 121.2(5)
C(13)–C(14)–C(15) 120.9(4)
C(10)–C(15)–C(16) 121.4(4)
Average
deviation
22
17
a Wavenumbers in cm−1
b Relative intensities in parentheses; w=weak, m=medium, s=
strong, vs=very strong. calc.intensities in km mol−1
.
.
a Estimated standard deviations in the least significant figure are
given in parentheses.
lies on the tail of a very strong band. For structure F
the calculated intensity of w10 makes it impossible for
w10 to be obscured in the experimental spectra. Hence
six w(CO) bands are expected for structure E, whereas
seven w(CO) bands are expected for structure F.
A unique assignment of the structure based on the
DMol calculations, which was possible in the case of
MnCo(CO)8(CNR), cannot be made for Mn-
Co(CO)7(CNR)2. As for the number of w(CO) bands,
the spectrum calculated for E tallies better with the
spectrum of 12; but the absolute values of wavenum-
bers of w(CO) for F tally a little better with the experi-
mental values.
3.4. Structures of MnCo(CO)8(2,6-xylyl)NC (1) and
MnCo(CO)7((2,6-xylyl)NC)2 (2)
Crystals of 1 (orange–red) and 2 (dark red), suitable
for an X-ray crystallographic study, were obtained
from pentane at −25°C. Bond angles and bond lengths
for 1 and 2 are given in Tables 11–14.
Table 13
Bond lengths (A) for 2
a
˚
Co(1)–C(6)
Co(1)–C(4)
Co(1)–Mn(1)
Mn(1)–C(2)
Mn(1)–C(8)
O(1)–C(1)
O(3)–C(5)
O(5)–C(5)
O(7)–C(7)
N(1)–C(9)
N(2)–C(18)
C(9)–C(14)
C(10)–C(16)
C(12)–C(13)
C(14)–C(15)
C(18)–C(23)
C(19)–C(25)
C(21)–C(22)
C(23)–C(24)
1.769(5)
1.774(4)
2.9035(7)
1.838(4)
1.924(4)
1.148(4)
1.138(5)
1.143(4)
1.138(5)
1.413(5)
1.402(4)
1.425(6)
1.485(7)
1.352(8)
1.485(7)
1.397(5)
1.499(5)
1.362(6)
1.499(5)
Co(1)–C(5)
Co(1)–C(7)
Mn(1)–C(1)
Mn(1)–C(3)
Mn(1)–C(17)
O(2)–C(2)
O(4)–C(4)
O(6)–C(6)
N(1)–C(8)
N(2)–C(17)
C(9)–C(10)
C(10)–C(11)
C(11)–C(12)
C(13)–C(14)
C(18)–C(19)
C(19)–C(20)
C(20)–C(21)
C(22)–C(23)
1.770(4)
1.800(5)
1.785(4)
1.841(4)
1.941(3)
1.139(4)
1.151(4)
1.141(5)
1.160(4)
1.153(4)
1.347(6)
1.449(7)
1.449(7)
1.379(6)
1.394(5)
1.400(5)
1.368(6)
1.379(5)
Table 11
Bond lengths (A) for 1
a
˚
Mn(1)–C(5)
Mn(1)–C(8)
Mn(1)–C(9)
Co(1)–C(3)
Co(1)–C(2)
O(1)–C(1)
O(3)–C(3)
O(5)–C(5)
O(7)–C(7)
N(1)–C(9)
1.775(5)
1.845(5)
1.936(4)
1.778(7)
1.782(6)
1.131(5)
1.140(6)
1.152(5)
1.142(5)
1.157(4)
1.385(5)
1.390(5)
1.363(6)
1.392(6)
Mn(1)–C(6)
Mn(1)–C(7)
Mn(1)–Co(1)
Co(1)–C(1)
Co(1)–C(4)
O(2)–C(2)
O(4)–C(4)
O(6)–C(6)
O(8)–C(8)
N(1)–C(10)
C(10)–C(15)
C(11)–C(17)
C(13)–C(14)
C(15)–C(16)
1.840(4)
1.854(5)
2.870(1)
1.780(5)
1.797(5)
1.142(6)
1.140(5)
1.138(4)
1.149(5)
1.404(5)
1.387(5)
1.503(6)
1.367(6)
1.513(6)
C(10)–C(11)
C(11)–C(12)
C(12)–C(13)
C(14)–C(15)
a Estimated standard deviations in the least significant figure are
given in parentheses.
a Estimated standard deviations in the least significant figure are
given in parentheses.

K. Beck et al. / Inorganica Chimica Acta 288 (1999) 159–173
171
Table 14
tion geometry around the Co center is trigonal–bipyra-
midal, and octahedral around the Mn center. Both
complexes contain a single metal–metal bond and no
bridging ligands. The overall symmetry is C1 for 1 and
Cs for 2. To the best of our knowledge, these are the
first crystal structures reported for an unbridged Mn–
Co carbonyl complex. The isocyanide ligand on Mn in
1 is in an equatorial position. The two isocyanides on
Mn in complex 2 are in cis-equatorial position to each
other. This is similar to those in 1,1-Mn2(CO)8((t-
butyl)NC)2 [35] and Re2(CO)10−n(CNR)n (n=1, R=t-
butyl; n=2, R=2,6-xylyl) [5b]. The probable reason
for this is that CO is a very strong p-acceptor. There-
fore two CO ligands trans to each other compete for
electron density, weakening the metal–C bond. CNR is
a considerably weaker p-acceptor; hence it is less elec-
tron-withdrawing than a CO from the opposite CO
group, stabilizing the metal–carbonyl bond.
Bond angles (°) for 2a
C(6)–Co(1)–C(5)
C(5)–Co(1)–C(4)
C(5)–Co(1)–C(7)
94.7(2)
C(6)–Co(1)–C(4)
C(6)–Co(1)–C(7)
C(4)–Co(1)–C(7)
C(5)–Co(1)–Mn(1)
96.6(2)
99.8(2)
114.6(2)
78.68(12)
90.46(13)
90.5(2)
126.5(2)
114.4(2)
C(6)–Co(1)–Mn(1) 169.5(2)
C(4)–Co(1)–Mn(1)
C(1)–Mn(1)–C(2)
C(2)–Mn(1)–C(3)
C(2)–Mn(1)–C(8)
C(1)–Mn(1)–C(17)
C(3)–Mn(1)–C(17)
C(1)–Mn(1)–Co(1) 170.03(12) C(2)–Mn(1)–Co(1)
C(3)–Mn(1)–Co(1) 82.84(12) C(8)–Mn(1)–Co(1)
C(17)–Mn(1)–Co(1) 96.17(9)
C(17)–N(2)–C(18)
O(2)–C(2)–Mn(1)
O(4)–C(4)–Co(1)
O(6)–C(6)–Co(1)
N(1)–C(8)–Mn(1)
C(10)–C(9)–C(14)
C(9)–C(10)–C(11)
C(11)–C(10)–C(16) 119.9(5)
C(13)–C(12)–C(11) 126.6(6)
C(13)–C(14)–C(9)
C(9)–C(14)–C(15)
C(19)–C(18)–C(23) 123.9(3)
C(23)–C(18)–N(2) 117.2(3)
C(18)–C(19)–C(25) 121.8(3)
C(21)–C(20)–C(19) 121.6(4)
C(21)–C(22)–C(23) 121.3(4)
C(22)–C(23)–C(24) 121.3(4)
80.91(12) C(7)–Co(1)–Mn(1)
92.8(2)
94.5(2)
89.9(2)
91.3(2)
C(1)–Mn(1)–C(3)
C(1)–Mn(1)–C(8)
C(3)–Mn(1)–C(8)
90.1(2)
175.56(14)
C(2)–Mn(1)–C(17) 173.0(2)
91.14(14) C(8)–Mn(1)–C(17)
84.44(13)
80.42(12)
97.13(10)
177.8(3)
178.9(4)
176.7(3)
173.0(3)
177.5(4)
118.5(4)
117.2(4)
121.6(4)
C(8)–N(1)–C(9)
O(1)–C(1)–Mn(1)
O(3)–C(3)–Mn(1)
O(5)–C(5)–Co(1)
O(7)–C(7)–Co(1)
C(10)–C(9)–N(1)
N(1)–C(9)–C(14)
C(9)–C(10)–C(16)
C(12)–C(11)–C(10) 115.1(6)
C(12)–C(13)–C(14) 119.3(6)
C(13)–C(14)–C(15) 122.2(5)
N(2)–C(17)–Mn(1) 174.6(3)
176.7(3)
177.2(3)
173.5(3)
179.4(5)
173.9(3)
124.3(4)
118.5(5)
˚
The Mn–Co bond lengths of 2.870(1) A in 1 and
˚
2.9035(7) A in 2 are similar to the Mn–Mn bond length
˚
in Mn2(CO)10 (2.895(1) A) [36] and 1,1-Mn2(CO)8((t-
116.1(5)
121.7(4)
˚
butyl)NC)2 (2.924(1) A) [35]. In contrast, the Co–Co
˚
bond length in Co2(CO)8 (2.528(1) A) [37] and in
C(19)–C(18)–N(2)
118.9(3)
˚
Co2((t-butyl)NC)8 (2.4557 A) [38] is much shorter. This
C(18)–C(19)–C(20) 115.5(4)
C(20)–C(19)–C(25) 122.7(4)
C(22)–C(21)–C(20) 120.7(4)
C(22)–C(23)–C(18) 116.9(3)
C(18)–C(23)–C(24) 121.8(3)
is partly due to the two bridging ligands, but even for
the ‘high temperature’ isomer of Co2(CO)8,
(CO)3CoCo(CO)3 with six terminal carbonyls, the Co–
˚
Co bond length was calculated to be 2.634 A [39].
a Estimated standard deviations in the least significant figure are
given in parentheses.
Although the bond length of MnCo(CO)9 is unknown,
it can be assumed that the substitution of two CO by
two CNR ligands does not increase the bond length
dramatically. For example, the difference in the Re–Re
bond length between Re2(CO)10 and Re2(CO)8((2,6-xy-
ORTEP drawings for 1 and 2 are presented in Figs. 6
and 7, respectively. For both complexes the coordina-
Fig. 6. ORTEP drawing of complex 1.
Fig. 7. ORTEP drawing of complex 2.

172
K. Beck et al. / Inorganica Chimica Acta 288 (1999) 159–173
˚
lyl)NC)2 is only 0.007 A [5b]. Mn–Co bond lengths
reported in the literature are only for molecules con-
taining bridging ligands. For example [CoMn(CO)6-
to a staggered geometry. Although in Co2(CO)8 the
Co–Cterminal–O angles are in fact closer to 180°
(174.9(1)–178.9(1)°) [37], supporting this reasoning,
the Co–Ceq–O angles in HCo(CO)4 are even smaller
(172.6(1.2)°) than in 2 [44]. However, the COeq lig-
ands in HCo(CO)4 are bent towards the fifth ligand
H, and not away as in 2. In compound 1 the Co–
Ceq–O angle opposite the CNR ligand on Mn is
closer to 180° (177.9(4)°) than the other two Co–
Ceq–O angles (174.1(4) and 174.0(5)°), although the
Mn–Co bond length in 1 is shorter than in 2. There-
fore steric interactions seem unlikely to be the cause
for the Co–Ceq–O angles in 1 and 2 to be unequal.
It remains unclear why in complexes 1 and 2 two of
the three equatorial carbonyls on Co are more bent.
(Ph2C4Me2)] has a Mn–Co bond length of 2.5488(8)
˚
A
[40], [s2-N, s2-N%, h2-CꢀN-[glyoxalbis(t-butyl-
imine)]]hexacarbonyl cobalt manganese has a Mn–Co
2
˚
bond length of 2.639(3) A [41], and MnCo(CO)3(m -
CO)2(m2-Ph2PCH2PPh2)2 has a Mn–Co bond length
˚
of 2.726 A [42].
Surprisingly the CNR group in 1 and both CNR
groups in 2 are in eclipsed position relative to CO
groups on Co. This indicates very little steric interac-
tion between CNR on Mn and CO on Co.
The Mn–CO distances in 1 [Mn–COax: 1.775(5) A;
Mn–COeq: 1.854(5) A, 1.845(4) A, 1.840(4) A] and 2
[Mn–COax: 1.7854(4) A; Mn–COeq: 1.838(4) A,
1.841(4) A] are longer, than those observed for 1,1-
Mn2(CO)8((t-butyl)NC)2 [Mn–COax: 1.748(5) A; Mn–
COeq: 1.824(5) A] [35]. The Mn–CNR distances in 1
˚
˚
˚
˚
˚
˚
˚
˚
˚
4. Conclusions
˚
˚
˚
(1.936(4) A) and 2 (1.924(4) A, 1.941(4) A) are
shorter, than in 1,1-Mn2(CO)8((t-butyl)NC)2 (1.944(4)
A) [35]. The reason for this is probably the enhanced
p-backbonding ability of aryl isocyanides compared
to alkyl isocyanides [43].
The chemistry of MnCo(CO)9 towards CNR is
closely related to that of Mn2(CO)10. The substitution
site on MnCo(CO)9 is the Mn center. Similar to
Mn2(CO)10, the number of coordinated CNR per Mn
never exceeds two. The CNR ligands are always in
equatorial position, terminally coordinated and, if
more than one, are in cis position to each other for
both Mn2(CO)10 and MnCo(CO)9 derivatives.
The calculation of IR spectra employing DFT
methods proved to be a useful tool in the assignment
of CO stretching bands for MnCo(CO)9. A unique
assignment of structure for the complexes Mn-
Co(CO)9−n(CNR)n, based on the compound’s calcu-
lated IR spectra, was possible only for n=1. It
appears that the higher the degree of substitution of
CO by CNR is, the less definitive the results of the
DMol calculations are. We are currently investigating
the reactivity of the complexes MnCo(CO)9−n(CNR)n
towards alkynes and alkenes, under Pauson–Khand
reaction conditions.
˚
The Co–COeq distances in 1 (1.778(7), 1.782(6) and
˚
˚
1.797(5) A) and 2 (1.774(4), 1.770(4) and 1.800(5) A)
˚
are shorter than those in HCo(CO)4 (1.818(3) A) [44]
and in Co2(CO)8 (1.827(2) A) [37], supporting the
˚
statement that the electron density around the Co is
increased compared to Co2(CO)8. Increased electron
density increases the Co“CO p-backbonding; hence
the Co–CO distances decrease. However the Co–
˚
˚
COax bond in 1 (1.780(5) A) and 2 (1.769(5) A) are
˚
longer than in HCo(CO)4 (1.764(10) A) [44]. This
might be attributed to the fact that the structure of
HCo(CO)4 was determined by electron diffraction,
and not by X-ray diffraction.
Less Mn“CO p-backbonding means a shorter C–O
bond length, because less electron density is pushed
into the p* orbital of CO. Therefore the MnC–O
bond lengths in 1 and 2 might be expected to be
shorter compared to Mn2(CO)10. However, the ob-
served average MnC–O bond lengths for 2 and
˚
Mn2(CO)10 are the same (1.142 A) [36], whereas the
˚
average MnC–O bond lengths in 1 (1.145(5) A) are
5. Supplementary material
only slightly longer than in Mn2(CO)10. The differ-
ences in the MnC–O bond lengths are within the
experimental error and no definite conclusions can be
made.
Most of the M–C–O angles in 1 and 2 are close
to 180°. The two Co–C–O angles opposite the two
CNR ligands are smaller (173.0(3) and 173.5(3)°) than
the third equatorial Co–C–O angle (177.5(4)°). This
could indicate a slight steric interaction between the
COs and the eclipsing CNR ligands; however, the en-
ergy involved is apparently not large enough to lead
Crystallographic data (excluding structure factors)
for the structures reported in this paper have been
deposited with the Cambridge Crystallographic Data
Centre as supplementary publication no. CCDC-
103060 (Compound 1) and CCDC-103061 (Com-
pound 2). Copies of the data can be obtained free of
charge on application to CCDC, 12 Union Road,
Cambridge CB2 1EZ, UK (fax: +44-1223-336 033;
E-mail: deposit@ccdc.cam.ac.uk). Structure factors
are available upon request from the authors.

K. Beck et al. / Inorganica Chimica Acta 288 (1999) 159–173
173
[17] N.J. Coville, P. Johnston, A.E. Leins, A.J. Markwell, J.
Organomet. Chem. 378 (1989) 401.
Acknowledgements
[18] G. Sbrignadello, G. Bor, L. Maresca, J. Organomet. Chem. 46
(1972) 345.
[19] (a) DMol and Insight II are commercially available from Molec-
ular Simulations, 9685 Scranton Road, San Diego, CA 92121-
3752. (b) DMol user manual, p. 85.
Crystallographic data were collected through the
Ohio Crystallographic Consortium, funded by the Ohio
Board of Regents 1995 Investment Fund (CAP-075),
located at the University of Toledo, Instrumentation
Center in A&S, Toledo, OH 43606. We thank Professor
Thomas L. Beck for useful conversations. JAKB sin-
cerely thanks Dr Ewa Skrzypczak-Jankun (University
of Toledo) for assistance in using the CCD diffractome-
ter system. KB gratefully acknowledges a Millenium
Petrochemicals graduate fellowship.
[20] S.H. Vosko, L. Wilk, M. Nusair, Can. J. Phys. 58 (1980) 1200.
[21] A.D. Becke, Phys. Rev. A38 (1988) 3098.
[22] J.P. Perdew, Y. Wang, Phys. Rev. B45 (1992) 13244.
[23] C. Lee, W. Yang, R.G. Parr, Phys. Rev B37 (1988) 785.
[24] (a) A. Klamt, G. Schu¨u¨rmann, J. Chem. Soc., Perkin Trans. II
(1993) 799. (b) J. Andzelm, C. Ko¨lmel, A. Klamt, J. Chem. Phys.
21 (1995) 9312. (c) A. Klamt, V. Jonas, T. Bu¨rger, J.C.W.
Lohrenz, J. Phys. Chem. A 102 (1998) 5074.
[25] The SMART v4.05 and SAINT v4.05, Siemens Analytical X-ray
Instruments, Inc., Madison, WI.
[26] SADABS, G.M. Sheldrick, University of Go¨ttingen, Germany.
[27] SHELXTL v5.03, G.M. Sheldrick, University of Go¨ttingen, Ger-
many, and Siemens Analytical X-ray Instruments, Inc.,
Madison, WI.
[28] J.K. Shen, Y.-C. Gao, Q. Zhen, F. Basolo, Organometallics 8
(1989) 2144, and leading references therein.
References
[1] (a) W.A. Herrmann, B. Cornils, Angew. Chem., Int. Ed. Engl.
36 (1997) 1048. (b) W.R. Moser, D.W. Slocum (Eds.), Homoge-
neous Transition Metal Catalyzed Reactions, American Chemi-
cal Society Advances in Chemistry Series, vol. 230, Washington,
DC, 1992.
[29] H.K. Van Dijk, J. van der Haar, D.J. Stufkens, A. Oskan, Inorg.
Chem. 28 (1989) 75.
[2] S.A. Lanemann, G.G. Stanley, in: W.R. Moser, D.W. Slocum
(Eds.), Homogeneous Transition Metal Catalyzed Reactions,
American Chemical Society Advances in Chemistry Series, vol.
230, Washington, DC, 1992, p. 349.
[3] F.A. Cotton, G. Wilkinson, Advanced Inorganic Chemistry, 5th
Edn., Wiley, New York, 1988, p. 255 and references therein.
[4] (a) M.O. Albers, N.J. Coville, S. Afr. J. Chem. 35 (1982) 139. (b)
G.W. Harris, N.J. Coville, J. Cryst. Spectros. Res. 19 (1989) 451.
[5] (a) G.W. Harris, N.J. Coville, Organometallics 4 (1985) 908. (b)
G.W. Harris, J.C.A. Boeyens, N.J. Coville, Organometallics 4
(1985) 914. (c) G.W. Harris, N.J. Coville, J. Cryst. Spectros. Res.
19 (1989) 451.
[6] Y Yamamoto, H. Yamazaki, Inorg. Chem. 17 (1978) 3111.
[7] G.K. Barker, A.M.R. Galas, M. Green, J.A.K. Howard, F.G.A.
Stone, T.W. Turney, A.J. Welch, P. Woodward, J. Chem. Soc.,
Chem. Commun. (1977) 256.
[8] (a) K.K. Joshi, P.L. Pauson, Z. Naturforsch. 76 (1962) 565. (b)
T. Kruck, M. Hofler, Chem. Ber. 97 (1964) 2289.
[9] E.S. Moosberry, R.K. Sheline, J Chem. Phys. 56 (1972) 1852.
[10] G. Sbrignadello, Inorg. Chim. Acta 48 (1981) 237.
[11] E.E. Ernstbrunner, M. Kilner, J. Chem. Soc., Dalton Trans.
(1976) 417.
[30] R.B. King, M.S. Saran, Inorg. Chem. 11 (1972) 2115.
[31] (a) L. Fan, T. Ziegler, J. Phys. Chem. 96 (1992) 6937. (b) W.
Thiel, M. Bu¨hl, Inorg. Chem. 36 (1997) 2922. (c) A. Be´rces, J.
Phys. Chem. 100 (1996) 16538. (d) V. Jonas, W. Thiel,
Organometallics 17 (1998) 353.
[32] N.J. Gould, D.J. Parker, Spectrochim. Acta, Part A 31 (1975)
1785.
[33] G. Bor, J. Organomet. Chem. 65 (1974) 81.
[34] P.M. Treichel, in: F.G.A. Stone, R. West (Eds.), Advances in
Organometallic Chemistry, vol. 11, Academic Press, New York,
1973, p. 24.
[35] D.J. Robinson, G.W. Harris, J.C.A. Boeyns, N.J. Coville, J.
Chem. Soc., Chem. Commun. (1984) 1307.
[36] M Martin, B. Rees, A. Mitschler, Acta Crystallogr., Sect. B 38
(1982) 6.
[37] P.C. Leung, C. Coppens, Acta Crystallogr., Sect. B 39 (1983)
535.
[38] W.E. Carrol, M. Green, A.M.R. Galas, M. Murray, W.T.
Terence, A.J. Welch, P. Woodward, J. Chem. Soc., Dalton
Trans. (1980) 80.
[39] E. Folga, T. Ziegler, J. Am. Chem. Soc. 115 (1993) 5169.
[40] F.W.B. Einstein, P. Manning, L.K. Peterson, K.G. Tyes, Inorg.
Chim. Acta 111 (1986) L49.
[41] L.H. Staal, G. van Koten, K. Vrieze, J.A. Cras, W.P. Bosman,
Inorg. Chem. 20 (1981) 555.
[12] R. Obrecht, R.H. Herrmann, I. Ugi, Synthesis (1985) 400.
[13] R. Appel, R. Kleinstu¨ck, K.-D. Ziehn, Angew. Chem., Int. Ed.
Engl. 10 (1971) 132.
[14] W. Hieber, O. Vohler, G. Braun, Z. Naturforsch., Teil B 13
(1958) 192.
[42] D.J. Elliot, D.G. Holah, A.N. Hughes, J.J. Vittal, P.J. Pud-
dephatt, Organometallics 12 (1993) 1225.
[15] P.A. Leach, S.J. Geib, J.A. Corella II, G.F. Warnock, N.J.
Cooper, J. Am. Chem. Soc. 116 (1994) 8566.
[16] R.B. King, Organometallic Synthesis, vol. 1, Academic Press,
New York, 1965, p. 174.
[43] P.M. Treichel, in: F.G.A. Stone, R. West (Eds.), Advances in
Organometallic Chemistry, vol. 11, Academic Press, New York,
1973, p. 26.
[44] E.A. McNeill, F.R. Scholer, J. Am. Chem. Soc. 99 (1977) 6243.
.