Heterosite effects in novel heteronuclear clusters [Os 2Ru(CO)11(PPh3)] and [Os2Ru(CO) 10(2-acetylpyridine-N-isopropylimine)]

FULL PAPER

Heterosite Effects in Novel Heteronuclear Clusters [Os₂Ru(CO)₁₁(PPh₃)] and

[Os₂Ru(CO)₁₀(2-acetylpyridine-N-isopropylimine)]

Frank W. Vergeer,^[a,e]Martin Lutz,^[b]Anthony L. Spek,^[b]Maria J. Calhorda,^[c,d]

[a]

Derk J. Stufkens,^[†][a]and Frantisek Hartl*

ˇ

Keywords: Cluster compounds / Density functional calculations / Time-resolved spectroscopy / Photochemistry / Spectro-

electrochemistry

A new synthetic route towards the mixed-metal cluster [Os₂-

Ru(CO)₁₂] is described together with the syntheses of its PPh₃

and iPr-AcPy (iPr-AcPy = 2-acetylpyridine-N-isopropylimine)

derivatives. The molecular structures of the novel clusters

[Os₂Ru(CO)₁₁(PPh₃)] and [Os₂Ru(CO)₁₀(iPr-AcPy)] were de-

termined on the basis of crystalline solid solutions of the

Os₂Ru and corresponding Os₃species. The structures reveal

that coordination of the Lewis bases occurs exclusively at the

ruthenium site of [Os₂Ru(CO)₁₂], which is in agreement with

density functional theory (DFT) calculations on several struc-

tural isomers of these compounds. According to the time-de-

pendent DFT results, the lowest optically accessible excited

state of [Os₂Ru(CO)₁₀(iPr-AcPy)] has a prevailing σ(Ru–Os2)-

π*(iPr-AcPy) character, with a partial σσ*(Ru–Os2) contri-

bution. In weakly coordinating 2-chlorobutane, the excited

state has a lifetime τ = 10.4 1.2 ps and produces biradicals

considerably faster than observed for [Os₃(CO)₁₀(iPr-AcPy) (τ

= 25.3 0.7 ps). In coordinating acetonitrile, the excited state

of [Os₂Ru(CO)₁₀(iPr-AcPy)] decays mono-exponentially with

a lifetime τ = 2.1 0.2 ps. In contrast to [Os₃(CO)₁₀(iPr-AcPy)]

that forms biradicals as the main primary photoproduct even

in strongly coordinating solvents, zwitterion formation from

the solvated lowest excited state is observed for the hetero-

metallic cluster. This is concluded from time-resolved absorp-

tion studies in the microsecond time domain. Due to the

lower tendency of the coordinatively unsaturated ⁺Ru(CO)₂-

(iPr-AcPy^·–/0) moiety to bind a Lewis base, the heteronuclear

biradical and zwitterionic photoproducts live significantly

shorter than their triosmium counterparts. The influence of

the weaker Os2–Ru(iPr-AcPy) bond on the redox reactivity is

clearly reflected in very reactive radical anions formed upon

electrochemical reduction of [Os₂Ru(CO)₁₀(iPr-AcPy)]. The

2–

dimer [^–Os(CO)₄-Os(CO)₄-Ru(CO)₂(iPr-AcPy)]₂is the only

IR-detectable intermediate reduction product. The dinuclear

complex [Os₂(CO)₈]^2–and insoluble [Ru(CO)₂(iPr-AcPy)]_nare

the ultimate reduction products, proving fragmentation of the

Os₂Ru core.

Germany, 2005)

nects two different core metal atoms, has mainly originated

from their possible application in both homogeneous and

heterogeneous catalysis.^[5]Apart from the general expecta-

tion that different transition metal atoms in close proximity

of each other could initiate novel reactions by synergistic

interactions, the different reactivity of adjacent metal

centres in mixed-metal clusters may provide additional bi-

or multifunctional activation pathways and increase the se-

lectivity of substrate–cluster interactions.^[6]Although catal-

ysis by mixed-metal clusters in a number of cases indeed

resulted in higher catalytic activity^[7,8]or different product

selectivity^[9,10]than observed for their homonuclear ana-

logues, the mechanistic insight into the role of the different

metal centres in the catalytic cycle is generally limited. This

problem mainly has its origin in scarcely available series of

isostructural mixed-metal clusters, which precludes a sys-

tematic study of their bonding properties and reactivity.

One of the best known series of mixed-metal clusters is

the family of group 8 triangular clusters [M₃(CO)₁₂] (M =

Fe, Ru, Os), where all possible metal combinations – except

[FeRuOs(CO)₁₂] – have been prepared.^[3,11]Within this

Introduction

In the last four decades, considerable efforts have been

made to develop general synthetic routes towards hetero-

nuclear transition metal clusters.^[1–4]The interest in this

type of clusters where at least one metal–metal bond con-

[†] Deceased.

[a] Van’t Hoff Institute for Molecular Sciences, University of Am-

sterdam,

Nieuwe Achtergracht 166, 1018 WV, Amsterdam, The Nether-

lands

Fax: +31-20-525-6456

E-mail: f.hartl@uva.nl

[b] Department of Crystal and Structural Chemistry, Utrecht Uni-

versity,

Padualaan 8, 3584 CH Utrecht, The Netherlands

[c] Instituto de Tecnologia Química e Biológica, Av. da República,

EAN,

Apart. 127, 2781-901 Oeiras, Portugal

[d] Departamento de Química e Bioquímica, Faculdade de Ciên-

cias, Universidade de Lisboa,

1749-016 Lisboa, Portugal

[e] Current address: Physikalisches Institut, Westfälische Wilhelms-

Universität Münster,

Wilhelm-Klemm-Strasse 10, 48149 Münster, Germany

2206

DOI: 10.1002/ejic.200401040

Eur. J. Inorg. Chem. 2005, 2206–2222

Heterosite Effects in Novel Heteronuclear Clusters

FULL PAPER

series, the clusters [Os₂Ru(CO)₁₂] and [Ru₂Os(CO)₁₂] are of

particular interest for our investigations, as the bonding

properties, photochemistry and redox behaviour of the

homonuclear dodecacarbonyl analogues and, in particular,

their α-diimine-substituted derivatives [M₃(CO)₁₀(α-diim-

Results and Discussion

Synthesis of [Os₂Ru(CO)₁₂] (1)

The heteronuclear cluster [Os₂Ru(CO)₁₂] (1) was synthe-

ine)] (M = Ru, Os), have been studied thoroughly in our sized from its homonuclear analogues [Os₃(CO)₁₂] and

groups.^[12–21]

[Ru₃(CO)₁₂] in a four-step reaction sequence (Scheme 1).

A continued systematic study of their heterometallic Starting from [Os₃(CO)₁₂], the linear cluster [Os₃(CO)₁₂-

Ru/Os derivatives is desired in order to reveal the influence (Br)₂] was obtained by oxidation with Br₂in refluxing

of the cluster core variation on the photochemistry and CH₂Cl₂[step (i)]. After the addition of Br₂, rapid removal

redox reactivity. However, one of the underlying reasons for of the heating bath is required in order to prevent the reac-

the lacking studies of [Os₂Ru(CO)₁₂] and [Ru₂Os(CO)₁₂] is tion of [Os₃(CO)₁₂(Br)₂] with additional Br₂producing at

the serious difficulty to obtain these clusters in the pure this stage undesired [Os₂(CO)₈(Br)₂] that is known to trans-

form. Although several synthetic procedures have been re- form into [Os₂(CO)₆(Br)₂] at elevated temperatures.^[27]Sim-

ported, they yielded either hardly separable mixtures of ilar oxidative addition of X₂(X = Cl, Br, I) to [Ru₃(CO)₁₂]

[Ru₃(CO)₁₂], [Ru₂Os(CO)₁₂], [Os₂Ru(CO)₁₂] and [Os₃- was reported^[28,29]to yield the mononuclear complex cis-

(CO)₁₂] in a nearly statistical 1:2:2:1 ratio^[22–24]or gave the [Ru(CO)₄(X)₂] as the initial product, which reflects the in-

desired pure heterometallic clusters merely in very low creasing stability of the metal–metal bonds towards oxi-

yields (Յ 5%).^[25,26]Development of efficient novel syn- dation on descending the periodic table. Unfortunately, the

thetic routes toward these species is therefore compulsory.

synthetic strategy depicted in Scheme 1 could therefore not

Aimed at performing a comparative study of the hetero- be applied for the synthesis of the isostructural cluster

nuclear cluster [Os₂Ru(CO)₁₀(α-diimine)] and its homonu- [Ru₂Os(CO)₁₂]. The controlled formation of the dinuclear

clear analogues [M₃(CO)₁₀(α-diimine)] (M = Ru, Os), a complex [Os₂(CO)₈(Br)₂] was achieved by slight modifica-

novel route for the synthesis of [Os₂Ru(CO)₁₂] (1) has been tion of the procedure described by Moss et al. [step (ii)].^[30]

introduced, producing the cluster in a reasonable yield. Visible irradiation of [Os₃(CO)₁₂(Br)₂] at λ Ͼ 420 nm was

Apart from the parent carbonyl cluster, we also report the found to prevent excessive formation of the mononuclear

syntheses and crystal structures of the substituted deriva- complex cis-[Os(CO)₄(Br)₂]. The latter compound is not

tives [Os₂Ru(CO)₁₁(PPh₃)] (2) and [Os₂Ru(CO)₁₀(iPr- only obtained as an unavoidable side-product formed dur-

AcPy)] (3) (iPr-AcPy = 2-acetylpyridine-N-isopropylimine), ing irradiation of [Os₃(CO)₁₂(Br)₂], but it can also be

schematically depicted in Figure 1. Both derivatives were formed from [Os₂(CO)₈(Br)₂] by UV/Vis irradiation of this

prepared in order to establish the preferential coordination complex in the presence of Br₂. In a third step, the powerful

site of different Lewis bases in substitution reactions with nucleophile Na₂[Ru(CO)₄] was prepared from [Ru₃(CO)₁₂]

[Os₂Ru(CO)₁₂]. The influence of the heteronuclear cluster by reduction with sodium in liquid ammonia [step (iii)].^[31]

core on the photo- and electrochemical reactivity of [Os₂- As [Ru(CO)₄]^2–is extremely air- and moisture-sensitive and

Ru(CO)₁₀(iPr-AcPy)] in comparison with the homonuclear readily converts to [Ru(CO)₄H]^–, the reduction of [Ru₃-

derivatives is also described. Density functional theory (CO)₁₂] was performed at reduced pressure (3×10^–4Pa) on

(DFT) calculations provide direct insight into the bonding a high-vacuum line. After the completion of the reaction,

properties of [Os₂Ru(CO)₁₁(PPh₃)] and [Os₂Ru(CO)₁₀(iPr- the product must be dried extensively, as small traces of

AcPy)] and support the discussion of the experimental re- ammonia proved to react instantaneously with [Os₂(CO)₈-

sults.

(Br)₂], reducing the yield of [Os₂Ru(CO)₁₂]. In order to pre-

vent thermal decomposition of Na₂[Ru(CO)₄] during the

drying process, the temperature was carefully kept below

273 K. Finally, Na₂[Ru(CO)₄] was extracted in thoroughly

dried THF and added to [Os₂(CO)₈(Br)₂] by vacuum fil-

tration, instantaneously producing [Os₂Ru(CO)₁₂] (1) in

reasonable yields (step (iv)).

Unfortunately, in some cases, cluster 1 was found to con-

tain variable amounts of [Os₃(CO)₁₂] and/or [Ru₂Os(CO)₁₂]

impurities. The latter cluster could result from incomplete

reduction of [Ru₃(CO)₁₂] by less than six equivalents of so-

dium. Thus, although the number of impurities and their

percentages are significantly reduced in comparison with

the established synthetic procedures,^[22–24]the purity of

cluster 1 and its substitution products remains difficult to

control, even when employing this novel synthetic route.

In order to establish the preferred coordination site in

cluster 1 for different Lewis bases (L), substituted deriva-

tives with monodentate PPh₃and chelating α-diimine = iPr-

Figure 1. Schematic structures of the investigated clusters [Os₂-

Ru(CO)₁₁(L)] [L = CO (1), PPh₃(2)] and [Os₂Ru(CO)₁₀(NN)] [NN

= iPr-AcPy (3)].

Eur. J. Inorg. Chem. 2005, 2206–2222

www.eurjic.org

2207

F. W. Vergeer, M. Lutz, A. L. Spek, M. J. Calhorda, Derk J. Stufkens, F. Hartl

FULL PAPER

Figure 2. Molecular structure of the cluster [Os₂Ru(CO)₁₁(PPh₃)]

(2) in the crystal. Displacement ellipsoids are drawn at the 50%

probability level. The site of Ru3 is occupied for 64% by Ru and

for 36 % by Os.

Scheme 1. Synthesis of [Os₂Ru(CO)₁₂] (1).

AcPy ligands were prepared by a thermal reaction route

involving precursor clusters [Os₂Ru(CO)_12–n(MeCN)_n] (n =

1 for L = PPh₃; n = 2 for L = iPr-AcPy). The substituted

products [Os₂Ru(CO)₁₁(PPh₃)] (2) and [Os₂Ru(CO)₁₀(iPr-

AcPy)] (3) (see Figure 1) were characterised by IR and

NMR spectroscopy and mass spectrometry. Their molecu-

lar structures were determined by single-crystal X-ray dif-

fraction and compared with the optimised geometries re-

sulting from DFT calculations.

Table 1. Selected bond lengths [Å] and bond angles [°] for cluster

2, with standard uncertainties in parentheses. For numbering see

Figure 2. The site of Ru3 is occupied for 64% by Ru and for 36%

by Os.

Ru3–Os1

Ru3–Os2

Os1–Os2

Ru3–P1

Ru3–C9

Ru3–C10

Ru3–C11

P1–C12

P1–C18

P1–C24

2.88393(19)

2.9147(2)

2.88275(16)

2.3683(7)

1.939(3)

1.940(3)

1.887(3)

1.842(3)

1.831(3)

1.826(3)

Os1–Ru3–Os2 59.620(4)

Os1–Os2–Ru3 59.660(4)

Ru3–Os1–Os2 60.721(5)

Ru3–P1–C12

Ru3–P1–C18

Ru3–P1–C24

C12–P1–C18

C18–P1–C24

C12–P1–C24

116.70(9)

114.07(9)

116.04(9)

104.64(14)

103.18(13)

100.30(13)

Molecular Structure of [Os₂Ru(CO)₁₁(PPh₃)] (2)

The crystal structure of cluster 2 is shown in Figure 2.

This structure was determined on a mixture of [Os₂Ru-

(CO)₁₁(PPh₃)] and [Os₃(CO)₁₁(PPh₃)] in a ratio of 0.64:0.36.

Selected bond lengths and angles are presented in Table 1.

The structure of 2 exhibits disorder amounting to a 36% π-accepting CO ligand by the much stronger σ-donor PPh₃,

Os occupancy at the Ru atom site, being significantly increasing the electron density on the metal core and re-

distorted from the D_3hsymmetry observed for [M₃(CO)₁₂] sulting in expansion. Like in the corresponding homonu-

(M = Ru, Os).^[32,33]Most importantly, the crystal structure, clear clusters [M₃(CO)₁₁(PPh₃)] (M = Os, Ru),^[35,36]the

in combination with (temperature-dependent) ³¹P NMR length of the metal–metal bond positioned cis to the PPh₃

spectra, reveals that the PPh₃ligand in [Os₂Ru(CO)₁₁- ligand [Ru–Os2: 2.9147(2) Å] is increased to a larger extent

(PPh₃)] occupies an equatorial position and is exclusively than the other metal–metal bond lengths [Ru–Os1:

coordinated at Ru. This is in contrast with the results of 2.8839(2) Å; Os1–Os2: 2.8827(2) Å]. This difference has

Pereira et al. who obtained an 1:2 mixture of osmium- and been attributed to steric interactions between the PPh₃li-

ruthenium-substituted isomers after the reaction of PPh₃gand and the cis-CO group on the adjacent metal

with the tetranuclear mixed-metal cluster [RuOs₃(μ-H)₂- atom.^[35,36]The steric and electronic effects are also re-

(CO)₁₃] under similar conditions.^[34]

flected in the marked shortening of the M–CO_eqbond cis

The geometry around the phosphorus atom in cluster 2 to the PPh₃ligand [Ru–C11 = 1.887(3) Å] compared to the

is essentially tetrahedral, the phenyl groups being slightly average length of these bonds in, for example, [Ru₃(CO)₁₂]

bent away from the Ru atom [C–P–C angles are in the range [average Ru–CO_eq: 1.921(5) Å]. This shortening is ascribed

100.30(13)–104.64(14)°]. The average metal–metal bond to increased π-back-bonding to the CO ligand, resulting

length (2.894 Å) is longer than the unsubstituted homonu- from the presence of the σ-donating phosphorus ligand at-

clear clusters [M₃(CO)₁₂] [M = Ru: 2.8542 Å;^[33]M = Os: tached to the same metal and to the sterically induced

2.8771(27) Å^[32]]. This is partly due to the substitution of a lengthening of the Ru–Os2 bond trans to this CO group.^[35]

2208

www.eurjic.org

Eur. J. Inorg. Chem. 2005, 2206–2222

Heterosite Effects in Novel Heteronuclear Clusters

FULL PAPER

The Ru–Os bond lenghts [2.9147(2) and 2.8839(2) Å] nicely adjacent metal atom are much less pronounced compared

fit within the values reported for the corresponding metal– to the bulky PPh₃ligand. Apart from this feature, the twist-

metal bond lengths in the homonuclear analogues [M₃- ing of the Os(CO)₄units in 2a is much smaller than in the

(CO)₁₁(PPh₃)] [M = Os: 2.918(1) and 2.891(1) Å, M = Ru: experimental structure. This is reflected in the orientation

2.907(3) and 2.876(3) Å]. This comparison is possible even of the Os–CO_axbonds in 2a, which are almost perpendicu-

though the reported Os–Ru distances do not correspond to lar to the Os₂Ru plane (dihedral angles C1–Os1–Os2–C5 =

“pure” [Os₂Ru(CO)₁₁(PPh₃)], but to a blend of Os₂Ru and 0.3° and C2–Os1–Os2–C6 = 0.5°). The larger distortion in

Os₃in solid solution (vide supra). The distortion from the the experimental structure has its origin presumably in the

D_3hsymmetry, manifested by the twisting of the Os(CO)₄phenyl groups of the PPh₃ligand. In the solid state, these

units, is reflected in the C_ax–Os1–Os2–C_axdihedral phenyl rings interact with the phenyl rings of the two neigh-

angles.^[35]Just like for the metal–metal bond lengths, the bouring clusters by four edge-to-face interactions to give a

values for C1–Os1–Os2–C5 (3.5°) and C2–Os1–Os2–C6 one-dimensional chain. Similar interactions, for example,

(12.8°) closely resemble those reported for the homonuclear have been reported for the cluster [Os₃(CO)₁₁{P(p-

analogues: 5.4° and 14.4° for Ru and 4.4° and 12.3° for Os, C₆H₄F)₃}].^[44]Besides the crystal packing effects, not incor-

respectively.

porated in the geometry optimisation of 2a, the twist of the

The presence of [Os₃(CO)₁₁(PPh₃)], which was refined M(CO)₄units also depends on the selected computational

with a substitutional disorder model for the Ru/Os3 atom basis set.^[19]

site, has most likely its origin in the formation of [Os₃-

(CO)₁₂] as a side-product during the preparation of 1 (vide

supra). This is also reflected in the mass spectra of some of

the [Os₂Ru(CO)₁₂] samples where a peak at m/z 908 is as-

cribed to the molecular ion of [Os₃(CO)₁₂]. The presence of

ca. 35% [Os₃(CO)₁₁(PPh₃)] is also reflected in the ³¹P{H}

NMR and FD mass spectra of 2, indicating that the ob-

served ratio between the Os₃and Os₂Ru phosphane-substi-

tuted clusters in the crystal structure is similar in the gas

phase and in solution.

Density Functional Study of [Os₂Ru(CO)₁₁(PPh₃)] (2)

The density functional theory (DFT, ADF pro-

gramme^[37–43]) study of cluster 2 was performed with the

aim to learn, whether the theory indeed predicts [Os₂-

Ru(CO)₁₁(PPh₃)] to be most stable with the PPh₃ligand

coordinated at the ruthenium centre at an equatorial site.

The cluster [Os₂Ru(CO)₁₁(PH₃)] (2a) served as a model for

2. Some features of the optimised geometry of 2a slightly

deviate from those of the crystal structure of 2 (Table 2,

Figure 3). First of all, the calculated metal–metal bond

lengths are slightly longer than the experimental ones (up

to 0.05 Å), similar to the situation for [Os₃(CO)₁₂].^[19,20]The

Ru–Os bond cis to the PPh₃ligand is calculated to be signif-

icantly longer than the other two metal–metal bonds, in

agreement with the crystal structure. This trend points to

the role of electronic effects apart from the expected steric

influence. For, the differences in the metal–metal bond

lengths should be smaller in the model 2a, as the steric in-

Figure 3. Optimised geometries and relative energies of the model

teractions of the PH₃ligand with the cis-CO group on the cluster 2a and its virtual isomers 2b–e.

Table 2. Comparison of selected ADF/BP calculated bond lengths [Å] and angles [°] in cluster 2a with the experimental crystallographic

data.

Bond^[a]

Calcd.

Exp.

Angle^[a]

Calcd.

Exp.

Ru–Os1

Ru–Os2

Os1–Os2

Ru–P

2.913

2.956

2.933

2.339

2.88393(19)

2.9147(2)

2.88275(16)

2.3683(7)

Os1–Ru–Os2

Os1–Os2–Ru

Ru–Os1–Os2

60.0

59.3

60.8

59.620(4)

59.660(4)

60.721(5)

[a] See Figure 2.

Eur. J. Inorg. Chem. 2005, 2206–2222

www.eurjic.org

2209

F. W. Vergeer, M. Lutz, A. L. Spek, M. J. Calhorda, Derk J. Stufkens, F. Hartl

FULL PAPER

In addition to the optimised model 2a, the relative ener-

gies of several structural isomers 2b–e have also been calcu-

lated (Figure 3). The isomers 2a and 2b have in common

that the PH₃ligand is coordinated to ruthenium, but they

differ in the coordination of the ligand relative to the Os₂Ru

plane (equatorial and axial, respectively). In isomers 2c–e,

the PH₃ligand is coordinated at one of the osmium atoms,

2e being the only model with the PH₃ligand coordinated

in an axial position. Isomers 2c and 2d have the PH₃ligand

coordinated in an equatorial position, but they differ in the

orientation of the PH₃ligand relative to the ruthenium cen-

tre. Importantly, isomer 2a appeared to be more stable by

11.4 kJ·mol^–1than isomer 2d, which proves the preferable

coordination of the PH₃ligand at ruthenium. Isomer 2c,

with the PH₃ligand perpendicular to the Os1–Os2 bond, is

in turn slightly more stable than isomer 2d, where the posi-

tion of the PH₃ligand presents almost a continuation of

the Os1–Os2 bond. The isomers 2b and 2e, with the PH₃

ligand occupying an axial site at Ru and Os, respectively,

are highest in energy.

Figure 4. Molecular structure of the cluster [Os₂Ru(CO)₁₀(iPr-

AcPy)] (3) in the crystal. Displacement ellipsoids are drawn at the

50% probability level. The site of Ru1 is occupied for 89% by Ru

and for 11% by Os.

Although the difference in relative energy between the

model structures 2a and 2c is relatively small, no experi-

mental evidence has been obtained for [Os₂Ru(CO)₁₁-

(PPh₃)] clusters with the phosphane ligand coordinated to

osmium. However, as the relative energies calculated by

DFT refer to isolated gas-phase molecules at zero Kelvin

temperature, with PH₃as the model for the PPh₃ligand,

the obtained values merely reflect a trend in the thermo-

dynamic stability of the different isomers in solution at

293 K.

In order to investigate whether a chelating α-diimine li-

gand also prefers coordination at the ruthenium site of 1,

we synthesized the cluster [Os₂Ru(CO)₁₀(iPr-AcPy)] (3) and

determined its molecular geometry. Apart from its struc-

ture, the major interest in cluster 3 stems from our aim to

reveal the influence of the heteronuclear cluster core on the

photo- and redox reactivity of this compound. For a proper

evaluation of this effect, the homonuclear clusters [Os₃-

(CO)₁₀(α-diimine)] were used as a reference.

Table 3. Selected bond lengths [Å] and angles [°] for cluster 3, with

standard uncertainties in parentheses. For numbering see Figure 4.

The Ru1 site is occupied for 89% by Ru and for 11% by Os.

Ru1–Os3

Ru1–Os2

Os2–Os3

Ru1–N1

Ru1–N2

Ru1–C9

Ru1–C10

N2–C11

C11–C12

C12–C13

C13–C14

C14–C15

C15–C16

2.8694(4)

2.9159(4)

2.8815(2)

2.141(4)

2.151(3)

1.846(5)

1.855(5)

1.343(5)

1.375(6)

1.380(6)

1.385(6)

1.383(6)

1.470(6)

N1–C16

N2–C15

1.307(5)

1.358(5)

Os3–Os2–Ru1

Os2–Os3–Ru1

Os3–Ru1–Os2

N1–Ru1–Os3

N2–Ru1–Os2

N1–Ru1–C9

N2–Ru1–C9

N1–Ru1–N2

N1–C16–C15

N2–C15–C16

59.328(8)

60.934(8)

59.738(8)

157.77(10)

101.98(9)

97.02(16)

171.47(16)

75.25(13)

115.7(4)

115.2(4)

Just as for cluster 2, the observed substitutional disorder

due to a small amount of the triosmium analogue most

likely has its origin in the formation of [Os₃(CO)₁₂] as a

side-product during the preparation of 1. Apart from the

crystal structure, the presence of a small amount of

[Os₃(CO)₁₀(iPr-AcPy)] is also reflected in the FAB⁺mass

spectrum of 3 where peaks at m/z 986.9 and 957.9 can be

ascribed to [M⁺– nCO] (M = [Os₃(CO)₁₀(iPr-AcPy)]; n =

Molecular Structure of [Os₂Ru(CO)₁₀(iPr-AcPy)] (3)

The molecular structure of cluster 3, determined by sin-

gle-crystal X-ray diffraction, is shown in Figure 4. Selected

bond lengths and angles are collected in Table 3. As in com-

pound 2, the corresponding Os₃compound, viz. [Os₃-

(CO)₁₀(iPr-AcPy)], was also present in the crystal structure

as a solid solution (occupancy 0.89:0.11 for Os₂Ru:Os₃).

Importantly, the crystal structure reveals that the iPr-AcPy

ligand in [Os₂Ru(CO)₁₀(iPr-AcPy)] is again bound at the

ruthenium atom. Similar to 2,2Ј-bipyridine (bpy) and N,NЈ-

diisopropyl-1,4-diaza-1,3-butadiene (iPr-DAB) in the

closely related [Os₃(CO)₁₀(α-diimine)] clusters,^[45,46]the

asymmetric iPr-AcPy ligand is coordinated to a single metal

centre in a chelating fashion. The nitrogen atom of the pyri-

dine ring and the imine nitrogen occupy axial and equato-

rial positions, respectively.

1

1, 2). Moreover, the H NMR spectrum of 3 in CDCl₃dis-

plays small signals at δ = 9.50 (d, 1 H), 8.02 (d, 1 H), 7.86

(dd, 1 H), 7.11 (dd, 1 H) and 4.44 (m, 1 H) ppm, pointing

to the presence of ca. 10–15% [Os₃(CO)₁₀(iPr-AcPy)] (esti-

mated from signal integrals). This proves that the ratio be-

tween [Os₂Ru(CO)₁₀(iPr-AcPy)] and [Os₃(CO)₁₀(iPr-AcPy)]

in solution is very close to that in the solid state.

Density Functional Study of [Os₂Ru(CO)₁₀(iPr-AcPy)] (3)

DFT calculations (ADF programme^[37–43]) were per-

formed in order to obtain more insight into the bonding

2210

www.eurjic.org

Eur. J. Inorg. Chem. 2005, 2206–2222

Heterosite Effects in Novel Heteronuclear Clusters

FULL PAPER

properties of 3 and the influence of the Ru atom on the is the replacement of the isopropyl group by a hydrogen

character of the frontier orbitals. The cluster [Os₂Ru- atom, shortening the Ru–N1 bond as a result of different

(CO)₁₀(H-PyCa)] (H-PyCa = pyridine-2-carbaldehyde-imine) steric and electronic effects.

(3Ј) was taken as a model, with the isopropyl and imine

From the ground-state DFT calculations, the composi-

methyl groups of the iPr-AcPy ligand replaced by hydrogen tion of the molecular orbitals of 3Ј has been obtained. The

atoms. In order to determine whether the theoretical calcu-

lations indeed predict the coordination of iPr-AcPy at the

ruthenium centre instead of the osmium sites, the relative

Table 5. Characters and one-electron energies of selected frontier

orbitals of [Os₂Ru(CO)₁₀(H-PyCa)] (3Ј) as calculated by the ADF/

BP method (L = LUMO, H = HOMO).

energies of the structural isomers 3Ј and 3ЈЈ were calculated,

with 3Ј having the α-diimine coordinated at ruthenium and

3ЈЈ at osmium (Figure 5). In both model systems, H-PyCa

is coordinated with the pyridine ring in the axial position.

A recent DFT study of the related cluster [Os₃(CO)₁₀(H-

PyCa)]^[18]has shown that this geometric isomer is more

stable than those with the pyridine ring equatorially bound

or with both nitrogen atoms coordinated in the equatorial

plane.

MO

E [eV] Ru

Os1

Os2

H-PyCa CO

111a L+2 –2.70

110a L+1 –3.04

0.5

23.4

8.4

25.5

24.4

38.3

52.8

35.6

0.1

7.3

1.1

1.0

21.9

16.1

7.0

9.6

0.6

4.5

6.5

16.7

17.1

3.2

10.9

19.8

87.7

9.5

64.7

24.9

3.4

7.8

5.8

6.0

44.6

14.4

25.4

26.7

30.2

17.1

26.5

109a

108a

L

H

–3.68

–5.28

107a H-1 –5.77

106a H-2 –5.84

105a H-3 –6.22

104a H-4 –6.46

1.8

Geometry optimisation of both isomers with DFT re-

vealed that 3Ј is more stable than 3ЈЈ by 22 kJ·mol^–1. This

is in agreement with the experimental structure, showing

that also the iPr-AcPy ligand prefers coordination at the

ruthenium site of [Os₂Ru(CO)₁₂]. The geometry of model

3Ј is in good agreement with the experimental structure of

3 (Table 4, Figure 5).

The relatively large difference between the lengths of the

two Ru–N bonds in 3Ј (2.05 vs. 2.17 Å) indicates that the

α-diimine ligand is coordinated in a more asymmetric fash-

ion than in the experimental structure. The probable reason

Figure 5. Optimised geometries and relative energies of the model

clusters 3Ј and 3ЈЈ.

Figure 6. 3D representations of the LUMO (a), HOMO (b),

HOMO-1 (c) and HOMO-2 (d) of [Os₂Ru(CO)₁₀(H-PyCa)] (3Ј).

Table 4. Comparison of selected ADF/BP-calculated bond lengths [Å] and angles [°] in cluster 3Ј with the experimental crystallographic

data.

Bond^[a]

Calcd.

Exp.

Bond/angle^[a]

Calcd.

Exp.

Ru–Os3

Ru–Os2

Os3–Os2

Ru–N1

2.896

2.928

2.913

2.049

2.168

1.356

1.385

1.405

1.383

1.412

1.426

2.8694(4)

2.9159(4)

2.8815(2)

2.141(4)

2.151(3)

1.343(5)

1.375(6)

1.380(6)

1.385(6)

1.383(6)

1.470(6)

N1–C16

N2–C15

1.314

1.382

1.307(5)

1.358(5)

Os3–Os2–Ru

Os2–Os3–Ru

Os3–Ru–Os2

N1–Ru–Os3

N2–Ru–Os3

N1–Ru–N2

59.5

60.5

60.0

152.9

96.8

75.2

116.5

113.9

59.328(8)

60.934(8)

59.738(8)

157.77(10)

97.41(9)

75.25(13)

115.7(4)

Ru–N2

N2–C11

C11–C12

C12–C13

C13–C14

C14–C15

C15–C16

N1–C16–C15

N2–C15–C16

115.2(4)

[a] See Figure 4.

Eur. J. Inorg. Chem. 2005, 2206–2222

www.eurjic.org

2211

F. W. Vergeer, M. Lutz, A. L. Spek, M. J. Calhorda, Derk J. Stufkens, F. Hartl

Table 6. TD-DFT calculated lowest-energy singlet excitation energies [eV] and oscillator strengths (O.S.) for [Os₂Ru(CO)₁₀(H-PyCa)].

FULL PAPER

Transition

Composition

Energy

[eV]

Wavelength

[nm]

Exp.

λ_max

O.S.

[a]

[nm]

1

2

3

4

5

6

7

8

63% (HǞL); 26% (H-1ǞL)

55% (H-1ǞL); 23% (H-2ǞL);

63% (H-2ǞL); 28% (HǞL+1)

49% (HǞL+1); 16% (H-3ǞL); 11% (HǞL)

76% (H-3ǞL)

85% (HǞL+2)

64% (H-4ǞL); 25% (H-1ǞL+1)

46% (H-2ǞL+1); 27% (H-1ǞL+1)

2.09

2.16

2.24

2.46

2.63

2.72

2.84

2.90

594

573

553

504

471

456

436

428

581

0.085

0.021

0.010

0.050

0.027

0.020

0.011

0.038

520

[b]

434

[a] Absorption maxima of [Os₂Ru(CO)₁₀(iPr-AcPy)] observed in hexane at 298 K (Figure 8). [b] Nonresolved.

contribution of the relevant atomic wavefunctions to the Table 7. Near-UV and visible absorptions of [Os₂Ru(CO)₁₀(iPr-

AcPy)] (3) in different solvents.

frontier orbitals is given in Table 5, with the HOMO (H)

and LUMO (L) being in boldface. Three-dimensional plots

of the three highest occupied molecular orbitals (HOMO,

HOMO-1 and HOMO-2) and of the lowest unoccupied

molecular orbital (LUMO), are depicted in Figure 6. Notice tetrahydrofuran

that the cluster has a different orientation in Figure 6a–d.

Solvent

λ [nm]

hexane

2-chlorobutane

323 (sh), 434 (sh), 520 (sh), 581

322 (sh), 453 (sh), 562

323 (sh), 456, 535

acetonitrile

322 (sh), 457, 522

The HOMO of 3Ј has large contributions from both the Ru

and Os2 centres, while Os1 is almost not involved. Accord-

In order to evaluate the solvatochromism of the lowest-

energy absorption band, we have plotted the experimentally

determined transition energies against the empirical solvent

parameter E*_MLCTof Manuta et al. (Figure 7), which is

based on the solvatochromism of the lowest MLCT band

of [W(CO)₄(bpy)].^[47]The plot could be well fitted by linear

regression. Its slope reflects the degree of solvatochromism,

while the intercept at E*_MLCT= 0 corresponds to the exper-

imental transition energy (205.3 kJ·mol^–1, 582 nm) extrapo-

lated to nonpolar isooctane interacting similarly weakly

with both ground and excited states. The obtained extrapo-

lated value is in good agreement with the calculated maxi-

mum of the lowest-energy transition (594 nm, Table 6) and

the observed maximum of the lowest-energy absorption

band of cluster 3 in hexane (581 nm, Table 7).

ingly, the HOMO has a σ-bonding character with respect

to the Ru–Os2 bond [σ(Ru–Os2)]. All three metal centres

participate in the HOMO-1 that is mainly σ-bonding with

regard to the Os1 and Os2 centres and will be denoted as

σ(Os1–Os2). The Os2 centre is almost not involved in the

HOMO-2 that receives large contributions from Ru, Os1

and the carbonyl ligands, being best described as a σ(Ru–

Os1) bonding orbital.

The LUMO of 3Ј mainly consists of the lowest π*(H-

PyCa) orbital, while the LUMO+1 is delocalised over the

cluster carbonyl core. Based on the contribution of the

atomic wavefunctions to the frontier orbitals, the HOMO–

LUMO transition has predominant σ(Ru–Os2)-to-π*(α-di-

imine) character. A comparison of the frontier orbitals with

those of [Os₃(CO)₁₀(H-PyCa)]^[18]shows that the HOMOs of

both clusters are very similar, with only one specific metal–

metal(α-diimine) bond involved in the bonding interactions.

Notably, the ruthenium centre in 3Ј contributes slightly

more (25%) to the HOMO than the corresponding osmium

centre in [Os₃(CO)₁₀(H-PyCa)] (18%). The contributions of

the π*(H-PyCa) orbital to the HOMO and the LUMO of

both compounds are nearly identical. The excitation ener-

gies and the oscillator strengths of the low-lying electronic

transitions of 3Ј were calculated using TD-DFT and are

presented in Table 6.

Electronic Absorption Spectra

Figure 7. Solvatochromic behaviour of the lowest-energy absorp-

tion band of [Os₂Ru(CO)₁₀(iPr-AcPy)], based on the empirical sol-

vent scale E*_MLCTaccording to Manuta et al.^[47]The solvents (with

E*_MLCTin parentheses) are CCl₄(0.12), toluene (0.30), CHCl₃

(0.42), THF (0.59), CH₂Cl₂(0.67), acetone (0.82) and acetonitrile

(0.98).

The electronic absorption spectra of cluster 3 are charac-

terised by a dominant lowest-energy band showing negative

solvatochromism (Table 7) typical for charge transfer (CT)

from transition metals to α-diimine ligands.

2212

www.eurjic.org

Eur. J. Inorg. Chem. 2005, 2206–2222

Heterosite Effects in Novel Heteronuclear Clusters

FULL PAPER

The experimental UV/Vis spectrum of cluster 3 in hexane ficiently in open-triangle photoproducts such as biradicals

and the corresponding spectrum simulated with SWizard^[48]and zwitterions.

are shown in Figure 8. A comparison between the theoreti-

cal and experimental data (Table 6, Table 7 and Figure 8)

shows that, besides the lowest-energy absorption band,

there is also a good agreement between the calculated max-

Time-Resolved Absorption Spectroscopy of Cluster 3 in

Noncoordinating Solvents

ima of the higher-lying transitions (Table 6), the simulated

SWizard spectrum, and the higher-lying band maxima of

were recorded in 2-chlorobutane (2-ClBu). Spectral changes

cluster 3 in hexane (Figure 8). As the absorption features in

the visible region are generally broad and poorly resolved,

the lowest-energy absorption band of 3 most likely consists

of several allowed transitions of a prevailing charge transfer

character, which is in line with its solvatochromism. A sim-

Picosecond transient absorption (TA) spectra of cluster 3

following the 505 nm excitation were monitored in the re-

gion 510–710 nm. The spectra measured at 1–40 ps after the

130 fs laser pulse are depicted in Figure 9. Kinetic traces

were obtained by plotting against time the ΔA values

averaged over a small nanometer range. Due to the poor

ilar solvent dependence is observed for the higher-lying

quality of the spectra, reliable lifetimes could only be ob-

shoulder at 520 nm that corresponds with the calculated

tained from the decay in the region 565–580 nm.

charge-transfer transition 4 (Table 6). The weak shoulder

on the low-energy side of the intense near-UV absorption

(434 nm in hexane) most likely consists of two differently

composed transitions that mainly originate in the cluster

core and are directed to the LUMO+1 (transitions 7 and 8,

Table 6). Although the calculated excitation energies of the

three different groups of electronic transitions (cluster

model 3Ј) slightly deviate from the experimental values

(cluster 3), the calculated oscillator strengths compare rea-

sonably well with the observed band intensities and clearly

predict that the lowest-energy, predominantly HOMO–

LUMO transition, is the most intense one in the visible re-

gion. Due to their relatively small oscillator strengths and

overlap, transitions 5 and 6 (Table 6) do not appear as dis-

Figure 9. Transient difference absorption spectra of cluster 3 in 2-

tinct bands in the UV/Vis spectrum of 3.

ClBu, recorded at time delays of –4 (baseline), 1, 3.5, 8.5, 16 and

31 ps, after 505 nm excitation (130 fs FWHM).

The TA spectrum at t_d= 1 ps (Figure 9) shows a bleach

at about 570 nm, very close to the maximum of the ground-

state absorption of the cluster in this solvent (562 nm), and

an absorption with a maximum at 645 nm. The bleach

partly disappears with a lifetime of 10.4 1.2 ps. The re-

maining TA spectrum obtained at t_d= 31 ps, shows a ble-

ach at 580 nm, which is approximately 47% of the initial

signal and does not change significantly in the ps time do-

main. Thus, at least part of the transient species does not

regenerate the parent cluster but instead converts into a sec-

ond species. This is also evidenced by the long-wavelength

absorption that transforms on the same time scale as the

decay of the bleach into a much broader absorption without

Figure 8. Electronic absorption spectrum of [Os₂Ru(CO)₁₀(iPr-

AcPy) in hexane at 293 K (−) together with the simulated SWizard

spectrum (···) and the major electronic transitions of cluster 3Ј, as

calculated with the ADF programme (see Table 6).

a distinct maximum. The ps TA spectra strongly resemble

those of [Os₃(CO)₁₀(iPr-AcPy)] in 2-ClBu.^[16]This observa-

tion also implies that the 11% impurity of the latter cluster

in the studied sample of 3 (vide supra) can hardly be distin-

Although the experimental results do not reveal whether guished by the TA method in this solvent.

the lowest-energy charge transfer transition originates from

In agreement with the results of the TD-DFT study, the

a molecular orbital with a predominant d_π(Ru) character initially observed broad transient absorption above 600 nm

or a delocalised σ(Ru–Os2) character, the TD-DFT results is assigned to an excited state having a predominant σπ*

are clearly in favour of the latter assignment. The lowest- character. In this state, one electron has been transferred

energy transition therefore belongs to a σπ* or sigma-bond- from the σ(Ru–Os2) bonding orbital (vide supra) to the

to-ligand charge transfer (SBLCT) transition, causing sig- lowest π* orbital of the iPr-AcPy ligand. Such absorptions

nificant weakening of the Ru–Os2 bond in the excited state. are characteristic for complexes in metal-to-α-diimine ex-

Visible excitation of 3 is therefore expected to result ef- cited states^[49–52]and for α-diimine radical anions contain-

Eur. J. Inorg. Chem. 2005, 2206–2222

www.eurjic.org

2213

F. W. Vergeer, M. Lutz, A. L. Spek, M. J. Calhorda, Derk J. Stufkens, F. Hartl

FULL PAPER

ing at least one aromatic group.^[49,53,54]The remaining tran-

sient absorption in 2-ClBu is very similar to that observed

in the ns TA spectra of 3 (vide infra) and is assigned to the

biradical [^·Os(CO)₄–Os(CO)₄–⁺Ru(CO)₂(α-diimine^·–)], in

accordance with the results for [Os₃(CO)₁₀(iPr-AcPy)].^[16]

This assignment has also been verified by recording the ps

TA spectra at 250, 500 and 750 ps delays after the laser

pulse. These spectra do not differ from that obtained at t_d

= 31 ps, thereby proving that the biradicals formed directly

from the σπ* excited state do not disappear on the picose-

cond time scale. Although the TA spectra of 3 in 2-ClBu

are very similar to those of the corresponding triosmium

cluster, the excited-state lifetime of the former complex is

significantly shorter (Os₃: τ = 25.3 ps vs. Os₂Ru: τ =

10.4 ps). This may be due to the fact that the Os2–Ru(α-

diimine) bond is weaker and, thus, the energetic barrier for

the biradical formation is lower. Assuming similar molar

absorbance for the homo- and heteronuclear biradicals, the

higher intensity of the remaining transient signal after the

decay of the excited state (Os₃: 30% vs. Os₂Ru: 47%) also

points to a more facile formation of the open-structure

photoproducts in the case of cluster 3.

The fate of the biradicals was studied in several weakly

coordinating solvents (2-ClBu, THF, acetone) by nanose-

cond transient absorption (ns TA) spectroscopy. Spectra of

3 in acetone were obtained by excitation at 532 nm and the

spectral changes were monitored in the 350–800 nm region.

The difference absorption spectra, measured 0–900 ns after

the laser pulse, are depicted in Figure 10. Kinetic traces

were recorded after excitation at 532 nm and probed at

560 nm (bleach). The resulting lifetimes are collected in

Table 8, together with the values for the corresponding

homonuclear cluster [Os₃(CO)₁₀(iPr-AcPy)].

Figure 10. Nanosecond transient difference absorption spectra of

cluster 3 in acetone, recorded in the time interval t = 0–900 ns after

the 532 nm laser pulse; the time delay between the spectra is 100 ns.

and ns time domain. The decay of the nanosecond transient

species in acetone almost completely regenerated the parent

cluster with a lifetime of 228 ns. On changing the solvent

to THF and 2-ClBu, the lifetime decreased to 95 and 27 ns,

respectively. This result demonstrates that the lifetime of the

biradical follows the coordinating ability of the solvent, in-

creasing in the order 2-chlorobutane Ͻ THF Ͻ acetone. A

further increase in the lifetime was found when 3 was irradi-

ated in 2-ClBu in the presence of 1.0 m MeCN. The kinetic

trace of this solution shows a decrease of the bleach on the

μs time scale (τ = 6.8 μs), but not its complete disappear-

ance. This means that the MeCN-stabilised biradicals did

not convert back to the parent cluster but transformed into

the corresponding MeCN-stabilised zwitterions, just as ob-

served for [Os₃(CO)₁₀(iPr-AcPy)].^[16]In the case of the tri-

osmium cluster, this process typically occurs for strong

Lewis bases (MeCN, pyridine) and firmly coordinating ole-

fins (styrene, octene).^[15]

The ns TA spectra of 3 in acetone reveal a strong bleach-

ing between 440 and 600 nm due to the disappearance of 3,

and transient absorption below 440 nm and in the long-

wavelength region. As stated above, the transient absorp-

tion is characteristic for α-diimine radical anions^[49–52]and

for α-diimine complexes in their metal-to-α-diimine excited

state, provided the α-diimine ligand bears at least one het-

eroaromatic group.^[49,53,54]As the ns TA spectra closely re-

semble those of [Os₃(CO)₁₀(iPr-AcPy)] in acetone,^[15]the

transient absorption is again assigned to the open-structure

Time-Resolved Spectroscopy in Strongly Coordinating

Solvents

In addition to the experiments in 2-ClBu, ps TA spectra

biradical [^·Os(CO)₄–Os(CO)₄–⁺Ru(Sv)(CO)₂(iPr-AcPy^·–)] of cluster 3 were also recorded in strongly coordinating

(Sv = acetone). In addition, the ns TA spectrum at t_d= 5 ns MeCN. The spectra measured at 1–5 ps after the 130 fs/

in acetone is very similar to the TA spectrum of 3 at t_d= 505 nm laser pulse are depicted in Figure 11. Kinetic traces

31 ps in 2-ClBu. This confirms that the biradicals formed were extracted by plotting against time the ΔA values

directly from the excited state, are present in both the ps averaged in the range 605–615 nm.

Table 8. Lifetimes [ns] of the solvent-stabilised biradical photoproducts of cluster 3 and the reference compound [Os₃(CO)₁₀(iPr-AcPy)],

obtained from their kinetic profiles probed at 560 nm. Standard deviations are given in parentheses.

Solvent

[Os₂Ru(CO)₁₀(iPr-AcPy)]

[Os₃(CO)₁₀(iPr-AcPy)]

2-chlorobutane

THF

27(4)

95(4)

22(3)

104(7)

Acetone

2-chlorobutane/MeCN (1.0 m)

228(8)

677(2)

6.8 (0.1)^[a]

13.5 (0.2)^[a]

[a] Lifetime in μs.

2214

www.eurjic.org

Eur. J. Inorg. Chem. 2005, 2206–2222

Heterosite Effects in Novel Heteronuclear Clusters

FULL PAPER

[Os₃(CO)₁₀(iPr-AcPy)], cluster 3 affords a single photopro-

duct that does not convert into another species in the time

domains studied. Based on our previous ns-μs TA studies of

biradical and zwitterion photoproducts of [Os₃(CO)₁₀(iPr-

AcPy)] in MeCN,^[12,15,16]the presence of 11% impurity of

the latter cluster in the investigated sample of 3 should re-

sult in ΔA ca. 0.002 in Figure 12, which is difficult to resolve

in the spectral noise.

Figure 11. Transient difference absorption spectra of cluster 3 in

MeCN, recorded at time delays of –1 (baseline), 1, 1.5, 2.5, 3.5 and

4.5 ps, respectively, after 505 nm excitation (130 fs FWHM).

The first TA spectra (Figure 11) are similar to those in

2-ClBu although the maxima of the bleach (550 nm) and

the transient absorption (630 nm) are somewhat shifted. Al-

though the TA spectra of 3 in MeCN resemble those ob-

tained for [Os₃(CO)₁₀(iPr-AcPy)] in this solvent,^[16]the ex-

cited states of both clusters behave differently. For [Os₃-

(CO)₁₀(iPr-AcPy)] the decay of the excited state in coordi-

nating MeCN is bi-exponential. The slower process (τ =

Figure 12. Transient kinetics at 560 nm measured for cluster 3 in

MeCN, following irradiation at 532 nm with a Nd:YAG laser (7 ns

FWHM, average of 10 shots at 10 s intervals, 2 mJ·pulse^–1).

In order to confirm that the single photoproduct is the

21.4 ps) has been ascribed to biradical formation, having a zwitterion

[^–Os(CO)₄–Os(CO)₄–Ru⁺(MeCN)(CO)₂(iPr-

lifetime similar to that observed in 2-ClBu.^[16]The faster AcPy)], the photoreaction of 3 in MeCN was followed with

process (τ = 2.9 ps) has been assigned to the heterolytic rapid-scan IR spectroscopy in the (sub)second time do-

cleavage of an Os–Os(α-diimine) bond from an MeCN-co- main. In addition, also pyridine was used as the solvent.

ordinated excited state with concomitant zwitterion forma- Cluster 3 was irradiated for 2 s with an argon-ion laser

tion; zwitterions were indeed observed in the ps TRIR spec- (514.5 nm, 150 mW) and the IR spectral changes in the CO-

tra of [Os₃(CO)₁₀(iPr-AcPy)].^[55]The excited state of cluster stretching region were monitored in both solvents on the

3 decays mono-exponentially in MeCN and its lifetime (τ = time scale of seconds to minutes. The difference IR spectra

2.1 ps) closely corresponds to the shorter 2.9 ps component of 3 in pyridine, measured 0–93 s after the laser pulse, are

of the bi-exponential decay of the excited state of [Os₃- depicted in Figure 13. Unfortunately, irradiation of 3 in

(CO)₁₀(iPr-AcPy)]. This points to the same process for both MeCN at 293 K only resulted in the formation of very weak

clusters and suggests that irradiation of 3 in neat MeCN transient bands at 1997 (vw), 1970 (s) and 1874 (s, br) cm^–1,

gives rise to very fast formation of zwitterions as the only that decayed with a lifetime similar to that of the zwitter-

photoproduct. As will be shown hereinafter, TA studies in ions formed upon irradiation of [Os₃(CO)₁₀(iPr-AcPy)] (τ =

the nano- and microsecond time domain and rapid scan 38 1 s).^[12]Based on the intensity of the transient signals

FTIR spectroscopic measurements confirm this presump- and the observed lifetime, this transient species is assigned

tion.

as the homonuclear zwitterion [^–Os(CO)₄–Os(CO)₄–

As described above, the kinetic trace of 3 in 2-ClBu in ⁺Os(MeCN)(CO)₂(iPr-AcPy)], originating from the 11%

the presence of 1.0 m MeCN showed a decrease of the ble- [Os₃(CO)₁₀(iPr-AcPy)] impurity in the sample (vide supra).

ach on the μs time scale (τ = 6.8 μs) that, in accordance Repeating the experiment at 273 K resulted in similar, more

with the results for [Os₃(CO)₁₀(iPr-AcPy)],^[16]was assigned intense product bands, whose decay was clearly bi-ex-

to the conversion of MeCN-stabilised biradicals into the ponential. Both the transient absorptions and the parent

corresponding MeCN-zwitterions. The MeCN-stabilised bi- bleaches rapidly decay within 5 s to approximately 15% of

radicals result in this case from the substitution of weakly their initial intensity. In a second, slower process, the re-

coordinating 2-ClBu by MeCN within the lifetime of the maining transient fully regenerates the parent cluster with

solvent-stabilised

biradicals

[^·Os(CO)₄–Os(CO)₄– a lifetime of 54.6 1.2 s, being very close to the lifetime of

Ru⁺(Sv)(CO)₂(iPr-AcPy^·–)] (Sv = 2-ClBu). In neat MeCN, the zwitterions formed upon irradiation of pure [Os₃-

however, the kinetic trace of 3 does not show any change (CO)₁₀(iPr-AcPy)] under these conditions. As the ampli-

in signal intensity on the nano- and microsecond time scale tudes of both processes nicely correspond with the observed

(Figure 12), indicating that the pathway for zwitterion for- ratio between [Os₂Ru(CO)₁₀(iPr-AcPy)] and [Os₃(CO)₁₀-

1

mation via the biradicals, similar to that observed for the (iPr-AcPy)] in the H NMR spectrum and crystal structure

triosmium analogue, does not exist. Thus, contrary to of 3, the faster process is assigned to the decay of

Eur. J. Inorg. Chem. 2005, 2206–2222

www.eurjic.org

2215

F. W. Vergeer, M. Lutz, A. L. Spek, M. J. Calhorda, Derk J. Stufkens, F. Hartl

FULL PAPER

the heteronuclear zwitterions [^–Os(CO)₄–Os(CO)₄–⁺Ru- the stronger coordination of the latter solvent. At the same

(MeCN)(CO)₂(iPr-AcPy)]. In agreement with the lifetime, time, the difference in the lifetimes of the heteronuclear and

the slower process is then attributed to the decay of the triosmium zwitterions has also increased significantly, the

homonuclear zwitterions originating from the [Os₃(CO)₁₀- latter one living in pyridine for more than 30 minutes. The

(iPr-AcPy)] impurity.

rapid-scan FTIR experiments thus clearly demonstrate that

the irradiation of [Os₂Ru(CO)₁₀(iPr-AcPy)] in strongly co-

ordinating solvents like MeCN and pyridine results in the

formation of solvent-stabilised zwitterions, with lifetimes

much shorter than those of the corresponding triosmium

zwitterions. The results of the combined time-resolved stud-

ies of cluster 3 are summarised in Scheme 2.

Stability of the Photoproducts

The observed photochemical reactivity of 3 differs signif-

icantly from that of the homonuclear clusters [Ru₃(CO)₈(μ-

CO)₂(α-diimine)].^[17]Although visible excitation of the lat-

ter clusters also resulted in the formation of open-structure

photoproducts like biradicals and zwitterions, the efficiency

of these processes is significantly lower due to the persistent

presence of bridging carbonyl ligands in the excited state.

Secondly, the triruthenium photoproducts could only be

stabilised at low temperatures or in the presence of strongly

coordinating Lewis bases or radical scavengers. Regarding

the formation of open-structure photoproducts, the photo-

reactivity of 3 resembles more closely that of its triosmium

analogue [Os₃(CO)₁₀(iPr-AcPy)], for which fairly stable sol-

vent-stabilised biradicals and zwitterions were observed

upon excitation into its lowest-energy electronic transi-

tion.^{[12,15,16,55]}The latter cluster therefore serves as a refer-

ence compound in order to evaluate the influence of the

heteronuclear cluster core in 3 on the photoreactivity.

Interestingly, the solvent-stabilised biradicals of parent

cluster 3 are significantly shorter-lived than their triosmium

counterparts (Table 8). This difference in the biradical life-

time is most striking in acetone, where the lifetime of the

heteronuclear biradical (τ = 228 ns) is merely 30% of the

value found for [Os₃(CO)₁₀(iPr-AcPy)] (τ = 677 ns). A much

smaller difference is observed in THF (Os₂Ru: τ = 95 ns vs.

Os₃: τ = 104 ns) while in weakly coordinating 2-ClBu the

biradical lifetimes of both clusters are almost identical.

These results clearly document that the difference in the

biradical lifetime increases with the coordinating ability of

the solvent. A similar trend applies for the lifetimes of the

solvent-stabilised zwitterions. Whereas the lifetime of the

heteronuclear zwitterion increases only slightly from a few

seconds in MeCN to 23 s in pyridine, the lifetime of the

triosmium zwitterion in pyridine amounts to more than 30

minutes compared to 38 s in MeCN. The shorter lifetimes

of the heteronuclear photoproducts in better coordinating

solvents indicate that the stabilisation of these products by

solvent coordination is significantly reduced compared to

the corresponding triosmium species. This is attributed to

the higher tendency of the coordinatively unsaturated

⁺Os(CO)₂(iPr-AcPy^·–/0) moieties in the photoproducts to

bind a Lewis base. A similar difference in Lewis base coor-

dination on descending a transition metal group in the peri-

Figure 13. Difference rapid-scan IR spectra of cluster 3 in pyridine

measured at time delays of 1.8, 6.6, 11.4, 18.6, 28.2, 40.2, 59.4 and

93.0 s after the 514.5 nm laser pulse.

Irradiation of 3 in strongly coordinating pyridine at

room temperature resulted in similar spectral changes as

observed in MeCN at 273 K (Figure 13). After excitation,

the first spectra display instantaneous bleaching of the par-

ent ν(CO) bands, together with transient absorption bands

at 2073 (vw), 1994 (w), 1971 (s), 1965 (sh), 1898 (sh) and

1875 (s, br) cm^–1. Both the transient bands and parent ble-

aches decay again bi-exponentially. The lifetime of the he-

teronuclear zwitterion increased from a few seconds in

MeCN to 23.0 0.3 s in pyridine, which is consistent with

Scheme 2. Schematic representation of the photoreaction pathways

established for the cluster [Os₂Ru(CO)₁₀(iPr-AcPy)] (3). The sepa-

rate reaction steps (A)–(E) depend on the solvent (Sv) used: (i) (A)

+ (D) in 2-ClBu, THF, acetone (293 K); (ii) (B) + (E) in MeCN;

(iii) (A) + (C) + (E) in 2-ClBu + 1.0 m MeCN.

2216

www.eurjic.org

Eur. J. Inorg. Chem. 2005, 2206–2222

Heterosite Effects in Novel Heteronuclear Clusters

FULL PAPER

odic table is observed, for example, for the coordinatively

unsaturated radicals [M(CO)₃(α-diimine)]· (M = Mn, Re; α-

diimine is, e.g., bpy, iPr-PyCa, iPr-DAB). Thus, while the

unsaturated radicals [Re(CO)₃(α-diimine)]· form at room

temperature fairly stable 18e paramagnetic species [Re(CO)₃-

(α-diimine)(Sv)]· (Sv = MeCN, PrCN),^[56–58]analogous

radicals were not observed for [Mn(CO)₃(α-diimine)]·.^[59]

Redox Behaviour of [Os₂Ru(CO)₁₀(iPr-AcPy)] (3)

A combined cyclic voltammetric and IR spectroelectro-

chemical study of cluster 3 was performed in order to inves-

tigate the influence of the heteronuclear cluster core on the

reduction pathway and stability of the reduction products.

The redox potentials are presented in Table 9.

Figure 14. Cyclic voltammogram of cluster 3 at T = 293 K (a) and

T = 200 K (b). Conditions: 10^–3m cluster in THF/10^–1m Bu₄NPF₆,

Pt disk microelectrode (0.42 mm²apparent surface area), v =

100 mV s^–1.

Table 9. Redox potentials of cluster 3 and its reduction products.

Cluster^[a][b]

E_p,c[V]^[c]

E_p,a[V]^[c]

3

–1.87 (irr)

–1.93 (irr)

+0.27 (irr)

+0.37 (irr)

–1.50 (irr)

–1.17 (irr)

–0.85 (irr)

3^[d]

Cyclic voltammetry thus documents that the closed-tri-

angle radical anion of 3 cannot be stabilised even at 200 K

and readily transforms into the open-structure dianion 3b^2–

(see Scheme 3). The latter dianion is also unstable and

could only be produced in detectable amounts at low tem-

peratures. This is in contrast with the results of [Os₃(CO)₁₀-

(iPr-PyCa)], where the oxidation of the open-structure di-

anion is already clearly visible at 293 K.^[53]Formation of

dimer [3b-3b]^2–, whose oxidation is observed at 200 K at

E_p,a= –0.85 V (O₂Ј), is known to take place by a rapid

nucleophilic attack of dianion 3b^2–at the yet nonreduced

parent cluster 3. A similar ECEC reduction path, for exam-

ple, has been reported for the homonuclear clusters

[Os₃(CO)₁₀(α-diimine)].^[53]Both 3b^2–and [3b-3b]^2–are

structurally related to the solvent-stabilised zwitterions

3b^{2– [d]}

[3b-3b]^2–

[3b-3b]^{2– [d]}

[a] Conditions and definitions: 10^–3mol·dm^–3solutions in THF

(containing 10^–1m Bu₄NPF₆) at 293 K, unless stated otherwise; Pt

disk microelectrode (0.42 mm²); v = 100 mV s^–1; redox potentials

vs. E_1/2(Fc/Fc⁺); E_p,c, cathodic peak potential for reduction of par-

ent cluster; E_p,a, anodic peak potential for oxidation of parent clus-

ter or its reduction products. [b] Assignments given in the main

text. [c] Chemical irreversibility denoted by (irr). [d] T = 200 K.

The cyclic voltammogram of 3 in THF at room tempera-

ture (v = 100 mV·s^–1) showed a chemically irreversible two-

electron reduction at E_p,c= –1.87 V (cathodic peak R₁, see

Figure 14a), most likely producing the open-core dianion

[^–Os(CO)₄–Os(CO)₄–Ru^–(CO)₂(iPr-AcPy)]^2–(3b^2–). Similar

dianions are formed upon reduction of the homonuclear

clusters [M₃(CO)₁₀(α-diimine)] (M = Ru, Os).^[17,53]At room

temperature, the dianion 3b^2–was, however, not detectable

on the reverse scan (vide infra). Instead, scan reversal be-

hind R₁resulted in the appearance of an anodic peak at

–1.17 V (O₂Ј) assigned, in accordance with the results for

[Os₃(CO)₁₀(iPr-PyCa)],^[53]to the oxidation of the cluster di-

2–

mer [^–Os(CO)₄–Os(CO)₄–Ru(CO)₂(iPr-AcPy)]₂([3b-3b]^2–)

containing an (iPr-AcPy)Ru–Ru(iPr-AcPy) bond. Ad-

ditional proof for this assignment was obtained from IR

spectroelectrochemical experiments (vide infra), where the

reduction of 3 resulted in the appearance of a ν(CO) band

pattern closely resembling that of the photogenerated zwit-

terions with an open cluster core. On lowering the tempera-

ture to 200 K, the reduction of 3 remained chemically

irreversible (Figure 14b). However, in contrast to the room-

temperature scan, the anodic sweep showed an additional

anodic peak at –1.50 V (O₂) assigned^[53]to the oxidation

of 3b^2–, the latter species being sufficiently stable at low

temperatures. The minor anodic process denoted with the

asterisk (Figure 14b) corresponds to the oxidation of unas-

signed species.

Scheme 3. Reduction path of cluster 3.

Eur. J. Inorg. Chem. 2005, 2206–2222

www.eurjic.org

2217

F. W. Vergeer, M. Lutz, A. L. Spek, M. J. Calhorda, Derk J. Stufkens, F. Hartl

FULL PAPER

[^–Os(CO)₄–Os(CO)₄–⁺Ru(Sv)(CO)₂(iPr-AcPy)], the pho- 2,2Ј-bipyridine (dmb), 2,2Ј-bipyrimidine (bpym)] forming

toproduct of 3 in coordinating solvents (Sv) (vide supra). fairly stable radical anions at moderately low temperatures,

For completeness, the oxidation of cluster 3 at the poten- the reduction of 3 remains chemically irreversible even at

tial E(O_m) is chemically irreversible, resulting in a second- 200 K. A comparison of the reduction potentials of the

ary product reducible on the reverse cathodic scan at the homonuclear [M₃(CO)₁₀(α-diimine)] clusters (M = Os,^[68]

potential E(RЈ_m). These processes were not studied in de- Ru^[17]) with that of 3 (–1.87 V) reveals that the π-acceptor

tail.

capacity of the iPr-AcPy ligand in 3 is similar to that of

In the course of corresponding IR spectroelectrochemical bpy (–1.85 V, M = Os, Ru) and dmb (–1.91 V, M = Ru).

experiments, the reduction of 3 in THF at 250 K resulted This implies that the higher reactivity of radical anions 3^·–

in the appearance of transient ν(CO) bands at 2049 (m), is not due to a limited donor power of the reduced iPr-

2009 (m), 1991 (m), 1965 (s), 1942 (sh) and 1864 (m) cm^–1, AcPy ligand but, probably, has its origin in the weak hetero-

similarly to the IR ν(CO) spectrum reported for the open- metallic Os2–Ru(iPr-AcPy) bond.

2–

core dimers [^–Ru(CO)₄–Ru(CO)₄–Ru(CO)₂(α-diimine)]₂

Just as for the corresponding [Os₃(CO)₁₀(iPr-PyCa)] clus-

(α-diimine = 2,2Ј-bipyridine, 2,2Ј-bipyrimidine).^[17]Accord- ter,^[53]no open-structure dianions were detected by IR spec-

ingly, these bands are assigned to the heteronuclear ana- troscopy on the spectroelectrochemical time scale of min-

2–

logue [^–Os(CO)₄–Os(CO)₄–Ru(CO)₂(iPr-AcPy)]₂([3b-3b]^2–, utes at 250 K, the dimer [3b-3b]^2–being the initially ob-

Scheme 3). No IR bands attributable to the open-structure served reduction product. The triangular open-structure

dianion 3b^2–were observed in the course of the reduction. units in dimer [3b-3b]^2–undergo fragmentation and linear

This means that 3b^2–is unstable on the spectroelectrochem- chains [Ru(CO)₂(iPr-AcPy)]_nare ultimately formed, to-

ical time scale at 250 K and readily reacts with the parent gether with the dinuclear complex [Os₂(CO)₈]^2–.

cluster 3 to form the dimer [3b-3b]^2–. Actually, the latter

species was also unstable and converted into another car-

bonyl product absorbing at 1938 (s) and 1863 (s, br) cm^–1.

Conclusions

In agreement with the literature^[31]and the results for

[Os₃(CO)₁₀(α-diimine)],^[60]this product is proposed to be

Crystal structures of the novel clusters [Os₂Ru(CO)₁₁-

the dinuclear complex [Os₂(CO)₈]^2–. As the highest-fre- (PPh₃)] (2) and [Os₂Ru(CO)₁₀(iPr-AcPy)] (3) clearly reveal,

quency ν(CO) band of the corresponding [Ru₂(CO)₈]^2–in agreement with DFT calculations on several structural

complex is found at significantly lower wavenumber (Ru: isomers, that both PPh₃and iPr-AcPy prefer coordination

1930 cm^–1vs. Os: 1940 cm^–1),^[31]the observation of at the ruthenium site of [Os₂Ru(CO)₁₂] (1). The Os2–Ru-

[Os₂(CO)₈]^2–provides another evidence that the α-diimine (iPr-AcPy) bond in cluster 3 is weaker than the correspond-

ligand in 3 is coordinated at ruthenium. Parallel to ing metal–metal(α-diimine) bonds in the homonuclear

[Os₂(CO)₈]^2–, reduction of [3b-3b]^2–most likely results in [Os₃(CO)₁₀(iPr-AcPy)] and doubly CO-bridged triruthen-

the formation of the polymeric chain [Ru(CO)₂(iPr-AcPy)]_n, ium analogues. This factor contributes to the observed dif-

as indicated by the formation of a blue film at the working ferences in the photo- and electrochemical reactivity of

electrode and in its vicinity. Similar species are also formed cluster 3 compared to the homonuclear cores; although, the

as the ultimate reduction products of the clusters [M₃(CO)₁₀- general course of the photo- and redox reactions remains

(α-diimine)] (M = Ru, Os), closely resembling [M(CO)₂- unchanged.

(bpy)]_n, generated by electrochemical reduction of mono-

Regarding the photochemical formation of biradical and

zwitterionic photoproducts, cluster 3 closely resembles its

nuclear complexes trans(Cl)-[M(CO)₂(bpy)(Cl)₂].^[61–64]

The interest in these open-chain polymers [M(CO)₂(α- homonuclear analogue [Os₃(CO)₁₀(iPr-AcPy)]. In weakly

diimine)]_n(M = Ru, Os) mainly derives from their electroca- coordinating 2-ClBu, the lowest σπ* excited state decays

talytic activity towards reduction of carbon dioxide.^[63,65–67]with a lifetime of approximately 10 ps, resulting in the for-

From the DFT results (Table 5, Figure 6) it is clear that mation of solvent-stabilised biradicals. In contrast to its tri-

the LUMO of cluster 3 is for about 65% the lowest π*(iPr- osmium analogue, irradiation of 3 in coordinating solvents

AcPy) orbital. Further, the LUMO has also got a nonnegli- exclusively results in heterolytic splitting of the Os2–Ru(iPr-

gible σ*(Ru–Os2) character. Single occupation of this or- AcPy) bond, producing solvent-stabilised zwitterions. Inter-

bital is therefore expected to result in largely α-diimine- estingly, the lifetimes of the biradical- and zwitterionic pho-

localised radical anion 3^·–with a weakened Ru–Os2 bond. toproducts of cluster 3 are significantly shorter than their

This bond becomes readily cleaved, similar to the photore- triosmium counterparts. This difference results most likely

activity of cluster 3 (vide supra). While the photochemical from a lower tendency of the coordinatively unsaturated

cleavage is promoted by the partial σσ* character of the ⁺Ru(CO)₂(iPr-AcPy^·–/0) moieties in the photoproducts to

excited state, the driving force behind the electrochemical bind a Lewis base. Electrochemically, the influence of the

cleavage is probably the pronounced donor nature of the weak Os2–Ru(α-diimine) bond is clearly reflected in the low

singly reduced iPr-AcPy ligand in 3^·–causing strong polaris- stability of the radical anions initially formed by one-elec-

ation of the Ru–Os2 bond nearly perpendicular to the iPr- tron reduction of 3.

AcPy plane. Unlike the corresponding homonuclear clus-

In general, the combination of experimental photo- and

ters [Os₃(CO)₁₀(iPr-PyCa)]^[53]and [Ru₃(CO)₈(μ-CO)₂(α-di- electrochemical techniques together with the theoretical

imine)]^[17][α-diimine = 2,2Ј-bipyridine (bpy), 4,4Ј-dimethyl- support from DFT calculations have provided a good in-

2218

www.eurjic.org

Eur. J. Inorg. Chem. 2005, 2206–2222

Heterosite Effects in Novel Heteronuclear Clusters

FULL PAPER

40%). IR (hexane): ν [ν(CO)] = 2118 (s), 2079 (vs), 2067 (s), 2063

˜

sight into the bonding properties of [Os₂Ru(CO)₁₀(iPr-

AcPy)]. Although the incorporation of a Ru centre into the

cluster core does not induce drastic changes in the photo-

and electrochemical reactivity pattern in comparison with

the triosmium derivative, more pronounced heterosite ef-

fects may, for example, be expected for mixed-metal clusters

combining transition metal centres from different groups of

the periodic Table (e.g., Os₂Pt, Os₂Rh).

(vs), 2052 (s), 2031 (s) cm^–1.

Synthesis of Na₂[Ru(CO)₄] and [Os₂Ru(CO)₁₂] (1): Two reaction

vessels, equipped with Rotaflo^®stopcocks and connected by a glass

frit, were placed on a standard high-vacuum line. Under a con-

tinuous N₂flow, [Os₂(CO)₈(Br)₂] (200 mg, 0.26 mmol) was placed

in one of the vessels while the other one contained [Ru₃(CO)₁₂]

(55.8 mg, 0.087 mmol) and metallic sodium (12 mg, 0.52 mmol).

Anhydrous NH₃(ca. 25 mL) was condensed in a separate vessel

onto metallic sodium and frozen, using a liquid nitrogen bath. Af-

ter closing the vessel containing [Os₂(CO)₈(Br)₂], NH₃was distilled

under vacuum onto the sodium/[Ru₃(CO)₁₂] mixture. Using an ace-

tone/dry ice bath, the solution was slowly warmed to 235 K and

stirred vigorously until the characteristic blue Na/NH₃solution

transformed into a yellow solution containing a white precipitate.

This usually required about 30 minutes of reaction time. The mix-

ture was stirred for another 30 minutes and remaining sodium,

which splashed on the walls of the flask, was washed down by cold

spotting with glass wool drenched with liquid nitrogen. After 1 h,

NH₃was evaporated by further warming to 260 K and the remain-

ing cream-coloured solid was dried under high-vacuum conditions

for another 3.5 h at this temperature. Na₂[Ru(CO)₄] was then dis-

solved on the vacuum line in THF that had been pre-dried with

sodium/benzophenone and thoroughly degassed in four consecu-

tive freeze-pump-thaw cycles. The pale yellow solution was de-

gassed once more to remove the last traces of ammonia. Finally,

the THF solution was warmed to 273 K and added to [Os₂(CO)₈-

(Br)₂] by filtration through the frit connecting the reaction vessels.

Upon mixing, the solution turned red and gas was evolving. The

solution was stirred overnight at 293 K, followed by evaporation of

the solvent. Purification of the crude product was established by

column chromatography (activated neutral alumina, hexane/

CH₂Cl₂gradient elution). After precipitation from THF at 190 K,

the product was obtained as a yellow powder in yields varying be-

tween 86 and 135 mg (40–63% based on [Os₂(CO)₈(Br)₂]), de-

pending on the quality of Na₂[Ru(CO)₄] (vide supra). IR (hexane):

Experimental Section

Materials and Preparations: Solvents of analytical grade [Acros:

acetone, acetonitrile (MeCN), dichloromethane (CH₂Cl₂), hexane,

tetrahydrofuran (THF); Aldrich: 2-Chlorobutane (2-ClBu); Merck:

pyridine] were dried with sodium wire (hexane), sodium/benzophe-

none (THF), CaSO₄(acetone) and CaH₂(MeCN, CH₂Cl₂, 2-ClBu,

pyridine) and freshly distilled under nitrogen prior to use. [Ru₃-

(CO)₁₂], [Os₃(CO)₁₂] (Strem Chemicals), PPh₃, Br₂(Aldrich), ferro-

cene (BDH) and NH₃(Praxair) were used as received. Trimeth-

ylamine N-oxide, Me₃NO·2H₂O (Janssen), was dehydrated before

use by vacuum sublimation. The supporting electrolyte Bu₄NPF₆

(Aldrich) was recrystallised twice from ethanol and dried in vacuo

at 350 K overnight. Neutral aluminium oxide 90 (70–230 mesh,

Merck) and silica 60 (70–230 mesh, Merck) for column chromatog-

raphy were activated by heating in vacuo at 450 K overnight and

stored under N₂. Preparative TLC was performed on Silica Gel G

plates (20×20 cm, 1 μm, Analtech).

Synthetic Procedures: All syntheses were performed under dry ni-

trogen, using standard Schlenk techniques. The preparation of Na_2-

[Ru(CO)₄] and the consecutive coupling with [Os₂(CO)₈(Br)₂] were

performed on a high-vacuum line at reduced pressure (Ϸ10^–4Pa).

[Os₃(CO)₁₂(Br)₂],^[69][Os₂(CO)₈(Br)₂]^[30]and Na₂[Ru(CO)₄]^[31]were

prepared by modified literature procedures. [Os₂Ru(CO)₁₁(MeCN)]

and [Os₂Ru(CO)₁₀(MeCN)₂] were prepared by similar procedures

ν [ν(CO)] = 2068(s), 2036 (s), 2015 (m), 2004 (m) cm^–1. UV/Vis

˜

as used by Foulds et al. for the syntheses of [Ru₃(CO)_12–n

-

(CH₂Cl₂): λ = 280 (sh), 329, 382 (sh) nm. MS (FD): m/z = 817

(MeCN)_n] (n = 1, 2).^[70]Both clusters were prepared in situ and

[M]⁺(calcd. 817.8). MS (EI): m/z = [M⁺– nCO] (n = 0–12).

only characterised by IR spectroscopy (vide infra).

Synthesis of [Os₂Ru(CO)₁₁(PPh₃)] (2): Me₃NO (5.5 mg, 0.07 mmol)

in MeCN (2 mL) was added to a solution of [Os₂Ru(CO)₁₂] (30 mg,

0.04 mmol) in THF (25 mL) at 200 K. After stirring the mixture

for approximately 30 minutes, IR spectra revealed almost complete

conversion to [Os₂Ru(CO)₁₁(MeCN)]. After addition of PPh₃

(10.6 mg, 0.04 mmol), the solution was warmed to room tempera-

ture. As IR spectra showed hardly any conversion after 60 minutes,

additional PPh₃was added (5.2 mg, 0.02 mmol) and the solution

was stirred for two days in the dark. After removal of the solvent

in vacuo, the residue was loaded on a silica 60 column packed in

hexane. Gradient elution with hexane/THF resulted in the clean

separation of three (yellow, orange and red) mobile bands. The yel-

low fraction was further purified by preparative TLC to remove

free PPh₃. Recrystallisation of the yellow residue from petroleum

ether 40–60 at 250 K yielded orange crystals of cluster 2 (15 mg,

40%). On the basis of their IR spectra, the orange and red fractions

likely contained higher substituted compounds, but no detailed

characterization of the mixtures was attempted.

Synthesis of [Os₃(CO)₁₂(Br)₂]: In a typical experiment, a solution

of [Os₃(CO)₁₂] (500 mg, 0.55 mmol) in CH₂Cl₂(350 mL) was

heated to reflux for 20 min. After addition of Br₂(40 μL,

0.78 mmol) the solvent was immediately removed in vacuo. The

pale yellow residue was dissolved in CH₂Cl₂/hexane, 1:4 (400 mL)

and precipitated at 190 K. The product was obtained as a pale

yellow powder (412 mg, 70%) and used in the synthesis of

[Os (CO) (Br) ] without further purification. IR (CH Cl ): ν

˜

2

8

2

[ν(CO)] = 2149 (vw), 2119 (s), 2062 (vs, br), 2030 (s), 2002 (w)

cm^–1.

Synthesis of [Os₂(CO)₈(Br)₂]:

A solution of [Os₃(CO)₁₂(Br)₂]

(350 mg, 0.33 mmol) in CH₂Cl₂(60 mL) was heated at 323 K to

dissolve all starting material. After cooling to room temperature,

Br₂(14.5 μL, 0.28 mmol) was added and the solution was irradi-

ated with a 125 W high-pressure Hg lamp using a λ Ͼ 420 nm

cut-off filter. The reaction was monitored by IR spectroscopy and

irradiation was stopped when no further increase of the product

ν(CO) bands was observed (ca. 85% conversion). After removal of

the solvent in vacuo, the crude yellow product was extracted with

hexane (5×10 mL). The combined fractions were filtered and the

solvents evaporated to dryness. The remaining solid was redissolved

in hexane containing a few drops of CH₂Cl₂and precipitated at

270 K. The product was obtained as a pale yellow powder (86 mg,

[Os Ru(CO) (MeCN)]: IR (CH Cl ): ν [ν(CO)] = 2104 (w), 2050

˜

2

11

2

(vs), 2040 (vs), 2019 (s, sh), 2008 (vs), 1984 (m) cm^–1.

[Os Ru(CO) (PPh )] (2): IR (CH Cl ): ν [ν(CO)] = 2107 (m), 2054

˜

2

11

3

2

(s), 2033 (s), 2017 (vs), 1999 (m), 1988 (m), 1977 (m), 1957 (w)

cm^–1. ¹H NMR (300 MHz, CDCl₃, 293 K): δ = 7.35 (m, 15

Eur. J. Inorg. Chem. 2005, 2206–2222

www.eurjic.org

2219

F. W. Vergeer, M. Lutz, A. L. Spek, M. J. Calhorda, Derk J. Stufkens, F. Hartl

FULL PAPER

H) ppm. ³¹P{H} NMR (300 MHz, CDCl₃, 293 K): δ = 29.6 ([Os₂-

Ru(CO)₁₁(PPh₃)]), –0.8 ([Os₃(CO)₁₁(PPh₃)]) ppm (ratio ca. 2:1).

MS (FD): m/z = 1142 [M]⁺(M = [Os₃(CO)₁₁(PPh₃)]) (calcd.

1141.9), 1054 [M]⁺(M = [Os₂Ru(CO)₁₁(PPh₃)]) (calcd. 1053.9) (ra-

tio ca.1:2).

group P2₁/c (no. 14); cell parameters:

13.9915(1) Å, 20.7504(2) Å;

a

=

9.1117(1),

b

=

c

=

β

=

113.6862(4)°; V

2422.54(4) Å³; Z = 4; ρ = 2.560 g cm^–3; F(000) = 1710. 43532 reflec-

tions were measured, 5551 reflections were unique (R_int= 0.0740).

An analytical absorption correction was applied (μ = 11.644 mm^–1,

0.20–0.78 transmission). 320 refined parameters, no restraints. R

(obsd. refl.): R₁= 0.0277, wR₂= 0.0419. R (all data): R₁= 0.0510,

wR₂= 0.0460. Weighting scheme w = 1/[σ²(F_o²) + (0.0147P)²],

Synthesis of [Os₂Ru(CO)₁₀(iPr-AcPy)] (3): To a solution of [Os₂-

Ru(CO)₁₁(MeCN)] at 270 K, freshly prepared from [Os₂Ru(CO)₁₂]

(100 mg, 0.12 mmol) and Me₃NO (18.6 mg, 0.25 mmol), a solution

of Me₃NO (13.5 mg, 0.18 mmol) in CH₂Cl₂was added dropwise.

After stirring for 45 minutes IR spectra revealed [Os₂Ru(CO)₁₀-

(MeCN)₂] as the main product. After addition of iPr-AcPy

(200 mg, 1.2 mmol) the solution was stirred overnight while warm-

ing to room temperature. Purification of the crude product over

silica using 2:3 hexane/THF as eluent gave [Os₂Ru(CO)₁₀(iPr-

AcPy)] as a purple powder (34 mg, 30%). Crystals were grown by

slow diffusion of hexane into a saturated solution of [Os₂Ru-

(CO)₁₀(iPr-AcPy)] in THF.

2

where P = (F_o+ 2F_c)/3. GoF = 0.996. Residual electron density

between –0.95 and 1.01 e/Å³.

CCDC-258240 (for 2) and -258241 (for 3) contain the supplemen-

tary crystallographic data for this paper. These data can be ob-

tained free of charge from The Cambridge Crystallographic Data

Centre via www.ccdc.cam.ac.uk/data_request/cif.

Spectroscopic Measurements: Electronic absorption spectra were re-

corded with a Hewlett–Packard 8453 diode-array spectrophotome-

ter and FT-IR spectra with a Bio-Rad FTS-7 (16 scans at 2 cm^–1

resolution) spectrometer. Rapid-scan FT-IR spectra were measured

with a Bio-Rad FTS-60A spectrometer (equipped with a dual-

[Os Ru(CO) (MeCN) ]: IR (CH Cl ): ν [ν(CO)] = 2079 (w), 2019

˜

2

10

2

(vs), 1984 (s), 1958 (m) cm^–1.

[Os Ru(CO) (iPr-AcPy)] (3): IR (THF): ν [ν(CO)] = 2082 (m), source rapid-scan 896 interferometer and a liquid-nitrogen cooled

˜

2

10

2028 (vs), 2002 (vs), 1990 (s), 1979 (s), 1962 (m), 1956 (sh), 1907

MCT detector) after excitation of the sample by the 514.5 nm line

1

(w) cm^–1. ¹H NMR (300 MHz, CDCl₃, 293 K) {for numbering

of a Spectra Physics Model 2016 argon-ion laser. H and ³¹P{H}

scheme see Figure 1; asterisks denote signals due to [Os₃(CO)₁₀(iPr-

NMR spectra were recorded with a Bruker AMX 300 spectrometer.

3

AcPy)] (ca.10–15%)}: δ = 9.50* (d, 1 H), 9.23 (d, J_H,H= 4.5 Hz, Field Desorption (FD), Fast Atom Bombardment (FAB) and Elec-

3

1 H, H⁶), 8.04 (d, J_H,H= 8.1 Hz, 1 H, H³), 8.02* (d, 1 H), 7.91

tron Impact (EI) mass spectra were collected with a JEOL JMS

SX/SX102A four-sector mass spectrometer. For synthetic purposes,

a Philips HPK 125-W high-pressure Hg lamp, equipped with ap-

propriate cut-off filters to select the desired wavelength region,

served as a light source.

3

(dd, J_H,H= 8.1, J_H,H= 7.5 Hz, 1 H, H⁴), 7.86* (dd, 1 H), 7.25

3

(dd, J_H,H= 7.5, J_H,H= 5 Hz, 1 H, H⁵), 7.11* (dd, 1 H), 4.44*

(m, 1 H), 3.73 (m, 1 H, CH(CH₃)₂), 2.66 (s, 3 H, N=C–CH₃), 1.40

[d, ³J_H,H= 4.8 Hz, 6 H, CH(CH₃)₂] ppm. MS (FAB⁺): m/z = 986.9

[M + H]⁺–CO (M = [Os₃(CO)₁₀(iPr-AcPy)]), 957.9 [M + H]⁺

–

Photochemistry: All photochemical samples were prepared under

nitrogen, using standard inert-gas techniques, and cluster concen-

trations typically between 10^–3–10^–4mol·dm^–3. Nanosecond tran-

sient absorption (ns TA) spectra were obtained by irradiating the

samples with 2 ns pulses of the 532 nm line (typically 5 mJ pulse^–1)

of a tunable (420–710 nm) Coherent Infinity XPO laser and using a

high-power EG&G FX-504 Xe lamp as probe light.^[17]Nanosecond

flash photolysis transient kinetics were measured by irradiating the

sample with 7 ns (FWHM) pulses of a Spectra Physics GCR-3

Nd:YAG laser (10 Hz repetition rate) and using a pulsed Xe lamp

perpendicular to the laser beam as probe light. The excitation wave-

length in this case (532 nm) was obtained by frequency doubling.

The 450-W Xe lamp was equipped with a Müller Electronik

MSP05 pulsing unit, giving pulses of 0.5 ms. A shutter, placed be-

tween the lamp and the sample, was opened for 10 ms to prevent

photomultiplier fatigue. Suitable pre- and post-cut-off filters and

band-pass filters were used to minimize both the probe light and

scattered light from the laser. The sampling rate was kept at a rela-

tively long time (10 s intervals) to prevent accumulation of possible

photo-induced intermediates. The light was collected in an Oriel

monochromator, detected by a P28 PMT (Hamamatsu), and re-

corded with a Tektronix TDS3052 (500 MHz) oscilloscope. The la-

ser oscillator, Q-switch, lamp, shutter and trigger were externally

controlled with a home made digital logic circuit, which allowed

for synchronous timing. The absorption transient were plotted as

ΔA = log(I_t/I₀) vs. time, where I₀is the monitoring light intensity

prior to the laser pulse and I_tis the observed signal at delay time

t.

2CO (M = [Os₃(CO)₁₀(iPr-AcPy)]), 925.91 [M + H]⁺(M = [Os₂-

Ru(CO)₁₀(iPr-AcPy)]) (calcd. 925.89), [M + H]⁺–nCO (n = 1–10)

(M = [Os₂Ru(CO)₁₀(iPr-AcPy)]).

X-ray Crystal Structure Determinations of Clusters 2 and 3: Inten-

sities were measured with a Nonius KappaCCD diffractometer

with rotating anode (Mo-K_α, λ = 0.71073 Å) at a temperature of

150(2) K up to a resolution of (sinθ/λ) = 0.65 Å^–1. The structures

were solved with Patterson methods (DIRDIF-97)^[71]and refined

with the programme SHELXL-97^[72]against F²of all reflections.

Non-hydrogen atoms were refined freely with anisotropic displace-

ment parameters, and hydrogen atoms as rigid groups. The draw-

ings, structure calculations, and checking for higher symmetry were

performed with the programme PLATON.^[73]The ruthenium sites

in both structures were partially occupied by osmium atoms. The

ruthenium and the corresponding osmium atoms were constrained

to the same coordinates and the same anisotropic displacement

parameters. Then the partial occupancies were refined with the cri-

terion that the total occupancy remains 1.0.

[Os₂Ru(CO)₁₁(PPh₃)] (2): C₂₉H₁₅O₁₁Os_2.36PRu_0.64; Fw = 1083.94;

yellow plates, 0.36 × 0.21 × 0.12 mm³; monoclinic, space group

C2/c (no. 15); cell parameters: a = 21.9824(1), b = 16.0859(1) Å,

c = 17.2605(1) Å; β = 103.7064(3)°; V = 5929.62(6) Å³; Z = 8; ρ =

2.428 g cm^–3; F(000) = 3996; 55381 reflections were measured, 6811

reflections were unique (R_int= 0.0519). An analytical absorption

correction was applied (μ = 10.526 mm^–1, 0.08–0.39 transmission).

398 Refined parameters, no restraints. R (obsd. refl.): R₁= 0.0196,

wR₂= 0.0425. R (all data): R₁= 0.0230, wR₂= 0.0434. Weighting

2

scheme w = 1/[σ²(F_o²) + (0.0165P)²+ 11.6188P], where P = (F_o

+

Picosecond transient absorption (ps TA) spectra were recorded

using the set-up installed at the University of Amsterdam.^[16]Part

of the 800 nm output of a Ti-sapphire regenerative amplifier

(1 kHz, 130 fs, 1 mJ) was focussed into a H₂O flow-through cell

(10 mm; Hellma) to generate white light. The residual part of the

2

2F_c)/3. GoF = 1.110. Residual electron density between –0.92 and

0.75 e/Å³.

[Os₂Ru(CO)₁₀(iPr-AcPy)] (3): C₂₀H₁₄N₂O₁₀Os_2.11Ru_0.89

; Fw =

933.61; red needles, 0.39 × 0.03 × 0.03 mm³; monoclinic, space

2220

www.eurjic.org

Eur. J. Inorg. Chem. 2005, 2206–2222

Heterosite Effects in Novel Heteronuclear Clusters

FULL PAPER

[2] W. L. Gladfelter, G. L. Geoffroy, Adv. Organomet. Chem. 1980,

18, 207–273.

[3] D. A. Roberts, G. L. Geoffroy, in Comprehensive Organometal-

lic Chemistry, vol. 6, Pergamon, Oxford, 1982.

[4] S. M. Waterman, N. T. Lucas, M. G. Humphrey, Adv. Or-

ganomet. Chem. 2000, 46, 47–143.

[5] P. Braunstein, J. Rosé, in Catalysis by Di- and Polynuclear Me-

tal Cluster Complexes (Eds.: R. D. Adams, F. A. Cotton),

Wiley-VCH, New York, 1998, pp. 443–508.

[6] P. Braunstein, J. Rosé, in Comprehensive Organometallic Chem-

istry II, vol. 10, Pergamon, New York, 1995.

[7] R. D. Adams, T. S. Barnard, Z. Li, W. Wu, J. H. Yamamoto, J.

Am. Chem. Soc. 1994, 116, 9103–9113.

800 nm fundamental was used to provide 505 nm (fourth harmonic

of the 2020 OPA idler beam) excitation pulses with a general output

of 5 μJ pulse^–1. After passing through the sample, the probe beam

was coupled into a 400 μm optical fibre and detected by a CCD

spectrometer (Ocean Optics, PC2000). The chopper (Rofin Ltd., f

= 10–20 Hz), placed in the excitation beam, provided I and I₀, de-

pending on the status of the chopper (open or closed). The excited

state spectra were obtained by ΔA = log(I/I₀). Typically two thou-

sand excitation pulses were averaged to obtain the transient at a

particular time delay.

Electrochemistry: Cyclic voltammograms (CV) of approximately

10^–3m parent cluster in 10^–1m Bu₄NPF₆electrolyte solution were

recorded in a gas-tight, single-compartment, three-electrode cell

equipped with platinum microdisc (apparent surface 0.42 mm²)

working, coiled platinum wire auxillary and silver wire pseudo-ref-

erence electrodes. The cell was connected to a computer-controlled

PAR Model 283 potentiostat. All redox potentials are reported

against the ferrocene-ferrocenium (Fc/Fc⁺) redox couple.^[74,75]IR

spectroelectrochemical measurements at variable temperatures were

performed with previously described optically transparent thin-

layer electrochemical (OTTLE) cells^[76,77]equipped with a platinum

minigrid working electrode (32 wires/cm) and CaF₂windows. The

potential during these measurements was controlled by a PA4 po-

tentiostat (EKOM, Czech Republic). In the spectroelectrochemical

experiments, cluster concentrations of 5×10^–3mol·dm^–3were used.

[8] M. A. Aubart, L. H. Pignolet, J. Am. Chem. Soc. 1992, 114,

7901–7903.

[9] S. Attali, R. Mathieu, J. Organomet. Chem. 1985, 291, 205–

211.

[10] I. Ojima, Z. Zhang, J. Org. Chem. 1988, 53, 4422–4425.

[11] L. J. Farrugia, in Comprehensive Organometallic Chemistry II,

vol. 10, Pergamon, New York, 1995.

[12] J. Nijhoff, M. J. Bakker, F. Hartl, D. J. Stufkens, W.-F. Fu, R.

van Eldik, Inorg. Chem. 1998, 37, 661–668.

[13] J. Nijhoff, F. Hartl, D. J. Stufkens, J. Fraanje, Organometallics

1999, 18, 4380–4389.

[14] J. Nijhoff, M. J. Bakker, F. Hartl, D. J. Stufkens, J. Organomet.

Chem. 1999, 572, 271–281.

[15] M. J. Bakker, F. Hartl, D. J. Stufkens, O. S. Jina, X.-Z. Sun,

M. W. George, Organometallics 2000, 19, 4310–4319.

[16] F. W. Vergeer, C. J. Kleverlaan, D. J. Stufkens, Inorg. Chim.

Acta 2002, 327, 126–133.

[17] F. W. Vergeer, M. J. Calhorda, P. Matousek, M. Towrie, F.

Hartl, Dalton Trans. 2003, 4084–4099.

[18] M. J. Calhorda, E. Hunstock, L. F. Veiros, F. Hartl, Eur. J. In-

org. Chem. 2001, 223–231.

[19] M. J. Calhorda, P. J. Costa, F. Hartl, F. W. Vergeer, C. R. Chi-

mie 2005, in the press.

[20] E. Hunstock, C. Mealli, M. J. Calhorda, J. Reinhold, Inorg.

Chem. 1999, 38, 5053–5060.

[21] E. Hunstock, M. J. Calhorda, P. Hirva, T. A. Pakkanen, Orga-

nometallics 2000, 19, 4624–4628.

[22] G. A. Battiston, G. Bor, U. K. Dietler, S. F. A. Kettle, R. Ros-

setti, G. Sbrignadello, P. L. Stanghellini, Inorg. Chem. 1980, 19,

1961–1973.

[23] R. P. Ferrari, G. A. Vaglio, M. Valle, J. Chem. Soc., Dalton

Trans. 1978, 1164–1166.

[24] B. F. G. Johnson, R. D. Johnston, J. Lewis, I. G. Williams, P. A.

Kilty, Chem. Commun. 1968, 861–862.

Computational Details: All density functional theory (DFT) calcu-

lations^[78]were carried out with the Amsterdam Density Functional

(ADF2000) programme.^[37–43]Vosko, Wilk and Nusair’s local ex-

change correlation potential was used.^[79]Gradient-corrected ge-

ometry optimizations^[80,81]were performed, using the Generalized

Gradient Approximation (Becke’s exchange^[82]and Perdew’s corre-

lation^[83,84]functionals). Relativistic effects were treated by the

ZORA method^[85]The core orbitals were frozen for Ru ([1–4]s, [2–

4]p, 3d), Os ([1–4]s, [2–4]p, [3–4]d), P ([1–2]s, 2p), and C, N, O (1s).

Triple ζ Slater-type orbitals (STO) were used to describe the valence

shells of H (1s), C and O (2s and 2p), P (3s and 3p), Ru (4d, 5s

and 5p) and Os (4f, 5d, 6s). A set of polarization functions was

added: H (single ζ, 2p, 3d), C, N, O (single ζ, 3d, 4f), P(single ζ,

3d, 4f), Ru (single ζ, 5p, 4f) and Os (single ζ, 6p, 5f). Full geometry

optimizations were performed without any symmetry constraints

on models based on the available crystal structures. The UV/Vis

spectrum was calculated using the SWizard programme,^[48]revision

3.7. The half-band widths, Δ_1/2,I, were taken to be equal to

3000 cm^–1, the default value of the programme. Three-dimensional

representations of the orbitals were obtained with Molekel.^[86]

[25] J. R. Fox, W. L. Gladfelter, G. L. Geoffroy, Inorg. Chem. 1980,

19, 2574–2578.

[26] H. Remita, R. Derai, M.-O. Delcourt, Radiat. Phys. Chem.

1991, 37, 221–225.

[27] J. R. Moss, W. A. G. Graham, J. Chem. Soc., Dalton Trans.

1977, 89–94.

[28] B. F. G. Johnson, R. D. Johnston, P. L. Josty, J. Lewis, I. G.

Williams, Nature 1967, 213, 901–902.

Time-dependent DFT (TD-DFT)^[87–89]calculations in the ADF

implementation were used to determine the excitation energies. In

all cases the ten lowest singlet-singlet excitation energies were calcu-

lated using the optimized geometries.

[29]

[30]

[31]

B. F. G. Johnson, R. D. Johnston, J. Lewis, J. Chem. Soc., A

1969, 792–797.

J. R. Moss, M. L. Niven, E. E. Sutton, Inorg. Chim. Acta 1988,

147, 251–256.

N. K. Bhattacharyya, T. J. Coffy, W. Quintana, T. A. Salupo,

J. C. Bricker, T. B. Shay, M. Payne, S. G. Shore, Organometallics

1990, 9, 2368–2374.

M. R. Churchill, B. G. DeBoer, Inorg. Chem. 1977, 16, 878–

884.

M. R. Churchill, F. J. Hollander, J. P. Hutchinson, Inorg. Chem.

1977, 16, 2655–2659.

L. Pereira, W. K. Leong, S. Y. Wong, J. Organomet. Chem.

2000, 609, 104–109.

M. I. Bruce, M. J. Liddell, C. A. Hughes, B. W. Skelton, A. H.

White, J. Organomet. Chem. 1988, 347, 157–180.

Acknowledgments

A. Hearley (University of Cambridge, UK) is gratefully acknowl-

edged for valuable discussions regarding the synthesis of [Os_2-

Ru(CO)₁₂]. This work was undertaken as part of the European col-

laborative COST project (D14/0001/99). Financial support was re-

ceived by F. W. V. and F. H. from the Council for Chemical Sciences

of the Netherlands Organization for Scientific Research (CW-

NWO, project No. 348–032).

[32]

[33]

[34]

[35]

[1] R. D. Adams, in Comprehensive Organometallic Chemistry II,

vol. 10, Pergamon, New York, 1995.

Eur. J. Inorg. Chem. 2005, 2206–2222

www.eurjic.org

2221

F. W. Vergeer, M. Lutz, A. L. Spek, M. J. Calhorda, Derk J. Stufkens, F. Hartl

FULL PAPER

[36] E. J. Forbes, N. Goodhand, D. L. Jones, T. A. Hamor, J. Or-

ganomet. Chem. 1979, 182, 143–154.

[37] E. J. Baerends, D. Ellis, P. Ros, Chem. Phys. 1973, 2, 41–51.

[38] E. J. Baerends, P. Ros, Int. J. Quantum Chem. 1978, S12, 169–

190.

[60]

[61]

F. Hartl, J. W. M. van Outersterp, unpublished results.

S. Chardon-Noblat, A. Deronzier, D. Zsoldos, R. Ziessel, M.

Haukka, T. Pakkanen, T. Venäläinen, J. Chem. Soc., Dalton

Trans. 1996, 2581–2583.

C. Caix-Cecillon, S. Chardon-Noblat, A. Deronzier, M.

Haukka, T. Pakkanen, R. Ziessel, D. Zsoldos, J. Electroanal.

Chem. 1999, 466, 187–196.

S. Chardon-Noblat, A. Deronzier, F. Hartl, J. van Slageren, T.

Mahabiersing, Eur. J. Inorg. Chem. 2001, 613–617.

F. Hartl, T. Mahabiersing, S. Chardon-Noblat, P. Da Costa, A.

Deronzier, Inorg. Chem. 2004, 43, 7250–7258.

S. Chardon-Noblat, A. Deronzier, R. Ziessel, Collect. Czech.

Chem. Commun. 2001, 66, 207–226.

S. Chardon-Noblat, A. Deronzier, R. Ziessel, D. Zsoldos, J.

Electroanal. Chem. 1998, 444, 253–260.

M.-N. Collomb-Dunand-Sauthier, A. Deronzier, R. Ziessel, In-

org. Chem. 1994, 33, 2961–2967.

J. W. M. van Outersterp, Ph. D. Thesis, University of Amster-

dam, Amsterdam, 1995.

B. F. G. Johnson, J. Lewis, P. A. Kilty, J. Chem. Soc., A 1968,

2859–2864.

G. A. Foulds, B. F. G. Johnson, J. Lewis, J. Organomet. Chem.

1985, 296, 147–153.

P. T. Beurskens, G. Admiraal, G. Beurskens, W. P. Bosman, S.

Carcia-Granda, R. O. Gould, J. M. M. Smits, C. Smykalla, The

DIRDIF-97 programme system, Technical Report of the Crys-

tallography Laboratory, University of Nijmegen, The Nether-

lands, 1997.

[62]

[39] E. J. Baerends, A. Bérces, C. Bo, P. M. Boerrigter, L. Cavallo,

L. Deng, R. M. Dickson, D. E. Ellis, L. Fan, T. H. Fischer, C.

Fonseca Guerra, S. J. A. van Gisbergen, J. A. Groeneveld, O. V.

Gritsenko, F. E. Harris, P. van den Hoek, H. Jacobsen, G.

van Kessel, F. Kootstra, E. van Lenthe, V. P. Osinga, P. H. T.

Philipsen, D. Post, C. C. Pye, W. Ravenek, P. Ros, P. R. T.

Schipper, G. Schreckenbach, J. G. Snijders, M. Sola, D. Swer-

hone, G. te Velde, P. Vernooijs, L. Versluis, O. Visser, E.

van Wezenbeek, G. Wiesenekker, S. K. Wolff, T. K. Woo, T. Zi-

egler, ADF 2000, University of Amsterdam, Amsterdam, 2000.

[40] C. Fonseca Guerra, O. Visser, J. G. Snijders, G. te Velde, E. J.

Baerends, in Methods and Techniques for Computational Chem-

istry (Eds.: E. Clementi, C. Corongiu), STEF, Cagliari, 1995,

pp. 303–395.

[41] C. Fonseca Guerra, J. G. Snijders, G. te Velde, E. J. Baerends,

Theor. Chem. Acc. 1998, 99, 391–403.

[42] P. M. Boerrigter, G. te Velde, E. J. Baerends, Int. J. Quantum

Chem. 1988, 33, 87–113.

[43] G. te Velde, E. J. Baerends, J. Comput. Phys. 1992, 99, 84–98.

[44] V. M. Hansen, A. K. Ma, K. Biradha, R. K. Pomeroy, M. J.

Zaworotko, Organometallics 1998, 17, 5267–5274.

[45] N. E. Leadbeater, J. Lewis, P. R. Raithby, G. N. Ward, J. Chem.

Soc., Dalton Trans. 1997, 2511–2516.

[46] R. Zoet, T. B. H. Jastrzebski, G. van Koten, T. Mahabiersing,

K. Vrieze, D. Heijdenrijk, C. H. Stam, Organometallics 1988,

7, 2108–2117.

[63]

[64]

[65]

[66]

[67]

[68]

[69]

[70]

[71]

[72]

G. M. Sheldrick, SHELXL-97, University of Göttingen,

Göttingen, Germany, 1997.

A. L. Spek, J. Appl. Crystallogr. 2003, 36, 7–13.

V. V. Pavlishchuk, A. W. Addison, Inorg. Chim. Acta 2000, 298,

97–102.

[73]

[74]

[47] D. M. Manuta, A. J. Lees, Inorg. Chem. 1983, 22, 3825–3828.

[48] S. I. Gorelsky, SWizard programme, revision 3.7, http://

www.obbligato.com/software/swizard/.

[49] B. D. Rossenaar, D. J. Stufkens, A. Vlcˇek Jr., Inorg. Chim. Acta

1996, 247, 247–255.

[50] B. D. Rossenaar, D. J. Stufkens, A. Vlcˇek Jr., Inorg. Chem.

1996, 35, 2902–2909.

[51] B. D. Rossenaar, E. Lindsay, D. J. Stufkens, A. Vlcˇek Jr., Inorg.

Chim. Acta 1996, 250, 5–14.

[52] M. P. Aarnts, D. J. Stufkens, M. P. Wilms, E. J. Baerends, A.

Vlcˇek Jr., I. P. Clark, M. W. George, J. J. Turner, Chem. Eur. J.

1996, 2, 1556–1565.

[75] G. Gritzner, J. Ku˚ta, Pure Appl. Chem. 1984, 56, 461–466.

[76] F. Hartl, H. Luyten, H. A. Nieuwenhuis, G. C. Schoemaker,

Appl. Spectrosc. 1994, 48, 1522–1528.

[77] M. Krejcˇík, M. Daneˇk, F. Hartl, J. Electroanal. Chem. Interfa-

cial Electrochem. 1991, 317, 179–187.

[78] R. G. Parr, W. Yang, Density Functional Theory of Atoms and

Molecules, Oxford University Press, New York, 1989.

[79] S. H. Vosko, L. Wilk, M. Nusair, Can. J. Phys. 1980, 58, 1200–

1211.

[53] J. Nijhoff, F. Hartl, J. W. M. van Outersterp, D. J. Stufkens,

M. J. Calhorda, L. F. Veiros, J. Organomet. Chem. 1999, 573,

121–133.

[54] A. Klein, W. Kaim, Organometallics 1995, 14, 1176–1186.

[55] F. W. Vergeer, C. J. Kleverlaan, P. Matousek, M. Towrie, D. J. [84] J. P. Perdew, Phys. Rev. B 1986, 34, 7406.

Stufkens, F. Hartl, Inorg. Chem. 2005, 44, 1319–1331.

[56] F. P. A. Johnson, M. W. George, F. Hartl, J. J. Turner, Organo-

metallics 1996, 15, 3374–3387.

[80] L. Versluis, T. Ziegler, J. Chem. Phys. 1988, 88, 322–328.

[81] L. Fan, T. Ziegler, J. Chem. Phys. 1991, 95, 7401–7408.

[82] A. D. Becke, J. Chem. Phys. 1988, 88, 1053–1062.

[83] J. P. Perdew, Phys. Rev. B 1986, 33, 8822–8824.

[85] E. van Lenthe, A. Ehlers, E. J. Baerends, J. Chem. Phys. 1999,

110, 8943–8953.

[86] S. Portmann, H. P. Lüthi, Chimia 2000, 54, 766–770.

[87] S. J. A. van Gisbergen, J. A. Groeneveld, A. Rosa, J. G.

Snijders, E. J. Baerends, J. Phys. Chem. 1999, 103, 6835–6844.

[88] S. J. A. van Gisbergen, A. Rosa, G. Ricciardi, E. J. Baerends, J.

Chem. Phys. 1999, 111, 2499–2506.

[57] G. J. Stor, F. Hartl, J. W. M. van Outersterp, D. J. Stufkens, Or-

ganometallics 1995, 14, 1115–1131.

[58] B. D. Rossenaar, F. Hartl, D. J. Stufkens, Inorg. Chem. 1996,

35, 6194–6203.

[59] B. D. Rossenaar, F. Hartl, D. J. Stufkens, C. Amatore, E. Mai-

sonhaute, J.-N. Verpeaux, Organometallics 1997, 16, 4675–

4685.

[89] A. Rosa, S. J. A. van Gisbergen, E. van Lenthe, J. A. Groenev-

eld, J. G. Snijders, J. Am. Chem. Soc. 1999, 121, 10356–10365.

Received: December 20, 2004

2222

www.eurjic.org

Eur. J. Inorg. Chem. 2005, 2206–2222

Article Doi

Heterosite effects in novel heteronuclear clusters [Os 2Ru(CO)11(PPh3)] and [Os2Ru(CO) 10(2-acetylpyridine-N-isopropylimine)]

DOI: 10.1002/ejic.200401040

Source and publish data:

Authors:

Article abstract of DOI:10.1002/ejic.200401040

Full text of DOI:10.1002/ejic.200401040

Products guided by the article

R&D Labs maybe for 869477-40-7

Relevant to this article

Hot Product