Ligand dependent structural changes in the acid-base chemistry of electron deficient benzoheterocycle triosmium clusters

Article abstract of DOI:10.1016/S0020-1693(02)00794-6

The reactions of the electron deficient benzoheterocycle clusters Os ₃(CO)₉(μ₃-η²-L-H)(μ-H) (L=phenanthridine, 2; 5,6-benzoquinoline, 3; quinoxaline, 4; 2-methyl benzimidazole, 5; 2-methyl benzotriazole, 6; 2-me

Full text of DOI:10.1016/S0020-1693(02)00794-6

Inorganica Chimica Acta 334 (2002) 343ꢀ354
/
www.elsevier.com/locate/ica
Ligand dependent structural changes in the acidꢀ
/
base chemistry of
electron deficient benzoheterocycle triosmium clusters
Edward Rosenberg a,*, Md. Joynal Abedin a, Dalia Rokhsana a, Alessandra Viale b,
Walter Dastru’ b, Roberto Gobetto b, Luciano Milone b,*, Kenneth Hardcastle c
a Department of Chemistry, University of Montana, Missoula, MT 59812, USA
b Dipartimento di Chimica IFM, Universita` di Torino, via P. Giuria 7-9, 10125 Torino, Italy
c Department of Chemistry, Emory University, Atlanta, GA 30322, USA
Received 19 October 2001; accepted 20 December 2001
Abstract
The reactions of the electron deficient benzoheterocycle clusters Os3(CO)9(m3-h2-LÃ
/
H)(m-H) (Lꢁ
/
phenanthridine, 2; 5,6-
benzoquinoline, 3; quinoxaline, 4; 2-methyl benzimidazole, 5; 2-methyl benzotriazole, 6; 2-methyl benzothiazole, 7; benzothiazole, 8;
2-methyl benzoxazole, 9; benzoxazole, 10) with n-butylamine and with the protic acids HBF4 and CF3CO2H are reported.
Complexes 2 and 3 behave very similarly to the analogous quinoline complex (Lꢁ
with similar spectral properties to 1 and undergoing simple protonation at the metal core with both HBF4 and CF3CO2H.
Complexes 4ꢀ10 on the other hand form an additional amine adduct with different spectral characteristics than 1ꢀ3 and in the case
of 5 this is the only detectable product. In addition, these compounds also form adducts with acetonitrile while 1ꢀ3 do not.
Compounds 4ꢀ9 also exhibit much more complex behavior in the presence of CF3CO2H than 1ꢀ3, giving trifluoroacetate adducts
/quinoline, 1) forming adducts with n-butylamine
/
/
/
/
/
subsequent to initial protonation. Although 4, 5 and 10 behave differently than 1 with n-butylamine and protic acids they form
exactly the same triphenyl phosphine adduct as 1 and 3 based on multinuclear NMR data. The solid-state structure for the amine
adduct of 2 is reported and suggested structures for the other amine and acid adducts, based on NMR data, are presented. The
reasons for the differences in the behavior of the three distinct groups of benzoheterocycles (1ꢀ
/
3, 4 and 5ꢀ10) are discussed. # 2002
/
Elsevier Science B.V. All rights reserved.
Keywords: Cluster compounds; Osmium; Heterocycles; Coordination chemistry
1. Introduction
state structures of 1ꢀ
bond lengths are remarkably similar throughout the
/
10 reveal that the metalÃ
/
ligand
˚
2.32(2) A)
We have been studying the chemistry of the electron
deficient triosmium clusters Os3(CO)9(m3-h2-LÃ
series (OsÃ
[5]. However, in our previous studies of the electro-
chemistry of these compounds, we found that 1ꢀ10 have
/
Nꢁ
/
2.10ꢀ
/
2.17(2); OsÃ
/
Cꢁ
/
2.25ꢀ
/
/
H)(m-H)
(Lꢁ
/
quinoline, 1; phenanthridine, 2; 5,6-benzoquino-
/
line, 3; quinoxaline, 4; 2-methyl benzimidazole, 5; 2-
methyl benzotriazole, 6; 2-methyl benzothiazole, 7;
benzothiazole, 8; 2-methyl benzoxazole, 9; benzoxazole,
10; Fig. 1) that result from the reaction of
Os3(CO)10(CH3CN)2 with a wide range of benzohetero-
cycles (all having a pyridinyl nitrogen b- to a CH on the
fused benzene ring) after photolysis of the initially
very different electrochemical properties [8]. Thus,
compound 1 exhibits two well-separated and chemically
irreversible 1eꢂ reduction potentials while the first
reduction potentials for 2 and 3 result in the formation
of stable radical anions. Compounds 5ꢀ10, on the other
/
hand exhibit two irreversible overlapping 1eꢂ reduction
potentials. The first reduction potentials correlate rea-
formed Os3(CO)10(m-h2-LÃ
/
H)(m-H) [1ꢀ/8]. The solid
sonably well with the metalÃ
transition in the 600ꢀ700 nm region, suggesting that
they are metal based, but the wide variation in the
overall electrochemical behavior of 1ꢀ10 suggests that
ligand orbitals are important in determining the chemi-
/
metal bond n to s*
/
* Corresponding authors. Tel.: ꢃ
4227.
E-mail address: rosen@selway.umt.edu (E. Rosenberg).
/
1-406-243 2592; fax: ꢃ1-406-243
/
/
0020-1693/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 0 2 0 - 1 6 9 3 ( 0 2 ) 0 0 7 9 4 - 6

344
E. Rosenberg et al. / Inorganica Chimica Acta 334 (2002) 343ꢀ354
/
Fig. 1. Structures of compounds 1ꢀ10 as determined by single crystal X-ray diffraction methods.
/
cal behavior of these systems despite their structural
similarities.
two or three heteroatoms, 5ꢀ
structural types react with triphenylphosphine to give
the adducts Os3(CO)9(m-h2-LÃ
H)(m-H)PPh3 (Lꢁquin-
/
10. All three of these
We recently completed a study of the reactions of 1
with amines and with protic acids [3]. We found that 1
was selective for primary amines and that it undergoes
simple protonation at the metal core with both HBF4
and CF3CO2H. The latter is known to form adducts
with related heterocycle triosmium clusters [9]. We
found that both the ability to coordinate amines (Lewis
acid properties) and the basicity of the metal core were
influenced by the nature of the substituent at the 5-
position of the quinoline ring. Electron donating groups
at this position enhanced the basicity of the metal core
and decreased its ability to bind amines [5]. In light of
/
/
oline, 11; 5,6-benzoquinoline, 12; quinoxaline, 13; 2-
methyl benzimidazole, 14; benzoxazole, 15). We pre-
viously reported the structures of 11 and 12 which have
the phosphine ligand in a radial position cis- to the
hydride and residing on the carbon bound osmium
atom, based on a solid state structure of 11 and the
almost identical 13C NMR of 11 and 12 (structure A,
Scheme 1; Table 1) [1,2,6]. A key feature of the 13C
NMR of 11 and 12 is the observation of three broadened
resonances attributable to the tripodal motion of three
of the four carbonyl groups of the Os(CO)4 moiety
(resonances g, h, i; Table 1). At low temperatures
these results and the electrochemical behavior of 1ꢀ10
/
discussed above, we thought it would be useful to survey
the reactions of these complexes with amines and with
protic acids in order to determine if the observed
differences in electrochemical behavior carry over to
(ꢂ
/
40 8C) the expected trans-13CÃ
/
13C coupling of 30ꢀ
/
35 Hz is observed between the resonances assigned to
their acidꢀbase properties and in order to compare their
/
reactivity with 1. We report the results of these studies
along with the related reactions of these complexes with
acetonitrile and triphenylphosphine.
2. Results and discussion
2.1. Reactions with triphenylphosphine, acetonitrile and
n-butylamine
The complexes 1ꢀ
structural types: (1) those related to quinoline, 1ꢀ
quinoxaline, 4; and (3) the 6,5-fused ring systems with
/
10 (Fig. 1) can be divided into three
/3; (2)
Scheme 1.

E. Rosenberg et al. / Inorganica Chimica Acta 334 (2002) 343ꢀ
/354
345
Table 1
Proton coupled 13C NMR data for the phosphine adducts of the electron deficient benzoheterocycle complexes
a
a
Chemical shifts are in ppm downfield relative to TMS, coupling constants in Hz at 25 8C. b From Ref. [1]. c From Ref. [6]. d JCH not observed
due to overlap of resonances.
carbonyls g and i confirming the presence of the
Os(CO)4 group. The 13C NMR spectra of compounds
13C NMR relative to CDCl3 (Tables 2 and 3). In the case
of 5, the light green solution shows evidence for the
1
presence of residual starting material in its H NMR as
well as the presence of two new hydride resonances at
13ꢀ
/
15 also exhibit these features and indeed have 13C
NMR spectra that are very similar to 11 and 12 (Table
1). On this basis we assign structure A to all of these
phosphine adducts (Scheme 1). In the case of benzox-
azole, a second minor isomer is observed as evidenced
by the observation of a second doublet in the hydride
ꢂ
/
12.15 and ꢂ12.40 ppm and associated new aromatic
/
and NH resonances. The hydride resonances of the two
new species present integrate in a ratio of 11.6:1 with the
hydride resonance attributable to 5. Using the starting
moles of 5 in the solution, the volume of acetonitrile and
the hydride integration ratios we can estimate the
equilibrium constant for the reaction of 5 with acetoni-
trile to be 0.68 Mꢂ1. The main feature of the proton
coupled 13C NMR in the carbonyl region is the
region at ꢂ
/
16.14 (J31Pꢀ1
/
Hꢁ14.2 Hz) along with
/
1
companion peaks to the major isomer in the H and
13C NMR. Although overlap of resonances and the
inseparability of the two isomers precluded the char-
acterization of this minor product we suggest that it has
the phosphine substituent on the nitrogen bound
osmium atom (structure A?, Scheme 1) as has been
previously reported for related nitrogen heterocycle
triosmium complexes [11].
observation of two closely spaced doublets (Ddꢁ
/1.4)
with relatively large coupling to the bridging hydride
(Table 3). This feature is also observed in the decacar-
bonyl of 10 where Ddꢁ/1.4 (Table 3). This small value
We have previously reported that 1 and its substituted
derivatives undergo ligand cleavage in acetonitrile at
for the Dd of the carbonyls trans- to the hydride is
indicative of a structure where the benzoheterocycle
bridges the same edge as the hydride and the acetonitrile
is coordinated to the unbridged osmium (structure B,
Scheme 1). This same feature was also observed for the
acetonitrile adduct of the imidoyl complex Os3(CO)9(m3-
70ꢀ80 8C in the presence of 1 atm of CO to yield
/
Os3(CO)12 and the free quinoline [4,5]. It seems reason-
able to suggest that the first step in this process is the
formation of an acetonitrile adduct. However, com-
pounds 1ꢀ
/
3 shows no evidence for adduct formation
h2-(CH2)3CÄ
obtained (Ddꢁ
/
NÃ
/
)(m-H), where a crystal structure was
when dissolved in acetonitrile, maintaining their green
color and exhibiting only minor shift changes in their 1H
NMR. In sharp contrast, compounds 5 and 10 undergo
an immediate color change, from dark to light green and
dark green to yellow, respectively, on dissolution in
acetonitrile and show significant changes in their 1H and
/
0.05) [10]. The assignments given in
Table 3 for the carbonyl groups of the acetonitrile
adducts of 4, 5 and 10 were made on the basis of
established trends in chemical shift and coupling con-
stant data for these types of cluster [3,10]. These
assignments are by no means unambiguous but the

346
E. Rosenberg et al. / Inorganica Chimica Acta 334 (2002) 343ꢀ354
/
Table 2
1H NMR of acetonitrile adducts of electron deficient benzoheterocycle clusters
a
a
Chemical shifts are in ppm downfield positive relative to TMS 25 8C.
From Ref. [5].
Minor isomer.
b
c
Table 3
Proton coupled 13C NMR of the acetonitrile adducts of the electron deficient benzoheterocycle complexes
a
a
b
Chemical shifts are downfield positive relative to TMS in ppm and coupling constants are given in Hz at 25 8C. Two axial carbonyls are
observed with the expected 13Cꢀ13
C satellites. Major isomer only.
/
c
general features as discussed above allow us to distin-
guish between the structural isomers possible (vide
infra). We cannot, of course, differentiate between
structures B and B? (Scheme 1) based on the current
data; B is chosen here based on analogy with previously
reported acetonitrile adducts [10]. Further evidence that
the structures of the decacarbonyl and acetonitrile
adduct of 5 are the same comes from the fact that
they both have the similar T1 values for the hydride in
acetonitrile (7.2 and 8.0 s, respectively). Compound 4
undergoes a color change from dark green to red (the
show broad hydride resonances at ꢂ
/
13.61 and ꢂ13.85
/
ppm in a relative intensity of 10:1 with companion
broad resonances in the aromatic region. The broadness
of these lines and the chemical shifts of the hydrides
suggest the population of isomers such as C/C? or D/D?
(Scheme 1, vide infra). This was not the case for 5 or 10.
It is important to note here that although 4ꢀ
adducts with acetonitrile while 1ꢀ3 do not, the rings are
not cleaved from the cluster at any significant rate when
heated in this solvent at 70ꢀ80 8C under 1 atm of CO as
are 1ꢀ3. This seems contradictory unless one considers
/10 form
/
/
/
1
same color as the decacarbonyl), and the H and 13C
the possibility that cleavage of the ligand is controlled
thermodynamically by the relative stability of the m3 and
NMR data support formation of an acetonitrile adduct
with structure B or B? (Tables 2 and 3). In addition to
one major adduct isomer and residual starting material
solutions of 4 in acetonitrile show the presence of three
minor isomers. One of the minor isomers shows a set of
sharp resonances partially overlapping with the major
m bonding modes for 1ꢀ
/
3 versus 4ꢀ10 (Fig. 2 represents
/
a qualitative picture of the situation). The reasons why
the m-bonding mode should be more stable thermody-
namically for 4ꢀ
/
10 than the m3-compared with 1ꢀ
/3 are
not apparent and must await a theoretical treatment of
the bonding in these systems. It is clear that all the m3-
complexes are kinetically stable and that adduct forma-
isomer and a hydride resonance at ꢂ
/
11.04 ppm and can
tentatively be assigned to B? [10]. The two other isomers

E. Rosenberg et al. / Inorganica Chimica Acta 334 (2002) 343ꢀ
/354
347
Fig. 2. Reaction profile diagram for ligand coordination and cleavage, illustrating the proposed difference in the relative stability of the m- and m3-
bonding modes for 1ꢀ10.
/
tion must precede ligand cleavage because cleavage of
1ꢀ3 does not proceed in solvents other than acetonitrile.
Compounds 1ꢀ10 all react with n-butylamine to form
yellow adducts, but here again there are significant
differences between the different structural types. In our
previous studies with 1, we suggested the isomeric
structures C and C? for the amine adducts with C being
the major isomer based on solution NMR data, T1
measurements and by analogy with related structurally
selected distances and bond angles in Table 5. The solid-
state structure of 16 reveals that the hydride and the
benzoheterocycle bridge different edges of the cluster,
with the amine ligand and the phenanthridine ring on
the same face of the cluster. The hydride bridges the
osmium atom coordinated to the amine and the osmium
atom bound to C(8) (quinoline numbering system) of
the phenanthridine ring. The overall structure is repre-
sented by structure D (Scheme 1). In light of the very
/
/
characterized ammonia adducts [10]. Compounds 2ꢀ
/
4
similar spectroscopic properties of 1ꢀ3, we must revise
/
appear to form the same two isomers but in the case of 2
and 3 the adducts precipitated from chloroform neces-
sitating measurement of the isomer ratios and equili-
brium constants in acetone (Table 6). As for 1, we
observe the formation of a single isomer and then the
gradual appearance of a second isomer. We previously
showed that this process involves formation of an
isomer with the amine ligand on the same face of the
cluster as the heterocycle followed by an intermolecular
equilibration to a mixture of isomers with the amine
ligand on either face of the metal triangle.
our earlier assignment of the amine adduct of 1 to
structure C and assign the solution structures of all three
amine adducts to an equilibrium mixture of D and D?.
The hydride was located along the longest edge of the
˚
Os(2)ꢁ3.0405(6) A) using the
metal triangle (Os(1)Ã
/
/
program HYDEX [12]. The phenanthridine ligand bridges
the second longest edge of the metal triangle (Os(2)Ã
/
˚
Os(3)ꢁ
shortest metalÃ
/
2.8873(6) A) while the unbridged edge has the
metal bond distance (Os(1)ÃOs(3)ꢁ
2.7810(5) A). The metal nitrogen bond distances are
essentially equal (Os(2)ÃN(1S)ꢁ2.209(7) and Os(3)Ã
2.195(7) A and are approximately the same as
the metal carbon bond distance with the phenanthridine
/
/
/
˚
/
/
/
˚
In the case of 2, we were able to isolate crystals of the
amine adduct, Os3(CO)9(m3-h2-LÃ
N(1B)ꢁ
/
/
H)(m-H)(n-C4H9N)
˚
(Lꢁ/phenanthridine, 16), suitable for a solid-state
ligand (Os(1)Ã
some longer than related distances in Os3(CO)9(m-h2-LÃ
H)(m-H)(L?) complexes [2,5]. The n-butylamine ligand
/
C(12B)ꢁ/2.190(9) A). These distances are
structural investigation. The solid-state structure of 16
/
is shown in Fig. 3, crystal data are given in Table 4 and

348
E. Rosenberg et al. / Inorganica Chimica Acta 334 (2002) 343ꢀ354
/
Table 5
Selected distances (A) and bond angles (8) for Os3(CO)9(m3-h -LÃ
H)(m-H)(n-C4H9N) (Lꢁphenanthridine, 16)
2
˚
/
/
Bond distances
Os(1)Ã
Os(1)Ã
Os(2)Ã
Os(1)Ã
Os(2)Ã
Os(3)Ã
/
Os(2)
Os(3)
Os(3)
C(12B)
N(1S)
3.0405(6) N(1S)Ã
2.7810(5) N(1B)Ã
2.8873(6) N(1B)Ã
2.190(9) C(1B)Ã
2.209(7) C(11B)Ã
2.195(7) C(12B)Ã
1.89(1)
/
C(1S)
1.50(1)
/
/
C(1B)
1.297(9)
1.46(1)
1.41(1)
1.40(1)
1.43(1)
/
/
C(13B)
/
/
C(2B)
/
/C(12B)
/
N(1B)
a
/C(13B)
OsÃCO
/
Bond angles
Os(1)ÃOs(2)Ã
Os(1)ÃOs(3)Ã
Os(2)ÃOs(1)Ã
Os(1)ÃOs(2)Ã
Os(3)ÃOs(2)Ã
Os(1)ÃOs(3)Ã
Os(2)ÃOs(3)Ã
Os(2)ÃOs(1)Ã
Os(3)ÃOs(1)Ã
OsÃCÃ
/
/
Os(3)
Os(2)
Os(3)
55.88(1) C(1S)Ã
64.85(1) C(1B)Ã
59.27(1) C(13B)Ã
89.7(2) C(13B)Ã
89.1(2) C(11B)Ã
86.8(1) C(1B)ÃN(1B)Ã
92.7(1) C(11B)ÃC(12B)Ã
/
N(1S)Ã
N(1B)Ã
N(1B)Ã
C(12B)Ã
C(12B)Ã
/
Os(2)
Os(3)
Os(3)
117.1(8)
118.2(8)
123.0(5)
125.4(7)
119.9(7)
118.2(8)
/
/
/
/
/
/
/
/
/
/
N(1S)
N(1S)
N(1B)
N(1B)
C(12B)
C(12B)
/
/Os(1)
/
/
/
/Os(1)
/
/
/
/
C(13B)
/
/
/
/
C(13B) 114.4(9)
/
/
91.3(2)
/
/
85.6(2)
a
/
/O
176.4(9)
Fig. 3. Solid state structure of Os3(CO)9(m-h2-LÃ
(Lꢁphenanthridine, 16).
/
H)(m-H)(n-C4H9N)
Numbers in parentheses are average standard deviations.
a
Average values.
/
and the phenanthridine ligand are perpendicular to the
plane of the metal triangle as can be seen from the OsÃ
/
OsÃN(1B), OsÃOsÃN(1S) and OsÃOsÃC(12B) angles
/
/
/
/
/
Table 4
Crystal data and structure refinement for 16
(Table 5). The n-butylamine ligand is disordered, axially
with a conformational population of 0.59:0.41 (Fig. 3).
Empirical formula
Formula weight
Temperature (K)
C26H20N2O9Os3
1073.02
100(2)
The formation constants for 1ꢀ3 evaluated after
/
isomer equilibration are quite similar as are the D/D?
isomer ratios and the hydride chemical shifts (Table 6).
This might be expected for these very similar hetero-
cycles. The 13C NMR patterns for the amine adducts of
1 and 2 are very similar and the T1’s of the hydrides are
almost identical and are significantly shorter than for 1
and 2 (4.0 and 5.2 s for 1 and 2, respectively; Tables 6
and 7). The short T1 values for these adducts are
attributable to enhanced dipolar relaxation indicative
of the close proximity of the amine to the hydride
ligand. In the 13C NMR of the carbonyl region of 13CO
enriched samples of the amine adducts of 1 and 2 no
˚
Wavelength (A)
0.71073
monoclinic
P2(1)/n
Crystal system
Space group
Unit cell dimensions
˚
a (A)
9.0881(8)
21.789(2)
13.9362(13)
90
˚
b (A)
˚
c (A)
a (8)
b (8)
g (8)
99.771(2)
90
3
˚
V (A )
Z
2719.6(4)
4
2.621
Dcalc (mg mꢂ3
)
Absorption coefficient (mmꢂ1
F(000)
)
14.038
1952
13Cꢀ13
evidence it is reasonable to propose the n-butylamine
adducts of 1ꢀ3 exist in solution as the same set of
isomers, D and D?.
Compounds 4ꢀ10 all show an increase in their
formation constants with n-butylamine relative to 1ꢀ
while the D/D? ratio remains about the same as in 1ꢀ
/
C coupling satellites are observed. Based on this
Crystal size (mm3)
0.14ꢄ
1.75ꢀ32.86
135h 5
205l 520
34 853
/0.05ꢄ
/
0.04
/
u Range for data collection (8)
Index ranges
/
ꢂ/
/
/
13, ꢂ
/335
/
k 532,
/
ꢂ/
/
/
/
Reflections collected
/
3
3
Independent reflections
9576 [Rint
94.5%
ꢁ0.1256]
/
/
Completeness to uꢁ
/
32.868
with the exception of 4. This increase in the formation
constant is consistent with the fact that 4ꢀ10 form
acetonitrile adducts while 1ꢀ3 do not. In addition, the
appearance of a third new hydride resonance is observed
at slightly lower fields (ꢂ11ꢀ12 ppm) for 5ꢀ10 (Table
6). In the case of 5, this is the only adduct hydride
resonance observed and therefore allowed us to examine
the structure of this isomer in detail. The 13C NMR of
Absorption correction
Max/min transmission
Refinement method
SADABS2
1.00000 and 0.721709
Full-matrix least-squares on F2
9576/0/357
/
/
Data/restraints/parameters
Final R indices [I ꢀ
/
2s(I)]
R1ꢁ
R1ꢁ
1.081
2.396 and ꢂ2.492
/0.0589, wR2ꢁ
/
0.0656
/
/
/
R indices (all data)
Goodness-of-fit on F2
/0.1343, wR2ꢁ
/0.0734
Largest difference peak and hole
ꢂ3
/
˚
(e A
)

E. Rosenberg et al. / Inorganica Chimica Acta 334 (2002) 343ꢀ
/354
349
Table 6
1H NMR, isomer ratios, hydride T1 and formation constants for n-butylamine adducts of electron defiicient benzoheterocycle clusters
a
c
Ring
No. Solvent
d(D)
d(D?)
d(C)
K(D/D?)
C/Dꢃ
/
D?
Hydride T1
Kform (Mꢂ1
)
Quinoline
1
2
CDCl3
ꢂ
/
13.39
13.23
13.33
13.55
ꢂ
/
12.76
12.56
12.73
13.05
2.8
2.2
1.4
1.5
44.4
18.3
Phenanthridine
5,6-Benzoquinoline
Quinoxaline
(CD3)2CO
(CD3)2CO
CDCl3
ꢂ
/
ꢂ
/
3
ꢂ
/
ꢂ
/
4
ꢂ/
ꢂ/
0.66
303
124
232
47
b
2-Me-benzimidazole
2-Me-benzotriazole
2-Me-benzothiazole
Benzothiazole
5
CD2Cl2
CDCl3
CDCl3
ꢂ
/
11.78
11.67
11.07
11.46
11.86
12.09
5.2, 6.9 (15.2)
6
ꢂ
/
13.67
13.46
13.52
13.60
13.63
ꢂ
/
13.15
12.86
12.98
13.06
13.15
ꢂ
/
1.56
2.59
2.54
2.0
0.69
1.61
0.23
2.0
7
ꢂ
/
ꢂ
/
ꢂ
/
8
CD2Cl2
CDCl3
CD2Cl2
ꢂ
/
ꢂ
/
ꢂ
/
274
310
1097
2-Me-benzoxazole
Benzoxazole
9
10
ꢂ
/
ꢂ
/
ꢂ
/
ꢂ/
ꢂ/
ꢂ/
1.09
0.51
a
Chemical shifts in ppm downfield positive from TMS 25 8C. b T1 for NH3 complex (ND3 complex). c Calculated using the relative intensity of
the hydride resonances of all observed isomers, the moles of starting complex and the amount of amine added.
Table 7
13C NMR data for amine adducts of the electron deficient benzoheterocycle complexes
a
a
b
c
Chemical shifts in ppm downfield positive relative to TMS 25 8C. From Ref. [3], Lꢁ
/
NH3. Lꢁn-BuNH2.
/
this adduct has a distinctly different pattern than that
observed for D in 1ꢀ3 with the carbonyl resonances
trans- to the hydride being much more well separated
(Ddꢁ4.6) suggesting that the two hydride bridged
osmium atoms are in a different environment than in
D. The T1 of the n-butylamine adduct of 5 is 5.2 s which
is considerably shorter than the T1 of the decacarbonyl
5 and its amine adduct are all longer than the
corresponding values for 1 probably reflects the replace-
ment of the hydrogen at C(2) with a methyl group. In
the case of 10, a fourth hydride resonance is observed at
/
/
ꢂ11.26 ppm very close the resonance we have assigned
/
to C or C? in a relative intensity of 1:1.2. It would appear
that both C and C? are populated in this case and this
isomer was included in the calculation of the formation
constant for the amine adduct of 10 (Table 6). That
isomers C and/or C? are preferred for 5 and populated
of 5 (11.5 s) that has structure B. The T1’s of 1ꢀ3 and
/
their amine adducts are, on the other hand, considerably
shorter and so we decided to evaluate the T1 values for
the NH3 and ND3 adducts of 5 in order to get a more
definitive idea of the proximity of the amine hydrogens
to the hydride. The values obtained of 6.9 and 15.2 s for
the NH3 and ND3 adducts, respectively, clearly indicate
that the amine ligand resides on an osmium atom bound
to the bridging hydride. Taken together with the 13C
NMR data, we suggest structure C or C? for the amine
adduct of 5. That the T1 values of 5, the decacarbonyl of
for 5ꢀ10 may attributed to contributions that make the
/
pyridinyl nitrogen more basic and therefore favor the
isomer where the acidic hydride bridges the edge of the
metal triangle to which both basic nitrogens are bound
(Scheme 2). In the case of the amine adduct of 5 we were
able to isolate analytically pure samples of the n-
butylamine
C4H9N) (Lꢁ
adduct
2-me-benzimidazole, 17) these crystals
Os3(CO)9(m-h2-LÃ
H)(m-H)(n-
/
/

350
E. Rosenberg et al. / Inorganica Chimica Acta 334 (2002) 343ꢀ354
/
2.2. Reactions with HBF4 and CF3CO2H
Compounds 1ꢀ3 undergo simple protonation with
/
1
HBF4 and CF3CO2H as evidenced from their H and
13C NMR (Tables 8 and 9). The solutions go from dark
green to orange on addition of both acids and the
resulting 1H and 13C NMR spectra are virtually
identical and can be assigned to structure E (Scheme
Scheme 2.
3) [3]. Compounds 4ꢀ10 also protonate with HBF4 but
/
with CF3CO2H yellow solutions are formed and the 1H,
13C and 19F spectra give strong evidence for the
formation of trifluoroacetate adducts (Scheme 3), con-
sistent with the greater tendency of these compounds to
form adducts with Lewis bases as described above [9].
were not suitable for solid-state structural analysis but
do corroborate the formulation of the adducts with the
proposed structure C.
Table 8
Hydride chemical shifts for initial protonation products of the electron
The changes in the spectra of 4ꢀ
to compound and will be discussed individually below.
In the case of 4ꢀ10, we observed no evidence of
/
10 vary from compound
a
deficient benzohetrocycle complexes
/
Ring
No. d(hydride)
d(hydride) Acid
protonation of the heteroatoms and this is particularly
surprising in the case of 4 but consistent with the
decreased basicity of the uncoordinated lone pair and
the enhanced basicity of the metal core in this complex.
Addition of a twofold excess of CF3COOH to a
CD2Cl2 solution of 4 results in an instantaneous color
Quinoline
Quinoline
1
1
ꢂ
/
13.70(d) (1.6)
13.66(d) (1.7)
13.71(d) (1.6)
13.70(d) (1.7)
14.08(d) (1.6)
14.10(d) (1.6)
13.71(d) (1.6)
13.36(d) (1.7)
13.22(d) (1.6)
13.50(d) (1.4)
13.41(d) (1.2)
13.43(d) (1.4)
13.38 (s)
ꢂ
/
11.65(d) HBF4
11.61(d) CF3CO2H
11.73(d) HBF4
11.73(d) CF3CO2H
11.74(d) HBF4
11.73(d) CF3CO2H
11.63(d) HBF4
12.61(d) HBF4
11.98(d) HBF4
12.42(d) HBF4
12.24(d) HBF4
12.61(d) HBF4
12.66 (s) HBF4
ꢂ
/
ꢂ
/
Phenanthridine
Phenanthridine
5,6-Benzoquinoline
5,6-Benzoquinoline
Quinoxaline
2
ꢂ
/
ꢂ
/
2
ꢂ
/
ꢂ
/
3
ꢂ
/
ꢂ
/
3
ꢂ
/
ꢂ
/
4
ꢂ
/
ꢂ
/
change from green to orangeꢀ
region of the 1H NMR consists of two broad resonances
at ꢂ12.24 and ꢂ14.08 ppm along with a small amount
of starting material at ꢂ12.33 ppm. As the temperature
is lowered to ꢂ50 8C these resonances sharpen and are
resolved into two sets of hydrides at ꢂ12.10 and ꢂ13.78
ppm and ꢂ12.38 and ꢂ14.88 ppm, in a relative
/
yellow. The hydride
2-Me-benzimidazole
2-Me-benzotriazole
2-Me-benzothiazole
Benzothiazole
5
ꢂ
/
ꢂ
/
6
ꢂ
/
ꢂ
/
/
/
7
ꢂ
/
ꢂ
/
/
8
ꢂ
/
ꢂ
/
/
2-Me-benzoxazole
Benzoxazole
9
10
ꢂ
/
ꢂ
/
ꢂ/
ꢂ/
/
/
/
/
a
Chemical shifts in ppm downfield positive from TMS; numbers in
parenthesis are coupling constants between the hydrides in Hz at
25 8C.
intensity of 1:2. This behavior can be understood in
terms of the averaging of two isomeric adducts F and F?
Table 9
Proton coupled 13C NMR of the protonated electron deficient benzoheterocycle complexes
a
a
b
Chemical shifts are downfield positive relative to TMS in ppm and coupling constants are given in Hz at 25 8C. Data from Ref. [3].

E. Rosenberg et al. / Inorganica Chimica Acta 334 (2002) 343ꢀ
/354
351
sharp hydride resonances at ꢂ
/
12.52 and ꢂ
/
13.45 ppm,
again almost identical to the hydrides observed with
HBF4. As for 5, the 19F NMR shows only free
trifluoroacetateꢀtriflouroacetic acid. Based on our ob-
/
servations so far it appears that adduct formation is
associated with the introduction of hydride fluxionality
while simple protonation results in rigid dihydride
species. This behavior is reminiscent of the addition of
phosphine to the rigid trihydride Os3(CO)8(m3-h2-
(CH2)3CÄ
/
NÃ)(m-H)3 where conversion of a m3-capping
/
ligand to a m-edge bridging mode results in inducing a
high degree of fluxionality to the hydride ligands [13].
The behavior of 8 in the presence of 1 equiv. of
CF3COOH is a composite of 4 and 6. At ambient
temperature a series of broad resonances is observed but
Scheme 3.
by hydride exchange. This is supported by the observa-
tion of two 19F NMR resonances at 1.81 and 0.66 ppm
relative to free CF3COOH at ꢂ50 8C, in the same
on lowering the temperature to ꢂ
hydride resonances can be seen at ꢂ
ppm, ꢂ12.79 and ꢂ14.71 ppm and ꢂ
ppm. A sharp peak is observed at ꢂ
attributable to unreacted 8. At ꢂ90 8C these reso-
nances shift somewhat and overlap but become sharper.
On standing the pair of resonances at ꢂ12.85 and
13.70 ppm increases in intensity and finally becomes
/
60 8C three pairs of
/
/
13.29 and ꢂ
/
13.68
relative intensity as the two sets of hydride resonances,
which average with each other but not with the free acid
as the temperature is increased. This behavior is very
similar to that observed for the related imidoyl adduct,
/
/
/
12.85 and ꢂ
/
13.70
/11.78 ppm
/
Os3(CO)9(m-h2-(CH2)3CÄ
/
NÃ)(m-H)2(CF3COO) that has
/
/
a solid state structure analogous to F [9].
Compound 5 undergoes only simple protonation with
CF3COOH as evidenced by the observation of two
ꢂ
/
the dominant species present in solution. This behavior
is paralleled in the 19F NMR where three resonances at
2.15, 2.31 and 2.34 ppm are initially observed at
sharp hydride resonances of equal intensity at ꢂ
/
12.63
and ꢂ
/
13.38 ppm. These shifts are almost identical to
ꢂ90 8C in addition to free trifluoroacetate at 0.0
/
those observed on reaction of 5 with HBF4 and the 19F
NMR spectrum shows only free trifluoroacetate ion.
The behavior of 6 with CF3COOH is distinctly
different than that of 4. Addition of 1 equiv. of this
acid results in the appearance of two broad hydride
ppm. On standing overnight the resonance at 2.15
ppm increases while the other two resonances disappear
and the trifluoroacetate peak diminishes in intensity.
These data can be understood in terms of initial
formation of F and F? which undergo interconversion
by hydride edge hopping on the NMR time scale
followed by gradual conversion to either G or G?
(Scheme 3).
resonances at ꢂ
intensity with about 50% of the starting material
remaining as a sharp resonance at ꢂ11.92 ppm. On
/
13.51 and ꢂ13.92 ppm of equal relative
/
/
standing overnight the relative intensity of the starting
material diminishes and the two broad resonances
The reaction of 9 with CF3COOH takes a similar
course but with some slight differences. As for 8
addition of 1 equiv. of CF3COOH leads to the forma-
tion of a dihydride species with broad resonances at
sharpen and shift to ꢂ
/
13.04 and ꢂ13.92 ppm. The
/
19F NMR initially shows a somewhat broadened free
triflouroacetate peak and a broad peak at 1.68 ppm that
sharpens and shifts to 1.95 ppm. Addition of a fivefold
excess of acid to the solution leads only to the
disappearance of the starting material and a slight
further shift of the hydride signals. The 19F NMR
shows free trifluoroacetate and a sharp signal at 1.95
ppm in the presence of excess acid. These data are best
understood in terms of formation of a adduct with
structure F or F? where formation of the adduct is slow
on the NMR time scale, followed by isomerization to G
or G? by an intermolecular (dissociative) process in the
intermediate exchange regime (Scheme 3). Apparently,
only one of the two possible isomers G or G? is
populated and the hydrides are fluxional at ambient
temperature.
ꢂ
/
13.58 and ꢂ
second broadened set of hydride signals on standing
overnight at ꢂ13.60 and ꢂ13.88 ppm. In contrast to 8
however, this latter set of hydrides is resolved into two
sets of hydrides at ꢂ13.35 and ꢂ13.83 ppm and ꢂ13.72
and ꢂ13.90 ppm in a ratio of 3:4 on lowering the
temperature to ꢂ
/
13.99 ppm that gradually convert to a
/
/
/
/
/
/
/
90 8C. This behavior is paralleled in
the 19F NMR where an initial peak at 1.75 ppm
gradually converts to a peak at 2.25 ppm on standing
at room temperature. Additionally, we were able to
observe two sharp doublets in the hydride region of the
1H NMR at ꢂ
/
12.61 and ꢂ13.43 ppm that we assign to
/
the presence of 9Hꢃ. Thus the formation of F and F? is
followed by conversion to G and G? where hydride edge
hopping is in the fast exchange regime for both sets of
isomers but where conversion of F to G is slow. That
9Hꢃ can be detected in this case during the conversion
Compound 7 like 5 undergoes simple protonation
with CF3COOH as evidenced by the appearance of two

352
E. Rosenberg et al. / Inorganica Chimica Acta 334 (2002) 343ꢀ354
/
of F to G provides good circumstantial evidence that
this conversion is intermolecular in nature. A similar
equilibrium between a protonated species and a trifluor-
oacetate adduct was reported for Os3(CO)9(m3-h2-
Particularly puzzling, and also intriguing, are the
factors governing the remarkable differences in the
behavior of 5 and 7 versus 4, 6 and 8ꢀ
trifluoroacetic acid. We are hopeful that theoretical
modeling of the bonding geometries of 1ꢀ10 currently
/
10 towards
(CH2)3CÄ
/
NÃ
/
)(m-H)2(CF3COO) [9]. Addition of more
/
CF3COOH to solutions of the adduct of 9 causes no
substantial changes in the NMR spectra but does result
in chemical shift changes of the 1H hydride, the aromatic
hydrogen resonances , the 13C carbonyl and the 19F
trifluoroacetate resonances. These shifts are probably
due to hydrogen bonding interactions with excess acid
but do not affect the interpretations presented here.
The reaction of 10 with CF3COOH takes a distinctly
different, yet simpler course than that of 9. In presence
of 1 equiv. of acid a single broad hydride resonance is
underway in our laboratories will further elucidate these
differences in reactivity.
The take home message of this detailed study is that
subtle differences in the nature of the heterocyclic ligand
can lead to dramatic differences in reactivity, even
though the bonding mode of the ligand to the cluster
in the entire class of molecules studied is virtually
identical.
observed at ꢂ
65 8C this broad resonance resolves into one major
pair of hydride resonances at ꢂ13.94 and ꢂ14.09 ppm
with very minor companion peaks at ꢂ12.18, ꢂ12.28,
13.15, ꢂ13.62, ꢂ13.67 and ꢂ14.95 ppm whose low
/
14.05 ppm at room temperature. At
3. Experimental
ꢂ
/
/
/
3.1. Materials
/
/
ꢂ
/
/
/
/
Compounds 1ꢀ
published literature procedures [4ꢀ
/
10, 11 and 12 were synthesized by
6]. The NMR sol-
intensities made it impossible to assign them as pairs.
Similarly, the observation of one major peak at 1.35
ppm and three minor peaks at 1.61, 1.67 and 2.65 ppm
/
vents CDCl3, CD2Cl2, CD3CN and (CD3)2CO (Norell)
were dried over molecular sieves (Type 4A, Mallin-
crodt). Trifluoroacetic acid (Aldrich) was distilled from
P2O5 before use. Perfluoroboric acid (Aldrich), n-
BuNH2 (Aldrich) and triphenylphosphine (Strem) were
used as received.
in the 19F NMR at ꢂ
65 8C confirms that the hydride
/
signals are due to adduct formation. Thus in the case of
10 only one isomer of the acid adduct is significantly
populated at ꢂ65 8C, most likely F, and at room
/
temperature the hydrides and the minor isomers are
exchanging on the NMR time scale.
3.2. Spectra and analyses
2.3. General conclusions
NMR spectra were measured on a Varian Unity Plus
400 MHz Spectrometer and on a JEOL-EX 400 MHz
These studies have revealed some significant differ-
Spectrometer. IR spectra were measured on a Perkinꢀ
/
ences in the chemistry of 1ꢀ
have a much greater tendency to undergo addition of
two electron donors than 1ꢀ3. This is reflected in the
/
10. First, complexes 4ꢀ
/
10
Elmer 1600 and a Thermo-Nicolet 633 FT IR Spectro-
meters. Schwarzkopf Microanalytical Laboratories,
Woodside, New York, performed elemental analyses.
/
larger formation constants of the amine adducts and in
the formation of adducts with the weaker donors
acetonitrile and trifluoroacetate. This can be rationa-
lized by the apparent greater thermodynamic stability
3.3. Synthesis of the phosphine derivatives 13ꢀ15
/
Compounds 4, 5 or 10 (95 mg, 0.1 mmol) were
dissolved in 20 mL CH2Cl2 and 26 mg (0.1 mmol)
PPh3 in 5 ml of the same solvent was slowly added to the
green solution under an atmosphere of nitrogen. The
for the m versus the m3 bonding mode for 4ꢀ
to 1ꢀ3 (Fig. 2). This most likely arises from the
contribution of resonance structures that make the
pyridinyl nitrogen a stronger donor in the case of 5ꢀ
10 (Scheme 2).
Second, reaction of 1ꢀ
/
10 relative
/
/
solution turned from green to yellowꢀorange immedi-
/
ately. After stirring for 15 min the solution was
evaporated, taken up in a minimum of CH2Cl2 and
/
10 with the bulky, soft electron
donor triphenylphosphine drives all the products of the
resulting adducts to a common structure while the
harder and less bulky ligands n-butylamine and trifluor-
oacetate result in a variety of isomeric adducts with a
range of fluxionality and stability. Indeed, although the
overall behavior of these electron deficient clusters
parallels the related m3-imidoyl triosmium clusters the
purified by thin layer chromatography using 20ꢀ
CH2Cl2ꢀC6H14 as eluent. A single orange band was
eluted yielding 110ꢀ120 mg (90ꢀ95%) of 13ꢀ15.
/
40%
/
/
/
/
3.4. Analytical and spectroscopic data for 13
Anal. Calc. for C35H21N2O9Os3P: C, 34.59; H, 1.74;
N, 2.30. Found: C, 34.14; H, 1.90; N, 2.33%. IR (n CO)
in CH2Cl2: 2091 (m), 2051 (s), 2010 (s), 1990 (s), 1931
structures of some of the adducts of 1ꢀ
doubt for lack of solid state structural data that were
available for the imidoyls [9ꢀ11,13].
/10 remain in
1
/
(w) cmꢂ1. H NMR (CD2Cl2): d 9.21 (d, 1H), 8.45 (d,

E. Rosenberg et al. / Inorganica Chimica Acta 334 (2002) 343ꢀ
/354
353
1H), 7.85 (d, 1H), 7.35ꢀ
/
7.15 (m, 16H), 6.9 (dd, 1H),
3.9.1. Measurement of the formation constants for the
amine adducts
Compounds 1ꢀ
into an NMR tube and 0.75 mL of CDCl3, CD2Cl2 or
(CD3)2CO were added by syringe. A 10ꢀ20-fold molar
ꢂ
/
11.92 (d, J31Pꢀ1
/
Hꢁ15.6 Hz, hydride).
/
/
10 (20ꢀ30 mg) was weighed directly
/
3.5. Analytical and spectroscopic data for 14
/
excess of n-BuNH2 (25ꢀ50 mL) was added slowly by
/
Anal. Calc. for C35H23N2O9Os3P: C, 34.53; H, 1.90;
N, 2.30. Found: C, 35.93; H, 2.40; N, 2.05%. IR (n CO)
in CH2Cl2: 2084 (m), 2040 (s), 2000 (s), 1985 (br), 1933
syringe, the tube was shaken vigorously and the 1H
NMR were measured immediately, 5 h later and after 24
h. Equilibrium constants were calculated from the
1
(w) cmꢂ1. H NMR (CDCl3): d 8.83 (s, br 1H), 7.35ꢀ
/
1
relative intensity of the hydride H NMR signals and
7.15 (m, 15H), 6.93 (d, 1H), 6.51 (m, 2H), 2.62 (s, 3H),
12.65 (d, J31Pꢀ1
the known concentrations of 1ꢀ10 and n-BuNH2.
/
ꢂ
/
/
Hꢁ15.1 Hz, hydride).
/
3.10. Crystal structure analysis of 16
3.6. Analytical and spectroscopic data for 15
A suitable crystal of 16, obtained by crystallization
from a CH2Cl2 solution containing an excess of n-
BuNH2, was mounted on a glass fiber using superglue.
The crystal was placed in a nitrogen gas stream at 110 K
on a Bruker D8 SMART APEX CCD sealed tube
diffractometer with graphite monochromated Mo Ka
radiation. A sphere of data was measured using a series
of combinations of f- and v-scans with 10 s frame
exposures and 0.38 frame widths. Data collection
indexing and initial cell refinements were all handled
using SMART [14] software. Frame integration and final
cell refinements were carried out using SAINT [15]
software. The final cell parameters were determined
from least-squares refinement on a maximum of 9999
reflections. The SADABS [16] was used to carry out
absorption corrections.
Anal. Calc. for C34H21NO10Os3P: C, 33.77; H, 1.75;
N, 1.15. Found: C, 33.37; H, 1.77; N, 1.00%. IR (n CO)
in CH2Cl2: 2087 (m), 2048 (s), 2000 (s), 1979 (br), 1941
(w) cmꢂ1. 1H NMR (CD2Cl2): d 8.29 (s, 1H), 7.15ꢀ
/
7.38
(m, 15H), 6.99 (d, 1H), 6.78 (d, 1H), 6.65 (dd, 1H),
13.10 (d, J31Pꢀ1
Hꢁ14.4).
ꢂ
/
/
/
3.7. Synthesis of the amine adducts 16 and 17
Compound 2 or 5 (50 mg, ꢁ
dissolved in 5 mL CH2Cl2 and 50 ml n-BuNH2 (ꢁ
/
0.05 mmol) were
20-
/
fold molar excess) in 5 ml C6H14 was added slowly to the
solution under a nitrogen atmosphere. The dark green
solution gradually turned to an amber yellow. After 1 h
the solution was concentrated to half volume under a
The structure was solved using direct methods and
difference Fourier techniques (SHELXTL, V5.10) [17].
Hydrogen atoms were placed in their expected chemical
positions using the HFIX command and were included
in the final cycles of least-squares with isotropic Uij’s
stream of nitrogen and stored at ꢂ
Yellow orange crystal of the adducts precipitated (20ꢀ
30 mg, 40ꢀ50% yield).
/
20 8C overnight.
/
/
related to the atoms ridden upon. The CÃ
/
H distances
3.8. Analytical and spectroscopic data for 16
were fixed at 0.93, 0.98 (CH4), 0.97 (methylene), or 0.96
˚
A (methyl). The position of the hydride was calculated
Anal. Calc. for C26H20N2O9Os3: C, 29.05; H, 1.88; N,
2.60. Found: C, 29.66; H, 2.20; N, 3.13%. IR (n CO) in
CH2Cl2: 2097 (w), 2073 (w), 2018 (s, br), 1972 (m), 1935
using the program HYDEX [12]. All non-hydrogen atoms
were refined anisotropically except for the n-BuNH2
ligand atoms. Scattering factors and anomalous disper-
sion corrections are taken from the International Tables
for X-ray Crystallography [18]. Structure, solution
refinement, graphics and generation of publication
materials were performed using SHELXTL, V5.10 [17]
software. Additional details of data collection and
structure refinement are given in Table 4.
1
(s, br) cmꢂ1. H NMR (CDCl3, major isomer only): d
9.10 (s, 1H), 8.60 (d, 1H), 8.27 (d, 1H), 8.10 (d, 1H), 7.99
(d, 1H), 7.85 (dd, 1H), 7.67 (dd, 1H), 7.31 (d, 1H), 6.78
(br, 2H), 4.58 (m, br, 1H), 4.04 (m, br), 3.40 (m, br, 2H),
2.88 (t, br, 3H), 2.76 (m, br, 2H), ꢂ
/
13.23 (s, br,
hydride).
3.9. Analytical and spectroscopic data for 17
4. Supplementary material
Anal. Calc. for C21H18N3O9Os3: C, 24.54; H, 1.85; N,
4.09. Found: C, 25.16; H, 2.22; N, 4.18%. IR (n CO) in
Crystallographic data for the structural analysis have
been deposited with the Cambridge Crystallographic
Data Centre, CCDC No. 158306 for compound 16.
Copies of this information may be obtained free of
charge from The Director, CCDC, 12 Union Road,
1
CH2Cl2: cmꢂ1. H NMR (CD2Cl2, major isomer only):
d 9.45 (s, br, 1H), 7.65 (dd, 1H), 6.94 (m, 2H), 2.75 (s,
3H), 2.12 (m, 2H), 1.56 (s, br, 2H), 1.38 (m, 2H), 0.89
(m, 2H), 0.56 (t, 3H), 0.25 (m, 2H), ꢂ11.78 (s, hydride).
/

354
E. Rosenberg et al. / Inorganica Chimica Acta 334 (2002) 343ꢀ354
/
Roy, S. Alam, K.A. Azam, Coord. Chem. Rev. 190ꢀ
975.
/
192 (1999)
Cambridge, CB2 1EZ, UK (fax: ꢃ44-1223-336-033; e-
mail: deposit@ccdc.cam.ac.uk or www: http://
www.ccdc.cam.ac.uk).
/
[6] R. Smith, E. Rosenberg, K.I. Hardcastle, V. Vazquez, J. Roh,
Organometallics 18 (1999) 3519.
[7] S.T. Beatty, B. Bergman, E. Rosenberg, W. Dastru’, R. Gobetto,
L. Milone, A. Viale, J. Organomet. Chem. 593ꢀ594 (1999) 226.
/
[8] E. Rosenberg, L. Milone, C. Nervi, D. Osella, J. Fiedler, Md.J.
Abedin, D. Rokhsana, Inorg. Chim. Acta 300 (2000) 766.
[9] E. Rosenberg, K.I. Hardcastle, S.E. Kabir, L. Milone, R.
Gobetto, M. Botta, N. Nishimura, M. Yin, Organometallics 14
(1995) 3068.
Acknowledgements
We gratefully acknowledge the support of the Na-
tional Science Foundation (E.R., CHE-9625367), The
University of Torino (E.R. and L.M.). Financial sup-
port from MURST and CNR (L.M. and R.B.) is also
gratefully acknowledged.
[10] E. Rosenberg, S.E. Kabir, M. Day, K.I. Hardcastle, E. Wolf, T.
McPhillips, Organometallics 14 (1995) 721.
[11] E. Rosenberg, D. Espitia, S.E. Kabir, T. McPhillips, M.W. Day,
K.I. Hardcastle, R. Gobetto, D. Osella, L. Milone, Organome-
tallics 10 (1991) 3550.
[12] A.G. Orpen, J. Chem. Soc., Dalton. Trans. (1980) 2509.
[13] E. Rosenberg, S.E. Kabir, M. Day, K.I. Hardcastle, Organome-
tallics 13 (1994) 4437.
References
[1] E. Rosenberg, S.E. Kabir, D. Kolwaite, K.I. Hardcastle, W.
Crosswell, J. Grindstaff, Organometallics 14 (1995) 3611.
[2] E. Rosenberg, E. Arcia, D.S. Kolwaite, K.I. Hardcastle, J.
Ciurash, R. Duque, R. Gobetto, L. Milone, D. Osella, M. Botta,
W. Dastru`, A. Viale, J. Fiedler, Organometallics 17 (1998) 415.
[3] E. Rosenberg, B. Bergman, A. Bar Din, R. Smith, R. Gobetto, L.
Milone, A. Viale, W. Dastru’, Polyhedron 17 (1998) 2975.
[4] B. Bergman, R.H. Holmquist, R. Smith, E. Rosenberg, K.I.
Hardcastle, M. Visi, J. Ciurash, J. Am. Chem. Soc. 120 (1998)
12818.
[14] SMART Version 5.55, Bruker AXS Inc., Analytical X-ray Systems,
Madison, WI, 2000.
[15] SAINT version 6.02, Bruker AXS Inc., Analytical X-ray Systems,
Madison, WI, 1999.
[16] SADABS, George Sheldrick, University of Gottingen
¨
,
Gottingen,
¨
Germany, 1996.
[17] SHELXTL V5.10, Bruker AXS Inc., Analytical X-ray Systems,
Madison, WI, 1997.
[18] A.J.C. Wilson (Ed.), International Tables of X-ray Crystal-
lography, vol. C, Kluwer Academic, Dordrecht, 1992, Tables
[5] J. Abedin, B. Bergman, R. Holmquist, R. Smith, E. Rosenberg, J.
Ciurash, K. Hardcastle, J. Roe, V. Vazquez, C. Roe, S. Kabir, B.
6.1.1.4 (pp. 500ꢀ
/
5020) and 4.2.6.8 (pp. 219ꢀ222).
/